IDENTIFYING ENERGY EFFICIENCY IMPROVEMENTS AND … · 2016-01-28 · * This document is fully electronically signed on 14/01/2016. TRACTEBEL ENGINEERING S.A. – Registered Office:

IDENTIFYING ENERGYEFFICIENCY IMPROVEMENTSAND SAVING POTENTIAL INENERGY NETWORKS,INCLUDING ANALYSIS OF THEVALUE OF DEMAND RESPONSE In support of the implementation of article 15 of the energy efficiency directive(2012/27/EU)

Final report

DG ENERGY - EUROPEAN COMMISSION |BELGIUM

RESTRICTED

18 December 2015

TRACTEBEL ENGINEERINGAvenue Ariane, 7 – 1200 Brussels - BELGIUMtel. +32 2 773 99 11 - fax +32 2 773 99 [email protected] TECHNICAL NOTE

Our ref.: GRIDEE/4NT/0364174/000/01

TS:

Imputation: P.007629/0004

RESTRICTED

Client :

Project : ENER/C3/2013-423 Grid Energy Efficiency

Subject : Final Report

Comments:

01 16/01/12 FIN *TE TEAM*ECOFYS TEAM*CRIGEN TEAM

*TE TEAM*ECOFYS TEAM

*S. Rapoport

00 16/01/12 FIN *TE TEAM*ECOFYS TEAM*CRIGEN TEAM

*TE TEAM*ECOFYS TEAM

*S. Rapoport

REV. YY/MM/DD STAT. WRITTEN VERIFIED APPROVED VALIDATED

* This document is fully electronically signed on 14/01/2016.

TRACTEBEL ENGINEERING S.A. – Registered Office: Avenue Ariane 7 – 1200 Brussels - BELGIUMVAT:BE 0412 639 681 – RPM/RPR Brussels: 0412 639 681 – Bank account - IBAN: BE74375100843707 – BIC/SWIFT: BBRUBEBB Th

isdo

cum

ent

isth

epr

oper

tyof

Trac

tebe

lEng

inee

ring

S.A.

Any

dupl

icat

ion

ortr

ansm

issi

onto

third

part

ies

isfo

rbid

den

with

out

prio

rw

ritte

nap

prov

al

IDENTIFYING ENERGY EFFICIENCY IMPROVEMENTS AND SAVING POTENTIAL IN ENERGY NETWORKS, INCLUDING

ASSESSMENT OF THE VALUE OF DEMAND RESPONSE, IN SUPPORT OF THE IMPLEMENTATION OF ARTICLE 15 OF THE

ENERGY EFFICIENCY DIRECTIVE (2012/27/EU)

GRIDEE/4NT/0364174/000/01 Ed. 2015/12/18 3/180 RESTRICTED

This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Identifying Energy Efficiencyimprovements and saving potential in

energy networks, including analysis of thevalue of demand response, in support of

the implementation of article 15 of theenergy efficiency directive (2012/27/EU)

Final reportDecember 2015





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Disclaimer

This document has been prepared for the European Commission however it reflects the

views only of the authors, and the Commission cannot be held responsible for any use

which may be made of the information contained therein.

© Tractebel Engineering, Ecofys 2015 by order of: European Commission





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

FOREWORD

This report has been prepared for the European Commission by a consortium composed ofTractebel Engineering and Ecofys.

The report has been organized in two parts.

The first part of the report focuses on the identification and assessment of measures to improvethe efficiency of electricity and gas networks.

The second part specifically focuses on the use of demand response to improve the efficiency ofelectricity networks, concerning reduction of losses and optimization of grid planning andoperation.

The project has been led by Stéphane Rapoport and Vincenzo Giordano (Tractebel Engineering),with the support of Georgios Papaefthymiou (Ecofys). The core team has included GregoireLejeune (Tractebel), Sébastien Leyder (Tractebel), Farid Comaty (Ecofys), Rena Kuwahata(Ecofys), Amelie Bonard (CRIGEN) and Pascal Vercamer (CRIGEN).

In addition to the core team valuable contributions have been made by several of our colleaguesat Tractebel Engineering and Ecofys.

The work has been followed by Mr Massimo Maraziti of the European Commission DG-Energy.During the study, a survey has been carried out among European DSOs, thanks to the facilitationrole of European associations EURELECTRIC, CEDEC, GEODE, EDSO4SG.

All conclusions are the responsibility of the analysis team and do not necessarily reflect the viewsof the European Commission, or any of the consulted stakeholders.

We are very grateful for all the valuable comments and suggestions received from the DG Energyofficers and the consulted organizations. Any remaining errors or omissions are the soleresponsibility of the analysis team.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

TABLE OF CONTENTS

ACRONYMS .................................................................................................................... 11

PART 1 - IDENTIFYING ENERGY EFFICIENCY IMPROVEMENTS AND SAVINGPOTENTIAL IN ENERGY NETWORKS............................................................................ 12

PART 1 - EXECUTIVE SUMMARY ....................................................................................... 13

1. INTRODUCTION (PART 1) .......................................................................................... 17

1.1. Objectives of this study .............................................................................. 17

1.2. The Energy Efficiency Directive .................................................................. 18

1.3. Reading guide to this document ................................................................. 19

1.3.1. Energy efficiency measures for electricity grids ....................................... 191.3.2. Energy efficiency measures for gas grids ................................................ 20

2. POTENTIAL OF ENERGY EFFICIENCY MEASURES ON LOSS REDUCTION INELECTRICITY GRIDS ................................................................................................. 21

2.1. Understanding energy losses in electricity networks ................................. 22

2.1.1. Type: Technical/non-Technical and Variable/Fixed .................................. 222.1.2. Voltage level: transmission and distribution networks .............................. 232.1.3. Components: Lines and Transformers .................................................... 242.1.4. Marginal vs Average losses ................................................................... 262.1.5. Role of DG and grid management ......................................................... 272.1.6. Reactive Losses ................................................................................... 282.1.7. Imbalances ......................................................................................... 292.1.8. Life cycle assessment: impact of losses on investment decisions ............... 29

2.2. Energy efficiency measures in electricity networks: categorisation andpotential ...................................................................................................... 29

2.2.1. Energy Efficient transformers ................................................................ 322.2.2. Increasing voltage level ....................................................................... 332.2.3. Increasing line capacity ........................................................................ 332.2.4. Feed-in control .................................................................................... 342.2.5. Transformers switching ........................................................................ 362.2.6. Network reconfiguration ....................................................................... 372.2.7. Conservation voltage reduction ............................................................. 372.2.8. Voltage optimisation ............................................................................ 382.2.9. Balancing loading of lines ..................................................................... 40

2.3. Regulatory regimes to promote loss reduction in electricity grids ............ 40





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

2.3.1. Regulatory incentives: basic options ...................................................... 402.3.2. State of play of regulatory mechanisms targeting loss reduction ............... 41

3. POTENTIAL OF MEASURES ON PLANNING/OPERATIONAL EFFICIENCY INELECTRICITY GRIDS ................................................................................................. 44

3.1. Impact of selected energy efficiency measures on operational efficiency 45

3.1.1. High efficiency transformers ................................................................. 453.1.2. Expanding line capacity ........................................................................ 463.1.3. Increasing voltage level ....................................................................... 473.1.4. Demand response ............................................................................... 493.1.5. Feed-in control of distributed generation & storage ................................. 503.1.6. Network reconfiguration ....................................................................... 513.1.7. Switching off redundant transformers .................................................... 523.1.8. Voltage optimisation ............................................................................ 533.1.9. Conservation Voltage Reduction (CVR) ................................................... 543.1.10. Balancing 3-phase loads ....................................................................... 55

4. COST-EFFECTIVENESS OF ENERGY EFFICIENCY MEASURES IN ELECTRICITY GRIDS ....... 56

4.1. Assessment of cost-effectiveness ............................................................... 56

4.2. Synthesis of the cost effectiveness of the measures .................................. 57

5. ASSESSMENT OF THE ENERGY EFFICIENCY POTENTIAL IN GAS GRIDS .......................... 61

5.1. Understanding Energy Losses in gas networks .......................................... 61

5.1.1. Mapping of losses in gas transmission network ....................................... 645.1.2. Mapping of losses in compression stations.............................................. 655.1.3. Mapping of losses in gas distribution network ......................................... 65

5.2. Energy efficiency measures in gas networks: categorisation and potential66

5.2.1. Gas transmission network: energy efficiency measures and reduction potential67

5.2.2. Compressor stations: energy efficiency measures and reduction potential .. 685.2.3. Distribution grids: energy efficiency measures and reduction potential ...... 69

6. ASSESSMENT OF COST-EFFECTIVENESS OF SELECTED ENERGY EFFICIENCYMEASURES - GAS ...................................................................................................... 71

6.1. Gas Transmission Network ......................................................................... 71

6.2. Gas Compressor Stations ............................................................................ 72

6.3. Gas Distribution Network ........................................................................... 74

PART 2 - ASSESSING THE POTENTIAL OF DEMAND RESPONSE TO INCREASE THEEFFICIENCY OF ELECTRICITY NETWORKS ................................................................... 76





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

PART 2 - EXECUTIVE SUMMARY ....................................................................................... 77

7. INTRODUCTION (PART 2) .......................................................................................... 84

8. OVERVIEW OF DEMAND RESPONSE CHARACTERISTICS ................................................ 85

8.1. Benefits of demand response across the power system value chain ......... 85

8.2. Factors affecting the value of Demand Response ...................................... 87

8.2.1. Type of loads ...................................................................................... 878.2.2. Characteristics of demand response....................................................... 888.2.3. Activation mechanisms of demand response ........................................... 898.2.4. Price-based mechanisms (implicit DR).................................................... 898.2.5. Volume-based mechanisms (explicit DR) ................................................ 90

9. VALUE STREAMS AND PROCUREMENT OPTIONS OF DEMAND RESPONSE FORDISTRIBUTION SYSTEM OPERATORS .......................................................................... 92

9.1. Value streams of demand response for distribution system operators ...... 92

9.2. Options for DSOs to procure demand response and capitalise the benefits94

9.2.1. Price-based procurement of demand response ........................................ 949.2.2. Volume-based procurement of demand response ...................................1019.2.3. Procurement options: Conclusions ........................................................105

10. ASSESSMENT OF THE VALUE OF DEMAND RESPONSE TO IMPROVE DISTRIBUTIONGRID EFFICIENCY AT PLANNING AND OPERATION ......................................................106

10.1. Objectives and global approach................................................................ 106

10.2. Building the simulation models ................................................................ 108

10.3. Definition of scenarios and simulation results ......................................... 113

10.3.1. Base case ..........................................................................................11410.3.2. Time-of-Use scenario ..........................................................................11510.3.3. Demand response (volume-based) .......................................................11710.3.4. Full-flex scenario ................................................................................121

10.4. Conclusions and key messages ................................................................. 123

11. REGULATORY EVOLUTIONS FOR DISTRIBUTION SYSTEM OPERATORS TOPROCURE AND ACTIVATE DEMAND RESPONSE ...........................................................126

11.1. Price-based approach: DSOs’ use of tariff signals .................................... 126

11.2. Volume-based approach: DSOs directly procure demand response ......... 128

11.2.1. Evolution of DSO remuneration based on TOTEX ...................................129





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

11.3. Role of DSO regarding ownership and use of flexibility assets ................ 130

11.4. Role of DSO as validator of DR transactions at distribution level ............ 131

11.5. Regulatory evolutions: conclusions and key messages ............................ 134

12. BIBLIOGRAPHY ........................................................................................................136

13. ANNEX I - PARAMETERS INFLUENCING THE IMPLEMENTATION AND THE COST-EFFECTIVENESS OF THE MEASURES – DETAILED ANALYSIS.........................................142

14. ANNEX II ENERGY EFFICIENCY MEASURES IN EUROPEAN MEMBER STATES:SURVEY ..................................................................................................................153

15. ANNEX III - SURVEY - QUESTIONNAIRE FORMAT ........................................................161

16. ANNEX IV - TOOL FOR RANKING MEASURES ACCORDING TO THEIR COST-EFFECTIVENESS ......................................................................................................166

17. ANNEX V- DETAILED SIMULATION RESULTS ...............................................................172





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

ACRONYMS

Capital Expenditure CAPEX

CHP Combined Heat and Power

Prosumers Consumers with local generation

DR Demand Response

DER Distributed Energy Resources

DG Distributed Generation

DMS Distribution Management System

DSO Distribution System Operator

EE Energy Efficiency

EV Electrical Vehicle

NPV Net Present Value

O&M Operation and maintenance costs

OMS Outage Management System

OPEX Operational Expenditure

PV Photo voltaic electricity generation

RES Renewable Energy Sources

SAIDI System Average Interruption Duration Index

SCADA Supervisory Control and Data Acquisition

TSO Transmission System Operator

VRES Variable Renewable Energy Sources





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

PART 1 - IDENTIFYING ENERGY EFFICIENCYIMPROVEMENTS AND SAVING POTENTIAL INENERGY NETWORKS





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

PART 1 - EXECUTIVE SUMMARY

Objective of the study

According to Article 15(2) of the Energy Efficiency Directive (2012/27/EU), Member Stateshave to ensure, by the 30th June 2015, that:

(a) an assessment is undertaken of the Energy Efficiency potentials of their gas andelectricity infrastructure, in particular regarding transmission, distribution, loadmanagement and interoperability, and connection to energy generating installations, includingaccess possibilities for micro energy generators;(b) concrete measures and investments are identified for the introduction of cost-effectiveEnergy Efficiency improvements in the network infrastructure, with a timetable for theirintroduction.

According to collected information at the moment of finalization of the interim report, someMember States are finalising their assessment, while others presumably are not.

The main objective of this study is therefore to support Member States which have not yet carriedout the assessment. The study also aims at defining common ground to carry out the review of theassessment reports that will be submitted by Member States to the EC.

Perimeter of the study

Energy savings/loss reduction. Loss reduction is an important aspect of efficiency in energygrids. To date, the electricity and gas consumption is responsible for 43% of the final energyconsumption in the EU and their infrastructure losses account for 1,5% and 0.3% respectively .Losses levels vary among Members States. One key way to improve energy efficiency is toreduce energy wastage. In the electricity sector this wastage is referred to as losses and in the gassector as shrinkage [2]. Energy efficiency potential therefore implies the minimisation of wastagein both gas and electricity transmission and distribution. The study will assess the potential ofselected measures to meet this objective.

Planning/Operational Efficiency - The directive also considers that, in assessing grid energyefficiency measures, planning/operational efficiency should also be considered. This entailsexploring opportunities to improve the efficient operation of available energy infrastructure andreduce the need for investing in new infrastructures1.

This study will particularly consider the planning/operational efficiency potential in electricitygrids, where, thanks to new “Smart” measures allowing active grid management (Smart Gridassets; flexibility of distributed energy resources), a higher potential for operation efficiency isavailable.

1 DG ENER guidance note http://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:52013SC0450&from=EN point 57 and 58, Chapter E





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

In particular, detailed analysis of the impact of demand response on energy efficiency (taking abroad view on energy efficiency and encompassing losses, and planning/operational efficiency)will be carried out in the second part of this study.

Methodological approach

The assessment of the energy efficiency potential in electricity and gas grids has been carried outaccording to the following four-step approach:

1) Identification of key measures improving the energy efficiency of gas and electrical grids.2) Identification and discussion of the potential of those measures in term of losses reduction,

through the identification the type of sources of losses in electricity and gas grids and amapping between the sources of losses and the corresponding measures that can mitigatethose losses

3) Discussion of the potential for planning/operational efficiency of each measure, lookingat aspects like investment deferral, RES hosting capacity, outage reduction, power quality

4) Identify the factors that affect the potential of each measure (e.g. load demand, pressurelevel etc.)

The analysis is supported by examples of the observed impacts of the measures based onliterature review and a data collection survey carried out with European system operators’associations (Eurelectric, EDSO, GEODE, CEDEC). Around 15 European distribution systemoperators have participated to the survey and shared their experiences and best practices (moredetails in ANNEX III and IV).

A key outcome of the study is the definition of a step-by-step methodology to guide MemberStates and system operators in the definition of energy efficiency measures for electricity andgas grids.

To this end an ad-hoc decision support tool has been set-up (ANNEX V).The goal of the tool isto help project promoters in selecting good candidate measures and refine their Energy Efficiencystrategy before carrying out full detailed cost-benefit analyses.

Some of the key outcomes and insights are summarized here below.

Grid Energy Efficiency - State of play in Member States

European associations report that most of their members have not yet been actively involvedby Member States in addressing the provisions of article 15.2. The level of mobilization in EUMember States in the implementation of the provisions of article 15.2 (definition ofinvestment plans to improve grid energy efficiency) appears to vary across Member States.Based on collected information, a number of Member States are working on the topic andwould likely publish their official assessments in the period after the completion of this study.According to information available at the time of publication of this study, Member Statesseem mainly to consider loss reduction initiatives in their assessments.Starting conditions of grids (e.g. voltage level; number of transformers) and potentials forenergy efficiency improvement vary widely. In particular, the level of losses variessignificantly among Member States.Loss reduction is typically taken into account in grid investment decisions, however often it isnot the main driver for ad-hoc investments.For electricity grids,





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

- System operators generally opt for traditional investments (replacement of assets;reinforcements etc.) to improve grid energy efficiency (mainly targeting loss reduction)

- Smart Grid-related EE measures (e.g. use of flexibility of DER) to pursue energyefficiency are less common; the existence of regulatory barriers hindering their viability.

Mapping of losses – electricity grids

A detailed categorization of losses in electricity grids have been carried out to map out whereand why losses occur in electricity grids.A list of factors that have a key impact on the level of losses in the system have beenidentified, including:- Loading (including peak demand)- Number of energized transformers- Lengths of the feeders- Level of power quality- Presence of distributed energy resourcesThe main potential for reduction of electricity grids is at distribution level, where the majorityof losses occur.The implementation of adequate metering systems at substations and at customers’ premises isan important first step to identify where losses are actually occurring in the distribution gridand to better target EE measures.

Mapping of losses – gas grids

A detailed mapping of losses in gas grids and compression stations has been carried out. Over15 sources of losses have been identified.A list of factors that have a key impact on the level of losses in the system have beenidentified and analysed, including:- Pressure level, volume of vented pipe- Length and age of the grid, pipes and joints materials- Number and efficiency of heaters, necessity of burners- Base or peak load compressor

Planning/Operational efficiency improvement potential

The potential benefits of each measure in term of planning/operation efficiency have beenanalysed, particularly for electricity grids where a higher potential of improvement exist. Thefollowing dimensions of planning/operational efficiency have been considered:- Investment deferral- RES hosting capacity- Security of supply (reduction of outages)- Power qualityThe factors impacting the effectiveness of each measure and the potential cost ranges havealso been analysed. The factors need to be tailored to specific grid conditions to assess thecost-effectiveness of individual measures in a given context.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Drivers and regulation

Measures which require the exploitation of DER flexibility have still limited diffusion inmany Member States, often just at pilot level. Their implementation would require thedefinition of new regulatory mechanisms over how DSOs can procure and activate DERflexibility. On this topic, a specific focus on Demand Response will be carried out in thesecond part of the study.Energy efficiency (and particularly loss reduction) is the main driver of investments whenthere is the need to comply with regulatory requirements and quality standards (e.g.EcoDesign directive). In general, loss reduction is considered as one of the variables of alarger investment optimization plan, but not the main driver for investments.The incentive to pursue specific investments for loss reduction is strictly related to thepresence of regulatory incentives. A wide variety of regulatory frameworks are present inEuropean Member States, which in some cases do not explicitly support loss reductionmeasures.

Catalogue of measures and decision-support tool

A detailed catalogue of energy efficiency measures together with an assessment of their cost-effectiveness has been defined for electricity and gas grids.As the potential of these measures is very much related to the specific local grid conditions,the analysis has focused on the assessment of key factors affecting the benefits and the costsof selected measures. System operators then need to customize these factors to their specificlocal grid conditions.An excel tool has been set-up to guide project promoters in the assessment of the cost-effectiveness of different energy efficiency measures according to their local grid conditions.The tool highlights the main factors affecting the benefits and cost of each measure andprovides examples for benchmark.

Cost-effectiveness of energy efficiency measures

A number of factors can greatly impact the cost-effectiveness of energy efficiency measures,such as:- Monetary value of losses, i.e. the average price of wholesale electricity (the higher, the

more cost-effective are energy efficiency measures)- Loss reduction is in many cases a side-benefit of investments driven by other needs. The

cost-effectiveness of energy efficiency improvements should thus be analysed within thecontext of a larger investment plan.

- Parallel grid developments (e.g. roll-out of smart meters; built-in smart functionalities inDG converters etc.) make the enabling infrastructure for certain EE measures alreadyavailable, reducing their implementation costs.

- Energy efficiency measures could have conflicting objectives (e.g. loss reduction andhosting capacity when implementing DG voltage optimization).

- External factors outside the control of system operators can also greatly impact thepotential of EE measures: e.g. the development and location of DERs.

A system approach to energy efficiency

A systemic approach should be adopted when assessing the impacts of the implementation ofa portfolio of EE measures. This could include:





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

- Synergies among EE measures, i.e. consider the whole EE program rather than individualEE measures. This includes the assessment of possible negative influence of parallelimplementation of measures.

- Synergies with on-going parallel investments in the energy system (e.g. installation ofsmart converters for DGs)

- Synergies with replacement/investment programs already scheduled (e.g. oversizing newlines taking into account the loss reduction potential).

1. INTRODUCTION (PART 1)

1.1. Objectives of this study

According to Article 15(2) of the EE Directive (2012/27/EU), Member States have to ensure, bythe 30th June 2015, that:

(a) an assessment is undertaken of the Energy Efficiency potentials of their gas andelectricity infrastructure, in particular regarding transmission, distribution, loadmanagement and interoperability, and connection to energy generating installations, includingaccess possibilities for micro energy generators;(b) concrete measures and investments are identified for the introduction of cost-effectiveEnergy Efficiency improvements in the network infrastructure, with a timetable for theirintroduction.

According to collected information at the moment of finalization of the interim report, someMember States are finalising their assessment, while others presumably are not.

The main objective of this study is therefore to support Member States which have not yet carriedout the assessment. The study also aims at defining common ground to carry out the review of theassessment reports submitted by Member States to the EC.

The analysis is supported by examples of the observed impacts of the measures based onliterature review and a data collection survey carried out with European system operators’associations.

Finally the study aims at integrating the result of the analysis in an easy-to use decision supporttool, to be used by system operators to screen most cost-effective energy efficiency investments.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

1.2. The Energy Efficiency Directive

The 2012 Energy Efficiency Directive (EED) establishes a set of binding measures to help theEU reach its 20% energy efficiency target by 2020 [1]. The Directive states that 1.5% of theaverage final energy consumption in the period 2010-2012 needs to be saved annually from 2014onwards till 2020. To achieve that, energy efficiency improvements throughout the whole valuechain of the energy sector are required, that is from generation-transmission-distribution toconsumption of any energy carrier to be delivered to the industry, building or transportationsector.

Article 15 of the EED targets specifically the transmission and distribution sector of electricityand gas and requires thus Member states to perform an assessment of the energy efficiencypotential of their respective infrastructure.

Loss reduction is a key aspect of energy efficiency in energy grids. To date, the electricity andgas consumption is responsible for 43% of the final energy consumption in the EU and theirinfrastructure losses account for 1,5% and 0.3% respectively2. Losses levels vary amongMembers States. One key way to improve energy efficiency is to reduce energy wastage. In theelectricity sector this wastage is referred to as losses and in the gas sector as shrinkage [2].Energy efficiency potential therefore implies the minimisation of wastage in both gas andelectricity transmission and distribution networks. Energy savings that can be provided byinvesting in measures aiming at reducing those grid losses can be significant, depending on thelocal situation.

Moreover, the directive also considers that grid energy efficiency measures should go beyondloss reduction and also take into account planning/operational efficiency. This entails exploringopportunities to improve the efficient operation of available energy infrastructure and reduce theneed for investing in new infrastructures3.

This study will particularly consider the planning/operational efficiency in electricity grids,where, thanks to new “Smart” measures allowing active grid management, a higher potential forimprovement is available. In particular, a detailed analysis of the impact of demand response onoperational efficiency will be carried out in the second part of this study.

2 Ecofys analysis from EuroStat database3 See DG ENER guidance note http://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:52013SC0450&from=EN point 57 an d 58, Chapter E





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

1.3. Reading guide to this document

For sake of brevity and readability, the main text reports the key results and messages of theanalysis.A more detailed description of all the analysis that has been carried out is reported in theAnnexes:

ANNEX I reports detailed analysis of the parameters affecting the potential and the cost-effectiveness of EE measures.ANNEX II and III report the results and the questionnaire of the stakeholders’ survey.ANNEX IV describes in detail the decision support tool (excel file) that synthetises theanalysis in a step-by-step methodology.

The interim report is organized in two main parts. A first part analyses the potential and the cost-effectiveness of energy efficiency measures in electricity networks, whereas the second partcarries out the same analysis for gas networks.

1.3.1. Energy efficiency measures for electricity gridsEnergy efficiency measures for electricity grids are analysed in chapters 2 and 3. Chapter 4provides a summary of the cost-effectiveness of the selected energy efficiency measures.The study considers three categories of energy efficiency measures for electricity grids:

Traditional ones (replacement/reinforcement)Network management (e.g. Smart Grid assets)Flexibility of distributed energy resources (Feed-in control).

Table 1 - Perimeter of the study for energy efficiency measures in electricity grids provides amapping of the perimeter of the study for the energy efficiency measures in electricity grids.

The second part of the study carries out a dedicated focus on the feed-in control category,especially on demand response. In particular, the use of DR flexibility in planning and operationwill be thoroughly explored.

EE Measures for electricitygrids

Grid energy efficiency – electricity grids

Loss reduction Planning/Operationalefficiency

Traditional(replacement/reinforcements)

Chapter 2 Chapter 3

Smart Network Management Chapter 2 Chapter 3

Feed-in control (flexibility ofdistributed energy resources)

Chapter 2 Chapter 3

Specific focus on Demand Response in second part of the report

Table 1 - Perimeter of the study for energy efficiency measures in electricity grids





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

1.3.2. Energy efficiency measures for gas gridsEnergy efficiency measures for gas grids are analysed in chapters 5 and 6. Selected measureshave been categorized as follows:

Measures for transmission gridsMeasures for compression stationsMeasures for distribution grids

The analysis has mainly focused on the potential for reduction of gas shrinkage in the gas system.

Table 2 reports the perimeter of the study for the energy efficiency measures in gas grids.

EE Measures for gas gridsPotential for reduction of gas

shrinkageCost-effectiveness of measures

Transmission grids Chapter 5 Chapter 6

Compression stations Chapter 5 Chapter 6

Distribution grids Chapter 5 Chapter 6

Table 2 - Perimeter of the study for energy efficiency measures in gas grids





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

2. POTENTIAL OF ENERGY EFFICIENCYMEASURES ON LOSS REDUCTION INELECTRICITY GRIDS

This chapter presents the qualitative and quantitative assessment of the energy efficiencypotential in electricity infrastructure with particular reference to losses.

Section 2.1 presents the key factors that are central to the understanding of how losses are createdin electricity networks. Based on this mapping, section 2.2 presents the key measures forincreasing energy efficiency, and key context indicators on how to assess their energy efficiencypotential. Figure 1 reports best estimations of the level of energy losses in electricity and gasgrids in a number of European Member States. The figure shows that the level of losses variessignificantly among Member States, hinting that potentials for energy efficiency improvement aredifferent and depends on local conditions (e.g. starting conditions of grids -e.g. voltage level;number of transformers- and potentials for energy efficiency improvement are different).

Figure 1 – Level of losses4 in electricity grids in selected EU Member States

4 Both technical and non-technical losses, based on the Ecofys analysis from EuroStat database





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

2.1. Understanding energy losses in electricitynetworks

Energy losses in electricity grids are created through a set of parallel mechanisms which affectthe different network components. In order to analyse the impacts and potential of energyefficiency measures, one should understand how losses are created through these mechanismsand how losses are mapped in electricity grids. Figure 2 presents the key factors on the mappingof losses (to allow the quantification of where losses occur in the system), and further presentsthe key mechanisms for losses creation (for a better understanding of how measures could lead tolosses reduction). In total there are 8 key dimensions to categorize energy losses in electricitynetworks, presented in the following sections.

Figure 2: Categorisation of losses in electricity networks: Mapping and key mechanisms [3], [2], [4]

2.1.1. Type: Technical/non-Technical and Variable/FixedElectrical losses can be divided into technical losses, which refer to energy transformed to heatand noise during the transmission and therefore physically lost, and non-technical losses, whichrefer to energy delivered and consumed, but for some reason not recorded as sales (such as theft).A general rule of thumb is that technical losses can rise up to 12%. When system losses exceedthis limit, they are most probably due to increased non-technical losses [5]. In general technicallosses can be split into [6]:





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Variable losses (also referred to as load or copper losses), occur due to the heating effect ofenergy passing through conductors in lines and cables and also in the copper in transformers.They vary in proportion to the square of the current and in proportion to the conductorresistance. In essence, measures to reduce variable losses can be understood under these twomain influencing factors and how they apply in the global system: they either aim to reducethe system power flows or to reduce the resistance of the transportation paths. In general,variable losses contribute roughly to two thirds to three quarters of the total power systemtechnical losses.Fixed losses (also referred to as non-load or iron losses), refer to energy needed for theenergisation of transformers5 or conductors. These losses are invariant with current anddepend mainly on the number of energised components. In this respect, measures to reducefixed losses mainly aim to reduce the number of energised components or to increase theirefficiency. In general they contribute to roughly one quarter to one third of the total networklosses.6

2.1.2. Voltage level: transmission and distribution networksElectrical losses are directly affected by the voltage level: increasing the voltage level leads to areduction of system currents and therefore of the respective energy losses. A consultation on thetreatment of losses by network operators was conducted by the European Regulators’ Group forElectricity and Gas7 for fourteen EU countries in 2008 and since, no update to the study has beencompleted8. Table 3 illustrates the percentage of losses per system level in selected EU countriesfrom [7]. The average losses in transmission networks (TNs) in EU vary between 1-2.6% whilefor distribution networks (DNs) between 2.3-13.4% [7]. This variation has mainly to do with thehistorical development and current state of the grids in each country with regard to age, designetc. In general, distribution networks present the highest losses and the highest energy efficiencypotential and therefore are the focus of this study.

5 This refers to the energy needed for the operation of the magnetic fields that enable the operation oftransformers.6 Another source of fixed losses are corona losses, which occur in high voltage lines. They vary withthe voltage level and the physical wire diameter and with weather conditions such as rain and fog andare generally a very small percentage of the overall system losses.7 Referred to today as the Council of European Energy Regulators

8 Eurostat provides data on for total system losses for each EU country in 2014 but the breakdown of thisdata does not exist





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Table 3: Percentage of losses in transmission and distribution networks in selected EU countries.

2.1.3. Components: Lines and TransformersPower lines and transformers are the primary components of the infrastructure in electricitynetworks. A key difference is that only variable losses occur in lines, while transformers areresponsible for both fixed and variable losses. Figure 3 presents how losses are distributedbetween these two key categories of components for the different voltage levels and different EUcountries. As can be seen, the majority of losses occur close to the customer, on low voltage (LV)networks followed by medium voltage (MV) grids. A high variation in energy losses intransformers can be observed, primarily due to the age of the assets.

Figure 3: Mapping of losses based on voltage level and components for different EU countries (FR: [8], [9]; GER: [10],[11], [12], [13], [14]; UK: [2], [5], [15]; EU: [16], GIS Data)





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

In Figure 3, the sum of the losses of the EHV and HV grids in each country falls in the generalrange of transmission losses for transformers and lines expressed in the top row of the figure. Forexample, in Germany, transmission losses sum to 37% of the total grid losses which falls in therange of 13% and 42%. The transmission losses are specifically high in this country compared toFrance and the UK, both at 25%. This is due to high international transits and internal transitsfrom north to south [17]. Similarly, the distribution of losses is not uniform across countries anddepends on the specific grid conditions. For example, in France and UK line losses are higher inthe MV grid compared to the LV grid, and in France, losses in LV transformers are higher than inLV lines, whereas in Germany losses are highest in LV lines. We discuss further the factors thataffect the losses in each component.

Transformers are the system components presenting both fixed and variable losses. Typically athird of the total system technical losses occur in transformers while fixed losses account forabout two thirds of total transformer losses. The operational efficiency of transformers dependson the loading. The peak efficiency of distribution transformers reaches 99.4% at approximately40-50% of rated capacity [18]. In the EU the average loading of distribution transformers inelectricity distribution companies is 18,9%, far below optimal, rendering their efficiency at98.38% [18]. Since transformers run round the clock for more than 40 years, a 0.1% marginalincrease in efficiency of a transformer population leads to significant loss reductions. [18]conducted a unique study in 2008 comparing the operating efficiency of newly purchased versusthe existing fleet of distribution transformers in electricity distribution companies across the EU.Figure 4 illustrates the results. On average, the efficiency of distribution transformers in the E.Uhas the potential to increase by 0.35%.

Figure 4 Efficiency of distribution sector transformer population and market [18]





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Lines

Power lines are the primary source of losses as indicated in Figure 3. Power line losses can beconsidered almost entirely as variable losses and are proportional to the conductor resistance andto the square of the current. Resistance of the line is equal to , where:

L is the length of the line, thus long lines have a higher R and higher lossesA is the cross sectional area , thus thicker lines lead to decreased line losses

the resistivity, which depends on the material of the line (copper or aluminium) and itstemperature, hence a heated line leads to higher losses

The current represents the power flow fluctuation and has a quadratic effect which is discussedbelow.

2.1.4. Marginal vs Average lossesThe currents in a system vary based on the system loading. The variable losses are the sum of thelosses for all operational instances, comprising of operational instances of high loading (marginallosses) and of normal loading (average losses). Average line losses are more often used whenestimating power losses, due to the fact that they are measured by meters, while measuringmarginal line losses requires advanced and expensive meters [19]. Since variable losses areproportional to the square of the current, marginal losses have a much higher contribution to thetotal sum. This is shown as an example in Figure 5 where the losses from 3 cases of systemloading are presented. In all cases the same energy is transported, but the loading profiles presenta different variability. Smoother loading profiles lead to significant losses reduction, up to 20%for a flat profile.

Figure 5 Impact of smoothing the system loading on the reduction of variable losses

Therefore, losses increase disproportionally when the system loading is higher, which is actuallywhen they cost more, since normally these are the cases when power prices are higher.Furthermore they create an increase in the system loading exactly at the times when there are lessavailable generation resources. In this respect, efficiency measures that help reduce the marginalloading of the system would have a higher contribution to the reduction of system losses [20].





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

2.1.5. Role of DG and grid managementIn the latest years, distribution networks experience increasing penetration levels of distributedgeneration (DG), in the form of small-to-medium size generating units connected in medium andlow voltage networks. A large share of these units consists of renewable generators, namely windturbines, CHP or PV panels. The general expectation has been that locating generation closer todemand could reduce losses since the distance over which electricity is transported is shortenedand the number of voltage transformation levels is reduced. [21] However, this is only true whenthe energy generated by DG and variable RES (VRES) units is consumed locally (synchronisedto load, local balancing). [22] In reality, such local balancing does not occur, either due to lack ofa proper operational market framework or due to the stochastic nature of the prime energy mover(in the case of VRES). This unbalanced operation can lead to increased network flows (and thuslosses), often translated into reverse flows from distribution networks to transmission systems.Figure 6 presents the expected effect of increasing shares of non-synchronised DG to distributionnetwork losses. In particular, losses are expected to decrease for low penetration levels, due to thefact that no significant reverse flows are expected but after a certain level they increase due to theincreased power flows they incur to the system. For example, [23] shows a study case in a LVgrid in Switzerland where grid losses reach a maximum reduction of 20% at 25% PV penetration.At a penetration level of 50%, they are equivalent to the case when there was no PV and furtherincrease if more PV panels are installed.

Figure 6: Impact of increasing DG penetration level and asynchronism of generation and demand to grid losses [12]





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

2.1.6. Reactive LossesCurrents in electricity systems comprise of the vector sum of two components, active(responsible for the transmission of active power) and reactive currents (responsible for theenergy alternately stored and released by inductors and/or capacitors). Reactive losses are due tothe circulation of reactive currents, which are increased when the power factor9 is deviating fromunity, and when voltage deviations are observed. Local voltage control actions (reactive powerinjection or absorption) are the most widely accepted means for supporting voltages and reducingreactive losses in distribution networks. In traditional distribution networks (comprising mainlyof loads), the voltage drops at the end of the line due the consumption of reactive power by theline when it is highly loaded. The placement of capacitors for local injection of reactive power tocompensate for those losses is the state of the art solution for voltage support. In [24] it isdemonstrated that the use of such device reduce reactive losses by up to 45% compared to thecase where reactive power is transmitted from an external grid to compensate for the voltagedeviation. In distribution networks comprising of loads and DG, situations of over-voltages areobserved if the DG output exceeds the load. In [25] a case study analyses the effect of DG on thevoltage deviations of a 10 kV test distribution feeder with five equally distributed load of 0.5MW each. Variable capacities of DG are connected to node 4.

.

Figure 7: Voltage profile of the test feeder for different generating scenarios [25]

Figure 7 demonstrates that all possible load/DG scenarios need to be studied to assess thedecrease of reactive losses in the system. In this specific case a 3.2 MW DG capacitysynchronised with the load would lead to almost no voltage deviation, hence minimizingreactive losses. However, in reality, utilities do not have control on the allocation or the size ofDG that will be installed in their distribution grid. Therefore different voltage support actionswith different impacts are used. We discuss them in the next section on voltage optimisation.

9 Power factor is a ratio between the real power and apparent power flowing through a conductor.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

2.1.7. ImbalancesElectricity networks comprise three phase systems which (especially in distribution grids) servesingle or dual phase loads. Imbalances in the loading between the three phases lead unavoidablyto increased currents in at least one phase, which increase marginal losses. Typically,imbalances occur on all parts of the low voltage network due to customers who use one or twophases with different load consumptions. In order to rebalance the network it is necessary toreconfigure phase loading to ensure that this captures the development of loads. This isnormally done based on maximum customer load and cannot capture the fluctuations of loadprofiles over time. In this respect, some imbalance is unavoidable [3].

2.1.8. Life cycle assessment: impact of losses on investmentdecisions

Losses are one key contributor to operational expenditures in power networks. Typically, optimalinvestment decisions translate in finding the components configuration that minimise the sum ofinitial investments and lifetime costs (Lifecycle costing – LCC). Due to the long lifetime of thenetwork assets, adopting a LCC approach that includes losses can steer decision making whenincluded in investment analysis. In the case of transformers, LCC shows that the purchase ofenergy efficient equipment could be optimal decision regardless of the higher initial capital costs.[18] Similarly, in the case of lines, and due to the characteristics of variable losses, a LCCoptimal decision leads to increased line capacities for reduction of losses. In distribution networkdesign, the inclusion of the cost of losses leads to much lower utilisation rates of the assets (thushigher installed capacities) in contrast to the practice of using the assets as much as possible toincrease capital efficiency. [26] Consequently, there is a trade-off between the cost of financingsurplus capacity and the cost of losses.

2.2. Energy efficiency measures in electricitynetworks: categorisation and potential

Figure 8 presents the key energy efficiency measures and their categorisation in two maincategories, a) the ‘traditional’ measures of component replacement, and b) measures that achievea reduction of losses by an improved management of the power system, either by controlling thenetwork infeeds (local balancing) or by grid management actions (such as networkreconfiguration, voltage control, or elimination of imbalances).





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Figure 8: Categorisation of key energy efficiency measures in electricity networks

The loss reduction potential of each measure can be in principle analysed by the factors presentedin the previous section. Table 4 presents a mapping of the energy efficiency potential of eachmeasure with respect to the categorisation of losses presented in section 2.1 (table refers todistribution networks, as they present the highest share of losses in the electricity system). As canbe seen, each measure affects a set of factors and therefore the potential of each measure willdepend on how losses are mapped in the specific grid. The table presents two dimensions whichtogether map the potential, firstly the applicability of each measure on different voltage levelsand secondly the effectiveness for loss reduction when the measure is applied. The combinationof the two dimensions allows the estimation of the potential, e.g. measures of high effectivenessbut of poor applicability would rank low on potential.

Concerning variable losses, a key differentiator is the point where the measure is applied. Systemloading at each voltage level corresponds to the sum of the loadings from the networks inunderlying voltage levels. Therefore, measures generally affect losses from the network levelwhere applied and upwards, making energy efficiency measures that apply to lower voltagelevels the most efficient. A typical example on this is demand response and demand flatteningactions: applying such measure to LV networks will lead to a flattening of profiles to the highersystem levels also, inducing a chain of loss reduction towards the upper levels. This also meansthat losses at the LV grid will only be influenced by measures at this voltage level, such ascustomer side demand response.

E-EE gridMeasures

Componentreplacement

EETransformers

Increase Linecapacity

IncreaseVoltage level

Systemmanagement

Feed-inControl

DG

VRES

Controllable(micro-chp)

DemandResponse

EnergyStorage

Grid Mgmt

NetworkMgmt

Trafoswitching

Networkreconfiguration

CVRVoltageControl

On Load TapChanging Trafo

Reactive powercompensation(e.g. DFACTS)

DER Voltagecontrol

Balance 3-phline loading





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Table 4: Mapping of the energy efficiency potential of each measure (focus on loss reduction in electrical distribution

grids)





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

2.2.1. Energy Efficient transformersThe key mechanism is the replacement of transformers with energy efficient ones10, which inmost cases are larger and with higher capital cost. As transformers are responsible for bothvariable and fixed losses in the system and their replacement is easier than changing cables orlines, this option presents a good loss reduction potential, and affects almost all factors of losses,as shown in Table 4. As shown in Figure 3 most losses in transformers occur in distribution grids,therefore the potential of the measure in lower voltage levels is higher.

The asset replacement strategy depends on the state of the population, characterized by the age,size and type of transformers. For an optimal investment decision the cost of losses should betaken into account in a full LCC analysis. Such LCC estimation shows that efficient transformershave a payback period of 2-13 years compared to less-efficient models when the cost of the lostenergy is included to the cost of purchase [5].

The lifetime of transformers is typically around 40 years. In many cases however, assets are stilloperational beyond this point, but with significantly lower efficiency levels. As shown in [18],20% of the oldest distribution transformers (installed before 1974) are responsible for 35% of thefixed losses and 30% of the variable losses in the EU, approximately a total of 38 TWh of lossesper year. Replacing them with energy efficient transformers can help reducing these losses by80% [5], leading to an annual saving of 30TWh, equivalent to the annual electricity consumptionof Denmark [27]. Furthermore, [28] quantified the impact of implementing the MEPS11 of theEcodesign Directive till 2025 and found that energy savings of 84 TWh could be achievedcompared to a business as usual scenario.

Western Power Distribution (DSO in UK) conducted a cost and benefit analysis including thevalues of losses to know whether their transformers needs to be replaced. [3] The outcome of thestudy shows that it is only beneficial to oversize their smallest units (pole or ground mounted)and to replace earlier than planned all pre-1958 ground mounted distribution transformers. Theexpected savings per annum reaches 2.76 GWh, 97% coming from the replacement of the oldtransformers.

1010 Energy efficient transformers have reduced variable and fixed losses. To reduce variable losses, the resistanceof the wires needs to be decreased, which can be done by increasing the cross sectional area or by using materialswith a lower resistance. To reduce fixed losses, one way is to use materials with better magnetic properties.

11 This option of the Mandatory Energy Performance Standards requires newly installed transformers to follow TopEfficient level of EN 50464-1 for Tier 1 (2015) and Least Life Cycle Cost for Tier 2 (2020).





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

2.2.2. Increasing voltage levelIncreasing voltage level reduces the currents required to distribute the same amount of electricity,increases current capacity of the grid and reduces substantially voltage drops and line losses. Itpresents further indirect benefits such as reduction of probability of short circuits, increase ofquality of supply and it increases the DG hosting capacity. The measure typically applies to MVnetworks, where multiple voltage levels are possible (e.g. 10kV and 20kV). In cases where theinsulation coordination allows both voltage levels, it can be applied by changing only thetransformers. In most cases however this is not the case and a full renewal of all networkcomponents is necessary, which implies high costs. The measure is particularly interesting toaging MV networks as it comes at little extra cost to up-rate the voltage [5].

As shown in Table 4, the applicability of the measure is mainly on MV grids and is deemedrather low since only a limited number of networks will fulfil the respective conditions forapplying this measure. On the other hand, the potential of the method on reducing variable lossesis high, in the cases this solution is chosen.

The measure was applied in large scale in Ireland for upgrading the MV rural networks from10kV to 20kV [29]. These networks were designed in the 50’s for lower typical householdconsumptions and load densities, and were severely loaded and experiencing very poor voltageregulation. The costs of 20kV conversion were little more than those of rebuilding in 10kV, yetthe voltage drop were halved, thermal capacity doubled and losses reduced by 75%.

2.2.3. Increasing line capacityIncreasing cross-sectional area of cables leads to reduction of variable losses. Typically, doublingcable rating leads to a reduction of losses by a factor of four. However, once a cable is installed,it becomes expensive to make alterations to the cable, since civil works cost element of replacinga cable outweighs any loss reduction benefits. The key opportunity to reduce losses thereforeexists at the time that the cable is initially installed or replaced [3]. An alternative approach is tomake use of high-temperature superconductors (HTS) which present no resistance when they arecooled down at -180°C and can carry five times the current of a conventional cable system withthe same outer dimensions. The losses from this system are due to the energy needed to operatethe cooling mechanism. [30]

As shown in Table 4, the applicability of cable up-rating is deemed medium, and the measureshould be applied in cases on new line installations or replacements. Applicability of HTS isdeemed rather low, mainly for cases of high currents that need to be transported over relativelyshort distances. Both measures have a high potential to reduce variable losses.

A CBA conducted by Western Power Distribution showed that is beneficial to uprate all theexisting MV underground cables from 6.6kV to 11 kV. The planning is to replace 700 km ofcables per year and savings are estimated at 3GWh per year. The study also showed that it is notbeneficial to uprate the 11kV network as these higher voltage network cables support adjacentnetworks in times of fault. [3]





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

A 1 km 10 kV HTS cable was installed in 2014 to replace several 110kV underground cablesconnecting two 10 kV substations in Essen Germany. The cost of the energy needed to cool thecable down to eliminate its resistance over its lifecycle is found to be 15% lower than the cost ofcompensating losses in conventional 110 kV cables. HTS are mentioned as the best technical andeconomically viable solution to avoid the necessary extension of the 110 kV grid in urban areas.[30]

2.2.4. Feed-in controlThis measure refers to applying control actions for local balancing of demand and supply indistribution grids, which leads to reduction of variable losses due to two key factors, thereduction of energy transportation distance (energy is consumed locally) and the reduction ofmarginal losses (by a flattening of flow profiles). This local balancing can be performed by threeoptions for control actions, of supply (DG), of demand (demand response actions) or of energystorage (can be considered as supply and demand)12. Under this measure the control of activepower is considered. (Control of reactive power of DGs is analysed under the VoltageOptimisation section).

The potential of this measure depends on three key factors, namely the ability for spatial andtemporal matching and the underlying technical limitations.

Spatial matching: it relates to the system architecture, namely DG penetration levels andrelative position of dispersed supply and demand; a general rule of thumb is that the measureis more effective when supply and demand are close and when it is of similar rating.Temporal matching: it refers to the operational framework that should coordinate localbalancing; such a framework is most of the times non-existent. Network losses are generallyreduced for low DG penetration levels, mainly due to the fact that independent of thecoordination scheme generation does not exceed demand. For higher penetration levelshowever, this lack of coordination translates into reverse power flows and to a gradualincrease of system losses (see Figure 6).Technical limitations: refer to the specific flexibility characteristics of the technology whichreveal the degree of local coordination that can be achieved. Such limitations include DGcontrollability, demand elasticity and potential for peak shifting/clipping and the size andflexibility of the storage devices.

12 New demand such as Electric Vehicles and Heat pumps are included in these categories, but with specificflexibility characteristics and limitations due to their primary operation.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

As can be seen in Table 4, although VRES are generally present in MV and LV grids, theireffectiveness to perform local balancing actions is deemed rather low, mainly due to the fact thatsuch actions translate into reduction/curtailment of active power production. Controllable DGpresent a medium effectiveness, since although they are technically capable, they cannot performsuch actions due to lack of enabling operational framework. Demand response presents a highapplicability and high effectiveness on reduction of technical losses, which however depends onthe technical potential of demand (peak shifting/clipping potential). This option presents also amedium potential for reduction of non-technical losses, since they are most of the time combinedto smart metering infrastructure that enables better monitoring of the creation of non-technicallosses. Finally, energy storage presents a high effectiveness on local balancing but a rather lowapplicability due to the low implementation of storage in distribution grids.

This impact is discussed in [31], where the potential of demand response from residentialcustomers in reducing marginal losses of a Swedish distribution network is investigated. Resultsshow that this reduction varies from 4%, (if 10% of the peak load is shifted) to a maximum of18%, if the load profile is flattened. The Model City Mannheim project proved effectively afterhaving developed, installed and tested smart control systems in 1000 households that the priceelasticity of customers to variable tariffs was 10% [32], which means that 10% of the demandcould be shifted when the price of electricity increased by 10%. In other terms, the 4% reductionof marginal losses mentioned in [31] is feasible. This translates into annual savings of 350 MWh,equivalent to the electricity consumption of 55 households in Sweden [33].

The IMPROGRESS project analysed the impact of distributed generation on three distributionnetwork in Germany, Spain and Netherlands [34]. The different characteristics of each area weretaken account: medium and low voltage networks, urban and semi-urban areas, number ofcustomer, shares of wind and PV and CHP and a forecast for load demand and DG penetrationtill 2020 was applied. The areas studied are Aranjuez in Spain (industrial, wind and CHP),Mannheim in Germany (residential, lots of PV roof and micro-CHP) and Kop van Noord (lots ofwind and CHP) in the Netherlands. The feed-in control action consisted on one hand of curtailingVRES or storing the excess power in an electric heat boiler and on the other shifting the demandfrom periods with low DG penetration to those where most CHP units were running. The studyshowed that if feed-in control is practiced instead of expanding the LV grid to cope with theforecast growth of DG until 2020, 97 €/kWDG in Netherlands can be saved, 77 €/kWDG inGermany and 73 €/kWDG in Spain. The potential savings calculation includes the Net PresentValue of the electrical losses that would have been caused by the increasing penetration levels ofDG in each distribution grids.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

In [35], energy storage is used to shift load from peak to off-peak period and the impact onreducing transmission and distribution losses is analysed. The authors developed a set ofnormalized charts to quantify the saved T&D losses as a function of the size of the energy storagedevice, its efficiency and the load shifting potential (i.e. the ratio of the off-peak to peak load).Based on an existing 1 MW battery storage system installed for shifting a peak load from 20 MWto 6 MW in a 12 kV distribution feeder in Virginia USA, the authors estimate that 1 to 3% ofT&D losses could be saved which translates in 180 to 330 MWh year of energy saved includingthe storage losses. The authors demonstrate that it is more beneficial to distribute small sizestorages at different locations rather than using a big size at a single site to maximize thereduction of T&D losses.

Within [14] the German Federal Ministry of Economics (BMWi) examines various measures toimprove the integration of VRES into the distribution grids. One of the key findings is that thecurtailment of 1% of the annual feed-in of VRES reduces the necessary grid extensions by 30%.A curtailment of 3% of the energy annually produced by VRES would be sufficient to avoid 40%of the required grid extensions. The effectiveness of curtailment measures rapidly decreases ifmore than 3% of the annual feed-in of VRES is curtailed.

2.2.5. Transformers switchingFixed losses can be reduced without replacing equipment by reducing the number of energisedtransformers in the system. This can be achieved by eliminating transformation steps or byswitching off transformers. Elimination of transformation steps could be achieved by directcoupling of higher voltage levels to lower ones without the use of intermediate transformers (useof a single transformation step). Alternatively, switching off transformers could be possible inperiods of low demand for configurations where multiple transformers are required in asubstation to meet peak load or for redundancy. This can be achieved via an ‘Auto Stop Start’System that allows to periodically de-energise redundant transformers in substations.

As shown in Table 4, the potential of the measure is high on reduction of fixed losses. In additionto reducing the fixed losses, this measure also reduces the variable (and marginal) losses byimproving of the loading of the energised transformers. The effectiveness of the measure indifferent voltage levels depends on the redundancy configurations used. In this respect, themeasure is especially effective in MV or LV networks where redundant transformers are often inuse.

Case studies show a 75% loss reduction in MV networks, equivalent to the loss reductionachieved by upgrading the 11kV network to 20 kV, however at about 1% of the upgrade cost.The project is in execution phase led by the Scottish and Southern Electric Power Distributionand will end in 2018 to validate the scale of loss reductions (640 MWh/year) and to confirm thatrepeated switching did not impact the asset lives and the supply security. [36].





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

2.2.6. Network reconfigurationThe measure involves changing network topology at the distribution level dynamically by usingautomated switching actions. Typical implementations are Smart Feeder Switching (SFS) andAdvanced Distribution Management Systems (ADMS). The configuration of the network has aneffect on losses in terms of the distance the electricity is transported. MV networks are typicallyconfigured as open loops and are controlled in order to be able to isolate faults and restore power.As the demand changes spatially and in time, it is often the case that the configuration of anetwork is not the optimal one for the specific demand situation. There might therefore be somescope for reducing losses by reconfiguring the network, to provide shorter, more direct paths towhere highest demand is situated. SFS and ADMS are typical implementation for performingautomated switching actions on distribution feeder systems enabling a reduction of distributionsystem losses [37], [15], [38].

As shown in Table 4, the measure involves reduction of variable losses mainly in MV networks,where such configurations are employed and its potential is defined as medium. The potentialdepends on the reconfiguration frequency and the spatial and temporal variations of powerdemand. A limitation may be if the switch-disconnector equipment is designed for a limitednumber of operations per lifetime (e.g. 1.000) [15], which limits the applicability of controlactions. Typically, for systems with high penetration levels of intermittent DG (especially windgenerators), which experience higher and random spatial and temporal demand variations, higherswitching frequencies are necessary.

Results from field pilot projects reported in [15] show that network reconfiguration may lead to atheoretical reduction of losses of 40% for an hourly reconfiguration (which however cannot beapplied due to limitations of the switching equipment), and up to 20%, 10%, 4% if thereconfiguration is made on a weekly, seasonal or yearly basis respectively.

2.2.7. Conservation voltage reductionConservation Voltage Reduction (CVR) is a reduction of active power delivered to demandresulting from a reduction of feeder voltage within the operational ranges. CVR schemestypically contain two fundamental components: a) reactive power compensation, achievedthrough the operation of shunt capacitors in order to maintain the power factor at the substationtransformer within a prescribed band, and b) voltage optimization, achieved through theoperation of substation voltage regulators in order to regulate the voltage at specific end of linepoints within a prescribed range. In this way the peak load and the annual energy consumptionare reduced. [9] When applied on LV networks, the measure can be applied by changing settingsat the primary substation [3].





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

The applicability of the method depends on the type of load in the system. The method typicallyapplies well to ‘voltage dependent’ loads such as filament lamps and resistive heating13.However, increasingly more load is ‘voltage independent’, as it is fed via a switched mode powersupply, which effectively changes its impedance based on voltage (such as HF fluorescent, LED,PCs VFD fed motors). Therefore lowering voltage may not, in all circumstances, lead to thedemand savings as desired and could actually increase losses. [2] Therefore, the measure isconsidered of low applicability and medium effectiveness, as presented in Table 4.

According to [39], utilities applying CVR may expect to see 1% reductions in electricityconsumption for every 1% reduction in voltage. Detailed analysis of CVR schemes in US showthat they can lead to peak demand reduction of 1 to 2.5% [39] and annual energy reduction of 4%[9] Western Power Distribution is planning a voltage reduction trial in the South Wales area ontheir MV networks from 11.4kV to 11.3kV and anticipates reduction of 95MWh which willreduce the customer bills by a calculated €13M each year and carbon emissions by 41 000 tonnesa year. [40]

2.2.8. Voltage optimisationVoltage optimisation refers to reduction of reactive losses by the application of local voltagecontrol actions (reactive power injection or absorption) aiming at local correction of the powerfactor. The placement of capacitors for local injection of reactive power is the state of the artsolution for voltage support for passive networks (networks including mainly loads). Indistribution networks comprising of loads and DGs, the fluctuating generation profiles lead tosituation of over-voltages (when local generation exceeds demand) or under-voltages (insituations of high demand and no DG production). Voltage support actions should therefore beperformed dynamically and react to the conditions in each operational instance in order tooptimise the feeder voltage profiles.

This is made possible through recent improvements in sensors, communications, controlalgorithms, and information processing technologies that allow monitoring voltage levelsthroughout the distribution system. This information is sent to devices that can adjust voltageregulating equipment and capacitor banks on distribution feeders in near-real time enabling quickadjustments in response to constantly changing load and voltage conditions. Options to performsuch actions are installing smart reactive power compensation devices (e.g. D-FACTS, see [41]),using on-load tap changing (OLTC) voltage control actions at the distribution transformers [42],or by using the capabilities of converters for power factor correction [12] in the case of converter-connected DG14.

13 As the resistance of these devices is largely fixed, applying a lower voltage reduces the currentdrawn, less power is transferred and hence overall load is reduced.14 Typically converters can adjust the power factor at the point of connection, by changing the injectionor consumption of reactive power.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

As shown in Table 4, there are difference between the applicability and effectiveness of theoptions. Smart reactive power compensation devices are applied mainly on MV grids but theirapplicability is considered medium mainly due to cost limitations. Their effectiveness is high,since they can be applied to specific grid points. The applicability of OLTC is consideredmedium and their effectiveness is also deemed medium, since their operation can be limited incases when multiple feeders are connected to the transformer (voltage control actions fromOLTC affect all feeders at the same time and this may lead to conflicts). [25] DER voltagecontrol is considered as a highly effective option with medium applicability mainly due to thefact that DER should be converter connected and because the effectiveness of controlling voltageprofiles by controlling reactive power from DG is best applied to feeders with low resistance toreactance (R/X) ratio [23]. As shown in [23] reactive power control from PV panels in a LV grid(<1kV) in Switzerland would not be effective if the line diameters connecting the panels to thefeeder were not larger than 50 mm2, i.e. the R/X ratio lower than 5.

In [39], the results from 26 voltage optimisation programs are reported. The voltage optimisationanalysis focused on CVR and on reduction of energy losses along feeders, by automatedswitching of capacitor banks. Results reported that for the 31 feeders under investigation half ofwitnessed line loss reductions in the range of 0% to 5%, and 5 feeders experienced lossreductions greater than 5%. Feeders with the worst baseline power factors (i.e., those with thehighest amount of inductive loads) showed the greatest reductions in line losses.Overcompensation was observed in some feeders, which resulted in line loss increases.

In [12] a study case is presented, investigating the role of reactive power compensation comingfrom DG on decreasing grid losses in the North of Germany. The analysis involves two MV gridareas of similar size in load and number of connected customers but with different DGpenetration levels with respect to load demand (Area 1 with 45% and Area 2 with 146%). TheDG mix analysed consisted of 90% wind power plants and 10% biomass plants. The resultsshows that grid losses would decrease by 15% in Area 1 if DG plants were adapting their powerfactor between a ±0,05 window but would increase by 3% in Area 2. The results confirm that theeffectiveness of reactive power provision from DG plants is dependent on the grid topology andDG penetration level.

In [42] a study case is presented, investigating the role of dynamically controlled OLTC ondecreasing grid losses. Six additional wind generation sites of 14 MW are connected to a genericMV radial distribution system in the UK which needs to supply an annual energy demand of 215GWh and a peak load of 38MW. If the wind generation sites are operating at a fixed unit powerfactor, the OLTC helps reduce grid losses annually from 2,8% to 2.3%, saving 1 GWh per year,equivalent to the average annual consumption of 240 houses in the UK [43]. [14] states that a fullpenetration of OLTC in Germany would reduce the average yearly grid expansion costs by 10%.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

2.2.9. Balancing loading of linesAs mentioned, imbalances are a common phenomenon in LV networks where single- or double-phase customers are connected to the three-phase system. These imbalances lead to having onephase carrying higher currents, giving rise to variable losses. The measure is to statically ordynamically transfer loads from one feeder to another to balance the total load across multiplefeeders and transformers. As the demand changes in time, optimally balancing should beperformed in short time periods and reacting to demand changes. [2]

The applicability of the measure is considered high and is limited to LV networks, as presented inTable 4. Its effectiveness is considered medium, depending on the frequency of balancing thatcan be achieved.

As discussed in [3], imbalances in the LV network can lead to neutral currents of around 35% ofthe phase current. The annual potential savings from imbalance correction in this respect are inthe order of €70 to €110 per kilometre of LV network. An analysis of the effect of load balancingon the reduction of power losses in 12 feeders of a LV distribution power network in Russia ispresented in [44]. The results show that the reduction of technical power losses after loadbalancing varies from 0% to 16% depending on each feeder configuration. A total of 13 MWhcould be saved per year equivalent on average to the annual consumption of 100 customers. It isimportant to note that this specific regional power network already experienced high distributionlosses of 18% in 2007. Reference [45] suggests the use of three-phase balancing PV units that areable to inject more power in one phase than in the other phases compared to single-phase PVunits that inject only in one phase. The case study of typical Belgian LV network proves that thethree phase balancing PV units could help decrease grid losses by 30% saving 7 GWh per year.

2.3. Regulatory regimes to promote loss reductionin electricity grids

2.3.1. Regulatory incentives: basic optionsVarious regulatory incentives exist to motivate network operators in investing in energyefficiency (particularly loss reduction) in their networks. Some key options are reported below.

2.3.1.1. TECHNICAL STANDARDS FOR CONDUCTORS AND TRANSFORMERS

A very effective measure that leads to an overall improvement in efficiency is setting mandatorystandards on a European level. In such cases, the national regulatory design should allow a higherCAPEX allowance to grid operators as savings in OPEX will over-compensate them in the longrun. Based on the EU Ecodesign Regulation, new power transformers installed after July 2015need to meet a minimum energy efficiency standard.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

2.3.1.2. FINANCIAL INCENTIVES

In general financial incentives to increase efficiency can be of various forms, e.g. grants, loans ortax breaks. They can also be set via lower interest rates for energy-efficient investments. Tofurther motivate and simplify investments, specific energy-efficiency budgets can be set up.Given that the transmission and distribution networks are regulated industries, the mostappropriate way to deliver financial incentives would seem to be via the existing regulations.

Another option regulators can use is setting individual targets for network losses for gridoperators which can be a mechanism that financially penalises or rewards deviations from thetarget. This can be a strong incentive for grid operators to improve in operational performancewhile leaving them free space to decide on the most effective measure.

2.3.1.3. OBLIGATION OR CERTIFICATE SCHEMES

With obligations or certificates for energy efficiency and energy savings, grid operators are givenspecific shares of losses (and savings) that they have to meet, just like the option of financialincentives. In many schemes, there is also the option to trade certificates to meet the obligation.In case of not conforming to the obligation, penalties (financial or not) can be set to promoteactions to meet the obligation which is a key difference to financial incentives.

2.3.1.4. VOLUNTARY AGREEMENTS

Regulators can start by setting voluntary agreements with grid operators which would comprisenon-binding guidelines e.g. for a maximum share of grid losses in power transmission anddistribution. It can acquire the market participants’ attention on the issue and provide alignmentson grid operation.

2.3.1.5. INFORMATION CAMPAIGNS TO OVERCOME LACK OF MOTIVATION

There are two potential areas where lack of information might influence: with the regulators notunderstanding completely the case for energy efficiency and not setting the proper regulatoryframework and with grid operators not having all the information for a business decision. Aninformation campaign about best practices set up by the regulator can overcome such barriers.

2.3.2. State of play of regulatory mechanisms targeting lossreduction

The following key insights regarding regulatory mechanisms to support grid energy efficiencymeasures (particularly regarding loss reduction) have been derived from the survey withEuropean system operators carried out within this study (see ANNEX II):

The incentive to pursue specific investments for loss reduction is strictly related to thepresence of regulatory incentives. A wide variety of regulatory frameworks are present inEuropean Member States, which in some cases do not explicitly support energy efficiencymeasures.Energy efficiency (and particularly loss reduction) is the main driver of investments whenthere is the need to comply with regulatory requirements and quality standards (e.g.EcoDesign directive). In general, loss reduction is considered as one of the variables of alarger investment optimization plan.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Energy Efficiency measures which require the exploitation of DER flexibility have stilllimited diffusion in many Member States, often just at pilot level. Their implementationwould require the definition of new regulatory mechanisms over how DSOs can procure andactivate DER flexibility.

Table 5 presents a synthesis of the different mechanisms in force among the respondents.

Bonus/malus following the levelof losses

DSOs from the Slovak Republic, Sweden (for 2016), theUK and Portugal report that they have specific incentivesthat directly target loss reduction, in form of bonuses orpenalties depending on the perfomances.

Cost reduction (Revenue capregulation)

DSOs from Bulgaria and The Netherlands report that theyare incentivised to perform losses reductions since it is partof their costs, and since the regulation targets a generalreduction of the cost. Loss reduction is therefore not a firstdriver if other cost reduction opportunities are found.

Maximum loss threshold DSOs from Bulgaria, Czech Republic, France andGermany report that they have an obligation to respect agiven loss threshold, overpassing them leading to audit ornon-recognition of the costs. However, this gives noincentives to go further if their current loss level is under thelimit.

Benchmark among other DSOs DSOs from Spain report that loss reduction is addressed by abenchmarking of all Spanish DSOs on their performance toreduce losses.

Table 5 - Regulatory incentives for energy efficiency measures highlighted by survey respondents

More generally, it is clear that proper regulatory incentives are necessary to stimulate lossreduction investments. Absence of such regulation often brings the opposite effect, and acts asdisincentive for network operators. There are two main elements in the tariff regulations whichmay act as an incentive or disincentive to invest in energy efficiency measures of the grid:

1) the way the tariff is calculated, which determines the amount that can be charged to theconsumer; and

2) Whether there is an incentive regulation for minimising operational costs.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Tariffs are usually calculated on the basis of all relevant operational costs of transmission anddistribution of electricity. In this respect, grid operators can transfer the costs for purchasingenergy to make up for losses directly to the consumers. A clear disincentive to invest in lossreduction measures is typically the case when the regulatory framework does not introduce a capto the amount of losses that can be transferred to the customer, combined to financial penaltiesfor trespassing the limit. In such cases, infrastructure investment decisions are based on theminimisation of capital costs only, without including of the operational life cycle costs. Hencelow cost, and subsequently non-energy efficient equipment is chosen over advanced efficientequipment.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

3. POTENTIAL OF MEASURES ONPLANNING/OPERATIONAL EFFICIENCY INELECTRICITY GRIDS

In this chapter we analyse the impact of measures proposed in chapter 2 on the broader aspects ofefficiency in infrastructure design and operation.

In particular we consider planning/operational efficiency in terms of:

Optimization of investment planning - Traditionally, new capacity investments (gridreinforcement or expansion) have been directly related to the increase of demand. With thegrowing penetration of distributed generation, increasing the hosting capacity of DG is alsobecoming an important driver of grid planning decisions.

In this respect, the new options offered by active grid management hold a significant potential inoptimizing the long term planning process. In particular, the use of flexibility of distributedenergy resources is currently being explored as an alternative to traditional grid reinforcements15

16. A specific focus on the use of demand flexibility to optimize grid planning decisions will becarried out in the second part of this study.

Investments in assets replacements could also be optimized. Measures presented in chapter 2 canin fact contribute to decrease stressful conditions in the grid that could lead to faults ofcomponents and a reduction of their useful life.

Optimization of grid operation - The reduction of losses is one important aspect of theoptimization of grid operation. It has been extensively analysed in chapter 2 and will not be takeninto account in this chapter. Other value streams for optimization of grid operation (OPEXreduction) include:

Reduction of outage times (e.g. via implementation of network automation)Improvement of quality of supply and reduction of customer complaintsReduction of maintenance costs (e.g. by minimizing stressful operating conditions for gridassets)…

In the following sections, we analyze in detail the factors affecting the cost-effectiveness of eachof the proposed measures considering key aspects of planning/operational efficiency.

Implementation examples and best practices collected in the stakeholder survey (ANNEX III)and in related literature are used for illustration.

15 Regulatory Recommendations for the Deployment of Flexibility, EC Smart Grid Task Force Expert Group3, January 2015

16 Study on the effective integration of Distributed Energy Resources for providing flexibility to theelectricity system, Report to The European Commission, April 2015https://ec.europa.eu/energy/sites/ener/files/documents/5469759000%20Effective%20integration%20of%20DER%20Final%20ver%202_6%20April%202015.pdf





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

3.1. Impact of selected energy efficiency measureson operational efficiency

3.1.1. High efficiency transformers

Planning/Operational efficiency Assets to deploy

Investment planning Grid operation Hardware Software Communication

Technologystandardization

Lighter weight

Technologystandardisation

Reduced fault rate

High efficiencyTransformers

Possible benefits on planning/operational efficiency

Benefits on investment planning

Some benefits could be achieved in term of investmentoptimization. Indeed, limiting the number oftechnologies present on the grid may lead to strongerpurchasing positions in the market, depending of thelocal market organisation, and may also lead to O&Mcost reduction (less various types of operation).

In case of pole-mounted transformers, lightertransformers might also reduce the need forinvestments in pole replacements.

Benefits on grid operations

Aging transformers might present a higher fault rate than up-to-date transformers. Replacingmost problematic assets will save O&M cost. According to the economic assessment carried outby the Asia-Pacific Economic Cooperation [46], replacing transformers in order to comply withcurrent performance standards leads to long-term savings, including additional operation costsaving (in each of the 20 studied countries, positive net present values were observed in their2016-2030 plans).

Enexis (The Netherlands) is replacing olderMV/LV transformers with existing but morerecent transformers (left over from otherregular replacements, such as capacityincrease). 150 transformers have been alreadyreplaced out of their 53000 MV/LVtransformers, at an average cost of 5300€/transformers. 0.1% of loss reduction isestimated to results from those replacements,which shows a significant potential since only0.3% of the assets have been replaced.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Assets to deploy - Investments costs

The costs of a transformer have a wide range of variability depending on the power rating. In[47],a study from the Indian government, it was showed that cost differences are high betweentheir “1 star” and the “5 star” transformers of small power rating (50% of difference for a 25 kVAtransformer) but those difference get smaller at higher power (27% for a 200 kVA transformers,for a cost estimated at 3390 €17 for 5 star transformer). However, according to a study carried outin the Asia Pacific Economic Copperaiton [46], it appears that, installed capacity being equal, thecost-effectiveness of switching to higher standard is higher when replacing small transformers.Cost-effectiveness can be maximized if the measure is implemented within a planned assetreplacement campaign. For instance, when replacing transformers at the end of their useful life,CEZ (DSO in Czech Republic), has found it economically convenient to opt for upper classtransformers despite a 20% higher price, thanks to energy efficiency improvements.

3.1.2. Expanding line capacity



Standardization Reduced SAIDI

Improve voltage quality

Cables None None


Benefits on investment planningSome optimization of investment planning can be achieved via the limitation of the number ofthe cables sizes in use. In fact this can lead to standardization of equipment on the network, andthus to an average purchase price reduction.

Benefits on grid operationsOversizing the conductors decreases the impedance of the line, resulting in lower voltage drop.Therefore, a positive impact on voltage quality of the grid should be obtained.

17 Change rate $ to €: 1$ = 0.894 € (11/05/2015)





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

A limitation of the current flowing through the line could reduce the probability of default (andthe aging of the line). Moreover, by standardizing the line type, higher standardization of fieldworks could be achieved, reducing the overall O&M costs.


It is difficult to give an accurate cost for such a measure,since it strictly depends on the implementationconditions (overhead lines or underground cables, size,voltage level, type of conductors, etc.). However, thecost of oversizing line sections can be relatively small inthe case of installation works (particularly in case ofcables) which are already planned. Western PowerDistribution (see [3]) decided to size up their currentstandard LV cables to discontinue their smallest-sizeMV cables. The study shows that sizing up 95mm²cables to 185mm² adds a cost of 14€/m18, whileexcavation costs range between 70€/m and 140€/m.

3.1.3. Increasing voltage level



Deferred capacityinvestment

Life time extension

Technologystandardisation

Increase DG hostingcapacity

Improvement of voltagequality

Reduction of DGcurtailment

Improvement of continuityof supply

Fitted lines cables

HV/MVsubstations:

TransformersMV cells(protections,switches…)Rails

MV/LVsubstations:

MV cells(protections,switches…)

Transformers

None None

18 Change rate £ to €: 1£=1.3793 € (11/05/2015)

Východoslovenská distribu ná (DSOfrom Slovak republic) has increased themaximal conductor section of their LVnetwork from 70mm² to 150mm². While thefirst driver of this measure was to reachhigher voltage quality, they expect anominal losses reduction of up to 20% atpeak load, and 5% on average (dependingon the intensity of operation). The increasein investment was minimized and managedby process changes in the procurementprocedure. The asset costs represent up to10% of the project costs. However, a fullroll-out was deemed to be too costly, due tocosts of other grid elements such as poles.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval



Since increasing the voltage level reduces the current flowing through the conductors to meet thesame power, networks reinforcement can be deferred. This is particularly true if only a limitednumber of assets need to be replaced. In their plan (see in [48]) to convert their rural 10kVnetwork to 20kV, the Irish DSO operator ESB Networks aims to double their thermal capacitywhile reducing peak losses. Future investments to expand the grid could also be reduced thanksto an optimization of the number of substations.

As shown in [49], power line aging is mainly due to itstemperature profiles along its lifetime. This temperatureprofile depends on the ambient air, the weather, type ofsoil, etc. but also the temperature generated by the carriedcurrent. Generally, less heavily utilized assets presentlower risks of overheating. This could extend theirlifetime.

A higher voltage level might also increase DG Hostingcapacity.

Finally, reducing the number of voltage levels across the grid allows stronger standardization ofnetwork technologies. As discussed in the previous section, this can lead to benefits both in termsof CAPEX and OPEX.


A higher voltage level typically leads to better voltage quality (less significant voltage drop alongthe lines - for example [48] reports that a 10kV to 20kV conversion can halve the voltagedrop/rise for a given power withdrawn/injected).

According to [50], a decentralized grid (more small transformers closer to the consumer), besideleading to high loss reduction (47% in reported case studies), also improves the reliability of thesystem (failure hours/years 4-5 times lower).


The cost of this measure strictly depends on the specific implementation scenario. A detailedstudy must be carried out to estimate the full cost of such a measure (e.g. assessment of whichadditional assets need to be replaced to sustain the new voltage level etc.).

A plan for voltage level increase can also be pursued by targeting a progressive increase of theshare of the HV/MV grid with respect to the LV grid when undertaking new grid investments. Intheir replies to the survey, the system operators HEDNO (Greece) and EDP (Portugal) indicatethat their plan of loss reduction foresees a progressive increase of the number of MV/LVsubstations.

Vattenfal (Sweden) is carrying out anoptimisation of voltages levels in theirregional grids. The analysis is performedarea per area, taking into account thenetwork topology and the load level. 5 to15% of loss reduction has been observed inareas where voltage increase has beenadopted.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

3.1.4. Demand response



Line extended lifetime


DG hosting capacityincrease

SAIDI reduction

Lower DG curtailment

Smart Meter

Sensors for feedersmonitoring

Ripple controlinjector

Radio

Optic fibre

PLC

GSM



The implementation of demand response allows shaving peak loads, reducing the need of gridcapacity increase and possibly extending the lifetime of the assets (by reducing their utilisation).Demand response could also increase the DG hosting capacity of the grid, by allowing localbalancing of DG units injection. Grid reinforcements might therefore be deferred/avoided.

Detailed analysis of the impact of demand response on grid planning will be carried out in thesecond part of this study.


Demand response can be used to prevent loads in feeders or other assets to reach critical valuesthat might determine asset faults. In this way, the average SAIDI can be reduced. Moreover,flexible load control synchronized with local DG units could reduce to a reduction of DG energycurtailment due to over/under voltage or to line overloading.


The on-going implementation of smart metering infrastructure in most European Member Statesis providing the enabling infrastructure to carry out demand response activation, increasing thecost-effectiveness of this measure for energy efficiency.

Other necessary assets to deploy include grid sensors to increase the observability of the grid,particularly in terms of feeder loading, to adapt the DSOs activation of demand response to actualgrid conditions.

Finally, another possible option for DSOs to activate demand response consists in adopting Time-of-Use tariffs based on ripple control. In this case, DSO’s might have to invest in local rippleinjectors.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

3.1.5. Feed-in control of distributed generation & storage





Customers’ investments

DG hosting capacityincrease

Power Quality problemsreduction

Voltage problems reduction

Lower restoration operations

Converter with“smart”functionalities

Smart Meter

Storage system

Extra connectionslines

Radio

Optic fibre

PLC

GSM



An efficient feed-in control (including reactive power management) can lead to capacityinvestment deferral, by reducing the peak demand from the MV/LV substation. Reducing peakdemand can also typically reduce the aging of the component.


As seen in [51], efficient DG converters might providecompensation in term of harmonics, voltage, reactivepower and phase balancing control. This might reducecomplaints from clients and the number of fieldinterventions.


In the case of a control strategy with directcommunication link between DG units and the DSO(for DSOs to send set-points for the DG controlsystem), an important cost item is related to thecommunication infrastructure between the DG unitsand the DSO.

It is worth mentioning that DG converters on the market are increasingly equipped with built-inSmart Functionalities, with limited cost premiums. Retrofitting costs need to be foreseen forolder-generation converters to enable them for more advanced feed-in control schemes.

ERDF (France) is studying theimplementation of local voltage regulator toperform reactive power management of DGon the MV grid. The choice of the finalsystem has been based on a trade-off betweenthe gain on hosting capacity and the gain onvoltage losses, and its optimized value of thereactive to active power ratio and a reactivepower regulation based on voltage. Thesystem is currently in implemented two caseson their MV grid. The expectation is areduction of 2-3% in losses, and to an overallcost reduction of 5% compared to a case withno reactive power control, thanks to avoidedreinforcement. OPEX reduction is alsoexpected, since the system self-adapts whennew DG are connected.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Storage system may also be coupled with DG units to provide a better feed-in control. In thatcase, the cost accrues to the DG owner. According to [52], storage systems could also be installedby DSOs when demand growth is uncertain, as a temporary solution to defer investment, lettingtime to assess which reinforcement are really needed

3.1.6. Network reconfiguration





Reduced fault rate

Reduced restoration time

Improved voltage quality

Remotecontrolledswitches andreclosers insubstation

Feedersmonitoringsystem

AdvanceSCADA/DMS

Controlsoftware

Radio

Optic fiber

PLC

GSM


Benefits on investment planningThe drivers of automatic grid reconfiguration are typically loss reduction and reduced number ofinterrupted customers. However, a better balance between the different feeders reduces the peakson those lines, therefore potentially reducing the aging of the lines, and reducing replacementcosts.

Benefits on grid operationsIncreasing the number of remotely controlled switches reduces the number of substations to beinspected when a fault occurs. In a study ( [53]) carried out by the US Department of Energywithin 5 utilities, the use of remotely controlled switches reduced the number of interruptedcustomer by 35%. The use of automatic control for fault management leads this number to 55%.By better balancing the load in the feeders, voltage drop can also be reduced.

According to [54], a simulation of an automation algorithm to control the AES Sul grid (Brazil)showed that a 8.2% reduction of interrupted customer per year can be achieved. Moreover, thesolution prevents fault occurrences by reconfiguring the grid before a critical value is reached.The benefits are particularly high in the case of rural networks.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval


The main cost item of the measure is the installation of remotely controlled switches andreclosers and (if not already available) of the associated communication infrastructure. Accordingto [37], various U.S. utilities report that they have implemented this measure at a cost rangingbetween 18.750 € and 35.750 € per installed switch, including software, monitoring andcommunication system. For example, Adams-Colombia Cooperative mentioned a cost of 20.400€ for an overhead switch and 35.750 € for an underground one. Voltage level is one of the keydrivers of the cost value.

However, it is possible to implement network reconfiguration manually, without relying on aremote communication system, and still capture interesting benefits, particularly on lossreduction. Enexis (DSO in The Netherlands) (manually) reconfigured the open points in 300 MVloops out of 44000km of their MV grid, achieving a loss reduction of 0.5%, at a reported cost of670€/loop (according to their replies to the survey).

3.1.7. Switching off redundant transformers



Asset extended lifetime Remotelycontrolledtransformersswitches

AdvancedSCADA/DMSControlsoftware

RadioOptic fiberPLCGSM



Switching off redundant transformers reduces their use and aging, possibly deferring needs forreplacements.


Switching off/on redundant transformers can be performed manually. However, this criticallyincreases the restoration time in case of the loss of the energized transformer. Implementation ofremote and/or automatically controlled switches is thus to be preferred. The cost of transformerswitches may vary significantly, depending on their power rating.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

3.1.8. Voltage optimisation




Increase networkcapacity

Increase DG hostingcapacity


Voltage problemsreduction

Facilitate DG integration

D-Facts: MV/LVSTATCOM,SSSC…

Capacitor banks

On-Load Tapchangetransformers

Voltage sensors

Smart Meters

AdvancedSCADA/DMS

Controlsoftware(centralizedsystem)

Radio

Optic fibre

PLC

GSM



Optimization of reactive power flows allows betterexploitation of line capacity, reducing lines loading andpossibly increasing their useful life. Voltage optimisationcan also compensate the voltage rise/drop due to DGinjection, allowing more units to be connected on the grid.In [55], it was shown that voltage control through on-loadtap changing transformer was cost-efficient to increase thegrid hosting capacity, deferring expensive linereinforcement.

Benefits on grid operationsReactive power management systems can be equipped tocompensate harmonics injections.

By regulating the voltage level of the grid, complains due to over/under-voltage can be mitigatedand DG curtailment reduced.


Different options exist to carry out voltage optimization. In a substation, On-Load Tap Changing(OLTP) transformers can be installed to keep voltage within an acceptable range whileoptimising the operation of the grid. This solution can be more cost-effectively pursued byforeseeing an OLTP when replacing an existing transformer.

CEZ (Bulgaria) plans to installcapacitors and harmonics filteringequipment in two 110kV/20kVsubstations and in one 110kV/20kVsubstation. CEZ estimates an investmentcosts 176k€/substation of expense andCEZ expects a saving of 474k€/annum,coming from less power qualityproblems to be compensated to thecustomers, less complaints over voltagefluctuation, reduction of losses,reduction of equipment overheating andtheir aging.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Devices injecting or absorbing reactive power can also be installed in various places of the grid.Ideally, the best reactive power reduction and power factor correction is performed by placingcompensation devices the closest to the load. However, due to installation and operation cost, atrade-off is to be found between cost, loss reduction and voltage constraints [56]. The length ofthe network and the voltage stability of the system are therefore key parameters.

Different technologies (fixed capacitors banks, STATCOM, etc. ) can be adopted. It should benoticed that, according to [57], self-commutated compensators have the highest impact on losseswhile having a lower cost than synchronous condenser or static compensator technologies.Manually controlled/remote controlled capacitor banks could also be considered. The cost of thesystem will depend on the KVAr rating of the installed devices.

To ensure more granular voltage optimisation, feeders voltage sensors or Smart Meters can alsobe installed to adjust the voltage regulation to the actual grid voltage profiles.

3.1.9. Conservation Voltage Reduction (CVR)



Increased networkcapacity

Extended lifetime

Reduced fault rate


On-load tapchangingtransformers

Voltage regulator

Capacitors banks

Voltage sensors

Smart meters

AdvancedSCADA/DMS


Radio

Optic fiber

PLC

GSM

Expected benefits on operational efficiency

Benefits on the investment planning

CVR main goal is to reduce energy consumption. Results of such a measure depend on variousconditions such as consumers’ load, CVR technology used etc. On average, a peak reductionbetween 0.5% and 4% (depending on the feeder) can be expected, as seen in Denmark ( [58]) andin the US ( [59] [60]). With a lower peak, the utilization of the line is also reduced, possiblyextending its lifetime.


As stated in the previous point, the peak reductions due to CVR reduces the utilisation of the lineand its aging. In principle, the probability of default and the number of interventions could alsobe reduced.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval


Conservation Voltage Reduction (CVR) exists in various forms of implementation and is highlyscalable (e.g. it can be implemented feeder by feeder). [60] describes some practical cases in theUS that have shown its low cost to achieve significant benefits (total energy saving cost between0.01 and 5 cents/kWh following Bonneville Power Administration). To provide an order ofmagnitude, Adams-Columbia have deployed a CVR on 10 feeders (including 10 voltageregulators, 30 monitoring sensors, 10 controlled capacitors banks and 40 solid state variablecapacitors) for 157k€, including software.

3.1.10. Balancing 3-phase loads





Reduced neutral fault rate

Power Qualityimprovement

LV phases switchesin MV/LVsubstation

Feeders loadmonitoring system

Three-phasesconnections

Phase switches atconnections

Smart meters

AdvancedSCADA/DMS


Radio

Optic fibre

PLC

GSM

Expected benefits on operational efficiency

Benefits on the investment planning

Unbalanced feeders lead to higher currents flowing in one of the single-phase lines, which mightreach its limit faster than the others. Balancing phases might thus extend the lifetime of the lines.


Phase balancing reduces intervention and complaints due to Power Quality problems. Problemsresulting from current in neutral lines can be reduced (up to 35% of the phase current couldcirculate in several lines according to [3]), since no zero current is flowing in a perfectly balancedsystem.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval


Various techniques can be considered to perform phase balancing.

1) Phase-switching between single-phase line can already be performed manually on a regularbasis (seasonal for instance) to create a more balanced three-phase system, based on longterm feeders load measurement. It is clearly the cheapest approach but the impact is limited.

2) The above technique can also be performed automatically, based on load measurements tooptimise the impact.

3) Switches at connections: a more effective phase balancing can be performed by dynamicallyswitching the phase a user is connected to, according to the user’s load. This requires to havea three-phase connection for every single-phase user, and to have Smart Meters to measurethe current. This technique is however much more expensive than the previous one.

4) Balancing DG injection: according to [61], replacing three-phase inverters by three single-phase inverters can control the injection phase by phase to balance the network.

4. COST-EFFECTIVENESS OF ENERGYEFFICIENCY MEASURES IN ELECTRICITYGRIDS

The goal of this chapter is to analyze in detail the cost-effectiveness of Energy Efficiency forelectricity grids. The excel tool described in annex IV provides a step-by-step decision support tocarry out the assessment.

4.1. Assessment of cost-effectiveness

Cost-effectiveness of each measure is defined as the ratio between the associated benefits andcosts. This entails looking at the benefits and the implementation costs associated to each energyefficiency measure. Apart from reducing energy losses, ad discussed in previous chapters, it isconsidered that EE measures could also provide additional benefits in terms of reduction ofOPEX (e.g. restoration time, power quality issues etc.) for the DSO and of avoidance of gridinvestments (CAPEX reduction).

The analysis consists in individual assessment of all potential benefits and of implementationcosts of energy efficiency measures. The output is a ranking of individual energy efficiencymeasures according to their cost-effectiveness.

The proposed methodology is synthetized in Figure 9.

The assessment of the cost-effectiveness of energy efficiency measures has the followingcharacteristics:





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Focus on individual EE measures – The focus is on detailed assessment of individual EEmeasures. The analysis does not consider the perspective of a larger investment plan whichalso includes a set of EE measures.Assessment of grid energy efficiency benefits according to the following key dimensions- Reduction of grid losses- Deferral/avoidance of grid investments- Increase of RES hosting capacity- Reduction of outages (security of supply)- Power QualityAnalysis of implementation costs and of factors affecting costs - The assessment of theinvestment costs of each measure is done considering the capital and operational costs and thefactors affecting them. Moreover, we also look into options for optimization of investments(e.g. replacing assets at the end of useful life).

Figure 9 - Methodology for assessment of cost-effectiveness of individual EE measures

4.2. Synthesis of the cost effectiveness of themeasures

Table 6, Table 7, Table 8 – provide a synthetic qualitative indication of levels of benefits andcosts of the different measures. The assessment assumes the stand-alone implementation of agiven measure (no synergies with other measures or other investments) on a per-unit basis.

Clearly, as mentioned, the cost-effectiveness of each measure strongly depends on the specificgrid conditions where it is implemented. Provided qualitative value are only intended to provideinitial guidance in comparing the different measures.

Ranking of EE measures





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Component replacement

Measure Lossreduction

Deferredinvestment

REShostingcapacity

Securityof supply

Power qualityimprovement

CostType Level

EETransformers

++/+++ +/0 + + -Delta cost in case ofplanned replacement /new installation

-Full cost in case ofproactive replacement

Low tohigh

Oversizing linecapacity

+++ ++ + -Delta cost in case of

planned replacement /new installation

-Full cost in case ofproactive replacement

Low tohigh

Increase voltagelevel

+++ +++ ++ + -Limited adaptation (incase of already voltagecompliant equipment)

-Massive replacementof all equipment

Low tovery high

Table 6 – Efficiency & cost of component replacement measures





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Grid management


Deferredinvestment

REShostingcapacity

Securityof supply


CostType Level

Networkreconfig.

++ + ++ ++ -Mainly software in caseof existing remoteswitches

-Full installation ofsoftware and remoteswitches

Low tohigh

Transformerswitching

+/+++ ++ - -Labour work only incase of manual switching

-Full installation of newremote switches

Very lowtomedium

Voltage opt. +++/++ + +++ +++ - Mainly software in caseof existing equipment

-Full new equipmentdeployment (e.gcapacitor/OLTP)

Mediumto high

CVR ++ ++ ++ ++ - Mainly software in caseof existing equipment

-Full new equipmentdeployment (e.gcapacitor/OLTP)

Medium

Bal. 3-ph.Load

++ + + ++ -Labour work only incase of manual switching

-Full installation of newremote switches

Low tohigh

Table 7 – Efficiency & cost of grid management measures

Feed-in control


Deferredinvestment

REShostingcapacity

Securityof supply


CostType Level

Demandresponse

+++ +++ ++ ++ Depending on theregulation

Low tomedium

DG control +/++ ++ ++ ++ ++ Depending on theregulation

Low tomedium

Table 8 –– Efficiency & cost of feed-in control measures





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Hig

hef

ficie

ncy

trans

form

ers

Expa

ndin

glin

eca

paci

ty

Incr

ease

volta

gele

vel

Net

wor

kre

conf

igur

atio

n

Switc

hing

-off

redu

ndan

ttra

nsfo

rmer

s

Dem

and

resp

onse

Feed

-inco

ntro

l&st

orag

e

Vol

tage

optim

isat

ion

Con

serv

atio

nV

olta

geR

egul

atio

n

Bal

anci

ng3-

phas

elo

ad

Factorsimpactingtheefficiencyof themeasure

Load rate X X X X X X X X XAge of the assets X XFault rate X X X X XHigh harmonics problems X XTransformers power rating XGrid length X XVoltage issues X X X X XFeeders load unbalance XNumber of switches XAverage restoration time XPresence of a repartition grid XNumber of transformers in parallel XDG curtailment level X XDistance between DG and consumersPower factor XDG penetration rate XNumber of voltage level XVoltage level XNeutral conductors issues X

Factorsimpactingthe cost ofthemeasure

Overhead underground grid X X X X X X XUrban rural network X X X X XTransformers accessibility XAge of the assets X XOther replacement/expansion plans X X X X X X X XTechnology standardization X X XAsset fitted for an upper voltage XGrid expansion / load growth XVoltage level X XDG integration XPenetration of the communicationsystem X X X X X

Flexibility market XSmart Meter roll-out plan X X X XDG incentive to customer XCombination with CVR XLV consumers’ density XLV substation density X

Table 9 – Factors impacting the overall efficiency and the cost of the measures





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

5. ASSESSMENT OF THE ENERGY EFFICIENCYPOTENTIAL IN GAS GRIDS

5.1. Understanding Energy Losses in gas networks

The natural gas transmission and distribution system consists of a complex network of pipelines.As detailed in Table 10, in the European Union, it represents more than 2150 thousand kilometresof pipelines.

Table 10: Length of gas transmission and distribution pipelines in Europe (Source: Gas Infrastructure Europe)

In order to ensure that natural gas remains pressurized during its transmission (and more rarelyduring its distribution), compression, by means of compressor stations, is required periodicallyalong the pipelines.

The following simplified diagram shows a typical natural gas transmission and distributionnetwork.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Figure 10: Typical natural gas transmission and distribution network

Energy losses due to gas transmission and distribution have been considered in this study to benatural gas shrinkage (gas losses and own consumption) and own electricity consumption. Figure11 reports the losses level among representative Member States.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Figure 11 – Level of losses19 in gas grids in selected EU Member States

Compressor stations are the largest sources of gas use on a gas distribution and transmissionnetwork. Electricity consumption can be important as well on compressor stations, whenelectricity is used to run compressors.

Gas losses can arise from multiple causes. They have been grouped in three categories:

Vented emissions: Direct gas releases to the atmosphere of natural gas resulting fromequipment design, regular process operations, maintenance activities, or emergency releases(e.g. process designed flow to the atmosphere through seals or vent pipes, equipmentblowdown for maintenance or an emergency, and direct venting of gas used to powerequipment (such as pneumatic devices)).Fugitive emissions: Unintentional leaks of gas from piping and associated equipmentcomponents (e.g. leaks from valve seals, packing or leaks resulting from corrosion, faultyconnections). Such emissions can occur on many different parts of the network and are verydifficult to estimate.

19 Both technical and non-technical losses, based on the Ecofys analysis from EuroStat database





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Emissions due to incidents: Inadvertent damages to pipelines may also cause gas emissions.A mapping of energy losses due to gas distribution and transmission is presented in the threefollowing sections.

5.1.1. Mapping of losses in gas transmission networkGas transmission is achieved by means of high pressure underground pipelines. These pipelinesare usually made of steel that is coated and cathodically protected20.

Figure 12 reports the mapping of mechanisms leading to losses in gas transmission network.These losses can be either natural gas emissions from pipelines (vented, fugitive or caused byincidents), electricity used for pipeline cathodic protection or gas to fuel heaters at regulationstations.

Figure 12: Mapping of losses in gas transmission network

20 Cathodic protection is a technique of running an electric current through the pipe to prevent corrosion and rusting.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

5.1.2. Mapping of losses in compression stationsAlong the transmission (and sometimes the distribution) networks, compressor stations are placedat regular interval. When gas enters the compressor station, it is compressed by either natural gasfuelled engines or turbines, or electric motors.

Figure 13 reports the mapping of mechanisms leading to losses in gas compressor stations. Fuelused to compress gas is the most important source of gas and electricity consumption. Gas lossescan be due to vented emissions from compressors, valves and pneumatic-driven devices orfugitive emissions.

Figure 13: Mapping of losses in gas compressor stations

5.1.3. Mapping of losses in gas distribution networkNatural gas is distributed to residential and small industrial customers through medium to lowpressure underground pipelines (distribution mains and service lines). Pipeline materials can bevery different depending on the localization, the pressure and the age of the network. Forexample, oldest distribution mains were constructed from cast iron pipes and yarn/lead joints orunprotected steel. New materials are mainly cathodically protected steel and plastic (e.g.polyethylene).





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Figure 14 reports the mapping of mechanisms leading to losses in gas transmission grids. Lossescan be due to natural gas emissions and gas or electricity consumption. Gas emissions can varywidely depending on the pipeline material as fugitive emissions from older networks are muchhigher than more recent ones. Gas or electricity consumption can differ from one network toanother, depending on situations (e.g. existence of cathodically protected steel, heaters atregulation stations).

Figure 14: Mapping of losses in gas distribution grids

5.2. Energy efficiency measures in gas networks:categorisation and potential

Energy efficiency measures in gas networks are considered in this study to be minimisation ofgas losses and own gas and electricity consumption. The three following sections present variousenergy efficiency measures associated to transmission network, compressor stations anddistribution network.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

5.2.1. Gas transmission network: energy efficiency measures andreduction potential

Table 11 reports the list of measures addressing the sources of loss in gas transmission networkidentified in the previous section. It also describes the reduction potential of each measure andidentifies the key factors affecting this potential.

Opportunities to improve energy efficiency measures in gas distribution networks are mainlyrelated to natural gas losses. Measures to reduce vented gas during maintenance operation (e.g.allow customers to use natural gas in order to lower the pressure in the affected section of pipe orgas recompression) can potentially lead to substantial minimisation of losses. For example, a80% reduction of gas losses can be achieved when gas is recompressed instead of vented beforemaintenance ( [62]).

Category oflosses Type of losses Associated measures Qualitative

potential Key factors

Natural gaslosses

Maintenanceoperations

Reduction of vented gas ++Pressure level,volume of ventedpipe

Recompression ++Pressure level,volume of ventedpipe

Permanent lossesdue to age and gasgrid type

Inspection andmaintenance programs + Length and age of the

network

Third party damages

Mapping, relation withcontractors, emergencycall centers…

+

Length of thenetwork, pressurelevel, sensitivity ofthe pipe area (urbanor rural, worksarea…)

Automatic shut-offvalves +

Length of thenetwork, pressurelevel, sensitivity ofthe pipe area (urbanor rural, worksarea…)

Gas orelectricity ownconsumption

Fuel to heatfacilities (atregulation stations)

Improved burner designand avoid heating whenpossible

+Number andefficiency of heaters,necessity of burners

Electricity forcathodic protection (-) (-) (-)

Table 11: Mapping of energy efficiency measures –Transmission





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

5.2.2. Compressor stations: energy efficiency measures andreduction potential

Table 12 reports the list of measures addressing the sources of loss in compressor stationsidentified in the previous section. It also describes the reduction potential of each measure andidentifies the key factors affecting this potential.

Category oflosses

Type of losses Associated measures Qualitativepotential Specific key factors

Naturalgaslosses

Compressors starts /shutdowns andblowdowns

Optimal strategies forcycling on- and off-line +++ Base or peak load

compressorReduction of vented gas (nodepressurizing aftershutdown…)

+++Base or peak loadcompressor

Pressure safetyvalves and gas-drivenpneumatic devices

Replacement of high bleeddevices + Number of high bleed

devicesInstallation of compressedair/electric systems ++ Number of gas-driven

pneumatic devicesFugitive emissionsfrom compressors(seals…)

Inspection and maintenanceand replacement of seals ++

Inspection frequency,type of seals to bereplaced

Fugitive emissionsfrom equipment’s notin compressionservice

Inspection and maintenance + Inspection frequency

Gasownconsumption

Engines(reciprocatinginternal combustion)

Optimization of equipmentdispatch

++ Number and capacity ofcompressors(Pressure and flowregulation)

Increased efficiency (retrofitcontrol)

++ Compressor efficiency(before and after engineretrofit)

Replacement with a moreefficient unit

+++ Compressor efficiency(before and after enginereplacement)

Gas turbines

Optimization of equipmentdispatch

+ Number and capacity ofcompressors(Pressure and flowregulation)

Replacement with a moreefficient unit

+ Compressor efficiency(before and after gasturbine replacement)

Table 12: Mapping of the energy efficiency potential of each measure –Compression

Measures to improve energy efficiency in compressor stations can be either technical solutions(e.g. replacement of seals) or optimized work practices (e.g. optimal strategies for cyclingcompressors on- and off-line).





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Various solutions can be implemented to reduce gas losses. For example, a Marcogaz report ([62]) describes the replacement of pneumatic actuators with devices operated by compressed airin a compressor station in Italy that enabled gas emissions to be reduced by 45%. Anotherexample is the reduction of gas emissions when taking compressor off-line. This can be achievedby keeping compressors pressurized (estimated emission savings: around 30% for a base loadcompressor ( [63])).

Gas and electricity consumption can be reduced when replacing old engines or gas turbines withmore efficient units. Another option is to optimize equipment dispatch (e.g. optimisation ofpipeline system operations in order to operate compressors at their nominal capacity and avoid asmuch as possible compressor starts and shutdowns).

5.2.3. Distribution grids: energy efficiency measures and reductionpotential

Table 13 reports the list of measures addressing the sources of loss in gas distribution networkidentified in the previous section. It also describes the reduction potential of each measure andidentifies the key factors affecting this potential.

The most efficient way to reduce gas losses is to replace or repair/reline old grey cast iron pipeswith yarn/lead joints, as they are responsible for the highest emissions. For example, a programof replacement of old pipes in the UK is expected to achieve 10 times lower emissions thanpreviously ( [62]).

Third party damages to gas distribution pipes can be another important cause of gas losses. Safetyis the primary driver to reduce these gas losses. Emissions can be reduced when implementingmeasures to prevent damages (e.g. enhancement of the grid mapping accuracy or training andtechnical guidance provided to contractors in charge of digging works). For example, reductionof the number of third party damages (with gas emissions) of around 30% during the period2010-2014 has been achieved in a French gas network (GrDF, 2015).





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Category oflosses

Type of losses Associated measures Qualitativepotential Key factors

Naturalgaslosses

Pneumatic emissions Reduction of vented gas + Number of gas-drivenpneumatic devices


Reduction of vented gas + Pressure level, volume ofvented pipe

Recompression + Pressure level, volume ofvented pipe

Permanent losses dueto age and gas gridtype

Inspection and maintenanceprograms ++ Length and age of the

network

Pressure management NA Length and age of thenetwork

Permeability ofmaterials and joints

Replacement orrepair/relining of old pipes +++

Length and age of thegrid, pipes and jointsmaterials

Third party damages

Mapping, relation withcontractors, emergency callcenters, protection of pipesin sensitive areas, shut-offvalves

++

Length of the grid,pressure level, sensitivityof the pipe area (urban orrural, works area…)

Gas orelectricity ownconsumption

Fuel to heat facilities(at regulationstations)

Improved burner design andavoid heating when possible +

Number and efficiencyof heaters, necessity ofburners

Electricity forcathodic protection (-) (-) (-)

Table 13: Mapping of the energy efficiency potential of each measure –Distribution network





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

6. ASSESSMENT OF COST-EFFECTIVENESS OFSELECTED ENERGY EFFICIENCY MEASURES -GAS

In the following we analyze more in detail the cost-effectiveness of energy efficiency measureswith most relevant potential at transmission, distribution and compressor station level.

The excel tool described in ANNEX V provides a step-by-step decision support to carry out theassessment.

6.1. Gas Transmission Network

Table 14 describes the type of implementation costs and an estimation of cost levels associated tomeasures with a significant reduction potential in gas transmission network.

Avoiding emissions during maintenance has significant energy efficiency potential. However,some measures can be quite expensive and may be not economically relevant in all cases (e.g.recompression of gas before maintenance with a mobile compressor unit). Less costly measures,with minor facility modifications, can also be implemented (e.g. use inert gases and pigs toperform pipeline purges instead of venting).

Additional benefits are less greenhouse gas emissions.

Categoryof losses

Type oflosses

Associatedmeasures

Qualitativereductionpotential

Type of costs Costlevel

Naturalgas losses


Reduction ofvented gas ++

Capital costs (minor facilitymodifications), operating costs(maintenance), labour costs

+

Recompression ++ Capital costs (pump-downcompressor), labour costs +++

Table 14: Type of implementing costs and estimated cost levels – Transmission network

Examples of implementation costs are given in Table 15. These figures give an order ofmagnitude and should be considered with caution since situations may differ from one network toanother.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Measure Capital costs estimation Operating costs estimationReduction of vented gas (Useinert gases and pigs toperform pipeline purges)

- $500 (for 2 miles of 10-inch diameterpipeline) ( [63],2011)

Reduction of vented gas(Inject blowdown gas into lowpressure mains or fuel gassystem)

$1,000 (minor facilitymodifications: additionalpiping) ( [63],2011)

-

Recompression (mobilecompressor unit)

200,000 € (for one case with 630,000 Nm3 natural gas saved from

venting) ( [62])

Table 15: Implementation costs of measures – Examples – Transmission network

6.2. Gas Compressor Stations

Table 16 describes the type of implementation costs and an estimation of cost levels associated tomeasures with a significant reduction potential in compressor stations.

Possible measures that can be applied within compressor stations can vary widely. Some of themcan be implemented with no (or nearly no) costs (e.g. avoid compressor depressurization aftershutdown). Conversely, an engine replacement with a more efficient unit represents a veryexpensive investment.

Overall, these measures can contribute to improve operational efficiency, less greenhouse gasemissions and cost savings due to reduction of gas losses and own gas consumption.

Category oflosses Type of losses Associated

measures



Natural gaslosses

Compressorsstarts /shutdowns andblowdowns

Optimal strategiesfor cycling on- andoff-line

+++ Labour costs +

Reduction of ventedgas (nodepressurizing aftershutdown…)

+++

No costs (with theoption: nodepressurizing aftershutdown) or minorcapital and labour costswith other options

+

Pressure safetyvalves and gas-drivenpneumaticdevices

Installation ofcompressedair/electric systems

++Capital costs(compressed air/electricsystems)

++

Fugitiveemissions fromcompressors(seals…)

Inspection andmaintenance andreplacement ofseals

++

Capital costs (seals),operating costs(maintenance), labourcosts

++





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Category oflosses Type of losses Associated

measures



Gas ownconsumption

Engines(reciprocatinginternalcombustion)

Optimization ofequipment dispatch

++ Labour costs +

Increased efficiency(retrofit control) ++

Capital costs (minorengine modification),operating costs(maintenance), labourcosts

++

Replacement with amore efficient unit +++

Capital costs (newcompressor), operatingcosts (gas or electricity +maintenance), labourcosts

+++

Table 16: Type of implementing costs and estimated cost levels – Compressor stations

Examples of implementation costs are given in Table 17. These figures give an order ofmagnitude and should be considered with caution since situations may differ from onecompressor station to another.

Measure Capital costs estimation Operating costs estimationOptimal strategies for cycling on-and off-line

- Labour (very variable costsdepending on availableequipment and optimizationpossibilities)

Reduction of vented gas (nodepressurizing after shutdown)

- -

Reduction of vented gas (Keepcompressor pressurized and routegas to fuel system or install staticseal)

< $5,000 for one compressor ( [63], 2006)

Installation of compressedair/electric systems

$45,000 to $75,000 (per facility) -

Inspection and maintenance(identification of leaks and cost-effective repairs)

- $26,248 (average value, for onecompressor station) ( [63],2003)

Replacement of seals (compressorrod packing system replacement)

$5,346 ($2008) (percompressor) (US EPA, 2014)

-

Replacement with a more efficientunit (case of an electric compressor)

$6,050,000 (per 5 reciprocatingcompressors replaced) (EPAGas STAR Program, 2011)

$6,200,000 (per 5 reciprocatingcompressors replaced) (EPAGas STAR Program, 2011)

Table 17: Implementation costs of measures – Examples – Compressor stations





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

6.3. Gas Distribution Network

Table 18 describes the type of implementation costs and an estimation of cost levels associated tomeasures with a significant reduction potential in gas distribution network.

As stated previously, replacement of grey cast iron pipes with polyethylene pipes can reducesignificantly gas emissions. This measure can be very expensive, especially in urban areas. Thisshould, however, be put in balance with associated benefits (higher safety level, less greenhousegas emissions, cost savings due to reduced gas losses and lower repair costs over the nextdecades).

Less costly measures can be implemented as well (e.g. prevention of third party damages,improvement in inspection and maintenance programs) and will contribute to a greater safety ofthe network, less greenhouse gas emissions and cost savings due to reduction of gas losses.

Categoryof losses

Type of losses Associatedmeasures



Naturalgas losses

Permanent lossesdue to age and gasgrid type

Inspection andmaintenanceprograms

++Operating costs(maintenance),labour costs

+

Permeability ofmaterials and joints

Replacement orrepair/relining ofold pipes

+++

Capital costs (newpipes or liners),operating costs(maintenance),labour costs

+++

Third party damages

Mapping, relationwith contractors,emergency callcenters, protectionof pipes insensitive areas,shut-off valves

++

Mainly labourcostsCapital costs (forprotection of pipesand shut-offvalves)

Variable

Table 18: Type of implementing costs and estimated cost levels – Distribution network

Examples of implementation costs are given in Table 19. These figures give order of magnitudeestimation and should be considered with caution since situations may differ from one network toanother.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Measure Capital costs estimation Operating costs estimation

Improvement in inspectionand maintenance programs(increased walking survey)

-

$11,000 for a pipeline system of2,900 miles (labour costs: $10,000and additional equipment: $1,000) ([63], 2011)

Replacement of grey castiron pipes with PE pipes

200,000 € (per km) ( [64])> $1,000,000 (for 1 mile of cast iron main replacement) ( [65])

Insert gas main flexibleliners (for cast iron andunprotected steel)

$10,000 (for 1 mile of cast ironmain and 1 mile of unprotectedsteel service lines) ( [63], 2011)

Mapping, relation withcontractors, emergency callcenters

-Mainly labour (Very variable costsdepending on implementedmeasures)

Shut-off valves (for servicelines)

$6,300 (for 350 valves installed) ([63], 2011) -

Table 19: Implementation costs of measures – Examples – Distribution network





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

PART 2 - ASSESSING THE POTENTIAL OF DEMANDRESPONSE TO INCREASE THE EFFICIENCY OFELECTRICITY NETWORKS





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

PART 2 - EXECUTIVE SUMMARY

As described in the interim report, this study particularly considers the planning/operationalefficiency potential in electricity grids, thanks to new “Smart” measures (Smart Grid assets;flexibility of distributed energy resources).

Accordingly, the focus of this second part of the report is to analyse the potential of demandresponse (DR) in bringing efficiency in electricity grids for Distribution System Operators(DSOs).

Key messages of the analysis are grouped into the following three main domains:

A. VALUE STREAMS AND PROCUREMENT OPTIONS OF DEMAND RESPONSEFOR DISTRIBUTION SYSTEM OPERATORS

There are two main categories of DR benefits for DSOs: avoided CAPEX and avoidedOPEX

- Avoided capital expenditure (CAPEX): DSOs need to expand grid capacity andupgrade transformers to cope with congestions. Depending on the demand and renewableenergy sources (RES-E) growth in an area, employing DR in the planning phase offerDSOs a valuable solution to asset replacement. This strategy may postpone the need toinvest in infrastructure upgrades and allow using the assets for their full lifetime,reducing short-term expenditures.21

- Avoided operational expenditure (OPEX): local balancing leads to a more balancedline loading and to a subsequent reduction of distribution grid losses which form a keyoperational expenditure for DSOs. Moreover, by relieving grid congestion, DR allowsRES-E generation to be better integrated in the system. This means that curtailment ofRES-E can be avoided, and hence compensation payments for the curtailed renewableenergy are reduced (in countries where it is foreseen).

It is worth mentioning that, while DR can also be used to optimise operation of the wholesale andbalancing markets, DSOs cannot capitalise on these benefits, as unbundling provisions restricttheir participation to these markets as flexibility provider.

21 Grid assets must be replaced at the end of their lifespan regardless of the situation. Since large part of theinvestment costs are related to installation, when replacing assets that have reached their lifetime, placing somelarger cables brings no significant additional cost. In this respect, CAPEX benefits from DR should be considered withrespect to the remaining lifetime of the assets.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

There are two main approaches for DSOs to procure demand flexibility: tariff-based(implicit) DR and volume-based (explicit) DR

Tariff-based - Implicit DR

Under this approach, DSOs use tariff signals to trigger demand flexibility that is in line with theneeds of the distribution system. DR is implicit in the sense that a predefined volume of DR isnot agreed before the DR activation, rather it results from the users’ response to the tariff signal.In this context, and throughout this report, we refer to “tariffs” as the regulated component of thefinal price, related to distribution grid charges.

Basically the options are to vary the distribution grid tariff based on a) fixed timeframes (time ofuse pricing); b) changing timeframes (dynamic pricing); c) changing timeframes and location(locational dynamic pricing). The different options are best suited to different purposes.

The limit of implicit DR is that it does not provide a guarantee to DSOs regarding the actual levelof DR that will be activated in response to tariff signals.

Volume-based - Explicit DR

Under this approach, the DSO (financially) compensates a DR service provider to activate aspecific quantity of DR.

When the DSO identifies a need/potential benefit for DR in its network, it could in principleprocure it by:

1) making a direct contract with the DR provider with customised conditions;2) issuing a call for tenders for the DR service required, whenever it is required; or3) establishing a local market for DR providers connected to the network with local price.

Which option is most efficient and effective depends on the number of DR providers that areconnected to the network (liquidity of the market). Forming bilateral agreements are mostefficient when there are only few, and forming a local market can be efficient when there aremany market players.

Reliability of DR delivery should be ensured, for example by applying penalties if there is adeviation between services sold and delivered (except in an emergency). However, it must alsobe kept in mind that very strict reliability standards will discourage service providers fromparticipating in such a DR scheme.

The correct implementation of explicit DR requires the definition of a mechanism to verify ex-post that the agreed amount of DR has been actually activated. Typically this is done via thedefinition of a baseline which represents the load profile had the DR event not taken place.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

B. ASSESSMENT OF THE VALUE OF DEMAND RESPONSE TO IMPROVEDISTRIBUTION GRID EFFICIENCY AT PLANNING AND OPERATION

An extensive simulation analysis has been carried out to assess how the integration of DR inDSOs planning and operation could allow finding the optimal trade-off between CAPEX(investment in equipment/asset replacement and upgrade) and OPEX spending (in particular costof thermal losses), particularly in presence of a growing pentration of distributed generation.

Tool for optimization of DSOs’ planning and operation in presence of flexibility

Tractebel’s proprietary tool Smart Sizing has been used for this analysis. The tool allows a globaloptimization of DSOs’ OPEX and CAPEX by selecting the optimal mix of traditional andflexibility-based investments. The optimization tool takes into account the following costs:

OPEX:

Technical losses costsActivation cost of flexibility (e.g. procurement of volume-based flexibility by the DSO)

CAPEX

Network components (e.g. lines, substations)ICT infrastructure for activation of flexibility

Grid models

Four different grid models have been built, iin terms of grid configuration, load patterns anddistributed generation profiles. They are intended to provide a constrasting set of operatingconditions to test the robustness of the results:

Distribution network in urban area located in Northern Europe;Distribution network in rural area located in Northern Europe;Distribution network in urban area located in Southern Europe;Distribution network in rural area located in Southern Europe.

Scenarios representing available flexibility

For each model we have run four main scenarios with growing availability of flexibility:

1) Base case scenario – In this scenario, planning and operation rely only on traditionalinvestments, without the option of DR.

2) Time-of-Use scenario - In this scenario, it is assumed an (exogenous) reshaping of the loadprofiles in the simulation thanks to the introduction of Time of Use (ToU) tariffs. The tooldoes not optimize the load profiles, it just considers the reshaped load profiles as input to thesimulation.

3) Demand response scenario – In this scenario, the tool is able to arbitrate betweeninvestments in traditional assets or DR (considering associated ICT investment costs andactivation costs ).

4) Full flex scenario: Demand response and DG curtailment options– In this scenario,DSOs can opt for two competing sources of flexibility: (explicit) DR and DG Curtailment.

Within each scenario a sensitivity analysis has been carried out by varying (homotetically) thepenetration level of the different types of DG (PV, wind, CHP) in Low Voltage (LV) andMedium Voltage (MV).





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Main messages

In the following we provide a summary of the main messages resulting from the simulationanalysis, which are valid on average on a given grid. Deviations are possible on specificfeeders/substations depending on local specificities.

Demand response is effective in reducing the total cost of the distribution grid

Demand response is an effective means to reduce the TOTEX (Total Expenditure) of thesystem (CAPEX+OPEX), when integrated in grid planning and operation:- Via reduction of direct and inverse peaks in the network and thus the need for

infrastructure costs (CAPEX).- Via reduction of technical losses (OPEX) by performing a local balancing with DG and

reducing the flows in the system.The reduction of TOTEX is achieved by finding the optimal balance between CAPEX andOPEX (in certain conditions, the level of technical losses could also increase, if off-set by adecrease of CAPEX). Of course, regulation should specifically incentivize DSOs to reduceTOTEX (please refer to chapter 5).

If DSOs do not receive the right incentives to reduce losses, then the use of flexibility couldmainly target CAPEX reductions (i.e. size assets to their capacity limits) and possibly lead

to higher loss volumes.

The simulation analysis shows that DR activation targets specifically loss reduction only ifsufficiently high costs for losses are considered.Regulation needs to pay attention to the costs of losses paid by the DSOs (or to provide theright loss penalties) to avoid that DSOs have an incentive to use flexibility to downsize assets(reduction of CAPEX), even if this leads to an increase of losses.

Explicit (Volume-based) Demand Response can bring higher benefits to the systemcompared to implicit (tariff-based) Demand Response.

Both implicit and explicit DR can reduce the total cost of the system. Both approaches canalso be implemented together to maximize the impact (even if the two mecanisms can be incompetition with each other, as the presence of ToU reduces the amount of flexibilityexploitable by explicit DR).Volume-based (explicit) DR can bring higher benefits as it can be more precisely activatedaccording to DSO’s needs.For a volume-based flexibility market to develop, DSOs should make clearly accessible theinformation about their flexibility needs (present and future). This would allow to reveal theavailability of flexibility resources and contribute to creation of a flexibility market.

Time of Use (ToU) tariff can provide significant system benefits, but they need to becarefully tailored to specific local conditions.

The impact of ToU tariffs on TOTEX depends on the level of penetration of DGs and on localconditions.In general, ToU can provide significant benefits and facilitate the integration of DG at a lowercost when the off-peak period is synchronized with DG generation. A location-basedcomponent in the ToU could help better tailor the tariff design to local conditions.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Poorly designed ToU timeframes, under certain conditions, could also have a negative effect.For example, with high DG penetration, simple day-night ToU might actually worsen thecosts of DG hosting capacity.

The cost/benefit ratio of Demand Response for DSOs depends on local conditions

On the benefit side:

The value of DR for DSOs strongly depends on the level of penetration of DG, on theconsumption patterns and on the grid characteristics.Typically, the simulations shows that a limited number of activation of DR per year canalready provide significant benefits for the system

On the cost side:

A key factor affecting the business case of DR for DSOs is the availability and cost of an ICTinfrastructure (smart meters).In those regions where smart meters are already being deployed (driven by policy decisions),the business case for DR is stronger both for residential and commercial/industrial customers(as DR programs do not need to include ICT infrastructure costs)In those regions where the cost of the ICT infrastructure should be borne by the DR program,it is recommended to focus on DR at MV level get most of the benefits.Demand Response is always part of the best flexibility mix to decrease the total cost of

distribution grid planning and operation

With limited penetration of DG, DR is the most economically viable flexibility option to drivedown the system cost. (This holds on average on a given grid. Deviations are possible onspecific feeders/substations depending on local specificities. For example in rural areas, withlimited load, curtailment of DG could be the best flexibility option to solve congestions onspecific feeders).With very high penetration of DG, reverse peak (peak of reverse power flows) becomes themain driver to size grid assets. Beyond that DG penetration level, generation curtailmentbecomes an economically viable option together with DR.

Demand Response has typically a positive impact on grid technical losses, but reduction oflosses is not the main driver for its activation

Reduction of losses is typically not the main driver of the activation of DR, rather a side-effectof the activation of DR to reduce the CAPEX.Higher loss reduction level is typically achieved in rural areas, given the structure of ruralgrids (longer feeders, overhead lines), for which cost of losses has a higher share of theTOTEX.

Demand flexibility allows an higher DG hosting capacity at lower cost

DR increases the synchronicity of the demand with DG and mainly reduces CAPEX.The positive impact of DR on DG hosting capacity becomes increasingly significant after acertain DG penetration level.In fact, a limited penetration of DG, even without any flexibility, could actually decrease theTOTEX of the system compared to the case without any DG (thanks to DGs balancing localdemand and reducing network flows).





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Strong penetration of residential PV will drive more rapidly the need for flexibility

A lack of variety in the types (PV, wind, CHP) and connection level (MV, LV) of DG has anegative effect on grid efficiency, driving up CAPEX (reinforcements) and OPEX (technicallosses).This means that in those areas where a single source of DG is mainly present (e.g. areas withstrong residential PV), demand flexibility can particularly provide positive system impact andincrease grid efficiency. This is of course dependent on the availability of sufficient level ofdemand flexibility (in which case DG curtailment could be the main flexibility option toexplore).

C. REGULATORY EVOLUTIONS FOR DISTRIBUTION SYSTEM OPERATORSTO PROCURE AND ACTIVATE DEMAND RESPONSE

Regulatory evolutions should ensure that tariff signals provide sufficient incentives to gridusers to adapt their demand profiles to grid conditions

Under an implicit (tariff-based) DR approach, the regulator should define the required changes innetwork tariffs to incentivise DR, by changing specific components of the tariffs (e.g. highercapacity charges), changing time intervals for ToU, introducing location-based components orallowing DSOs to adapt tariffs to grid conditions.

However, since distribution grid tariffs constitute just a component of the entire electricity pricepaid by a consumer, in cases they are low they might fail to form an effective market signal togrid users to change their consumption to meet grid needs.

DSOs can directly procure DR if DSOs remuneration is based on TOTEX

Under a volume based DR approach, DSOs procure DR services and remunerate the providedservice. Incentives to optimise operation need to be created by appropriate regulatory adjustmentsin order to allow DSOs recovering the costs of procuring DR. This can be done by allowing DSOremuneration to reflect the DSOs’ TOTEX (CAPEX+OPEX).

Regulation should ensure a neutral role for DSOs in the activation of flexibility, in order toensure market-based competition among different flexibility resources

As shown in the simulation analysis, economic benefits are achieved by enabling competitionbetween flexibility options for the provision of services at local or global level. To enable amarket-based approach, transparency and neutrality of DSOs is key. In this respect policyquestions are related to:

whether DSO should be allowed to own flexibility assets (e.g. storage)How to address the risk that, if most of the value of flexibility is controlled directly by DSOs(e.g. ownership of storage means; direct bilateral contracts with grid users), a local flexibilitymarket (targeting also other value streams, e.g. BRP; TSOs..) could never develop.How to ensure DSO transparency in validating flexibility transactions at local level which aredriven by global market signals (e.g. flex transactions for TSOs)





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

In case of conflicts for flexibility activation, local distribution needs should prevail oversystem transmission needs, unless when system security is at stake.

Flexibilities in the distribution grid can be used for multiple purposes, a) in the system level(participation to wholesale, intra-day and balancing markets) b) in transmission network level(alleviation of transmission congestions) and c) in the distribution grid (alleviation of distributioncongestions). Congestions at any network level could “block” the provision of flexibility servicesfrom the local level to the higher system levels. Respectively, conflicts on the activation offlexibility between TSO and DSO are possible. Therefore, a prioritisation between local andglobal requirements is needed, and local needs should prevail, unless when system security is atstake and alternative options are not viable. In this respect, DSO/TSO coordination should beapplied to avoid such conflicts. To achieve this coordination a new process could be defined, theDSO congestion forecast linked to the respective process at TSO level.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

7. INTRODUCTION (PART 2)

As described in the first part of the report, this study particularly considers theplanning/operational efficiency potential in electricity grids, thanks to new “Smart” measuresallowing active grid management (Smart Grid assets; flexibility of distributed energy resources.

Accordingly, the focus of this second part of the report is to analyse the potential of demandresponse in bringing efficiency in electricity grids for Distribution System Operators (DSOs).

The analysis has been organized in four main sections:

a) Synthesis of DR drivers, characteristics and value streams

- Brief review of benefits of demand response for the whole power system- Review of main factors affecting the value of demand response

DR characteristics (pre-activation time, duration, ..)Type of consumers, type of loads etc..

- Activation mechanisms for demand response (Price-based/Implicit and volume-based/explicit)

b) Synthesis of the value of DR for DSOs

- Focus on the benefits of demand response specifically for DSOs- Definition of DSO’s options to procure Demand Response for their needs: tariff-based

(implicit) demand response and volume-based (explicit) demand response

c) Simulation analysis of DR impact on grid efficiency at planning and operation

- Definition of four different network models to represent contrasted conditions (urban/rural;North/south)

- For each model, definition of four different scenarios, each with a growing level ofavailable flexibility for the DSOs in planning and operating its grid:

Base case: no flexibilityToU scenario: Demand flexibility triggered by (Static) ToU tariffsVolume-based DR scenario: DSO has access to volume-based demand responseFull flex scenario: DSO has access to two options of flexibility: (volume-based)demand response and DG curtailment

- The goal of the simulation analysis is to assess how the integration of flexibility optionscan reduce the CAPEX (investment in assets) and OPEX (losses) investments of DSOs, inpresence of a growing penetration of DG.

d) Regulatory evolutions to promote DR, particularly for grid efficiency

- Four specific regulatory matters have been analysedEvolutions of grid tariffs to trigger (price-based) demand response in line with theneeds of the system at local levelEnabling DSOs to integrate (volume-based) demand-response for planning andoperation. This requires the introduction of suitable procurement options (local markets,tender etc) and of specific incentives to DSOs to use these options (e.g. TOTEXremuneration).





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

8. OVERVIEW OF DEMAND RESPONSECHARACTERISTICS

This chapter describes briefly the main characteristics of Demand Response (DR).

The goal is to describe the general context of current DR implementations before focusingspecifically on how DR can be used by system operators to increase the planning and operationefficiency of their grids. This will be the subject of the subsequent chapters.

8.1. Benefits of demand response across thepower system value chain

There are many potential benefits arising from DR across the whole power system value chain.Figure 15 shows a wide spectrum of potential DR benefits, according to three main categories,namely benefits for the system, the society and other, non-monetized ones.

Figure 15 - Classification of demand response benefits across the whole power system value chain

A few examples of current valorization of DR for different system services are reported below:

Ancillary services

DR is contracted by system operators to provide ancillary services to the power system (e.g.frequency regulation, tertiary reserve). The capacity of DR contracted in these markets/programsis on the rise in many countries, as an alternative to the use of traditional generation resources. Afew examples are reported below:

France - 400MW contracted DR by RTE (French TSO) for balancing market (tertiaryreserve) [66]. Since July 2014, DR can also access the frequency reserve market (10 MWminimum bid).





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Italy - OOver 4000MW of Interruptible contracts for Sicily and Sardinia have been contracted(based on auctions) by Terna (Italian TSO)Belgium - 350 MW of DR were contracted in the ancillary reserves market in 2014

Capacity Market

These markets value the contribution of DR resources to meet system peak demand and thusensure the security of supply of the system. The integration of DR resources can reduce the needfor investments in generation peak units. A few examples are reported below:

PJM (USA) - 10GW of DR (mainly from Commercial & Industrial –C&I- sites) have beencontracted on PJM capacity market in the latest auction for 2017/2018. To provide anindication of the value associated to the DR, the flexibility service provider EnerNOC ismanaging 4GW of DR from C&I sites on PJM capacity market for total value of 185M$. [67][68]Belgium – Since 2014 demand resources (with capacity larger than 50MW) are contractedwith 1 year contracts for provision of strategic reserveFrance and the UK are also about to launch capacity markets which foresee the participationof DR resources.

Distribution network

The integration of DR in distribution grid operations is starting to emerge in Europe, particularlyin pilot projects. In the US, particularly in regions where integrated utilities operate, DRprograms are in place to solve distribution congestion issues.

California (USA): in 2013, 923 MW of DR (direct load control) had been contracted bydistribution utilities to address congestions at distribution level.

Additionally, DR is also currently used as a resource by market actors (e.g. BalancingResponsible Parties) to balance their portfolio.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

8.2. Factors affecting the value of DemandResponse

In the following, we describe a number of key variables affecting the potential and value of DR.

8.2.1. Type of loadsType of consumers

DR potential vary largely among sectors: Industrial (e.g. paper, steel, chemical industries etc.),commercial (hospitals, commercial sites, ..), residential (at household level).

Currently, DR comes mainly from commercial and industrial sector (lower implementation costper MW of flexible demand). Residential DR is expected to increasingly become cost-effectivewith the deployment of smart meters and the introduction of smart appliances (e.g. smartthermostats) able to automatically respond to control signals, without the need for finalconsumers to manually adapt their consumption.

Type of demand response resources

The flexibility potential of prosumers/consumers depends on the type of loads and generationassets on their premises. A useful categorization is provided in [69]:

1) Storable loada) Battery storageb) Electric Vehicles (smart charging)c) Thermal storage (e.g. HVAC, heat pumps)d) ..

2) Non-storable loada) Shiftable load, for example:

i) Flexible industrial processesii) Dishwashers, laundry,..iii) ..

b) Curtailable load, for example:i) Non-essential lightingii) ..

c) Base loadi) Essential services/Processesii) ..

3) Self-generationa) Use of generation at consumers’ premises to modify net-load profile (e.g. via PV; CHP)





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Load patterns

Load patterns affect the potential to provide DR services. For example, in summer the potentialfor DR of residential heating is zero. Another example is reported in Figure 16, where the loadpatterns of a residential water heater has already been optimized to peak/off-peak tariffs (seegreen profile) and thus does not leave any room for additional flexibility during the day.

Figure 16 – Example of load shapes for HVAC in office building and residential thermal loads

8.2.2. Characteristics of demand responseIn markets where DR is sold as a ‘product’, a number of characteristics are introduced to defineits ‘quality’ and thus its corresponding value.

Key parameters include:

Availability (e.g. weekday; summer period)Maximum possible number of DR activations per year (e.g. maximum 10 demand responseevents per year)Hours of day required to respond (e.g. 10am – 18 pm)Maximum duration of DR event (e.g. 4 hours)Time to respond (e.g. 4 hours after instruction)

For example, the UK TSO National Grid’s program to procure short-term balancing reserves(STOR), prescribed the following requirements for the provision of additional active power fromgeneration and/or demand reduction [70]:

3MW or more of generation or steady demandDeliver within 4 hours from instructionProvide for at least 2 hours20 hours recoveryAbility to provide STOR at least 3 times a week





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Figure 17 reports the requirements for different DR products (Limited DR, Extended SummerDR, Annual DR) for the PJM capacity market in the US22 [71]:

Figure 17 - Requirements of DR products on the PJM capacity market [PJM2012] -

8.2.3. Activation mechanisms of demand responseActivation mechanisms of demand response can be categorised into price-based (implicit) andvolume-based (explicit) procurement.

8.2.4. Price-based mechanisms (implicit DR)

Under price-based procurement, DR is considered “implicit”, in the sense that the activated levelof flexibility is not stated in advance, but results from the response of consumers to price signals.

The main parameters to consider are:

Price : Variable or Fixed (known in advance)Timing slots : variable or fixedOccurrence: every day or on critical eventCapacity/Energy : Charges depending on cumulated energy consumption or on power level

The combinations of these parameters create a wide variety of tariff schemes. Most relevant onesare briefly described below:

Inclining Block Rates

Fixed tariffs are applied by trench of consumed energy, hence the xxkWh are at x€/kWh, thefollowing yykWh at y€/kWh, etc.

Time-of-Use

Fixed periods of low and base tariffs all over the year. Typical example are day-night tariffs

22 DY stands for Delivery Year





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Peak Time rebate

The base tariff remains the same for all timeframes. However, if consumption is reducedduring critical events, customers can earn rebates

Critical peak pricing (CPP)

A high price is applied in a predefined number of days. The days where the CPP will beapplied are communicated in advance

Variable peak pricing

Mix of ToU and Real-Time pricing, where off- and on-peak periods are fixed, but the price ofthe on-peak period may vary depending on market conditions.

Real Time Pricing

Price varies hourly or sub-hourly, on a day-ahead or hour-ahead basis.

Each pricing scheme has advantages and drawbacks. From a wider perspective, they can beranked according to the level of Risk and Reward they provide. The higher the inherent risk, themore rewarding they can be for consumers who adapt their energy usage patterns (see Figure 18).

Figure 18 - Risk/Reward levels are adaptable to consumer's risk aversion [72]

8.2.5. Volume-based mechanisms (explicit DR)

Under price-based mechanisms, there is no guarantee that a predetermined amount of DR will beactivated. On the other hand, in volume based DR a predetermined volume (MW) of DR iscontracted. In this case, DR is considered “explicit”, in the sense that the activated level offlexibility needs to be explicitly defined in advance.

A key aspect of volume-based DR is the important role that control systems should play toautomatically control the load and make it “dispatchable’, in response to dispatch signals (e.g.from an aggregator).

Volume-based DR products are typically contracted on centralized organized markets (capacitymarkets, ancillary services, energy markets etc.), depending on the specific regulation in force.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Verification of activated DR

Volume-based demand response requires the definition of a mechanism to evaluate ex-post thelevel of DR that has been activated. For illustration purposes, three examples are provided belowrelated to the baseline calculation (i.e. the profile the load would have had, if the DR event hadnot taken place) in the PJM market (US), NEBEF market (France), STOR program (UK):

PJM (USA) [71]

Different approaches are used by PJM to calculate baseline of remotely metered sites. Thegeneral principle is that the error of the baseline (relative root mean square ) must be less than20% so that the site can actually be paid for load reductions .

The most commonly used baseline (about 85 % of participants) for participants in "PJMeconomic programs” is the average consumption measured during the same hours as those ofthe event during the four days 5 previous busiest days of activation.

Different methods are used for the different types of demand management capacity markets:

- Direct Load Control (DLC): In this case there is no need for tele-reading of meters or touse a baseline. An impact study is carried out ex-ante to assess the volume of DR activatedby sending the DLC signal (i.e. testing the functioning of the load control loop)

- Firm Service Level (FSL): No use of baseline is needed, the verification consists inmonitoring (via remote metering) that the load has well been reduced to a predefined level.

- Guaranteed Load Drop (GLD): the load must in this case be reduced by a predeterminedamount (Guarantee) and a baseline must be calculated based on remote meteringmeasurements. Adopted baselines could be based on comparable days or the day before /after

STOR (UK) [70]

Participants to the short term reserve program (STOR) of the UK TSO (National Grid) weretypically on call roughly between 7:00-13:00; 16:00-22:00.

The contribution of participating customers was judged relative to their level of consumptionimmediately prior to the dispatch signal.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

9. VALUE STREAMS AND PROCUREMENTOPTIONS OF DEMAND RESPONSE FORDISTRIBUTION SYSTEM OPERATORS

This chapter focuses specifically on the use of DR by DSOs. We analyse the key DR valuestreams for DSOs and present the options DSOs have to procure DR and capitalise on some ofthese value streams.

9.1. Value streams of demand response fordistribution system operators

From the perspective of a power system operator, DR can be a useful tool to correct imbalancesbetween power supply and demand and to relieve congestions in the grid. Specifically, thefollowing functions can be considered:

Reduce peak power demand, thereby potentially reduce line overloads and relieve congestionsIncrease demand in order to complement production of renewable energy sourced electricity(RES-E) to reduce reverse power flows and avoid curtailmentInfluence power flow in the network (quantity and direction) to balance 3-phase loading andto eliminate the need for network reconfiguration

Figure 19 shows a general overview of benefits that can be derived from DR in the power systemin general, in terms of optimising operation and planning.

Figure 19 – Grid efficiency value streams deriving from DR for the power system





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

DSOs are facing a twofold challenge. On the one hand distribution grids should cope withdemand growth and changing demand patterns (due to the incorporation of new demand such aselectric vehicles and heat pumps as well as increasingly wholesale-price sensitive demand). Onthe other hand, they should be able to host increasing shares of DG and especially renewableenergy sources (RES-E). DSOs are therefore faced with increased grid congestions caused eitherdue to overloads from new demand or due to reverse power flows from increased RES-E shares.DR can provide cost effective solutions for both planning and operational purposes to relievesuch grid congestions by reducing demand in case of overloads or increasing demand to counter-balance reverse power flows. In this respect, there are two main categories of DR benefits forDSOs:

Avoided capital expenditure (CAPEX): DSOs need to expand grid capacity and upgradetransformers to cope with congestions. Depending on the demand and renewable energysources (RES-E) growth in an area, employing DR in the planning phase offers DSOs avaluable solution to asset replacement. This strategy may postpone the need to invest ininfrastructure upgrades and allow using the assets for their full lifetime, reducing short-termexpenditures.23

Avoided operational expenditure (OPEX): local balancing leads to a more balanced lineloading and to a subsequent reduction of distribution grid losses which form a key operationalexpenditure for DSOs. Moreover, by relieving grid congestion, DR allows RES-E generationto be better integrated in the system. This means that curtailment of RES-E can be avoided,and hence compensation payments for the curtailed renewable energy are reduced (incountries where it is foreseen).

It is worth mentioning that, while DR can also be used to optimise operation of the wholesale andbalancing markets (as discussed in the previous chapter), DSOs cannot capitalise on thesebenefits, as unbundling provisions restrict their participation to these markets as flexibilityproviders.

23 Grid assets must be replaced at the end of their lifespan regardless of the situation. Since large part of theinvestment costs are related to installation, when replacing assets that have reached their lifetime, placing somelarger cables brings no significant additional cost. In this respect, CAPEX benefits from DR should be considered withrespect to the remaining lifetime of the assets.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

9.2. Options for DSOs to procure demand responseand capitalise the benefits

Figure 20 shows an overview of DSOs’ options to procure DR within its control area, classifiedinto price-based and volume-based options24. The individual options are explained in more detailin the subsequent sections.

Figure 20 – Overview of transaction options between DSO and DR service providers

9.2.1. Price-based procurement of demand responseDSOs can use price-based incentives to procure DR. This means using the regulated componentof the final price which reflects distribution grid charges (hereafter denoted, “grid tariffs”) toincentivise grid users to adjust their behaviour to meet grid conditions.

Figure 21 lists the options to use grid tariffs to trigger demand flexibility. Basically the optionsare to vary the distribution grid tariff based on a) fixed timeframes (time of use pricing); b)changing timeframes (dynamic pricing); or c) changing timeframes and location (locationaldynamic pricing). The different options are best suited to different purposes.

24 It is worth stressing that when referring to “price-based” options, since we are referring to DSOs, in this contextwe refer to the “regulated component” of the final price for consumers (in most cases this refers to the distributiontariff component of grid tariffs). In countries where market is not yet liberalized (where retail prices are stillregulated) “tariffs” refer to the whole price.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Figure 21 – Options on how to vary distribution grid tariffs to incentivise DR

9.2.1.1. TIME OF USE (TOU) TARIFFS

ToU tariffs apply fixed price differentiation of the network charges based on time. Consumers ofelectricity are charged cheaper tariffs during low-demand periods (typically night time), whilemore expensive tariffs apply during high-demand periods. The same tariff is charged to allconsumers in the DSO control area. This tariff structure is useful for reducing system peakdemand, when it is known to occur regularly around the same time every day.

Example of application: ToU tariffs have been used in France since 1960’s. Approximately 80%of the hot water tanks in France are programmed to start during off peak, which is estimated toshift 9 TWh of energy annually, corresponding to approximately 3 GW of installed peak loadcapacity. Synergy effects are reduced investment needs in the grid of more than 100 MEUR/year.[73]

Potential Challenges:

One observed side-effect from the example in France, was the so-called “bounce back effect,”where all DR switch on simultaneously at the beginning of the cheaper tariff period, causing anew peak. Pilot projects (e.g. NiceGrid project in France25) show that this can be successfullydealt with by using more than one off-peak period (shown in Figure 22).

25 http://www.nicegrid.fr/





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Figure 22 – Example of application of time of use tariffs by ERDF in France [74]

ToU tariffs are suited to ‘traditional’ power systems, where the daily patterns of (net) demand iswell defined, and the peaks occur consistently in the same timeframe. However, increasing sharesof Variable RES (VRE) in distribution systems alter the net demand patterns and the peaks do notalways occur in the same timeframe. This would make the ToU tariffs ineffective, and reduce thepredictability on what tariff should be applied when.

The same ToU tariffs apply to all consumers in the control area by the DSO, and act to solvesystem peak demand. However, this does not guarantee a solution to local congestion issues. Forexample, if there is a local congestion within the DSO area that requires a decrease in generationoutput during the time when there is a system peak and instead demand reduction is encouraged,a conflict between local grid conditions and system needs arise. In that event, solving the localproblem should be prioritised but this is not possible with ToU tariffs.

9.2.1.2. DYNAMIC TARIFFS

Dynamic tariffs apply price differentiation based on the current operational state of the system.This means that, if the system has abundant supply, the price signal is lower to encourage moreuse of electricity, and if there is a supply deficiency the price signal is increased to deter the useof electricity.

The difference between the time of use (ToU) tariffs and the dynamic tariffs is mainly that theToU tariffs originate in fixed peak/off peak hours (easy to forecast even with long planninghorizon) whereas the dynamic tariffs consider the current state of the system, not necessarilyfollowing the traditional peak/off-peak pattern. This makes them suitable for systems with VREpenetration where the timing of the system peaks is varying from day to day.

The same tariff is charged to all consumers in the DSO control area, therefore, the price signalscan be coupled with the state of the system and with global system needs, which can be definedas either:

a) TSO system imbalance: Higher or lower price can be applied during timeframe wheresupply-demand imbalance is forecasted (i.e., high consumption or high generation); or





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

b) TSO regional imbalance: variable price can be applied when there is a foreseen transmissiongrid congestion, based on congestion forecasts at a regular intervals (e.g. hourly).

Example of application: Dynamic tariffs are proposed by the Ecogrid EU26 market concept,which is illustrated in Figure 23. In this example, when there is an imbalance (type a or b above)detected in the system, the price would be set by a “real-time market operator,” which might bethe TSO or DSO (1). Therefore, depending on the state of the system, the real-time marketoperator will increase or decrease the tariff in order to encourage consumers in the control area ofthe DSO to decrease or increase electricity consumption. The price would be published andretrieved by the market actors, which include the DR service providers (2). If no imbalance orcongestion exists, the real-time price will be equal to day-ahead price. The participating actorswould then adjust (3) the consumption according to the price signal and thus assist in maintainingthe power system balance. The “real-time market operator” can monitor the impact to verify thatequilibrium was established (4). The price would be updated frequently, every 5 minutes, toutilise the potential for a dynamic response.

Figure 23 – The Ecogrid EU market concept architecture: prices set by the TSO and its impact on the market actors [75]


Even if the grid tariff is coupled with a price that represents the state of the system at a TSOlevel, (i.e., wholesale price, balancing price) and acts to solve system imbalance, it does notguarantee a solution to local congestion issues. For example, if there is a local congestion withinthe DSO area that requires a decrease in demand during the time when there is a systemimbalance that signals the demand to increase instead, a conflict between local grid conditionsand system needs arise. In such cases the local problem should be prioritised, which is notpossible with this tariff system.

26 www.eu-ecogrid.net/





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Distribution grid tariffs in most countries constitute a part of the whole electricity price [76]. Insome cases, the impact it has on the end price-signal to customers could be limited to the pointthat it becomes ineffective to steer changes in consumption behaviour.

9.2.1.3. LOCATIONAL DYNAMIC TARIFFS

Locational dynamic tariffs apply price differentiation based on time and location. This meansthat, dynamic pricing is applied, but the price can be different for service providers at differentlocations within the control area of the DSO. Therefore, a variable price is applied based onwhere distribution grid congestion appears.

Example of application: Locational dynamic tariffs are proposed by the Ecogrid EU marketconcept27, which is explained in the example in the previous sub-section. A full-scaledemonstration of the concept is foreseen in a complete distribution system (Bornholm island),owned and operated by the local DSO, Østkraft, in Denmark28. The demonstration is run in threephases, as shown in Figure 24. Basically system balancing, shown as Production & Consumptionin the centre of the figure, will consider three elements, in different phases. In the first phase, theprovision of global balancing services by DER and DR using energy and ancillary service pricesignals is demonstrated. This will use the framework for price signals as described in the examplein the previous sub-section on dynamic pricing (the market perspective). In the second phase,grid supervision and management technology is added to automate the operation of thedistribution system (the network perspective). In this phase, local congestions will be identifiedand flagged. In the third and last phase, locational pricing is developed based on the localcongestion signals generated in the previous phase (combination of both market and networkperspective). Flexible resources would then be solicited with price signals to alleviate gridcongestion.

27 http://www.eu-ecogrid.net/ecogrid-eu/the-ecogrid-eu-market-concept

28 The Bornholm distribution system is part of the Nordic interconnected system and fully integrated inthe Nordic electricity market.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Figure 24 – The Ecogrid EU demonstration phase [75]


DSO would have to anticipate local congestions and set a price signal based on that forecast. Thismeans that DSOs should develop new tools and processes.

9.2.1.4. POWER BAND CONTRACT

Power band contracts apply a price differentiation on the energy charge component based on themean power of the peak hour. This means that the higher the power demand of the peak hour,the higher the tariff for the customer. This dictates the EUR/kWh rate that a consumer has to payfor a given period (e.g., could be applied for the entire consumption in a year, or only the hourscorresponding to that peak hour). This method encourages reduction of peak power by thecustomer and essentially creates a dynamic link between the energy and capacity chargecomponents.

Example of application: Results from a pilot study in Finland [77] are shown in Figure 25 andFigure 26. Normally, a customer is billed based on the main fuse size. For a main fuse of 3 x25 A, the largest power band (utilised capacity) measured as an hourly power would be approx.17 kW (red line in Figure 25). With a flat-rate EUR/kWh tariff which is indifferent to the level ofhourly consumption, there is no incentive to reduce consumption in a specific hour over another.However, if the tariff is differentiated across the coloured bands (increasing with increasedpower), the customer would be incentivised to minimise peaks and try to keep the hourlyconsumption as low and even as possible. Figure 26 shows how the consumption pattern of acustomer was changed over a period of three years since their electricity bills were changed to apower band contract. It can be seen that both the mean and peak power consumption havedecreased.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Figure 25 – Hourly electrical energy consumption of a Finnish household with a flat rate electricity tariff. Superimposedwith potential power bands [modified from [77]]

Figure 26 - Hourly electrical energy consumption over three consecutive years of a Finnish household with a power bandcontract. [77]





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval


Consumer has to monitor/study his own consumption and understand when peak consumptionsoccurs, and how to avoid it, which complicates the applicability of the approach.

9.2.2. Volume-based procurement of demand responseDSOs can use volume-based incentives to procure DR. This means that the DSO will(financially) compensate a DR service provider to activate a specific quantity of DR.

Figure 27 – Volume-based procurement options of DR

Figure 27 lists the transaction options for DSOs to procure a certain volume of DR. When theDSO identifies a need/potential benefit for DR in its network, it can procure it by:

making a direct contract (bilateral agreement) with the DR provider with customisedconditions;issuing a call for tenders for the DR service required, whenever it is required; orestablishing a local market for DR providers connected to the network with local price.

Which option is most efficient and effective depends on the number of DR providers that areconnected to the network (liquidity of the market). Forming bilateral agreements are mostefficient when there are only few, and forming a local market can be efficient when there aremany market players.

Bilateral agreements and tenders establish a direct transaction between the DSO and local DRservice provider based on pre-defined conditions and price, while a local market would mean thatthe DSO would source DR from a market which sells standardised products.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

DSOs as regulated entities have a non-discriminatory obligation and must carefully observemarket power effects. Therefore, it is important to ensure sufficient liquidity in a market.Furthermore, reliability of DR delivery should be managed, for example by applying penalties ifthere is a deviation between services sold and delivered (except in an emergency). However, itmust also be kept in mind that very strict reliability standards will discourage service providersfrom participating in such a DR scheme.

Once a volume-based procurement method is established, the DSO could activate the actualservice in either of the following ways (see Figure 28):

Curtail, which means uncompensated automatic control of DR-assets to curtail on pro-rataacross all assets; orRe-dispatch, which is a more a market based solution, to compensate DR who can offer theservice at cheapest cost.

Figure 28 – Curtailment and re-dispatching process aggregator DSO [78]

9.2.2.1. BILATERAL AGREEMENTS

Bilateral agreements are contracts signed between the DR service provider and DSO. Thecontract would specify the conditions when and how the DR service provider would eitherincrease or decrease consumption. Key benefit is the ability to customise the servicerequirements.

Such contracts may be standardised for simplicity, but should retain the freedom to formulate theterms in order to fit the needs and preferences of the involved parties, e.g., in theory they couldhave any time horizon or format. This is one of the benefits of bilateral contracting, comparedwith other procurement designs. Such flexibility allows bilateral contracts to be used for a widerange of services.

One example of a standardized approach to contracts is to revise the calculation method for thefixed charge component of the distribution tariff, taking into account flexibility obligations. Thiswould apply to users (DG and consumers) connecting to the distribution network, paid eitherdirectly to the DSO (i.e., customer has separate contracts to utility as energy supplier and DSO asconnection supplier) or paid by consumers to utilities, which then pay the DSOs.


The nature of the product specifications in bilateral agreements makes it hard for traded servicesto compare with other similar non-standardised products. Without clarity on how to developcompetitive products, it could discourage liquidity in the market. Furthermore, it may be seen bythe regulator as going against the transparency and market based approached directed by the EU.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Because the products are unique and not standardised, the value of the product becomessubjective. The seller and buyer of the service need to have a good understanding of the serviceoffered, and also have an agreement about its value.

9.2.2.2. TENDERS

Call for tenders is a special procedure for generating competing offers from different bidders.The DSO would announce the framework of the service contract to all interested serviceproviders. They would then assess the offers and offer binding service contracts to those thatcomply with the stated requirements. Major benefit is the ability to customise the servicerequirements.

The content of the contracts would be similar to that of a bilateral agreement, in that it wouldspecify the conditions when and how the DR service provider would either increase or decreaseconsumption. The main difference is that with a tender, the content of the contracts are moretransparent compared to a bilateral agreement, due to the tendering process.

The framework of a contract may include the following [78]:

Type of DR product (firm or optional, then notice of exercise)Volume and duration;Specified location;Condition of delivery;Invoicing and payment;Measurement;Penalties for failure to deliver;Non-performance due to force majeure,Confidentiality obligations;Governing law and arbitration.

Example of application: In the United Kingdom, UK Power Networks has set-up bilateralcontract following a call for tenders with aggregators Flextricity, EDF Energy and Enernoc forthe provision of DR service to limit power congestions in selected distribution grid areas andavoid reinforcements [79] [80]. UK Power Network’s business plan forecasts savings of around£40 million [€52M] from DR schemes from 2015 until 2023 [81].


To some degree, the potential challenges are similar to those faced by bilateral agreements. Thetransparency of product description may be improved in the case of an open tender. However, theability to customise the technical requirements means that there will be a lack of standardisationacross different DSO areas. This could inhibit liquidity and competition from supply side.Furthermore, the administrative burden to manage a call for tender could be considered high for aDSO.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

9.2.2.3. LOCAL MARKETS

National markets in the form of system-wide balancing are already used for contracting DR.However, no markets exist for solving issues from the DSO perspective only.

It could be possible to make use of this existing framework and treat the activation of DR forDSO purposes as a “bonus” which is granted depending on the localisation of the flexibility andits activation according to DSO’s needs

It is also possible to have local markets set up by the DSO or another regulated agent, toexchange services locally. The design of the market could be very similar to the existingwholesale markets. The local market could be obtained by applying a market splitting mechanismfrom the TSO level market in case of a local congestion or problem. A local market approachwould be considered in the case of interconnected networks with a lot of local resourcesconnected to them and thus possible liquidities on the market.

Example of application: Local markets at pilot level (e.g. NiceGrid project in France; E-telligence project in Germany).

The eTelligence project in the Cuxhaven region of Germany [82] developed and tested a complexICT-based system to balance the intermittent feed-in from wind energy generation by integratingelectricity into the grid and a regional market, while simultaneously enabling a more activeparticipation of small end-consumers into power markets of the region. A digital energymarketplace was created, to enable small and medium-sized power-generating systems, as well asmedium to large electricity consumers, to trade energy products in a fully automated way. Themarketplace considers the availability of regional products under realistic conditions (grid29 andgeneration capacity) and links together producers, commercial consumers with flexible loads andenergy service providers. A link to the wholesale market was also included, to ensure sufficientliquidity on the marketplace. Some of the market participants included EWE NETZ GmbH,various cogeneration plants and a virtual power plant consisting of a PV system, a wind farm andtwo refrigerated warehouses.


The more local a market is the less liquid it becomes. Besides, even if a local market is created, ifthe DSO becomes the single service buyer, the local market would yield the same result as a callfor tenders. Therefore, the choice for a DSO between a call for tender or holding his own marketshould be made based on the possible liquidities on the market and especially the market optionshould be considered if there are other actors willing to buy the DR service. The only other actorthat would be willing to buy local DR services, besides the network operator, would be theproducers with regulated tariffs. Furthermore, in a meshed network with interconnectionsbetween different DSOs, it would be possible for a DSO to become buyer on his neighbouringDSO.

29 In order to understand the network behaviour, network measurement data (active and reactive power, voltages,currents, frequency) was collected from 100 substations in the distribution network, saved and exportedon request. The data was recorded for the participation of the network in the regional marketplace. [100]





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

If a DSO contracts DR to solve local constraints within a short timeframe, it will have to ensurethat the total balance at the point of connection with the TSO is not compromised, otherwise thepower flows at the TSO level and global system balance may be disrupted.

Furthermore, the administrative burden to manage a market would be considered high, especiallyfor small DSOs.

9.2.3. Procurement options: ConclusionsStatic priced-based procurement options like the ToU tariff and power band contracts could berelatively easy to implement. ToU tariffs would be defined ex-ante by the regulator, based on ananalysis of historical data about peak demand. Power band charges would be applied ex-post, andcould be integrated into existing billing processes.

Dynamic tariffs require the definition of a “real-time market operator” (i.e., DSO), and this entitywould have to set up a platform to send a signal to DR service providers about the tariff changein a given time interval depending on grid conditions. This setup is complicated compared tocurrent practices, but would be fairly simple to implement from a technological point of view,which would require a one-time effort to set up the system, and then could run as an automatedprocess.

Locational dynamic tariffs, which are based on the state of the system in terms of gridcongestions close to real-time, would require much more frequent forecasting and monitoring ofthe power flows in the system compared to current practice by TSOs and DSOs. Thereforeintensive efforts would need to be invested to realise locational dynamic tariffs.

Bilateral agreements and tenders establish a direct transaction between the DSO and local DRservice provider based on pre-defined conditions and price, while a local market would mean thatthe DSO would source DR from a market which sells standardised products. However, thecreation of local markets or even tendering would require liquidity in the market. Furthermore,the additional administrative burden could be difficult for small DSOs to manage. In this respect,bilateral agreements and tendering are probably the most accessible and effective options fromthe perspective of a DSO.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

10. ASSESSMENT OF THE VALUE OF DEMANDRESPONSE TO IMPROVE DISTRIBUTION GRIDEFFICIENCY AT PLANNING AND OPERATION

10.1. Objectives and global approach

The main objective of this chapter is to study, via simulation analysis, how flexibility canimprove the efficiency of distribution networks, with an increasing DG penetration levels, whilerespecting operation rules (Min – Max voltage limits, conductors and transformers load rate, N-1security).

Specifically, we assess how the integration of DR in DSOs planning and operation could allowfinding the optimal trade-off between CAPEX (investment in equipment/asset replacement andupgrade) and OPEX spending (in particular cost of thermal losses) (see Figure 29).

Figure 29 – Planning of distribution grids accounts for security constraints and OPEX/CAPEX trade-off

Figure 30 summarizes the three-step approach that has been used in the analysis.

Figure 30 – Followed approach for studying benefits brought by flexibility





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Modelling approach: “Hosting capacity of existing grid” vs “Greenfield ideal network”

To carry out this kind of analysis, several studies follow a classical “hosting capacity” approach.This consists in finding the maximum amount of DG that an existing grid can integrate safelywithout operational security violation.

In this study, we follow a “greenfield” approach, which allows exploring the benefits of acomplete new approach of DSO grid planning and operation which includes flexibility. It consistsin designing the ideal network structure to integrate a long term target level of DG at least cost(CAPEX+OPEX), given the load, generation and grid characteristics in specific regions. Thisallows drawing general conclusions about the value of integrating DR into DSOs’ planning andoperation.

Smart sizing simulation tool

The Smart Sizing tool, developed by Tractebel in Smart Grid pilot projects like PlanGridEV30 andGredor31, has been used for this study. It allows modelling both technical (e.g. networkcomponents, DG) and economic aspects (CAPEX and OPEX) when planning investments in thedistribution grid.

Figure 31 summarizes the Smart Sizing tool. In the model, the following cost items are included:

OPEX:

Technical losses costsActivation cost of the flexibility (e.g. procurement of volume-based flexibility by the DSO)

CAPEX

Network components and installation (e.g. lines, substations)ICT infrastructure for activation of flexibility

This tool allows a global optimization of OPEX and CAPEX by selecting the optimal mix oftraditional and flexibility-based investments.

30 www.plangridev.eu

31 https://gredor.be





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Figure 31 - SMART SIZING TOOL: global optimization of OPEX and CAPEX

10.2. Building the simulation models

The following realistic four distribution system models have been built based on consultants’own collected data and experience in previous assignments.

Distribution network in urban area located in Northern Europe;Distribution network in rural area located in Northern Europe;Distribution network in urban area located in Southern Europe;Distribution network in rural area located in Southern Europe.

These models are intended to provide a contrasting set of operating conditions to test therobustness of the simulation results. Realistic load and decentralised generation data profiles,coherent with grid characteristics and geographical locations, have been considered.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Figure 32 - Four distribution systems models

The modelling phase has been organized in four specific steps:

a) Modelling the grid structureb) Modelling distributed generationc) Modelling the loadd) Analysis of representative days of the year

a) Modelling the grid structure

The specificities of the grid are represented by parameters like

Type of conductor (overhead/underground)Network structure (radial/meshed)Average feeder lengthsAverage number of feeders per substations

The associated costs of the assets (e.g. cost of cables per km) have then been defined.

b) Modelling the distributed generation

Figure 33 summarizes the characteristics of DG in each model: the type (PV, wind, CHP)and the associated voltage level (LV, MV, connection to HV/MV substation).

This repartition has been done based on the information given in [24]. Accordingly, urbanareas are mainly characterised by PV connected in LV while rural areas host a larger portionof wind connected in Medium Voltage.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Figure 33 - Characteristics of generation for each model – Peak load and share across voltage levels

Figure 34 reports selected generation profiles, representative of local conditions (level of sunirradiation, wind speed, district heating system, and individual cogeneration system). Urban areasare mainly characterized by PV connected in LV (low voltage) and in MV (medium voltage),while rural areas are characterised mainly by wind connected in MV. Figure 35 reports the shareof DG across voltage levels in the four models.

Figure 34 - Characteristics of generation for each model – Generation profiles





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Figure 35 - Voltage level of connection of Distributed Generation

c) Modelling load

Figure 36 reports the characteristics of the load considered for each model, in terms of peakand load repartition across voltage levels.

Figure 36 – Characteristics of load for each model – Peak load and share across voltage levels

Figure 37 summarizes the different hourly load curves that have been considered to reflect thelocal specificities of each model (density of customers, categories of customers, economicalsectors). The type of houses, heat & cooling types, thermic insulation, and seasonal variationsimpact the residential load pattern.

The level of LV and MV commercial load depends of the area density (urban vs rural).Industrial loads are mainly located in urban areas and connected in MV.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Figure 37 - Characteristics of generation for each model – Load profiles

d) Modelling representative days

To reduce computational complexity, simulations have been run on a sub-set of daysrepresentative of the annual behaviour of the distribution system (Figure 38).

A total of 10 days (non-consecutive) have been selected. We have considered days with“typical conditions”, and days with “extreme conditions”. The latter are days in whichgeneration and load profiles are particularly constraining for the grid in terms of peak load orreverse peak.

Days with “Typical conditions”Winter weekday, summer weekday, inter-season weekday;Winter weekend, summer weekend, inter-season weekend;

Days with “Extreme conditions”Total peak load (net load) and reverse peak (net load);LV peak load (net load) and LV reverse peak (net load).

The probability of occurrence of the different days has then been considered to weigh thesimulation results associated to the corresponding day.

Figure 38 - Selection of 10 days to be representative of the annual behaviour





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

10.3. Definition of scenarios and simulation results

For each network model defined in the previous section, four different scenarios have beendefined according to the level of flexibility at DSOs’ disposal:

1) Base case scenario – In this scenario, planning and operation rely only on traditionalinvestments, without the option of demand response. The impact of the total cost of thesystem is then studied for the 4 models (North Urban, South Urban, North rural, South rural)when increasing the Decentralised Generation (DG) penetration level.

2) Time-of-Use scenario - In this scenario, it is assumed a (exogenous) reshaping of the loadprofiles in the simulation thanks to the introduction of ToU tariffs. The tool does notoptimize the load profiles, it just considered the ToU-based load profiles as input to thesimulation.

3) Demand response scenario – In this scenario, the tool is able to arbitrate betweeninvestments in traditional assets or demand response (ICT investment costs and activationcosts).

4) Full flex scenario: Demand response and DG curtailment options– In this scenario,DSOs can opt for two competing sources of flexibility competing on a cost basis: DemandResponse and Curtailment of DGs.

Within each scenario, the behaviour of the system has been assessed by increasing progressivelyand homotetically the installed capacity of the different types of DG connected at LV and MV(Figure 39).

In each scenario, DG penetration level varies from 0% to 400% of peak demand (consideringwith 50% increment steps).

The same process is repeated for each of the four network models (North urban; North rural,South Urban; South Rural) leading to a total of 16 simulation models (Figure 40). Consideringvariations of parameters within each scenario (e.g. penetration level of DG), over 150 simulationcases have been analysed.

Figure 39 – Simulation scenarios for each network model





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Figure 40 – Global picture of the built models and studied scenarios to study all flexibility options

10.3.1. Base caseThe base case does not consider any type of flexibility. The increase of DG penetration leads toreverse power flows that can lead to a reverse peak. When the reverse peak is higher than thedirect peak load, it will size the system and will therefore directly impact the CAPEX (seeANNEX V for additional detailed simulation results).

Figure 41 shows the evolution of the TOTEX (CAPEX + OPEX) from 0% to 400% DGpenetration level. The TOTEX is expressed in percentage of the case with 0% of DG penetrationof the same model (it means that by example, north rural case with generation is compared withthe north rural case without generation).

Key observations:

The increasing penetration of DG has first a positive impact on the TOTEX up to a minimumcost value (optimal point). In fact, at first DG penetration allows to reduce flows in thedistribution grid (reduction of losses and of need for reinforcements).Beyond a certain DG penetration level (break-even point) TOTEX becomes actually higherthan in the case without any DG: this is due to the need to increase CAPEX to reach higherDG hosting capacity.DG penetration level corresponding to the optimal and the break-even points strongly dependson the characteristics of the load, of the share of generation and of the characteristics of thenetwork- For example; in proposed case studies optimal cost and break-even points are reached

quicker in urban areas. This is due to the fact that reverse power flows in LV (which havethe highest investment impact as they affect the whole distribution system) first occur inurban areas where it is assumed that a higher share of DG is connected in LV.

In rural areas, break even points are reached for a higher penetration level of DG compared tothe urban cases.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

- Rural areas are mainly characterised by overhead lines which are cheaper thanunderground conductors. The optimal investment strategy is thus to oversize conductors toreduce losses.

- Rural grids are more loss-driven, since the losses account for a higher share of the totalcost than in urban grids.

The cost of losses (€/MWh) strongly affects the optimal balance between CAPEX and OPEX.With higher cost of losses, the investment decisions are steered toward oversizing the grid(thus higher CAPEX) to minimize the OPEX of losses.For a given DG penetration level, variety in the DG mix (in terms of voltage connection andtype) provides positive impacts on distribution system costs compared to the case of one typeof DG connected at a single voltage level (e.g. only PV connected in LV)- Single connection point: In general, DG all connected at LV determines cost impact on all

upstream levels (both for CAPEX and OPEX)- Single DG type (e.g. only PV or wind): the concentration of generation in the same time

periods could lead to an increase of the reverse peak and thus to higher infrastructureinvestments.

Figure 41 – Evolution of the total cost for the 4 distribution systems

10.3.2. Time-of-Use scenarioIn the “Time of Use scenario”, three types of ToU tariffs with different on peak/off-peak timeperiods have been considered (Figure 42; Figure 43):

Night/Day: Off-Peak from 8pm to 8amMid-day: Off-Peak from 1pm to 4pmMulti-period: Off-Peak from 8pm to 8am and from 1pm to 4pm





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Figure 42 – Different ToU periods considered

For sake of simplicity; it has been assumed that the impact of ToU tariffs lead to a transfer of 5%of the energy from the on-peak to the off-peak period. The optimized load profile is then used asan input for the Smart Sizing simulation to assess the delta TOTEX compared with the base casescenario.

Key observations are:

Night/day ToU (Figure 43)

Positive impact on TOTEX for low penetration of DG, thanks to the reduction of peakdemand and associated CAPEX investmentsWith higher DG penetration, TOTEX starts to actually increase compared to the base case.Thi is due to the decrease of the load during the PV peak, leading to higher reverse flows

Mid-day ToU (Figure 43)

Reduction of TOTEX is achieved beyond a certain threshold of DG penetration, when higherpeak consumption provides benefits in balancing peak DG injection.

Multi-period ToU

It allows combining the benefits of both Night/Day and Mid/day ToU cases in a wide varietyof conditions (rural/urban, north/south, different DG penetration levels). Results are availablein Annex I.

Simulation results show that a well-defined ToU tariffs can indeed provide benefits in terms ofCAPEX and OPEX for the distribution grid. The level of impact strongly depends on the specificcharacteristics of the grid and of the load/generation conditions.

02468

10121416

0 10 20

Load

[MW

]

Night/day ToU

02468

10121416

0 5 10 15 20Lo

ad[M

W]

Midday ToU

LoadwithoutTOULoadwithTOU





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

A key aspect is to well design time periods in a way to incentivize a synchronization of local loadand generation. Under certain conditions, ToU might actually lead to negative TOTEX impacts,due to an increase of reverse power flows to a higher level than actual peak demand.

Figure 43- TOTEX reduction with Night/Day and Mid-day ToU tariffs

10.3.3. Demand response (volume-based)Under this scenario, DSOs have the option to select two types of investments:

Investments in traditional grid assetsInvestments in Demand Response

Two costs are associated to Demand Response:

Demand Response CAPEX: cost of deploying the ICT infrastructure to activate DemandResponse (e.g. Smart Metering infrastructure)Demand Response OPEX: cost of activation of demand response (which includes all the costsborne by aggregators to provide demand response services: e.g. energy boxes, aggregationplatforms etc.)

Thus, the optimization strategy aims at selecting the investments that minimize the total CAPEX(traditional assets + ICT assets) and OPEX (cost of losses + costs of DR activation).

In the following we assess how the availability of DR can reduce DSOs’ system costs comparedto the base case.

Clearly, these benefits will strictly depend on the costs (CAPEX and OPEX) associated to DR.Therefore in the following we will carry out two different types of analysis:

Case 1: Sensitivity analysis on DR activation cost (DR OPEX)

DR OPEX: Analysis of 4 different activation costs for DR,DR CAPEX: Negligible ICT infrastructure costs. This corresponds to a case in which smartmeters are already in place.

Case 2: Sensitivity analysis on DR ICT investment cost (DR CAPEX)

DR CAPEX: Use of two different ICT infrastructure costs of DR,

-2%

-1%

0%

1%

2%

3%

4%

0% 50%

100%

150%

200%

250%

300%

350%

400%

Night/day ToU

-2%

-1%

0%

1%2%

3%

4%

0% 50%

100%

150%

200%

250%

300%

350%

400%

Midday ToU

North rural

South rural

North urban

South urban





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

DR OPEX: Fixed activation cost of DR .

The proposed two cases represent the two extreme of cost allocation of the ICT infrastructure toDR programs: in the first case DR program does not borne any cost (e.g. because smart metersare supposed to be already in place) whereas in the second case it includes the whole ICTinfrastructure cost. This allows having robust results, as, depending on the conditions in differentcountries, intermediary cost conditions will hold.

Case 1 – Demand response scenario - Sensitivity analysis on DR OPEX

In this case, the cost of DR is limited to the activation cost and does not consider investments inICT. Several activation costs have been tested, based on an average end-user price of electricityof 200 €/MWh32 and are presented in the Table 20.

Nearly free 1€/MWh

Equivalent to a 10% reduction 20€/MWh


Equivalent to a 40% reduction(=Night tariff)

80€/MWh


Table 20 - Simulated Demand Response activation cost

Figure 44 reports the TOTEX variation as compared to the case without DG and withoutflexibility, for different DR activation costs.


DR has always a positive effect on the TOTEX in the different grid conditionsThe value of DR in terms of TOTEX reduction increases with DG penetrationEven in presence of relatively high DR activation cost (150€/MWh), DR brings benefits. Thishints that a business case for DR services to DSOs could be reasonably pursued.DR brings more values in the urban case, where DG is mainly represented by PV connected inLV.

32 Source : https://ec.europa.eu/energy/en/statistics/market-analysis





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

ovalFigure 44 - TOTEX in % for different DR activation costs as compared to the case without DG and without flexibility

The implementation of DR could also lead in some circumstances to an increase of the losseswhile providing a reduction of the TOTEX (e.g. see north-urban case in Figure 45, where theoptimal investment strategy leads to privilege a minimization of CAPEX, at the expenses of ahigher OPEX).

With growing penetration of DGs, the benefis of DR in terms of reduction of total costs and oflosses become increasingly significant (compare Figure 45 and Figure 46 which refer to a DGpenetration equal to 100% of peak demand and to 400% of pea demand respectively).

Figure 45 – Decrease of TOTEX and decrease of losses for different at 100% DG penetration level





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Figure 46 – Decrease of TOTEX and decrease of losses for different at 400% DG penetration level

The amount of load which is shifted during the year is generally low, but it has a large impact forreducing the direct and/or the reverse peak..

The maximum daily DR (percentage of the load affected by DR activation) increases with theDG penetration, as shown on the charts of the Figure 47. With high level of DG penetration, themaximum daily demand response level is generally higher in urban areas. Again, this is due tothe different share of DGs among voltage levels and to the cost of cables (overhead/underground).

Figure 47 – Maximum daily load demand response (in percentage of the total load) – results for a 80€/MWh demandresponse cost





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Case 2 – Demand response scenario - Sensitivity analyses on DR CAPEX

In this case, we test the impact of DR for different level of ICT infrastructure costs (see Table 21)and a fixed DR activation cost of 80€/MWh .

LV grid MV grid

Low deployment cost 50 k€/MW 0.5 k€/MW

High deployment cost 200 €/MW 2 k€/MW

Table 21 - Simulated ICT deployment costs

The result of the simulations is presented at the Figure 48.


ICT costs seem in general to have a limited impact on overall TOTEX reduction.The ICT costs particularly influence the activation of DR at LV (residential consumers). In the“high ICT deployment cost” case, DR in LV is rarely activated.High ICT costs have a significant negative effect particularly in presence of high residentialPV penetration (urban areas) as they limit the availability of DR at residential level to balancelocal PV generation).

The decrease of ICT costs in MV has a higher positive impact on TOTEX in those cases where alarge part of DG is connected in MV (rural areas)

Figure 48 – TOTEX decrease as compared with the base case without flexibility for two ICT infrastructure deploymentcost

10.3.4. Full-flex scenarioUnder this scenario, DSOs can opt for two sources of flexibility for the optimization of theirplanning and operation: DR and generation curtailment. The goal of the scenario is to assess thebenefits of DR in a more constraining situation, where a competing source of flexibility(generation curtailment) is also at DSOs’ disposal.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

The compensation cost for the curtailed energy has been set at 120€/MWh, representing 150% ofthe average price of the energy sold on the wholesale market (80€/MWh33). The cost hypothesesare therefore the following:

DR activation cost 80€/MWh

ICT investment cost LV - MV 50€/MW – 0.05€/MW

Generation curtailment cost 120€/MWh

Figure 49 presents the TOTEX reduction in the full-flex scenario for different level of DGpenetration.

Figure 50 presents the activation level of DR and generation curtailment under different DGpenetration level.

Figure 49 - Total cost decrease in the full-flex scenario

Figure 50 - Use of Demand Response and generation curtailment

33 Source : https://ec.europa.eu/energy/en/statistics/market-analysis

0%1%2%3%4%5%6%7%8%

TOTE

Xde

crea

se(%

)*

100% DGpenetration

200% DGpenetration

300% DGpenetration

020406080

100

Ener

gy[M

Wh]

% DG penetration

South rural

0

1000

2000

0%10

0%20

0%30

0%40

0%Ener

gy[M

Wh]

% DG penetration

South urban

Generationcurtailment

DemandResponse





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval


DR is typically the preferred flexibility option: at first for reducing direct peak load and thenfor reducing inverse net peak load for high DG penetration levelsAt high level of DG penetration level, when reverse peak becomes the driver to size gridassets, generation curtailment becomes also a viable flexibility option.With low penetration of DG, DR is mainly used in MV rather than in LV for urban areas, dueto the low ICT costs considered in LV. This holds until a certain level of DG penetration (highreverse flows from PV require more DR in LV) (detailed results are available in Annex I).The cost of losses strongly affects the procurement strategy of DR. The simulation analysisshows that DR activation specifically targets loss reduction only with a sufficiently high levelof loss costs (see Figure 51 – DR impact on loss reduction considering two different loss costs (80€/MWh and120€/MWh)Figure 51, where a 50% increase in loss costs per MWh has been considered). Thisis particularly the case in rural areas where the overall investment strategy is more loss-drivenas discussed previously.

Figure 51 – DR impact on loss reduction considering two different loss costs (80€/MWh and 120€/MWh)

10.4. Conclusions and key messages

In the following we provide a summary of the main messages resulting from the simulationanalysis, which are valid on average on a given grid. Deviations are possible on specificfeeders/substations depending on local specificities.

Demand response is effective in reducing the total cost of the distribution grid

Demand response is an effective means to reduce the TOTEX (Total Expenditure) of thesystem (CAPEX+OPEX), when integrated in grid planning and operation:- Via reduction of direct and inverse peaks in the network and thus the need for

infrastructure costs (CAPEX).





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

- Via reduction of technical losses (OPEX) by performing a local balancing with DG andreducing the flows in the system.

The reduction of TOTEX is achieved by finding the optimal balance between CAPEX andOPEX (in certain conditions, the level of technical losses could also increase, if off-set by adecrease of CAPEX). Of course, regulation should specifically incentivize DSOs to reduceTOTEX (please refer to subsequent chapters).

If DSOs do not receive the right incentives to reduce losses, then the use of flexibility couldmainly target CAPEX reductions (i.e. size assets to their capacity limits) and possibly lead

to higher loss volumes.

The simulation analysis shows that DR activation targets specifically loss reduction only ifsufficiently high costs for losses are considered.Regulation needs to pay attention to the costs of losses paid by the DSOs (or to provide theright loss penalties) to avoid that DSOs have an incentive to use flexibility to downsize assets(reduction of CAPEX), even if this leads to an increase of losses.

Explicit (Volume-based) Demand Response can bring higher benefits to the systemcompared to implicit (tariff-based) Demand Response.

Both implicit and explicit DR can reduce the total cost of the system. Both approaches canalso be implemented together to maximize the impact (even if the two mechanisms can be incompetition with each other, as the presence of ToU reduces the amount of flexibilityexploitable by explicit DR).Volume-based (explicit) DR can bring higher benefits as it can be more precisely activatedaccording to DSO’s needs.For a volume-based flexibility market to develop, DSOs should make clearly accessible theinformation about their flexibility needs (present and future). This would allow revealing theavailability of flexibility resources and contributing to creation of a flexibility market.

Time of Use (ToU) tariff can provide significant system benefits, but they need to becarefully tailored to specific local conditions.

The impact of ToU tariffs on TOTEX depends on the level of penetration of DGs and on localconditions.In general, ToU can provide significant benefits and facilitate the integration of DG at a lowercost when the off-peak period is synchronized with DG generation. A location-basedcomponent in the ToU could help better tailor the tariff design to local conditions.Poorly designed ToU timeframes, under certain conditions, could also have a negative effect.For example, with high DG penetration, simple day-night ToU might actually worsen thecosts of DG hosting capacity.

The cost/benefit ratio of Demand Response for DSOs depends on local conditions

On the benefit side:

The value of DR for DSOs strongly depends on the level of penetration of DG, on theconsumption patterns and on the grid characteristics.Typically, the simulations shows that a limited number of activation of DR per year canalready provide significant benefits for the system





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

On the cost side:

A key factor affecting the business case of DR for DSOs is the availability and cost of an ICTinfrastructure (smart meters).In those regions where smart meters are already being deployed (driven by policy decisions),the business case for DR is stronger both for residential and commercial/industrial customers(as DR programs do not need to include ICT infrastructure costs)In those regions where the cost of the ICT infrastructure should be borne by the DR program,it is recommended to focus on DR at MV level get most of the benefits.

Demand Response is always part of the best flexibility mix to decrease the total cost ofdistribution grid planning and operation

With limited penetration of DG, DR is the most economically viable flexibility option to drivedown the system cost. (this holds on average on a given grid. Deviations are possible onspecific feeders/substations depending on localspecificities. For example in rural areas, withlimited load, curtailement of DG could be the best flexibility option to solve congestions onspecific feeders).With very high penetration of DG, reverse peak (peak of reverse power flows) becomes themain driver to size grid assets. Beyond that DG penetration level, generation curtailmentbecomes an economically viable option together with DR.

Demand Response has typically a positive impact on grid technical losses, but reduction oflosses is not the main driver for its activation

Reduction of losses is typically not the main driver of the activation of DR, rather a side-effectof the activation of DR to reduce the CAPEX.Higher loss reduction level is typically achieved in rural areas, given the structure of ruralgrids (longer feeders, overhead lines), for which cost of losses has a higher share of theTOTEX.

Demand flexibility allows a higher DG hosting capacity at lower cost

DR increases the syncronicity of the demand with DG and mainly reduces CAPEX.The positive impact of DR on DG hosting capacity becomes increasingly significant after acertain DG penetration level.In fact, a limited penetration of DG, even without any flexibility, could actually decrease theTOTEX of the system compared to the case without any DG (thanks to DGs balancing localdemand and reducing network flows).

Strong penetration of residential PV will drive more rapidly the need for flexibility

A lack of variety in the types (PV, wind, CHP) and connection level (MV, LV) of DG has anegative effect on grid efficiency, driving up CAPEX (reinforcements) and OPEX (technicallosses).This means that in those areas where a single source of DG is mainly present (e.g. areas withstrong residential PV), demand flexibility can particularly provide positive system impact andincrease grid efficiency. This is of course dependent on the availability of sufficient level ofdemand flexibility (in which case DG curtailment could be the main flexibility option toexplore).





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

11. REGULATORY EVOLUTIONS FORDISTRIBUTION SYSTEM OPERATORS TOPROCURE AND ACTIVATE DEMAND RESPONSE

This chapter presents the key regulatory evolutions that will allow DSOs to procure and activateDR. We identify four key needed regulatory evolutions, in terms of how to:

a. use tariff signals (price-based approach – implicit DR),b. directly procure DR (volume-based approach – explicit DR),c. avoid conflicts with use and ownership of flexibility assets andd. Validate transactions at distribution level.

11.1. Price-based approach: DSOs’ use of tariffsignals

In this section, we discuss the effectiveness of regulatory evolutions in using grid tariffs as alever for activating DR for DSOs’ needs.

Distribution grid tariffs are generally made of a number of components [83]:

Fixed component : fixed payment per customer per year (covers fixed costs like metering);Capacity component : euro per kW (measured capacity) or per kVA (connection capacity);Energy component: euro per kWh (proportionate, progressive or regressive).Other components: reactive power, renewable energy charge, etc

A grid tariff may consist of one or more of the above components. Figure 52 shows whichcomponents most commonly apply in European countries. As can be seen, key components arethe fixed and energy charges, followed by capacity and other charges. Reactive charges aremostly not applied.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Figure 52 – Distribution tariff components in EU countries by consumer group: Household [83]

Under a price-based approach for DR, the regulator should define the required changes innetwork tariffs to incentivise DR. The following main regulatory evolutions can be considered:

Regulator defines the share of the different components, e.g. by defining a higher capacitycomponent. For example, power band contracts (see section 9.2.1, Power band contract)modify the capacity charge component, while the fixed charge component could be subject tonegotiations as a contract clause in bilateral agreements.Regulator defines different time intervals for ToU tariffs: For example, by introducingtime intervals which are aligned to PV production for areas with high PV penetration.Regulator allows the DSOs to dynamically change the tariff according to gridconditions: in order to pursue this option, measures should be taken to manage conflicts withparticipation in wholesale energy market prices (distribution congestion needs vs wholesaleenergy market needs).

To allow demand response actions, distribution network tariffs should be able to form clearsignals. However, distribution tariffs are only a component of the entire electricity tariff paid by aconsumer. Figure 53 shows the distribution grid tariff component of electricity tariffs in differentEU Member States. It can be seen that distribution grid tariffs constitute between 8% and 45% ofthe entire electricity tariff. This means that the cost component that a DSO can influence may bequite small om some cases, therefore the price signal could be insufficient to trigger DR. Wherethe proportion is relatively large, distribution grid tariffs could be used as an effective price signalto activate DR for DSOs’ needs.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

ovalFigure 53 – Distribution grid tariff component of final electricity tariffs in EU Member States (2014 S2) [Ecofys analysis

based on data from Eurostat]34

11.2. Volume-based approach: DSOs directlyprocure demand response

In general, DSOs would be willing to pay for DR if the avoided costs are larger than the cost ofcontracting DR. Under a volume based procurement approach, DSOs procure directly therequired amount of DR services and remunerate the provided service.

There are three main value streams that could be considered for covering DR costs, whichhowever may not be sufficient to incentivise volume-based DR for the reasons discussed below:

1) Savings from avoided investment costs: these savings depend on the specific networkasset conditions and system evolution.

2) Savings from avoided payments for curtailing DG-RES to solve congestions: thesepayments depend on the regulatory framework and in some countries are not applicable.

Due to the abovementioned reasons, in most cases, incentives to optimise operation are eitherinsufficient or not clearly visible to DSOs in its current form. Additional regulatory adjustmentscould help to emphasise the incentives in optimising operation, and facilitate more freedom forDSOS on how to recover the costs of procuring DR. To incentivise DSOs to capture benefitsfrom optimising operation and planning, an effective solution would be to allow them to capture

34 Note that data from Eurostat does not report consistently what constitutes distribution grid chargeacross different countries. This means that, the renewables surcharge and some form of taxes like localtax may be included in the percentage calculation for some countries. In this case, it would notcorrespond to the actual percentage that the DSO could steer. Detailed review of each country in termsof their distribution grid structure would be prudent, but is out of scope of this report.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

both CAPEX and OPEX savings. This can be done by allowing DSO remuneration to reflectTotal Expenditure (TOTEX, which encompasses both CAPEX and OPEX).35

11.2.1. Evolution of DSO remuneration based on TOTEXRegulation based on TOTEX is suited to provide appropriate incentives for the use of “smarter”solutions. This means regulation focusing on total costs (TOTEX) rather than capital cost(CAPEX) and operational costs (OPEX) separately. Such a setup would allow more freedom tothe DSO to decide on approaches that could either affect CAPEX or OPEX, such as to use DERto solve network related problems rather than investing in additional network infrastructure.

Concretely, TOTEX regulation could allow DSOs to arbitrate between traditional gridinvestments (CAPEX) and procurement of flex-based solutions (OPEX).

Additionally, specific incentive regulation could be foreseen to steer DSOs’ CAPEX/OPEXinvestment decisions toward desirable performance objectives, like reduction of losses.

For example, to that end, suitable mechanisms could be the application of bonus/malus regulationfor a given loss target or to explicitly encourage energy efficiency investments in the regulatedasset base.

Example case: UK regulation

Regulation in the UK is seen as the most progressive in allowing TOTEX to be reflected indistribution grid tariffs and thus encouraging optimisation of operation as well as planning. Thekey characteristics are presented below:

Calculation Method of Distribution Grid Tariffs

Distribution Use of System Charges are calculated by applying the Common DistributionCharging Methodology (CDCM) which is approved by OFGEM. This methodology also impliesa cost component for losses but does not incentivise grid operators for a loss reduction.

Treatment of Losses in the price-control scheme

Distribution losses are calculated by OFGEM as an allowed loss percentage on an annual basisfor each distribution company. Distribution system operators are receiving a reward or penaltybased on the performance against the target. In 2009 the financial incentive for the reduction oftransmission losses has been 5p/kWh.

35 Another option is to allow DSOs to act as aggregators of DR, so that they can participate in wholesaleand balancing markets, where benefits are generated. However this option is not allowed due tounbundling rules.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Treatment of operational and capital costs

The regulation of both transmission and distribution system operators is designed with a certainamount of allowed capital and operational costs for each company. Operational costs ofdistribution grid operators are defined by the performance of their business operations andmaintaining their network. Capital costs are expenditure on investment in distribution assets witha long lifetime, such as underground cables, overhead electricity lines and substations. As aconsequence, DSOs are remunerated for investments in efficient grid infrastructure as long asthey are part of their forecasts in the business plans and at the same time they are incentivised tominimise operational costs. This design is an effective and efficient combination of output- andinput-based components to achieve loss reductions as it also takes into account the lifetimeoperational costs.

Conclusion

The overall regulatory design of the UK has incentives for loss reductions. The incentives fordistribution grid operators are clear, as the regulatory authority calculates allowed losspercentages annually. DSOs are penalised or rewarded based on their performance in losses.Thus, they have strong financial incentives to improve in distribution efficiency. At the sametime, they are rewarded for their announced capital costs and incentivised to minimiseoperational costs. This represents a combination of input- and output- based regulatory scheme.

11.3. Role of DSO regarding ownership and use offlexibility assets

DSOs have the best overview of resources in its local network area. If the DSO has the option tocontract DR in its area, they could be activated to manage congestions in the distribution grid. Incases when forecasted congestions do not finally take place, the DSO might be in the position tohave contracted flexibility which could be used for other purposes (surplus flexibility). In thiscase, an option could be that the DSO sells this flexibility to other entities acting like anaggregator. From a technical point of view, the following actions could be possible:

DSO can activate local DR to manage congestion in its own area, and where there is surplusflexibility, sell the service at the balancing market.DSO can activate local DR to manage congestion in its own area, and where there is surplusflexibility, it could then sell the DR service to the TSO, if the TSO is willing to pay for DR tosolve congestion.Alternatively, the DSO could sell the DR service to the DG in its area, to avoid curtailment, ifDG is willing to pay for DR to avoid curtailment, or DSO avoids paying compensation forcurtailment to RES-DG.

Since DSOs have the best overview on the grid situation, one could argue that they could makeoptimal use of flexibility. However, DSOs are regulated entities and their operations in terms ofcommercial transactions should be checked so that they do not impede competition. In thisrespect, in most cases DSOs shall not be able to use the surplus flexibility, since he is not allowedto trade in the market.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

A solution for allowing these resources to reach the market could be that the DSO frees thespecific market participants from their obligation, allowing them to reach other markets (use it orlose it concept). In this respect, the flexibility resources could further reach to additional revenuesfrom offering services to the higher system levels. In this concept, it would be critical that theDSO is responsible for approving activation of the specific DR that are in excess.

An application example where the DSO is controlling local flexibility for solving localcongestions exists in Italy. In the ENEL network, electrical storage systems (ESS) have beenidentified as a promising flexibility resource which generates several benefits. For example, thePOI project is used to demonstrate how DSOs could use flexibility to solve congestions in boththe distribution grid and transmission grid, and also serves the balancing market. [84]

11.4. Role of DSO as validator of DR transactions atdistribution level

Flexibility from DR (or generally from Distributed Energy Resources) in the distribution grid canbe used for three key purposes: a) in the system level (participation to balancing markets) b) intransmission grid level (alleviation of transmission grid congestions) and c) in the distributiongrid level (alleviation of distribution grid congestions). However, the way the resource is used bythe different levels may result in a conflict. This is why a prioritisation process is needed to useDR for optimisation of grid operation and planning. Conflicting interests of grid and marketbased use of flexibilities and between DSOs and TSOs need to be identified and access rightsneed to be defined.

Market and grid based control signals and access rights to local flexibilities

Local conditions inevitably deviate from the global conditions at system level. In such cases, thelocal grid situation could lead to different signals for flexibility services than the global marketsituation,. An example of this is when the system is short of balancing power, and at the sametime a congestion appears at the DSO level due to local overproduction. In this case, the globalsignal for DR is to reduce demand or increase generation, while the local conditions need theopposite to happen to alleviate the congestion, i.e. increase local demand or reduce localgeneration. In such cases, a clear prioritisation should be in place.

There are different options to coordinate market and grid based signals. One option is to limitcontrol signals, meaning that only one of either third market parties or grid operators get accessto the local flexibility. Another option is to prioritise one of the control signals, meaning thatboth, third party marketer and the grid operator get access to local flexibility but in cases ofconflicting signals, one of the signals is prioritised. The second option allows more efficient useof resources, as it is shared between two purposes, while the first option makes it mutuallyexclusive.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

For the second option, the priority of one of the signals has to be defined in advance. To ensuresecure grid operation, the alleviation of congestions has to be prioritised against market signals.36

When there is grid congestion, the control signal of the grid operator must be prioritised and thegrid operator has to remunerate the flexibilities used.

TSO/DSO coordination process for accessing DR: Role of DSO as validator

Conflicts can be avoided by applying the following prioritisation process, starting from locallevel and moving towards global (see Figure 54):

1) DSO decides to reserve or release DR flexibility based on local congestion forecast2) If no local congestions are foreseen, DR flexibility is handed over to TSO. This can be done

by:a) DSO contracts DR reserve and sells it to TSO (however DSO in effect acts as

aggregator)b) DSO lets DR providers (or aggregators) to offer directly their services to the TSO

3) TSO decides whether to use DR to resolve congestion or system imbalance.

Each step would be remunerated separately.

Figure 54 - Conflict of interest in using DR in distributed networks

36 This is also due to the fact that anyway, in cases of congestions, the grid capacity does not allowpower flows on the congestion direction, therefore the operation based on the conflicting market signalwould anyway not be feasible (e.g. reduce local demand (or increase local generation) on a situation oflocal overproduction).





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

The coordination process would essentially require a technical validation by DSOs (step 1mentioned above). This means that the DSO must decide what happens with the DR if networkconstraints render impossible the participation to global markets. In such a case, the DR serviceprovider should request a validation from the DSO before activation for system balancingpurposes.

Example case: DSO/TSO coordination for Switzerland

For the case of Switzerland, [85] recommended a coordination process as depicted in Figure 55.Currently, TSOs are the only system operators that prepare congestion forecasts. Forecasts of thenetwork status are reviewed in three time intervals, 2-days ahead of dispatch, 1-day ahead, andintraday. These activities define the need for flexibilities for a given timeframe, and can be linkedwith different markets in the respective timeframes in order to enable solutions. To ensure thatcongestions in the DSO areas are not compromised by the activities that currently exist, a newprocess which includes a DSO congestion forecast is proposed. For DSOs, this is a new processthat consists in identification of the local grid situation and in definition of the need for localflexibilities. This process should be linked and run in parallel to the Day-ahead TSO CongestionForecast. The DSO Congestion Forecast is a preventive action that reduces distribution gridconstraints and allows the early identification of critical situations in the distribution grid. Theassessment of the need for local flexibilities is based on the day-ahead and intraday marketresults. The process ends before the TSO Intraday Congestion Forecast.

Figure 55 – DSO Congestion forecast timeline





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

11.5. Regulatory evolutions: conclusions and keymessages

1. Regulatory evolutions should ensure that tariff signals provide sufficient incentives togrid users to adapt their demand profiles to grid conditions

Under an implicit (tariff-based) DR approach, the regulator should define the required changes innetwork tariffs to incentivise DR, by changing specific components of the tariffs (e.g. highercapacity charges), changing time intervals for ToU introducing location-based components orallowing DSOs to adapt tariffs to grid conditions.

However, since distribution grid tariffs constitute just a component of the entire electricity pricepaid by a consumer, in cases they are low they might fail to form an effective market signal togrid users to change their consumption to meet grid needs.

2. DSOs can directly procure DR if DSOs remuneration accounts for it (e.g. use ofTOTEX remuneration)

Under a volume based DR approach, DSOs procure DR services and remunerate the providedservice. Incentives to optimise operation need to be created by appropriate regulatory adjustmentsin order to allow DSOs recovering the costs of procuring DR. This can be done by allowing DSOremuneration to reflect the DSOs’ TOTEX (CAPEX+OPEX).

3. Regulation should ensure a neutral role for DSOs in the activation of flexibility, inorder to ensure market-based competition among different flexibility resources

As shown in the simulation analysis, economic benefits are achieved by enabling competitionbetween flexibility options (DR, storage, curtailment of DG) for the provision of services at localor global level. To enable a market-based approach, transparency and neutrality of DSOs are keyelements. In this respect policy questions are related to:

whether DSO should be allowed to own flexibility assets (e.g. storage)How to address the risk that, if most of the value of flexibility is controlled directly by DSOs(e.g. ownership of storage means; direct bilateral contracts with grid users), a local flexibilitymarket (targeting also other value streams, e.g. BRP; TSOs..) could never develop.How to ensure DSO transparency when taking the role of validator at local level of flexibilitytransactions which are driven by global market signals (e.g. flex transactions for TSOs)





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

4. In case of conflicts for flexibility activation, local distribution needs should prevail oversystem transmission needs, unless when system security is at stake.

Flexibilities in the distribution grid can be used for multiple purposes, a) in the system level(participation to wholesale, intra-day and balancing markets) b) in transmission network level(alleviation of transmission congestions) and c) in the distribution grid (alleviation of distributioncongestions). Congestions at any network level could “block” the provision of flexibility servicesfrom the local level to the higher system levels. Respectively, conflicts on the activation offlexibility between TSO and DSO are possible. Therefore, a prioritisation between local andglobal requirements is needed, and local needs should prevail, unless when system security is atstake and alternative options are not viable. In this respect, DSO/TSO coordination should beapplied to avoid such conflicts. To achieve this coordination a new process could be defined, theDSO congestion forecast linked to the respective process at TSO level.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

12. BIBLIOGRAPHY

[1] European Commission, «Energy Efficiency Directive,» 12 October 2012. [En ligne].Available: http://eur-lex.europa.eu/legal-content/EN/TXT/?qid=1399375464230&uri=CELEX:32012L0027. [Accès le January2015].

[2] OFGEM, «An assessment of the energy efficiency potential of the gas and electricityinfrastructure of Great Britain,» 2014.

[3] Western Power Distribution, «Losses Strategy,» 2015.

[4] Ecofys, «Flexibility options in Electricity Systems,» Berlin, Germany, 2014b.

[5] Waide Strategic Efficiency, «PROPHET II: The potential for global energy savings fromhigh-efficiency distribution transformers,» 2014.

[6] KEMA , «Analysis of Insufficient Regulatory Incentives for Investments into ElectricNetworks. An Update.,» 2009.

[7] ERGEG, « Treatment of Losses by Network Operators. doi:Ref: E08-ENM-04-03,» 2008.

[8] M. Oddi, F. Gorgette et G. Roupioz, «ERDF experience in reducing network losses,» chezProceedings 20th International Conference on Electricity Distribution CIRED, Prague,2009.

[9] K. Schneider, F. Tuffner, J. Fuller et R. Singh, «Application of Automated Controls forVoltage and Reactive Power Management - Initial Results,» U.S Department of Energy,2012.

[10] Bundesnetzagentur, Bundeskartellamt, «Monitoringbericht 2014,» Bonn, 2014.

[11] TEN Thüringer Energienetze GmbH, «Netzverluste,» 2014.

[12] W. Heckmann, L. Hamann, M. Braun, H. Barth, J. Dasenbrock, C. MA, T. Reimann et A.Scheidler, «Detailed analysis of network losses in a million customer distrbution gridwith higher penetration of distributed generatrion,» chez 22nd International Conferenceon Electricity Distribution, Stockholm, 2013.

[13] TenneT TSO GmbH, «Netzverluste,» 2014.

[14] BMWi, «Verteilernetzstudie,» 2012.

[15] M. Clemence, R. Coccioni et A. Glatigny, «Schneider Electric White Paper: How UtilityElectrical Distribution Networks can Save Energy in the Smart Grid Era,» SchneiderElectric, 2013.

[16] Eurelectric, «Power Distribution in Europe - Facts & Figures,» 2013.

[17] C.-A. Sylvie, «ERGEG public consultation on "Treatment of Losses by NetworkOperators,» EnBW Energie Baden-Württemberg, 2008.

[18] F. Topalis, W. Irrek et R. Targosz, «Strategies for development and diffusion of EnergyEfficient Distribution Transformers,» SEEDT Project, 2008.

[19] L. Fickert, W. Spitzl, E. Schmautzer, W. Brandauer, TU Graz et Wien Energie Stromnetz





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

GmbH, «Strategien zur Minimierung von Netzverlusten,» chez 6. InternationaleEnergiewirtschaftstagung an der TU Wien, Wien, 2009.

[20] J. Lazar et X. Baldwin, «Valuing the Contribution of Energy Efficiency to AvoidedMarginal Line Losses and Reserve Requirements,» Regulatory Assistance Project, 2011.

[21] S. Kalambe et A. Ganga, «Loss minimization techniques used in distribution network:bibliographical survey,» Renewable and Sustainable Energy Reviews 29, pp. 184-200,2014.

[22] G. Papaefthymiou, P. Zonneveld, E. Coster et L. van der Sluis, «Calculation of gridlosses with the aid of stochastic modeling of wind turbines and loads,» chez 16th PowerSystems Computation Conference. Vol. 2, 2008.

[23] C. Bucher, «Analysis and Simulation of Distribution Grids with Photovoltaics,» ETHZürich, Zürich, 2014.

[24] S. Abbott, B. Fox et D. Morrow, «Distribution Network Voltage SUpport UsingSensitivitzßbased Dispatch of Distributed Generation,» 2013.

[25] E. Coster, J. Myrzik, B. Kruimer et W. Kling, «Integration Issues of DistributedGeneration in Distribution Grids,» Proceedings of the IEEE, vol. 99, n° %11, pp. 28-39,2011.

[26] S. ur , G. Strbac et X. Zhang, «Effect of losses in design of distribution circuits,» IEEProceedings Gener.Transm. Distrib., , Vols. %1 sur %2148, No. 4, n° %14, 2001.

[27] IEA, «Electricity Information,» 2014.

[28] A. T. De Almeida, F. Martins et B. Santos, «Implemeting Directive 2009/125/EC of theEuropean Parliament and of the Council with regard to Ecodesign Requirements forPower, Distributions and Small Transformers,» ISR-University of Coimbra, April 2013.

[29] K. Niall et A. Walsh, «Maximising benefits from distribution losses management - AnESBN perspective,» Frankfurt, 2011.

[30] F. R. D. A. Merschel et N. K. I. o. T. Mathias, «The AmpaCity Project,» T&D WorldMagazine, 23 Dec 2013.

[31] E. Koliou , T. Eklund, L. Söder , C. Bartusch, K. Alvehag et R. Hakvoort, «EconomicImpact of Demand Response on Costs to Distribution System Operators,» KTH RoyalInstitute of Technology, Stockholm, 2013.

[32] MVV Energie AG, IWES, IBM, IFEU, IZES, Papendorf Software Engineering, Power PlusCommunication AG, Universität Duisburg-Essen, «Modellstadt Mannheim,» 2013.

[33] B. Bergesen, L. Groth, B. Langseth, I. Magmussen, D. Spilde et J. Toutain, «Energyconsumption 2012 - Household energy consumption,» Norwegian Water Resources andEnergy Directorate, 2012.

[34] C. L. U. F. M. I. E. ECN, «Market and regulatory incentives for cost efficient integrationof DG in the electricity system,» 2010.

[35] A. Nourai, V. Kogan et C. M. Schafer, «Load Leveling Reduces T&D Line losses,» 2007.

[36] Southern Energy Power Distribution, «Low Energy Automated Networks,» Low CarbonNetworks, 2014.

[37] NRECA, Pinney et CRN, «Costs and Benefits of Smart Feeder Switching,» 2013.

[38] D. Pinney, «Costs and Benefits of Smart Feeder Switching: Quantifying the operating





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

value of SFS,» National Rural Electric Cooperative Association, for U.S. DOE/NETL,2013.

[39] U.S Department of Energy, «Application of Automated Controls for Voltage and ReactivePower Management - Initial Results,» Smart Grid Investment Grant Program, 2012.

[40] Western Power Distribution, Accenture, «Low Voltage Network Templates Summaryreport,» 2014.

[41] D. Divan, «Improving power line utilization and performance with D-FACTS devices,»2005.

[42] L. F. Ochoa et G. P. Harrison, «Minimising Energy Losses: Optimal Accomodation andSmart Operation of Renewable Distributed Generation,» 2010.

[43] UK Government, «Domestic energy consumption in the UK between 1970 and 2013,»Department of Energy and Climate Change, 2014.

[44] D. K.Chembe, «Reduction of Power Losses Using Phase Load Balancing Method in PowerNetworks,» World Congress on Engineering and Computer Science Vol I, 2009.

[45] S. Weckx, C. Gonzales et J. Driesen, «Reducing Grid Losses and Voltage Unbalance withPV inverters,» chez PES General Meeting-Conference & Exposition IEEE, NationalHarbor, MD, 2014.

[46] Ernest Orlando Lawrence Berkeley National Laboratory, «Energy Efficiency Potental forDistribution Transformers in the APEC Economies,» 2013.

[47] L. B. N. Laboratory, «Cost-Effective Potential for Imporved Energy Efficiency of KeyElectrical Products in India,» 2006.

[48] European Commission, «Evaluation of Smart Grid projects within the Smart Grid TaskForce Expert Group 4 (EG4),» European Commission, 2013.

[49] J. Hechenbergerova, P. Musilek, E. Pelikan et D. Koval, «Identification of critical agingsegments and hotspots of power transmission lines,» chez 9th International Conferenceon Environment and Electrical Engineering , 2010.

[50] M. Lupescu, «Improving the energy efficiency by reducing the losses in the low-voltagepower distribution grid,» U.P.B. Sci. Bull., Series C, Vol. 72, Iss. 4, 2010 , 2010.

[51] N. Grass, G. Kerber, M. Sebeck et R. Keck, Grass, N.;Kerber, G.;Sebeck,M.;Keck,R..

[52] The Energy Storage Association, «T&D Upgrade Deferral,» [En ligne]. Available:http://energystorage.org/energy-storage/technology-applications/td-upgrade-deferral.[Accès le 24 04 2015].

[53] U.S. Department of Energy, «Fault Location, Isolation, and Service RestorationTechnologies Reduce Outage Impact and Duration,» 2014.

[54] D. Bernardon et L. L. Pfitscher, «Automatic reconfiguration of distribution networksusing smart grid concepts,» chez 10th IEEE/IAS International Conference on IndustryApplications (INDUSCON), 2012.

[55] M. Meuser et H. H.-J. Vennegeerts, «IMPROVED GRID INTEGRATION OF DISTRIBUTEDGENERATION IN EXISTING NETWORK STRUCTURES,» chez 21st InternationalConference on Electricity Distribution, Frankfurt, 2011.

[56] A. Shammah, A. Azmy et A. Abou El-Ela, «Optimal Sitting and Sizing of CapacitorsBanks in Distribution Networks using Heuristic Algorithms,» Journal of Electrical





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Systems, Vols. %1 sur %29-1, pp. 1-12, 2013.

[57] J. Dixon, L. Morán, J. rodirguez et R. Domke, «Reactive Power CompensationTechnologies, State-of-the-Art Review,» Proceedings of the IEEE, Vol. 93, December2005.

[58] « Article 15.2. Approach and examples from Denmark,» 2014.

[59] U. D. o. Energy, «Evaluation of Conservation Voltage Reduction on a National Level,»2010.

[60] NRECA-DOE smart grid demonstration project, «Costs and Benefits of ConservationVoltage Reduction – CVR Warrants Careful Examination,» 2013.

[61] S. Weckx, C. Gonzalez DeMiguel et P. Vingerhoets, «Phase switching and phasebalancing to cope with a massive photovoltaic penetration,» chez 22nd InternationalConference on Electricity Distribution, Stockholm, 2013.

[62] Marcogaz, «Reduction of methane emissions in the European gas industry,» 2008.

[63] US EPA, «Natural Gas STAR Program».

[64] Ecofys, «Economic Evaluation of Methane Emission Reduction in the Extraction,Transport and Distribution of Fossil Fuels in the EU,,» 2001.

[65] ICF, «Economic analysis of methane emission reduction opportunities in the U.S.onshore oil and natural gas industries,» 2014.

[66] GDF Suez, «Le retour de l'interruptibilité via RTE,» 2013. [En ligne]. Available:http://www.gdfsuez-flash-energie.fr/Le-retour-de-l-interruptibilite.

[67] PJM State & Member Training Dept., «PJM 101 - The Basics,» 2013.

[68] PJM Froward Market Integration, «PJM Manual 11 : Energy & Ancillary Services MarketOperations Revision : 60,» 2013.

[69] X. He, L. Hancher, I. Azevedo, N. Keyaerts, L. Meeus et J.-M. Glachant, «THINK:"Shift,Not Drift: Towards Active Demand Response and Beyond",» European UniversityInstitute, 2013.

[70] T. Jacobo, «Demand Response in European electricity markets,» Florence, 2013.

[71] PJM, «PJM Demand Side Response Intermeditate overview,» 2012.

[72] The_Brattle_Group, «Lessons from Demand Response: trials and potential savings forthe EU,» 2009. [En ligne].http://www.brattle.com/system/publications/pdfs/000/004/556/original/Lessons_from_Demand_Response_Faruqui_Harris_Oct_12_2009.pdf?1378772113.

[73] Sweco, Ecofys, Tractebel, PwC, «Study on the effective integration of Distributed EnergyResources for providing flexibility to the electricity system, Final report to The EuropeanCommission,» European Commission, 2015.

[74] O. Huet et S. Boyard, What Market Design Regarding Distribution And Active NetworkManagement ?, ERDF, 2014.

[75] Y. Ding, P. Nyeng, J. Østergaard, M. D. Trong, S. Pineda, K. Kok, G. B. Huitema et O. S.Grande, «Ecogrid EU e A Large Scale Smart Grids Demonstration of Real Time Market-based Integration of Numerous Small DER and DR,» chez ISGT Conference 2012, Berlin,2012.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

[76] F. I. Ecofys, «Strompreise und ihre Komponenten: Ein internationaler Vergleich,»Bundesministerium für Umwelt, Naturschutz und Reaktorsicherheit, 2014.

[77] LUT Energia, «Tarrif scheme options for distribution system operators,» LappeenrantaUniversity of Technology, Lappeenranta, 2012.

[78] University of Manchester & EDF SA, «Description of market mechanisms (regulations,economic incentives and contract structures) which enable acive demand participation inthe power system,» European Commission, 2011.

[79] UK Power Networks, «Demand side response,» UK Power Networks, [En ligne].Available: http://innovation.ukpowernetworks.co.uk/innovation/en/research-area/demand-side-response/. [Accès le 2015].

[80] UK Power Networks, «UK Power Networks - Press releases,» 12 October 2012. [Enligne]. Available: http://www.ukpowernetworks.co.uk/internet/en/news-and-press/press-releases/Contracts-to-reduce-Londons-peak-demand.html. [Accès le 2015].

[81] L. Dann, «UKPN saves £43m in upgrade work through DSR,» Utility Week, 16 November2015. [En ligne]. Available: http://utilityweek.co.uk/news/ukpn-saves-43m-in-upgrade-work-through-dsr/1189753#.Vlg3A3arTIV. [Accès le 2015].

[82] Grid Innovation, «eTelligence,» [En ligne]. Available: http://www.gridinnovation-on-line.eu/Articles/Library/ETelligence.kl#sthash.ynNANjQy.dpuf. [Accès le 2015].

[83] Mercados; refE; Indra, «Study on tariff design for distribution systems,» EuropeanCommission, 2015.

[84] C. Noce, ENEL plans for storage introduction in Italian distribution network, Enel SpA –Infrastructure and Network division, 2012.

[85] Ecofys, Swiss Economics, «Zukünftige Energiemärkte und die Rolle der Netzbetreiber,»Bundesamt für Energie (BFE), Bern, 2015.

[86] A. Al-Badi, A. Elmoudi et I. Metwally, «Losses Reduction In Distribution Transformers,»chez Proceedings of the International Multiconference of Engineers and ComputeScienstis Vol II, 2011.

[87] Electrical Engineering Portal, «Electrical Engineering Portal - Primary DistributionVotlage Levels,» 10 August 2012. [En ligne]. Available: http://electrical-engineering-portal.com/primary-distribution-voltage-levels. [Accès le April 2015].

[88] B. Sampath Kumar et Inamdar, «Loss Reduction by Optimal Placement of DistributedGeneration on A Radial feeder,» ACEEE Int. J. on Electrical and Power Engineering,Vol.02, No. 01, pp. 24-29, 2011.

[89] Synergrid, «Etude de Synergrid pour la réalisation de l’article 15.2 de la Directived’efficacité énergétique 2012/27/EU du Parlement Européen et le Conseil du 25 octobre2012,» 2015.

[90] Energeia, Department of Resources, Energy and Tourism, «Energy EfficiencyOpportunities in Electricity Networks - Findings of industry trials for the extension of theEEO program to network businesses,» Australian Government, 2013.

[91] Operation Technology Inc., «Optimal Capacitor Placement: Cost Benefits Due to LossReduction,» etap.com, Irvine.

[92] H. Saboori et H. Abdi, «Application of a Grid Scale Energy Storage System to Reduce





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Distribution Network Losses,» 2010.

[93] energynautics, «Kurzgutachten zur Eigenstromerzeugung in Rheinland-Pfalz,» 2014.

[94] NRECA; CRN - Cooperative Research Network, «Costs and Benefits of Smart FeederSwitching - Quantifying the Operating Value of SFS - Initial Findings,» 2013.

[95] PRCI, «Methods to Reduce the Greenhouse Gas Footprint from Pipeline Compressor andPump Stations,» 2014.

[96] Ecofys, «Methane from fugitive emissions,» 2009.

[97] E. Lakervi et E. J. Holmes, Electricity distribution network design, 2nd edition, London:Institution of Electrical Engineers, 1995.

[98] Ecogrid EU, «Ecogrid EU Market Concept,» 2011-2015. [En ligne]. [Accès le 2015].

[99] Energinet.dk, «EcoGrid EU Newsletter No1 - Power Market Oepration in Real Time,»February 2012. [En ligne]. Available:http://energinet.dk/SiteCollectionDocuments/Engelske%20dokumenter/Forskning/EcoGridEU_Newsletter_nr.%201_February%2029_2012.pdf. [Accès le 2015].

[100]

WattClarity, «Three sample approaches (of 173 worldwide) to integrating renewablesinto the grid,» GLOBAL ROAM PTY LTD, 15 January 2015. [En ligne]. Available:http://www.wattclarity.com.au/2015/01/three-sample-approaches-to-integrating-renewables-into-the-grid/. [Accès le 2015].





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

13. ANNEX I - PARAMETERS INFLUENCING THEIMPLEMENTATION AND THE COST-EFFECTIVENESS OF THE MEASURES –DETAILED ANALYSIS

13.1.1.1. HIGH EFFICIENCY TRANSFORMERS

Factors affecting benefits

Grid Parameters Efficiency impact

Transformer load rate The optimal efficiency of a transformer is reached at a loading of 30-40 ofrated capacity. The potential for energy savings can therefore higher by firstreplacing overloaded/underloaded transformers and oversize/downsize thenew ones to reach the optimal operating point (clearly, oversizing costs mustbe taken into account in the evaluation).

Age of the transformers Old transformers are the main contributors to losses, since they had to meetless stringent efficiency standards. The study on transformer replacementcarried the in the APEC ( [46]) regions show that for instance, an oldefficiency standard 25kVA transformer in Russia has a 98% efficiency,compared to 98.9% for 1500 kVA. New standards allow them to reachrespectively an efficiency of 99.5% and 99.7%).

Transformers fault rate Regularly defective transformers can be targeted to improve the quality ofsupply and reduce the O&M cost.

High harmonicsproblems

If a transformer encounters non-linear loading and therefore harmonicsproblems, losses due to Eddy currents might appear and reduce theefficiency of the transformers (Simulations carried out in [86] on 200 kVAtransformer showed an efficiency reduction from 98.1% to 96.5% due toharmonics). New transformers equipped with harmonics filters, better corestructure, etc. might improve the operating efficiency.

Factors affecting costs

Work constraints Cost impact

Transformersaccessibility

The conditions of installation works (outside, overhead, underground, in aprivate building, etc.) can have a significant impact on implementation costs.

Transformer powerrating

According to the study made on the APEC countries [46], extra-cost relatedto switching to higher standards is more cost-effective with smalltransformers. However, in case of proactive replacement, transformers ofmedium power rating replacement are more cost-effective than replacingsmall or very large ones.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Investmentoptimisation

Cost impact

Age of the transformer The residual value of an old transformer can be close to zero. Itsreplacement has thus a limited accounting impact.

Other replacementplans / newtransformers

As already explained, transformers can be replaced for various reasons, andit is cost-effective to seize those occasions to implement low-losstransformers. Moreover, the implementation of a new transformer can becoordinated with a voltage increase plan or plan to perform voltage control(by installing on-load tap changing transformers).

In general, changing the minimal requirements for all new transformersimprove the long term efficiency of the asset population

Technologystandardizationopportunities

Limiting the number of technologies present on the grid leads to O&M costreduction (less variability in the types of operation; lower cost inmaintenance and procurement)

13.1.1.2. EXPANDING LINE CAPACITY



Age of the network The use of new cables with up-to-date technological standards can reduce therisks of defaults/malfunctioning and improve continuity of supply.

Load rate Heavily loaded feeders generate higher losses. Heavily loaded feeders arealso the key target of replacement plans, to increase grid capacity. From anenergy efficiency perspective, the cost-effectiveness of this measure isclearly maximized when carried out within a replacement plan which aims atreducing congestions and takes also into account the minimization of losses.

Fault rate When replacing lines presenting a high fault rate, due to previous faults,isolation damages, etc., oversizing of the lines can be considered.



Overhead /underground grid

Excavation works cost are much higher (depending on the area) than layinglines overhead. All benefits being equal, targeting the replacement ofoverhead lines is more cost-effective.

Urban / rural areas Cables or overhead lines installation can be made more quickly and moreeasily in rural area, resulting in lower cost.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval


Cost impact

Age of the network Targeting the replacement of old cables is more cost-effective thanks to theirlower residual value.

Replacement planalready scheduled

As stated in the previous points, various drivers could lead to a linereplacement plan. Opportunities to introduce new line size standard have tobe taken into account. Synergies can here be found with an increase of thevoltage level, where lines replacement may be needed to fit the new voltagelevel.


Limiting the number of technologies present on the grid leads to O&M costreduction (less various types of operation; optimization of purchasingoperations)

13.1.1.3. INCREASING VOLTAGE LEVEL



Feeders length An increase of voltage level could be particularly useful in rural areas, whichtypically present longer feeders.

Feeders load rate Current in the conductors is directly proportional to the voltage level.Heavily loaded feeders generate the highest marginal losses and wouldbenefit the most from a voltage increase.

Fault rate Lines with higher fault rate could be firstly addressed when considering thismeasure.

Current voltageproblems

If there are known local voltage problems (due to DG for instance),increasing voltage on those feeders could reduce clients complains andO&M cost. If DG injection must be curtailed due to too high voltage rise,increasing line voltage would have a mitigation effect.

Factors affecting the costs


Number of alreadyfitted assets

A key point in the cost-effectiveness of the measure is the number of assets(feeders, transformers, etc.) to be replaced. If the insulation level of the assetis sufficient to stand higher voltages, the measure can be less costly todeploy. Otherwise the measure can be highly CAPEX intensive. In LV,going from a 230V to 400V network represent more significant cost (need ofa four-lines network, three-phases user installation adaptation, etc.)

Underground /overheadgrid

If the implementation of the measure requires asset replacement (andparticularly cable replacements), an underground grid will much moreexpensive to replace

Rural /urban area Asset replacement is easier and less costly in rural area, where the





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

accessibility of the installation is easier and where less work constraints aregenerally to be encountered.


Cost impact

Grid expansion / loadgrowth

If a grid expansion is envisaged to the future load growth, the voltage levelsupplying those zones has to be considered, especially if the expansion willrequire high feeder length. Feeders of higher voltage are cheaper (a 34.5kVfeeder can be 45% cheaper than a 12.5kV lines [87]), but on the othersubstations equipment is more expensive. Substation density per feeder is akey point when assessing the measure.

Age of the network Targeting assets at the end of their lifetime is more cost-effective thanks totheir lower residual value.


Limiting the number of technologies present on the grid leads to O&M costreduction (less various types of operation; optimization of purchasingoperations). A limitation of the number of voltages level also limits thenumber of transformers to be installed in the upper voltage level.

Combination withtransformers,substations or linesreplacement plans

All planned asset replacement actions represent an opportunity to increasethe number of asset fitted for higher voltages

DG integration plan If it is planned to invest in the grid to improve the hosting capacity ofdistributed generation, the opportunity to increase voltage level should beassessed.

13.1.1.4. DEMAND RESPONSE

Factors affecting the benefits


Feeder load rate Critically loaded feeders can be firstly address by such a measure.

DG curtailment Feeders where the DG’s are often curtailed will benefit from an interruptibleload, especially if curtailment compensation is to be paid by the gridoperator.



Flexibility market How the flexibility can be purchased is a key component, its price willimpact the day-to-day flexibility activation





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval


Cost impact

Smart Meter roll-outplan

When and to which extent the smart meters will be placed on the network isto take into account to define which feeders will be targeted first.

Penetration of thecommunication system

As the above, the extent to which the load can be easily controlled thanks toan existing communication tool impact the cost of measure deployment.

13.1.1.5. FEED-IN CONTROL OF DISTRIBUTED ENERGY GENERATION & STORAGE



Level of harmonics Harmonics injected in the network by consumer appliances can lead toPower Quality problems and to an increase of losses in the transformersIncentives to place efficient inverters can be made where high levels ofharmonics are met.

Feeder load rate Loss reduction is always most important to be performed on highly loadedfeeder. However, depending on the situation, DG can sometimes increase theload rate. Storage solution could also be envisaged.

Distance between DGand consumers

The placement on a DG unit on a feeder can have a significant impact on theincrease/decrease of losses. In [88], it has been demonstrated that, dependingon the load pattern, an optimal placement on a radial feeder can lead to a lossreduction between 71% and 86%.

Poor power factor Advanced inverter can enhance the Power Factor by controlling theinjection/absorption of reactive power. Incentives to promote those inverterswhere a local poor power factor is known can be made.

Voltage qualityproblems

Following the above, good inverters can control the reactive power andtherefore the voltage. Local voltage problems, especially those leading to abad DG integration, can be solved by good inverters.




If extra connections works are carried out, excavation works will increase thecosts.

Participation to the userasset cost

Depending on the local regulation, DSO can participate to the cost of theinverter to promote efficient equipment.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval


Cost impact

Smart Meter roll-outplan

With a complete access to the local consumption data, an inverter or itscentralised control can make the best forecast and control. A Smart Meterroll-out is a good opportunity to improve the grid efficiency.


The choice of the control solution (stand-alone, centralised or decentralised)can be made depending on the existing communication system on thetargeted grid.

13.1.1.6. NETWORK RECONFIGURATION



Number of switches /substation on a feeder

The impact in terms of reduction of losses and of restoration times highlydepends on the number of remote switches which are included in the gridreconfiguration. The optimal balance between number of switches and levelof benefits should be assessed depending on the specific grid conditions.According to the Synergird study [89], moving the open-point in the middleof the loop can lead to 66% of reduction compared to an end of loop open-point on a feeder with 4 substations, but this reduction can go up to 73% inthe case of 16 substations loop.

Current unbalancebetween feeders

A load repartition of 70%/30% between two feeders can lead to 1.4 times thelosses in a case of a 50%/50% repartition ( [90]). This makes it clear that thefirst feeders to be addressed by such a measure should be feeders with strongunbalances.

Current averagerestoration time

Feeders with an high restoration time, due to the distance between switches,the accessibility of the installations, the number of switches to be controlled,etc. highly benefits from the implementation of remote switches.EnergyUnited (North Carolina – see in [37]) reduced its Customer MinutesInterrupted (CMI) of 90% on specific feeders thanks to automatically remotecontrolled switches.

Current fault rate Feeders with a high fault rate would particularly benefit from this solution,reducing the total yearly interruption time.

Presence of sub-transmission /reparation grid

Repartition grids carry more load and impact more consumers than the finalgrid on a same voltage level.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval




Installation of switches on overhead lines does not represent major workInstallation cost for underground grid would be significantly higher.

Urban / rural areas Field deployment of remotely controlled switches is generally easier and lesscostly in a more rural area, where installations can be more easily accessible.

Voltage level Higher voltage levels require switches of higher insulation, which are moreexpensive.


Cost impact

Existing assets Feeders already equipped with remotely controlled switches would mainlyneed software adaptation to integrate new functionalities (e.g. for balancingload; for reducing restoration time). Investment will be much lower than inthe cases where the implementation and integration of new switches is to beforeseen.

Substation renovationplans

Old substations which must be restored can be easily targeted by such aplan.


The implementation cost of this solution could greatly be diminished if arobust communication system already exits.

13.1.1.7. SWITCHING OFF REDUNDANT TRANSFORMERS



Number oftransformers in parallel

Substations including more than two transformers working in parallel areexpected to have higher energy efficiency potential.

Load rate At a certain load rate level (according to [89]), it might become moreefficient in term on losses to connect both transformers to feed the load at thesame time. Load pattern of one substation must be measured to assess thebest transformers operation mode.

Low transformers faultrate / Low SAIDI

Switching off redundant transformers has the negative side effect to increasethe restoration time, since the transformer needs to be magnetised beforereaching its full operation mode. The measure will be more efficient in caseof feeders with a low transformer fault rate and a non-critical SAIDI. In thecase of manual switching, the restoration time and cost increase significantly.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval



Cost impact

Substation renovationplans

Transformers switches replacement will be optimised if other works are tobe carried out in the targeted substation.

13.1.1.8. VOLTAGE OPTIMISATION



Low power factor The lower the power factor, the higher the potential benefits of reactivepower compensation. Specific analysis shall be carried out to select mostsuitable feeders to be targeted.

Voltage stabilityproblems

Feeders where known voltage problems exist could be firstly targeted.

DG penetration rate Feeders with high levels of DG injection can be affected by voltageproblems and thus could represent good candidate for the implementation ofthe measure.

Low (advanced) DGpenetration rate.

If DG units on the feeders already performed voltage control, CVR would beredundant.

Feeders load rate Power factor correction is expected to be particularly effective on heavilyloaded feeders, leading to higher benefits in terms of reduction of losses andof congestions.

Number of voltage level If power factor correction is performed in an upper level of the grid, theeffect will impact all the downstream levels. Implementing reactive powercompensation on a few places on a sub-transmission grid can be more costeffective than implementing it in various locations at lower voltage level.

Length of the network Following the above, the longer the path between the reactive consumptionand the compensation device, the higher the losses.

Line specificcharacteristic

According to [23], a low R/X ratio leads to a less efficient reactive powercontrol by DG converters.



Underground/Overheadgrid

The installation cost is significantly higher in the case of an undergroundgrid. Installation of capacitors banks or voltage regulator on overhead linesdoes not involve major works.

Urban / rural grid Accessibility of the installation is easier in rural condition. Moreover, longfeeders in rural areas are the most subject to voltage drops.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Voltage level Voltage level has a significant impact on the cost of one capacitor.According to [91], purchase cost for 4.16kV capacitors is 20$/kvar and25$/kvar at 13.8kV.


Efficiency impact

Combination with CVR Assets for voltage optimisation at the LV level can also be leveraged toperform Conservation Voltage Reduction (CVR), as presented in thefollowing section.

Smart meter-roll outplan

Implementation of Smart Meters would greatly increase the observability ofthe grid (measures regarding voltage profile and reactive/active power flows)supporting the implementation of voltage optimization measures.

Substation renovationplan

Works to be carried out in substations can be opportunities to installcapacitors banks.


Depending on where Volt/VAR control assets will be installed and on theirnumber, the availability of a robust communication system is important, andmust be taken into account.

13.1.1.9. CONSERVATION VOLTAGE REDUCTION



Feeders load rate CVR is highly scalable. Targeting only specific highly loaded feeders canalready lead to a major losses reduction (the U.S. Department of Energydetermined in [59] that if only 40% of all feeders would have been equippedwith CVR, 80% of the target can be reached, and 37% of the objective whendeployed only on 10% of the feeder). The main impact of CVR can beachieved with a limited cost. Test can be made before deploying thesolution.

Voltage Stabilityproblems

Feeders encountering regularly voltage problems may be the first target of aCVR deployment plan.

Number of final LVcustomers

CVR objective is to reduce the end-user consumption by reducing voltagelevel (while remaining in the required nominal voltage range). Therefore, themore customers are connected on the LV, the more effective the impact willbe in term of losses reduction and on peak reduction, leading to customer’speak reduction and to investments deferral.




The installation cost will be significantly higher in the case of anunderground grid. Installation of capacitors banks or voltage regulator on





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

overhead does not represent major works.

Urban / rural areas Field deployment of CVR will be facilitated in a more rural area, whereinstallations can be easily accessible. Moreover, rural grids are usually notmeshed, allowing an easier voltage control.


Cost impact

Existing assets Feeders already equipped with capacitors banks, feeder voltage sensors oron-load tap changing transformers only needs automation to perform CVR.

Smart meter roll-outplan

End-users voltage monitoring facilitates and optimizes the implementation ofthe measure, and limits the number of voltage sensors to deploy.

Transformersreplacement plan

To be more effective, the first substation targeted by the measure might beones where transformers have to be replaced. On-load tap changingtransformers can therefore be installed.


In the case of centralized CVR, the implementation will be facilitated if arobust communication system already exits.

13.1.1.10. BALANCING 3-PHASE LOADS



Feeders load rate The higher the power demand on one feeder, the more the imbalance willhave an effect on the losses and on the cable capacity.

Neutral conductorsload rate

Current flowing in neutral conductors also results in losses in a 4-linessystem, and can cause protections to work in case of overloading. Feederswith frequently overloaded neutral conductors can be firstly addressed.

Neutral conductorsfault rate

As the above, fault related to the neutral conductors can be reduced, as nocurrent would normally flow in it.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Factors affecting the cost



Installing switches and replacing connections can be made cheaper in anoverhead grid.

Urban / rural grid Switches installations can be cheaper if an easy access is available.

LV consumers’ density The more LV consumers, the more adaptation work will be need to performan optimal balancing. Grid with less but more important connection willrequire less works, but might be less accurate.

LV substation density As the above, the number of LV substations is important to assess to whichextend and with accuracy the measure can be implemented.


Cost impact

LV substationrenovation plan

If LV substations are to be renovated, it presents an opportunity to installedphase switches.

Connectionsreplacements plan

Every connection replacement is an opportunity to increase the number ofthree-phases connection.

Smart Meters roll-outplan

Phase connection balancing requires measuring the instantaneous load. IfSmart Meter are to be installed, such measures will be provided.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

14. ANNEX II ENERGY EFFICIENCY MEASURES INEUROPEAN MEMBER STATES: SURVEY

In order to collect experiences and best practices from the field and to integrate the views of mainstakeholders in this study, a survey was carried out among stakeholders with respect to theimplementation of energy efficiency measures in gas and electricity grids. Mainly DSOs havebeen targeted, as most of losses occur at distribution level. European DSO associations(EURELECTRIC, GEODE, EDSO, CEDEC) have been involved to facilitate the distribution ofthe questionnaire

In particular, the survey aimed at collecting information about:

Their experience with the implementation of grid energy efficiency measuresThe Regulatory mechanisms in place in their country to incentivise the implementation ofenergy efficiency measures.Their view on the list of energy efficiency measures proposed in this study

The detailed description of the questions in the survey and the detailed list of respondents isreported in ANNEX I.

Despite numerous attempts it was not possible to have CEER’s view on the regulatorymechanisms to incentivise energy efficiency measures and their view on the level of mobilizationin the implementation of article 15.2 in their members’ countries.

A total of 16 replies have been collected, providing a good geographical coverage of EU MemberStates. Since the great majority of respondents were electricity DSOs, in this chapter we willmainly focus on the experiences of DSOs in electricity grids. Figure 56 reports the list ofrespondents.

Collected responses have provided very useful insights in the implementation of grid energyefficiency measures by DSOs in different Member States, facing different grid conditions.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Figure 56 - Overview of respondents to the survey

14.1.1. Level of mobilization in Member States regarding theimplementation of article 15.2

According to collected information, the level of mobilization in EU Member States in theimplementation of the provisions of article 15.2 (definition of investment plans to improve gridenergy efficiency) appears to be generally limited. European associations report that most of theirmembers have not yet been actively involved by Member States in addressing the provisions ofarticle 15.2. However, as was confirmed, a number of Member States are working on the topicand would likely publish their assessments in the period after the completion of this study.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Examples of available information from some Member States include:

UK:

OFGEM (UK regulator) is carrying out a study to comply with the requirements of theregulation [2]. A separate assessment of energy networks in Northern Ireland is beingprepared by the Northern Ireland Authority for Utility Regulation.In this context, OFGEM has launched a survey among UK system operators to collect bestpractices and information to- Assess the energy efficiency potentials of the gas and electricity infrastructure of Great

Britain- Identify a list of concrete measures and investments for the introduction of cost-effective

energy efficiency improvements in the network infrastructure, with a timetable for theirintroduction.

Findings have to be delivered before or on 30th June 2015 to the Secretary of State.FINLAND:

The proposal for the transposed legislation was discussed and handled in the FinnishParliament during autumn 2014 and the legislation has come in force beginning of 2015.In May 2015, the Finnish Ministry of Employment and Economy together with the EnergyAuthority have launched a survey on fulfilling the requirements of Article 15(2). Authoritieshave also invited other stakeholders (like Finnish Energy Industries and Finnish GasAssociation) to join the survey. The survey is carried out by Lappeenranta University ofTechnology (LUT) and will end in June 2015.The main topics of the survey are:- the technical losses and the overall efficiency of the system- the technical losses of the Finnish electric system are around 1%- so the focus will be more

on the overall energy efficiency of the system, including distributed generation, demandresponse and energy storage. New technology like smart meters, smart grids and LVDC(Low Voltage Direct Current) distribution system will have a big role.

- the concrete measures and investment will also be identifiedBELGIUM

In the context of the implementation of the article 15.2 by 30th June 2015, Synergrid(federation of electricity and gas network operators in Belgium) carried out an assessment ofthe energy efficiency potentials of gas and electricity infrastructure, together with the energyregulators (FORBEG, Forum for Regulatory Bodies).The scope of the study considers energy efficiency in a larger sense than just the reduction oflosses. In particular, the study looks at:- 1. Potential for reduced grid losses and reduced energy consumption by the network

operators.- 2. Potential to improve the efficient operation of available energy infrastructure, which in

turn could allow reducing the need for investing in new infrastructureCategories of measures that will be considered:- Investing measures by the network operators- Operational measures by the network operators- Changing the behaviour of the consumers, considering incentivizing Mechanisms (e.g.

tariffs by Time-of-Use, or flexibility vs. capacity tariffs)





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

A public (shorter) version of the report has been made available. The document will serve assupport for regulator to review the investment plans from the grid operators.

SWEDEN:

The Swedish Energy Agency has carried out a study to assess the energy efficiency potentialof gas and electricity grid in Sweden.The energy saving potential in grid infrastructure is considered to be relative low, althoughnot to be neglected- The main share of the loss reduction potential in electricity grids by 2030 could be

obtained by changing the location of electricity generation units. This is however outsidethe responsibility of grid operators.

- The potential that is achievable through measures conducted by the grid companiesthemselves is around 175GWh/year by 2020 and 400GWh/year by 2030. In relation tototal losses, these potentials are quite small: 4% by 2020 and 7% by 2030.

- Standards for new transformers through the EcoDesign Directive are expected toincentivise energy efficiency improvements.

- Losses in gas grids are considered very limited and therefore investments in energyefficiency measures would not be justified.

A key remark made by the study is that reduction of losses should not be the main driver ofgrid investments, as that can lead to sub-optimal solutions. It is important to adopt a systemsapproach for energy efficiency.In this respect, the introduction of regulatory incentives to improve more efficient networkoperations are expected to support investment plans which take also loss reduction intoaccount.

14.1.2. Summary of key insights of the survey

The following insights have been derived from the analysis of the survey respondents. Furtherliterature analysis has been carried out to further validate received replies:

State of play in Member States

Significant differences across Member States concerning the level of losses.Starting conditions of grids (e.g. voltage level; number of transformers) and potentials forenergy efficiency improvement are different. Transposition of best practices should be donewith care.

Drivers

Energy efficiency within larger grid investment decision plans - Energy efficiency istypically taken into account in grid investment decisions, however often it is not the maindriver. Energy efficiency is considered as one of the variables of a larger investmentoptimization plan. The importance of energy efficiency as a criterion in grid investmentdecisions clearly depends on the existence of regulatory incentives to reduce losses.Compliance with energy efficiency obligations: In specific cases, energy efficiencyinvestments (particularly concerning loss reduction) are undertaken to comply with regulatoryrequirements and quality standards (e.g. compliance with the EcoDesign directive leading toinstallation of higher efficiency transformers)





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Mapping of losses

Electricity losses mainly occur at distribution levelThe implementation of adequate metering systems at substations and at customers’ premises isan important first step to identify where losses are actually occurring in the distribution grid.The following factors have a key impact on the level of losses in the system- Loading (including peak demand)- Number of energized transformers- Lengths of the feeders and grid voltage levels- Level of power quality- Presence of distributed energy resources

Catalogue of energy efficiency measures

A variety of energy efficiency measures exist, and their potential is very much related to thespecific local grid conditions.In general, system operators opt for traditional investments (replacement of assets;reinforcements etc.) to improve grid energy efficiency.

Factors impacting grid energy efficiency potential outside control and responsibility ofsystem operators

Smart Grid-related EE measures (e.g. feed-in control) are less common: in many cases theright regulatory framework is not in place to make them a viable option.External factors outside the control of system operators can also greatly impact the potentialof EE measures: e.g. the development and location of DERs.

Cost-effectiveness

In general, a number of factors can greatly impact the cost-effectiveness of energy efficiencymeasures, including:

Monetary value of losses, i.e. the average price of wholesale electricity (the higher, the morecost-effective are EE measures)Loss reduction is a side-benefit of investments driven by other needs. Therefore the cost-effectiveness of an energy efficiency measure depends on additional benefits in terms ofCAPEX/OPEX reduction.When implementing EE measures, additional benefits in terms of reduction of CAPEX andOPEX should be considered. However, for some measures, a trade-off needs to be foundbetween improvement of energy losses and other benefits (e.g. trade-off with hosting capacityfor DER voltage measure).Parallel developments (e.g. roll-out of smart meters; built-in smart functionalities in DGconverters etc.) make the enabling infrastructure for certain EE measure already available,reducing their implementation costs.Most adopted EE measures concern implementation of higher efficiency assets (e.g. highercapacity cable; higher efficiency transformers) within a planned asset replacement strategy.An overall investment optimization strategy is therefore key to ensure that most cost-effectiveenergy efficiency measures are implemented.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

14.1.3. Energy efficiency measures implemented by surveyrespondents

Figure 57 provides a brief synthesis of grid energy efficiency measures implemented byrespondents. The survey results show that classical measures such as transformers replacementand line capacity increase are the most commonly adopted measures, as their implementation fitswith traditional replacement programs, where old or inefficient assets are targeted. Reactivepower Management is also well adopted, since voltage quality and distributed generation can alsobe addressed.

Figure 57 - Grid energy efficiency measures in electricity grids adopted by survey respondents

0

2

4

6

8

10

12

Num

bero

fDSO

adop

tingt

hem

easu

re

Number of DSOs adopting the measure





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Table 22 shows examples of achieved energy efficiency results thanks to implemented measures:

Measures Implementationexample

Highlights of lossreduction

Costs Additionalbenefits

High efficiencytransformers

VSD is replacing 100transformers/year since2011 – 10% have beenreplaced so far.CEZ Bulgaria replacedtheir four 110kV/20kVtransformers with thehighest losses.Enexis is replacing theiroldest MV/LVtransformers by morerecent re-usetransformers.

VSD expects 8.5% oflosses reduction.CEZ Bulgaria saved6.7 k€/year on lossesper replacedtransformer.In the UK, low losstransformers appearas one the mostefficient solution.

For CEZ CzechRepublic, costsincrease of atleast 20%.Enexis isreplacing oldMV/KVtransformers for5.3k€/transfo

Higher qualityof supply,loweroverheating

Increasing linecapacity

VSD increase theirmaximum cross sectionof their LV lines from70mm² to 150mm².

Belgian DSO’s aredeveloping a tool toincorporate losses in theselection of the minimalcable section.

VSD expects onaverage a 5% (locallyup to 20%) reductionof losses,

Higher voltagequality, lowerprobability offault, increasecapacity

Voltageoptimisation /Reactive powermanagement

CEZ Bulgaria equippedthree of their 110kV/10-20kV substation withcapacitors andharmonics filtering units.

CEZ Czech Republic isconducting a pilotproject where 20synchronous generatorsin 110kV receive setpoint in voltage andreactive power.

ErDF implementedVolt/Var control ofdistributedgeneration. 2-3% ofloss reduction isreported.

CEZ CzechRepublic performeda loss reduction of7% per line, leadingto 2.5% overallreduction.

CEZ Bulgariaequipped 3110kV/MVsubstation at acost of 528k€

DistributedGenerationintegration

Optimisationof the voltageprofile





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Measures Implementationexample

Highlights of lossreduction

Costs Additionalbenefits

Increasevoltage level

Vattenfal is currentlystudying area per area atoptimising the voltagelevel in their regionalgrid.

HEDNO and EDPincreased the number ofhigher voltagetransformers, thusincreasing the share ofhigher voltage portion ofthe grid.

Vattenfal expects a5% to 15% lossreduction from theincrease in voltagelevels, with locally25-30%

HEDNO and EDPincreased the numberof higher voltagetransformers, thusincreasing the shareof higher voltageportion of the grid.

ConservationVoltageReduction

Stromnetz Berlin hasalready conductedvoltage reduction

According to anOFGEM study, theincrease share ofvoltage-insensitiveloads is underminingthe potential of thismeasure.

Table 22 - Highlights of energy efficiency measures in electricity grids implemented by survey respondents





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

15. ANNEX III - SURVEY - QUESTIONNAIREFORMAT

A. Project CONTEXT and goal of the QUESTIONNAIRE

According to Article 15(2) of the Directive, Member States shall ensure, by 30 June 2015,that:- (a) an assessment is undertaken of the energy efficiency potentials of their gas and

electricity infrastructure, in particular regarding transmission, distribution, loadmanagement and interoperability, and connection to energy generating installations,including access possibilities for micro energy generators;

- (b) concrete measures and investments are identified for the introduction of cost-effective energy efficiency improvements in the network infrastructure, with a timetablefor their introduction.

The European Commission has recently launched a study to provide support to MemberStates in fulfilling the requirements of article 15.2 of the Energy Efficiency Directive- The study is titled “Measures for grid energy efficiency in gas and electricity grids”- The European Commission considers that cooperation of relevant stakeholders (system

operators, regulatory authorities) is a key element to maximize the pertinence and impactof the study.

The goal of this questionnaire is to collect data and best practices from systemoperators:- Description of energy efficiency measures for electricity grids (measures for reducing

technical losses)- Description of energy efficiency measures for gas grids (measures for reducing gas losses

and for optimizing use of energy in grid operations)

B. Survey Part 1 - Energy efficiency in electricity grids

1) Please provide a description of your specific grid conditions :

Average number of customers perKM2

Average length of feeders

Number of customers

Number of voltage levels in thegrid?

Length of the network per voltagelevel?

Number of transformers?

Average age of transformers?





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Level of losses on the grid in thelast five years?

2) Please provide a short description of the on-going and planned energy efficiencymeasures (measures for reducing technical losses):

Are there any regulatory incentivesto reduce grid energy losses?

Which energy efficiency measureshave been taken so far?

Which energy efficiency measuresare you planning?

3) Please find below a proposed list of energy efficiency measures. Could you pleaseprovide your comments?





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

4) Please provide a description, as detailed as possible, of energy efficiency measures youhave implemented or are planning to implement.

DESCRIPTION OF ENERGY EFFICIENCY MEASURES –ELECTRICITY GRIDS

Which measure? Please describe.

Level of penetration of the measure?(e.g. number of impacted feeders,transformers…)

Which losses reduction has beenobserved (%)?

Which factors have a highsensitivity on the cost-effectivenessof the measure? (e.g. age of asset,peak load,…)

Additional Benefits? (reduction ofCAPEX/OPEX,…)

Estimation of the cost of themeasure?

Which lessons learned to optimizethe implementation of this measure?

C. Survey Part 2 - Energy efficiency in gas grids

1) Please provide a description of your specific grid conditions by filling in this table:

Average number of customers perKm2

Number of customers

Number of pressure levels in thegrid?

Length of the network per pressurelevel?

Number of pressure reductionstations?





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Average age of gas heaters inpressure reduction stations?

Level of losses on the grid in the lastfive years?

2) Please provide a short description of the on-going and planned energy efficiencymeasures (measures for reducing gas shrinkage and for optimizing use of energy ingrid operations)

Are there any regulatory incentivesto improve grid energy efficiency(i.e. reduction of gas losses and ofuse of energy)?

Which energy efficiency measureshave been taken so far?

Which energy efficiency measuresare you planning?

3) Please find below a proposed list of energy efficiency measures. Could you pleaseprovide your comments?

Measures for reducing energy use for grid operation (e.g. optimization of compression; pressure profiles)

Gas heatingIncrease the use of multistage regulator lines, lines insulation, heat exchanger…Use of high-efficiency gas heaters in pressure reduction stationsUse of CHP

Cathodic protection Autonomous generation through solar panelsQuantity of gas delivered to LP users Optimization of the delivered pressure

Reduce transmission needs by using decentralizedgeneration

Biogas plantsStoragePower-to-gasMeasures for reducing gas losses

Scheduled for operationLower pressure through a closed by LP-gridOptimization of the valves placed on the grid to reduce the volume to evacuateIncrease the use of mobile flow interrupter

Unscheduled (leakages)

Replacement of old pipes to PE pipes% of steel pipes cathodically protectedNumber of valves on the networkUse of automatic flow-stopper on grid and connectionsLimitations of the operating pressure

Assessment of lossesSmart MetersFrequency of leakage detections





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

4) Please provide a description, as detailed as possible, of energy efficiency measures youhave implemented or are planning to implement. Please include additional slides todescribe additional measures.

DESCRIPTION OF ENERGY EFFICIENCY MEASURES –ELECTRICITY GRIDSWhich measure? Pleasedescribe.Level of penetration of themeasure? (e.g. number ofimpacted pipelines/pressurestations)Which losses reduction hasbeen observed (%)?Which factors have a highsensitivity on the cost-effectiveness of the measure?(e.g. pressure level, age ofasset,…)Additional Benefits?(reduction ofCAPEX/OPEX,…)Estimation of the cost of themeasure?Which lessons learned tooptimize the implementationof this measure?





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

16. ANNEX IV - TOOL FOR RANKING MEASURESACCORDING TO THEIR COST-EFFECTIVENESS

An excel tool has been set-up to guide project promoters in the assessment of the cost-effectiveness of different energy efficiency measures according to their local grid conditions. Thetool highlights the main factors affecting the benefits and cost of each measure and providesexamples for benchmark.

The goal of the tool is to help project promoters in selecting good candidate measures and comeup with an overall Energy Efficiency strategy.

This document is the support to an Excel tool, whose goal is to help the user to make up aqualitative cost-efficiency ranking. Those forewords aim to present the general methodologybehind this tool.

The analysis provided through the tool is divided in two mains steps:

1) Measure by measure impact and cost assessment2) Ranking of those measure in term of efficiency and cost

The figure here below shows the schematic organisation of the tool:





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Step 1 – Measure by measure analysis

The goal of this step is to assess, on one hand, the impact of one measure on the user’s specificgrid losses, investment planning and grid operations, and, on the other hand, to assess the cost ofits implementation. It is important to note here that all the final assessment is on the user’s hands,with strong guidelines provided by the Task 1 and 2, presented in this report. The analysis to bedone for one measure is presented in figure 2.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

ovalThe elements providing by the present report and being the guidelines of the tool are classified as

follow:

Grid factors: losses, CAPEX and OPEX reduction highly depends of the specific conditionsof each grid, such as the current technology in use, the fault rate, the average of assets,…Implementation constraints: present the elements of the grid how influence the cost of work,such as the situation of the lines (underground/overhead), facility of access, age of the assetsPlanned investments: list other investments which may provide opportunities for synergiesand reduce implementation costs.Asset list: list of the assets which should be implemented for the measure.

Step 1.1 – Impact assessment

The users will be guided in the tool following open question, based on open questions, in order toallow the user to build a complete analysis in function of his own grid factors. Each measureshould be assessed independently from the others, as well as each factor, following an exhibit aspresented down here:

MEASURE N Loss reduction Investment planningimpact

Grid operationimpact

Open question toguide expertsreflexions

Current load rate? Future capacityneeds?

Fault rate?

Possible penetration?

Expected impact … … …

Grid factors

Grid factors

Loss impact

Plannedinvestments

Implementationconstraints

Assets list

Costassessment

Invest. Planning impact

Grid op. impact

Grid factors





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

It is important to repeat at this point that the expected impact here is a totally free judgementfrom the user. It will serve as basis for the ranking of the measure in the next step.

Step 1.2 – Cost assessment

Secondly, the cost assessment of the measure has to be performed. As for the impact assessment,open questions guide the user to assess his costs in his own case, based on the local workconstraints, the asset list for a measure and the current investments already planned for theconcerned assets. It is especially for this last point that the final exhibit (presented here below) ofthis cost assessment make a parallelism between all the measure and the planned investment,asset by asset, in order the identify replacement already in the pipe and synergies betweenmeasures. If such synergies appear, the measure list should be adapted, and a new assessmentshould be done, taking this time the merged measures into account.

Investment currentlyplanned

Measure 1 Measure 2

Asset 1

Asset 2

Asset …

Step 2 – Ranking of the measures (pair-wise comparison)

Based on the step 1 analysis, the user should rank the measures between them, for each KPI andfor the cost. This is performed in the tool through a Multi-Criteria Analysis. The principle is thefollowing:

1) For one KPI, each measure is weighted in front of the others, between 0.1 (very lowefficiency compared to the other measure) and 9 (highly efficient).

2) A relative priority is then given for each measure (the higher the score, higher was theefficiency of the measure in front of all the others). An exemple of those two first step isgiven below

Measure 1 Measure 2 Measure 3 Measure… Relativepriorities

Measure1

1 1/3 1/2 … 0,3

Measure2

3 1 2 … 0,17





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

Measure3

2 0.5 1 … 0,04

Measure…

… … … 1 …

1) A priority should also be given to the KPIs. A MCA comparison is also perform betweeneach KPI. The ranking is made by taking into account the current level of losses, the targetedlevel, the regulatory framework, the presence of incentive aiming to reduce losses, CAPEX,OPEX… The relative priority calculated for each KPI is used to weight the differentpriorities given to one measure between them.

2) The sum of the weighted priorities gives the overall ranking of the efficiency of themeasures, such as showed in exhibit xxx.

KPI 1 KPI 2 KPI 3 Weightedpriorities

KPIs weights 0,5 0,2 0,3

Measure 1 0,3 0,02 0,4 0,29

Measure 2

Measure 3

Measure … … … … …

1) Following the same rules, a ranking of the costs is performed, and is used as qualitativeassessment of the costs.

2) The final cost-efficiency ranking is given by the ratio . It is important tonote the results provided here are qualitative, and should only be used to identify whichmeasure should be further investigate in more quantitative way. The graph provided herebelow give an exemple of the results, where “quick wins” and ‘must do” can be easilyidentified.





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

High efficiencytransformers

Increase of linescapacities

Increase of thevoltage level

Demand response

DistributedGeneration

Networkreconfiguration

Reactive powermanagement

CVR

Balancing 3 phases-loading

0,00 1,00 2,00 3,00 4,00

Final score/cost





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

17. ANNEX V- DETAILED SIMULATION RESULTS

17.1.1.1. REVERSE POWER FLOWS APPEAR WHEN INCREASING DG PENETRATIONRATE

17.1.1.2. IMPACT OF DG PENETRATION ON OPEX AND ON CAPEX





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

17.1.1.3. LOSSES COSTS IMPACT SIGNIFICANTLY THE SIZING OF THEINFRASTRUCTURE (CAPEX)

17.1.1.4. 100% PV CASE VS DIFFERENT DG TYPE SHARE AMONG THE VOLTAGE





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

17.1.1.5. COST EVOLUTION WITH NIGHT/DAY TIME-OF-USE

17.1.1.6. COST EVOLUTION WITH MIDDAY TIME-OF-USE





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

17.1.1.7. COST EVOLUTION WITH MULTIPERIOD TIME-OF-USE

17.1.1.8. LOSS REDUCTION THANKS TO A MULTIPERIOD TOU





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

17.1.1.9. LOW LEVEL OF SHIFTED ENERGY OVER THE YEAR PROVIDES ALREADY ATOTEX DECREASE

17.1.1.10. GENERATION CURTAILMENT





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

17.1.1.11. DETAIL COST REDUCTION WITH A MIXED FLEXIBILITY SOLUTION(CURTAILMENT + DEMAND RESPONSE)

17.1.1.12. PEAK REDUCTION ACHIEVED THANKS TO THE MIXED FLEXIBILITYSOLUTION

-2%0%2%4%6%8%

10%

CAPE

Xde

crea

se(%

)*

CAPEX

-10%-8%-6%-4%-2%0%2%4%

OPE

Xde

crea

se(%

)*

OPEX

0%

2%

4%

6%

8%

North rural South_rural North_urban South_urban

TOTE

Xde

crea

se(%

)*

TOTEX

100% DG penetration 200% DG penetration 300% DG penetration





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

17.1.1.13. USE OF DEMAND RESPONSE VS CURTAILMENT

17.1.1.14. USE OF THE DEMAND RESPONSE BETWEEN THE VOLTAGE LEVEL





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

17.1.1.15. TOTEX EVOLUTION FOR LOSSES COSTS AT 80€/MWH AND 120€/MWH

17.1.1.16. REDUCTION OF LOSSES DEPENDING OF THE PRICE OF THE LOSSES





This

docu

men

tis

the

prop

erty

ofTr

acte

belE

ngin

eerin

gS.

A.An

ydu

plic

atio

nor

tran

smis

sion

toth

irdpa

rtie

sis

forb

idde

nw

ithou

tpr

ior

writ

ten

appr

oval

17.1.1.17. COMPARISON OF THE TOTEX DECREASE BEWEEN THE BASE CASE AND100% PV ONE

TRACTEBEL ENGINEERING S.A.Avenue Ariane 71200 - Brussels - BELGIUMwww.tractebel-engineering-gdfsuez.com

Vincenzo GIORDANOtel. +32 2 773 83 92fax +32 2 773 88 [email protected]

A leading international engineering consultancy company, Tractebel Engineering is part of ENGIE*, an industrial groupwith the financial strength to address the challenges of the future. With approximately 4,400 people in 33 countries, weoffer life-cycle engineering solutions for energy, water and infrastructure clients. Services include a full range ofengineering assignments: Architect Engineer, Owner’s Engineer and Consulting Engineer. Our customers are private andpublic companies, as well as national and international institutions. *GDF SUEZ is now ENGIE