New Control Functions for 100% Renewable Generation: An …psc.rtpis.org/documents/Keynotes/2018 Clemson University... · 2019-05-19 · 1 MGC: Microgrid Controller 2 ESS: Energy

New Control Functions for 100% Renewable Generation:

An Industry PerspectiveUlrich Muenz Siemens Corporate Technology, Princeton, NJ

joint work with A. Mešanović, A. Szabo, D. Unseld, J. Bamberger, C. Ebenbauer, R. Findeisen

Restricted © Siemens siemens.com

Unrestricted © Siemens AG 2018

Sept 2018Page 2

Our global presence –

Partner to customers all over the world

A worldwide

presence

is the heart of the

Siemens brand – and

that goes for us as well.

This presence enables us

to quickly offer targeted

solutions that are tailored

to regional requirements.

Corporate Technology – worldwide locations

3 Employee figures: As of September 30, 2017

Corporate Technology –Our competence center for

innovation and business excellence3

4001,600patent experts

8,000 5,400employeesworldwide

softwaredevelopers

researchers


Sept 2018Page 3

Improving speed and focus with

“Company Core Technologies”

Siemens Company Core Technologies

Energy Storage

Simulation and

Digital Twin

Data Analytics,

Artificial Intelligence

Distributed Energy

Systems

Connected

(e)Mobility

Additive Manufacturing

Power Electronics

Blockchain

Applications

MaterialsAutonomous Robotics

Software Systems

and ProcessesFuture of Automation

Cybersecurity

Connectivity and

Edge Devices

Sie

men

s R

&D

Sp

en

din

g

FY14 FY17 FY18eCurrency: Billion EUR

4.0

5.2

~ 5.6

~+40%

~€500 million focused

investment in all innovation fields.


Sept 2018Page 4

Siemens Corporate Technology builds an Energy Living Lab

in Princeton, NJ to support energy R&D in the US

Evaluate new concepts

Prime research partner for

top universities and

national labs in US

Generate funding opportunities

Demonstration site for BUsContribute to Siemens CO2

footprint reduction

Provide uninterrupted power

supply

Increase level of comfort

Reduce energy costs

€€

Note: 1 MGC: Microgrid Controller 2 ESS: Energy Storage System

Siemens DESIGO

Building management

system for reduced

energy consumption

Siemens MGC1

Microgrid controller for

fully self-sufficient

microgrid operation.

Siemens ESS2

1000 kWh battery

storage system for

peak shaving.

Photovoltaic

System

836 kWp photovoltaic

system to reduce CO2

footprint.

Siemens Gas

Generator

400kW generators for

backup / peak shaving.

SINACON

Siemens inverter for

PV and ESS infeed

and to build up the grid

during outages.

Siemens

VersiCharger

Electric vehicle

chargers to supply

employees‘ vehicles


Sept 2018Page 5

High renewable generation is achieved today with hydro power,

but large scale hydro power is not available in many countries

Source: IEA: Energy Policies of IEA Countries: New Zealand 2017 Review

https://www.iea.org/publications/freepublications/publication/EnergyPoliciesofIEACountriesNewZealand2017.pdf


Sept 2018Page 6

High wind and PV generation is available everywhere,

but Wind & PV generation is still below 25%

Source: IEA: Energy Policies of IEA Countries: Australia 2018 Review

http://www.iea.org/publications/freepublications/publication/EnergyPoliciesofIEACountriesAustralia2018Review.pdf


Sept 2018Page 7

New Control Functions have to solve diverse challenges

to close the gap to 100% wind and solar generation

ChallengeNew Control

FunctionsNew HW

➢ Generation

meets load

➢ Long distance

power transfer

➢ Low inertia

grids

Storage

HVDC

Synchronous

Condensers

Demand

Side

Management

Robust

Optimal

Power Flow

Adaptive

Power

Oscillation

Damping10MW

Electricity generation from Wind & PV as a

percentage of the total generation in Island Grids

Japan

100MW 1GW 10GW 100GW

20%

40%

60%

80%

100%

?

UKIrelandOahu

Kodiak

Bonaire

King Island

EI Hierro

HawaiMaui

Power system size

in peak demand

Wind & PV

Gap

Data from:

RMI Renewable Microgrids: Profiles from islands and remote communities across the globe

https://www.hawaiianelectric.com/clean-energy-hawaii/clean-energy-facts/about-our-fuel-mix

https://www.renewable-ei.org/en/statistics/electricity/

https://www.gov.uk/government/collections/electricity-statistics

http://www.eirgridgroup.com/site-files/library/EirGrid/Generation_Capacity_Statement_20162025_FINAL.pdf

https://rmi.org/wp-content/uploads/2017/04/Islands_RenewableMicrogrids_Report_2015.pdf






Sept 2018Page 8


An Industry Perspective

New Control Functions

Robust Optimal Power Flow

Adaptive Power Oscillation Damping

Mitigate risks of

volatile generation

at minimal cost

Dynamically maximize

stability reserves

Cost &

constraint

Power

system

model

Uncer-

tainty

Control

action


Sept 2018Page 9






Cost &

constraint

Power

system

model

Uncer-

tainty

Control

action

Mitigate risks of

volatile generation

at minimal cost


stability reserves


Sept 2018Page 10

Volatile generation requires Robust Optimal Power Flow to

mitigate risks of volatile generation at minimal cost

Benefits

• Increased resiliency, e.g. for market clearance

• Fast response without communication

• Ready for brown-field applications

Status quo Change Solution

Robust OPF:

• Optimize both

setpoints and

droops/reserve to

enable fast power

flow adjustment

without fast

communication

• Recalculation of

R-OPF depending

on uncertainty

➢ Fast and scalable

optimization

algorithms.

Uncertain

power flows:

• Determined by non-

dispatchable, volatile

generation

• High line loading

because generation

far from load

➢ Fast adaptation of

power flow

required.

Predictable

power flows:

• Determined by

slowly changing,

dispatchable

generation

• Low line loadings

because generation

close to load

➢ OPF calculations

hours in advance.

4 7532

1

6

9

810

1

2

34

droop

droop

droop

droop droop

droop

Online Robust Optimal Power Flow

Generator Wind Park controller


Sept 2018Page 11

Robust Optimal Power Flow in a nutshell

Cost &

Constraint

Power

System Model

Control

Action

Uncertain

Renewable

Cost:• Disptach / redispatch costs

• Primary reserve costs

• Power line losses

Constraint:• DC voltage limits

• Generator active power limits

• Converter active power limits

• Power line limits

• Droop gain limits

• Frequency limits

AC/DC Converter parameters:

• VC0 – voltage reference

• P0 – active power reference

• kV, kP – voltage and frequency droop gains

Generators:

• P0: active power reference

• ΔPt: redispatch

• kP: primary reserve / droop

AC Grid:

DC Grid:

Uncertainty:

• active power infeed in a convex set

for t ∈ [t0, t0+T], e.g. T=15 min;

∙ 1 ∙ ∞


Sept 2018Page 12

Robust Optimal Power Flow combines uncertainties and

droop controllers in one optimization problem

❖ Minimize cost and maximize resiliency

• Power flow satisfies constraints for all variations

• Optimal usage of available resources (decentralized droop controllers) to minimize

cost and for fast reaction without (fast) communication

Robust OPF

Cost &

constraint

Power

system

model

Uncer-

tainty

Control

action

Cost &

constraint

Power

system

model

No

uncer-

tainty

No

control

action

Cost &

constraint

Power

system

model

Uncer-

tainty

No

control

action

Cost &

constraint

Power

system

model

Contin-

gency

No

control

action

OPF Stochastic OPFSecurity constraint OPF


Sept 2018Page 13

Robust Optimal Power Flow achieves primary reserve

sharing across asynchronous AC power systems

1 Load -100 MW at AC bus 6• Default droops lead to frequency limit violation in AC B

• Solved by slightly modifying droops of generators (AC A&C) and HVDC converters

➢ no re-dispatch required

2 Generation +200W at DC bus 5• Default droops lead to DC power line overloading between bus 5&6

• Solved by slightly modifying droops of generators (AC A&C) and HVDC converters

➢ no re-dispatch required

Cigré B4 DC Test Grid

• 3 AC and 3 DC grids

• Generation center AC A+B & Load center

AC B interconnected through DC system

• 4GW total generation and load

• Default droops: 5%

Cost

f / V / …Constraint Satisfied Constraint Violated

Operation with

high uncertainty

Operation with

high uncertainty

and redispatch

Operation with high

uncertainty and primary

reserve adaptation

3 Key Findings

• Robust OPF provides solution for complex primary reserve provisioning

• Improved resiliency at minimal cost


Sept 2018Page 14

Robust Optimal Power Flow is fast even for realistic power

systems sizes

Application Example: Danish 2020 Transmission System

1 Setup: Danish 2020 Transmission System

• 2 asynchronous AC systems interconnected through HVDC system;

• 300 buses, 4GW total generation and load.

2 Key Findings

• Robust OPF runs in less than 5 minutes for a 300 bus system with 200

uncertain infeeds

• Computation time is comparable for 1- and 2-norm undertainties, but

slightly larger for ∞-norm uncertainties


Sept 2018Page 15






Cost &

constraint

Power

system

model

Uncer-

tainty

Control

action

Mitigate risks of

volatile generation

at minimal cost


stability reserves


Sept 2018Page 16

High renewable integration requires Adaptive Power

Oscillation Damping to maximize stability reserves

Benefits

• Increased resiliency through stability reserve optimization;

• Low communication bandwidth required;

• Ready for brown-field applications.

Status quo Change Solution

Adaptive power

oscillation damping:

• Online estimation of

eigenmodes

• Online optimization

of PSS controller

parameters;

➢ Fast and reliable

optimization

algorithms.

Time-varying

eigenmodes[1]:

• Weather dependent

generation

• Faster power

system dynamics

because of PE

inverters;

• Reduced stability

reserve;

➢ Increased risk of

black-outs & poorly

damped oscillations.

Time-invariant

eigenmodes:

• Dispatchable

generation

• Power system

stabilizer (PSS)

damp oscillations

➢ Manual tuning of

PSS to damp

interarea modes.

[1] S. Al Ali, T. Haase, I. Nassar, and H. Weber, “Impact of increasing wind power generation on the north-south inter-area

oscillation mode in the European ENTSO-E system,” IFAC Proceedings Volumes, vol. 47, no. 3, pp. 7653–7658, 2014.


PSS

PSS

PSS

PSS

Communication

AC grid

Generator

Wind Park

PSS

Solar

PSS


Sept 2018Page 17

Tim

e D

om

ain

Optim

ized P

ara

ms

Adaptive Power Oscillation Damping in a nutshell

Power System

1110

0100

1110

0100

Modeling, calibration & verification

Power system

status and

forecasts

Optimized

controller

parameters

Fourier

Tra

nsfo

rm

Tim

e D

om

ain

Initia

l P

ara

mete

rs

Optim

ized F

ourier

Tra

nsfo

rm

Digital Twin of the Power System

Cloud: MindSphere

• Large Island (Europe) • Small Island (IREN2)

Nonlinear model: Linearized model: H-inf Optimization:


Sept 2018Page 18

Application example shows significant increase of power

oscillation damping for IEEE39 benchmark model

[2] P. Kundhu, Power System Stability and Control, McGraw-Hill, 1993.

[3] A. Moeini, I. Kamwa, P. Brunelle, G. Sybille, "Open Data IEEE Test Systems Implemented in SimpowerSystems for Education and Research in Power Grid Dynamics and Control," Power Engineering Conference (UPEC), 2015 50th International Universities,

1-4 Sept. 2015, Staffordshire University, UK. (https://www.mathworks.com/matlabcentral/fileexchange/54771-10-machine-new-england-power-system-ieee-benchmark)

[4] IEEE committee report, "Dynamic models for steam and hydro turbines in power system studies," IEEE Transactions on Power Apparatus and Systems, Vol. PAS-92, No. 6, 1973, pp. 1904-1915.

[5] "Recommended Practice for Excitation System Models for Power System Stability Studies," IEEE® Standard 421.5-1992, August, 1992.

IEEE 39 bus benchmark model

❖ Initial Parameters ❖ Optimized Parameters

❖ IEEE 39 bus model

from[3] with component

models from[4,5] and

PSS from[2].

❖ HVDC line between

buses 16 and 27.

❖ 216 states❖ 128 controller

parameters

Optimization

Problem

Characterization

1 8

10

2 3 5 4 7

6

9

=

˷˷˷

=

˷˷˷

C1

C2

1

2

3

4

56

7

8

9

10

11

12

13

14

1516

1718

19

20

2122

23

24

2526

2728 29

30

31 323334

35

36

37

38

39

Generator =˷˷˷ HVDC converter station


Sept 2018Page 19

Detailed power plant models can be optimized with this

approach

❖ 10 tunable controller parameters per generator❖ 19 states per generator

Decomposition of Generator

Synchronous

Generator

VT

Ɯ

Pm

Turbine +

Governor

Ɯ

Efd

AVR +

Exciter

Ɯ

PSS

VT

Vref

VPSS

Automatic Voltage Regulator

+ Exciter Model

PSS

Turbine + Governor


Sept 2018Page 20

Successful field test of Adaptive Power Oscillation

Damping in Wildpoldsried, Germany

❖ Optimization /wo Droops ❖ Optimization /w Droops❖ Initial Parameters ❖ Manually Tuned Parameters

❖ Status

• Modelling, calibration, and validation of simulations

• Validation of oscillation damping with standard HW

❖ Outlook

• Evaluation in customer projects

• Extension to larger island grids, e.g. Hawaii

• Evaluation for transmission systems

Load bank

Transformer

Batteries

Inv1

Inv2

Inv3

Inv4

Inv5

Inv6

time (s) time (s) time (s) time (s)

P (

kW

)

P (

kW

)

P (

kW

)

P (

kW

)


Sept 2018Page 21



10MW

Electricity generation from Wind & Solar as a

percentage of the total generation in Island Grids

Japan

100MW 1GW 10GW 100GW

20%

40%

60%

80%

100%

UKIrelandOahu

Kodiak

Bonaire

King Island

EI Hierro

HawaiMaui

Power system size

in peak demand

Wind & PV

Gap

Data from:

RMI Renewable Microgrids: Profiles from islands and remote communities across the globe





High renewable generation is achieved

today only with hydro power

Wind & PV generation is still below 25%

New Control Functions are needed to

bring Wind & PV up to 100%

Robust Optimal Power Flow mitigates

risks of volatile generation at minimal cost


dynamically maximizes stability reserves

1

2

3

4

5

https://rmi.org/wp-content/uploads/2017/04/Islands_RenewableMicrogrids_Report_2015.pdf






Sept 2018Page 22

Ulrich Muenz

Head of Research Group

Autonomous Systems and Control / US / CT RDA FOA ASY-US

755 College Road East

Princeton, NJ 08540

USA

Mobile: +1 609 216 0170

E-mail: [email protected]

siemens.com

Contact

mailto:[email protected]

Documents

New Control Functions for 100% Renewable Generation: An …psc.rtpis.org/documents/Keynotes/2018 Clemson University... · 2019-05-19 · 1 MGC: Microgrid Controller 2 ESS: Energy