Optimization and supervision of complex energy systems

Optimization and supervision of complex energy systems

Dr.Ing Dhaker ABBES

Professor-Researcher

Co-responsible of ESEA field

HEI-Lille

Yncrea- Hauts-DE-France

STA'2016

18/12/2016

2Plan

INTRODUCTION

TYPES OF OPTIMIZATION AND SUPERVISION

EXAMPLES AND CASE STUDIES

- Design and supervision of a hybrid wind-photovoltaic

system with batteries

- Design and supervision of a hybrid railway power station

- Day-ahead Optimal Operational Planning of Generators

CONCLUSION AND PERSPECTIVES

REFERENCES AND PUBLICATIONS

3INTRODUCTION

Context : Toward Smartgrids

Energetic optimization and reliability increasing in all types of

networks (terrestrials, offshore, on board, habitat, railway,…)

→ Network optimized management

Renewable Energy integration

Electrical vehicle integration

Demand management of electricity

Contribution of energy storage systems

4INTRODUCTION

Problematic :

How to integrate and manage all these sources of energy ?

5

Types of optimization and supervision

6Types of optimization and supervision

* Explicit Optimization

- Direct (from an explicit mathematical function to optimize)

Mono-objective optimization :


* Explicit Optimization

- Direct (from an explicit mathematical function to optimize)

Multi-objective optimization :


* Explicit optimization :

- Through the dynamic simulation of the system (example to minimize the energy in a building by dynamic thermal simulation of it.)

Source : Michael Wetter. GenOpt – A Generic optimization program. Juin 2001


• Implicit optimization :

- Mathematical model difficult to determine.

- Intuitive algorithms (examples: energy management of multi-source systems

with fuzzy logic, algorithms for maximizing photovoltaic or wind power MPPT, etc.).


• Supervision types :


• Real time supervision methodologies :

- Implicit methods using, for example, fuzzy logic, multi-agents systems or a combination of tools.→ Well adapted for complex systems where some quantities or

states are weakly foreseeable and therefore not well known(wind, lighting, network state, load,…).

- Explicit methods using objective function.→ Guarantee an optimal; for example; maximize renewable

energy production, …Difficult to implement in real time.

- Causal methodology

→ Inversion of power flows to determine reference power.

→ These require detailed models and good instantaneous

knowledge of flows and losses.

12

Case study 1 : Design and supervision of a hybrid wind-photovoltaic system

with batteries

Examples and case studies

13Examples and case studies

Case study 1 : Design and supervision of a hybrid wind-photovoltaic system with batteries



Example of explicit linear optimization:



Example of explicit linear optimization:

Linear optimization using Excel (SIMPLEX)


Example of non-linear explicit optimization:

- Proposed method based on a dynamic system simulation associated with an

optimization algorithm:

• LCC [€] : Life Cycle Cost

• EE [MJ] : Embodied Energy in the system

• LPSP [%] : Loss of Power Supply Probability




- Proposed method based on a dynamic system simulation associated with an

optimization algorithm:




A mono-objective formulation solved by the SQP algorithm(Newton-Wilson Algorithm)

Minimum LCC [€] or EE[MJ]




A mono-objective formulation solved by the SQP algorithm (Newton-Wilson Algorithm)





Finding the best compromise between LCC [€] , EE [MJ] and LPSP [%]

A multi-objective formulation resolved in two ways :

By a scalar method with unit weighting coefficients:

None of the "objective" functions is favored.

By a Pareto approach (notion of dominance) with the N.S.G.A-II method.




Finding the best compromise between LCC [€] , EE [MJ] and LPSP [%]

Optimization results in the case of a multi-objective optimization obtained by applying the scalar method

Representation of the compromise surfacePareto fronts in the case of a tri-objective optimization: LCC [€] vs EE [MJ] & LPSP [%]



Examples of extremal and intuitive optimization methods:

Example 1 : Tracking the maximum power point of a photovoltaic generator (MPPT)

Perturb & Observe methodP

V

P1

P2

P1

P2

P1 P2

V optimal

dP < 0

dP > 0

dP > 0

The system continuously adapts thevoltage across the PV array.

Irradiance change Changing the P-V characteristic

07,0 V

Initial voltage:

interest : improve system performance




Example 1 : Tracking the maximum power point of the photovoltaic generator (MPPT)

Perturb and Observe

Incremental Conductance

Parasitic Capacitance

Constant Voltage

Anti-islanding

Fuzzy MPPT

Etc. Incremental Conductance MPPT

http://vincent.boitier.free.fr/INSA/biblio_MPPT/Boitier_Maussion_MPPT_Vfinale.pdf

Principal used méthods

http://vincent.boitier.free.fr/INSA/biblio_MPPT/Boitier_Maussion_MPPT_Vfinale.pdf




Example 1 : Tracking the maximum power point of the wind turbine (MPPT)• In the absence of any knowledge of

wind turbine efficiency and tip speed

characteristics :

Extremal method such us fuzzy logic

• If wind turbine efficiency vs tip speed

characteristic is known :

Speed control :

Torque control :

http://ethesis.inp-toulouse.fr/archive/00000079/01/mirecki.pdf

http://ethesis.inp-toulouse.fr/archive/00000079/01/mirecki.pdf



Supervision

SOC and battery power output for 1 day.

Photovoltaic generator and wind turbine power output for 1 day.

26

Case study 2 : Design and supervision of a hybrid railway power station






Design with explicit nonlinear optimization:

Optimization solving algorithmSQP/GA

Extended optimization strategy (month/year)

Optimization problem & model formulation

· System description & model components· Specifications (decision variables, objective

functions & constraints)

Optimization results analysis

Extension of the temporal study domain

Optimal solution foundAcceptable convergence rate

Respect of variable constraints

Y

N

Organization chart of IFTEH sizing optimization method (Collaboration with the Optimization team)



Design with explicit nonlinear optimization:

Variables Constraints Objective function

𝐒𝐏𝐕, with 𝟎 < 𝐒𝐏𝐕 < 𝐒𝐏𝐕_𝐦𝐚𝐱𝐒𝐰, with 𝟎 < 𝐒𝐰 < 𝐒𝐰_𝐦𝐚𝐱

𝐏𝐦𝐚𝐱_𝐬𝐭𝐨, with 𝐏𝐬𝐭𝐨_𝐢, 𝐢 = 𝟏. . 𝟐𝟒

−𝐏𝐦𝐚𝐱_𝐬𝐭𝐨 ≤ 𝐏𝐬𝐭𝐨_𝐢 ≤ 𝐏𝐦𝐚𝐱_𝐬𝐭𝐨𝐄𝐦𝐚𝐱_𝐬𝐭𝐨

𝐄𝟎 = 𝐄𝐟𝐄𝐦𝐢𝐧_𝐬𝐭𝐨 > 𝟎

Minimizing total cost Min(C)

Description of optimization problem with SQP algorithm

Variables Objective function

𝐒𝐏𝐕, with 𝟎 < 𝐒𝐏𝐕 < 𝐒𝐏𝐕_𝐦𝐚𝐱𝐒𝐰, with 𝟎 < 𝐒𝐰 < 𝐒𝐰_𝐦𝐚𝐱

𝐏𝐬𝐭𝐨_𝐢, 𝐢 = 𝟏. . 𝟐𝟑with 𝐏𝐬𝐭𝐨_𝟐𝟒 = − 𝐢=𝟏𝟐𝟑 𝐏𝐬𝐭𝐨_𝐢

and −𝐏𝐦𝐚𝐱_𝐬𝐭𝐨 ≤ 𝐏𝐬𝐭𝐨_𝐢 ≤ 𝐏𝐦𝐚𝐱_𝐬𝐭𝐨

Minimizing total cost Min(C)

Description of optimization problem with genetic algorithm



Design with explicit nonlinear optimization:Solution found with SQP Solution found with GA

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24-1

-0.5

0

0.5

1

1.5

2

Time (hour)

Pow

er (

p.u.

)

PPV

Psto

Pgrid

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24-1

-0.5

0

0.5

1

1.5

2

Time (hour)

Pow

er (

p.u.

)

PPV

Psto

Pgrid

Generators and load power profiles for obtained optimal solutions

Optimization variable

(Spv)

Cost function (C) Evaluations

SQP 200000m² 39736k€ 5351

GA 221500m² 42342k€ 26100

Numerical results of optimization


Energy bill components:

• limitation of the exceeding subscribed grid power

• limitation of the adjustment gap

Energy management services are related to:

• energetic aspects (reliability, power quality…)

• environmental (favoring renewable consumption)

• economic aspects (electricity bill reduction)

Multi-criteria approach for HRPS supervision (with implicit optimization)

Short-term actions

Wholesale market

(purchased 24 h in advance)

Public electricity network

user tariff (10min index)

Adjustment mechanism

(gap between purchased

power and real consumption)



Methodology for HRPS energy management

STEP 1Work specifications

STEP 2Design of the supervisor

STEP 3Chart representation of operating modes

- Functional graphs -

STEP 4Determination of the membership functions

STEP 5Chart representation of fuzzy operating modes

- Operational graphs -

STEP 6Determination of the fuzzy rules

STEP 7Determination of indicators to measure

the achievement of objectives


Methodology for HRPS energy management (implicit optimization)

STEP 1Work specifications

Objectives Constraints Means of actions

Predictive mode – LONGT TERM

Reducing energy bill

(regarding short-term trades)

Trains consumption predictions

RES forecast

Electricity market fluctuations

Storage power (Psto-ref-lgt)

(Predictive reference

power)

Fuzzy Logic energy management – SHORT TERM

Limitation of subscribed power exceeding

Favoring local RES consumption

Ensuring storage system availability

Subscribed power

Storage limits

RES availability

Storage power (Psto-ref-sht)

(Predictive mode

adjustment)




STEP 2Design of the supervisor

FL energy management

+-

Ptrain

PRES

ΔPlocal

SOC

ΔPexcess

K1

K2

K3

K5

+-

Pgrid

Psubscribed Psto_ref_sht

Predictive mode

Long-term

Ptrain_predictive

Electricity cost

Psto_ref_lgt

HRPS supervision

PRES_forecast

K4

K1, K2, K3, K4, K5 = normalisation gains

To favor local RES consumption

To ensure the storage availability

To limit exceeding subscribed power

To reduce the electricity bill

the grid power excess amount

the power difference between train

consumption and RES production




STEP 3Chart representation of operating modes

- Functional graphs -

1. To limit exceeding grid power by load shedding2. To favor the RES local

consumption

3. To ensure storage discharging when near

the highest level of saturation

Storage level is Small Storage level is Big

Storage level is Medium Storage level is Medium

N1 N3

3. To ensure storage charging when near the

lowest level of saturation

N2

|

|

|

|

Inputs:• ΔPlocal (Negative-Big, Zero, Positive-Big)• SOC (Small, Medium, Big)• ΔPexcess (Negative-Big, Zero, Positive-Big)• Psto-ref-lgt (Negative-Big, Zero, Positive-Big)Output:• Psto-ref-sht (Negative-Big, Zero, Positive-Big)

STEP 4Determination of the membership functions

-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1

0

0.5

1

Plocal

Plo

cal

Negative-Big Positive-BigZero

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

0

0.5

1

SOC

SO

C

Small Medium Big

-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1

0

0.5

1

Pexcess

Pex

cess

Negative-Big Zero Positive-Big

-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1

0

0.5

1

Psto-ref-lgt

Pst

o-re

f-lgt

Negative-Big Zero Positive-Big

-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1

0

0.5

1

Psto-ref-sht

Pst

o-re

f-sht

Neg-Big Zero Pos-MedNeg-Med Pos-Big



STEP 5Chart representation of fuzzy operating modes

- Operational graphs -

STEP 6Determination of the fuzzy rules

|

|

|

ΔPlocal is PB or Z

ΔPlocal is NB

SOC is M & ΔPexcess is PB

Psto_ref_sht is PM

N1.11

Psto_ref_sht is Z

|

Psto_ref_lgt is NB Psto_ref_lgt is PB|

Psto_ref_sht is PB

|

|

Psto_ref_sht is NM

N1.12

Psto_ref_sht is NB

|

|

Psto_ref_sht is Z

|

|

N1.1

Psto_ref_lgt is Z Psto_ref_lgt is Z Psto_ref_lgt is Z Psto_ref_lgt is Z

Psto_ref_lgt is NB Psto_ref_lgt is PB

Mode Inputs Output

SOC ΔPlocal ΔPexcess Psto-ref-lgt Psto-ref-sht

N1.11

M not NB PB PB Z

M not NB PB Z PB

M not NB PB PB PB

N1.12

M NB PB NB NB

M NB PB Z NM

M NB PB PB Z

…

from 81 possible cases, only 30 fuzzy necessary rules

If SOC is M

and ΔPexcess_p.u. is NB (Pgrid<Psubscribed)

and ΔPlocal_p.u. is PB (Ptrain>PRES)

and Psto-ref-lgt_p.u. is NB (to charge)

then Psto-ref-sht_p.u. is NB (charge reference)




STEP 7Determination of indicators to measure

the achievement of objectives

Economic indicator (monthly component of the exceeding subscribed power)

Energy indicator (ratio between the locally consumed energy of RES and the produced one)

subscribedgridexcess

Xx

excesst

Tt

PPP

xPkCMDPSt

)(2

100(%)_

RES

nonconsRESRES

RESE

EEI

• five tariff time periods T

• α=0.3584€/kW

• kt (%) is a power coefficient for each time tariff period t

• Xt represents the index set x belonging to each time class t).



0 1 2 3 4 5 6 7-0.1

0

0.1

Po

we

r(p

.u.)

Psto-ref-lgt

0 1 2 3 4 5 6 70

0.5

1

Weekday

pric

e(p

.u.)

electricity price

FL energy management

+-

Ptrain

PRES

ΔPlocal

SOC

ΔPexcess

K1

K2

K3

K5

+-

Pgrid

Psubscribed Psto_ref_sht

Predictive mode

Long-term

Ptrain_predictive

Electricity cost

Psto_ref_lgt

HRPS supervision

PRES_forecast

K4

K1, K2, K3, K4, K5 = normalisation gains

To favor local RES consumption

To ensure the storage availability

To limit exceeding subscribed power

To reduce the electricity bill

Predictive mode storage reference power

0 1 2 3 4 5 6 70

0.2

0.4

0.6

0.8

1

Weekday

Po

we

r(p

.u.)

Ptrain

Ppv

Pwind

Production and consumption profiles

Depends on electricity price in short term trades

Storage system:

Csto = 5000 kWh

Psto-max = 1MW

ηcharge=90%, ηdischarge=90%

SOCmin, SOCmax

time response constant (τ=0.5s)


0 2 4 6 8 10 12 14 16 18 20 22 0 2 4 6 8 10 12 14 16 18 20 22

0

0.5

1

Po

we

r(p

.u.)

Psubsribed

Pgrid

0 2 4 6 8 10 12 14 16 18 20 22 0 2 4 6 8 10 12 14 16 18 20 220

0.5

1

SO

C(%

)

SOC

0 2 4 6 8 10 12 14 16 18 20 22 0 2 4 6 8 10 12 14 16 18 20 22-0.05

0

0.05

time(hours)

Po

we

r(p

.u.)

Psto-ref-lgt

Psto

storage power adjustment

storage system charges

when RES exceed train

consumption and when

economic interesting

adjustment of Psto-ref-lgt reference power

the subscribed power is reduced

thanks to RES local consumption

and storage management



Comparison of different supervision cases

Study case CMDPS IRES

Reference case 5338 € 0%

Case 1 (Predictive mode) 1024 € 96,5%

Case 2 (FLEM strategy) 1036 € 95,5%

Case 3 (Adjustment) 942 € 96,9%

well performance of FLEM

strategy compared to

predictive mode results

RES is locally consumed almost in totality

Subscribed power is reduced five times in HRPS supervision


41

Case study 3 : Day-ahead Optimal

Operational Planning of Generators



Scheme of Day-ahead Optimal Power Reserve Planning

Focus on the design of the MCEMS under particular constraints.

Uint commitment problem with dynamic programming is developed in order to reduce the

economic cost and/or CO2 equivalent emissions.

Ф(t)

Dynamic

programming

Optimization

Operational planning

of MGT

J

Day-ahead optimal operation planning Constraints

· Power balance · MGT efficiency · Maximization of RE · Battery energy

24 hou

r ahead

forecasting

and

pow

er reserve qu

alify

Criteria

· Objective function

? (t)

u(t)

x(t)

PR

L~

vP~

Electrical system modeling

Loads

PV based AG Technological

features

Economic cost

CO2 emissions MGT

how much power reserve from AG and how much from MGT ?

Case study 3 : Day-ahead Optimal Operational Planning of Generators


2. Non-linear Constraints Power reserve assessment with x % of LOLP;

Balancing between load demand and power:

Maximization of renewable energy usage: PV power is considered as a prior source;

Battery capacity limitation (more PV power, more battery storage !)

Different strategies:

1.During the day (use PV or store energy)

2.During the night (discharge battery)

0))()(()()()( =)( _

N

1n 1

_ ·

tPttPtPtPt iMGT

M

i

inAGresL

MGT corresponding inequality constraint: )(%100)()(%50 max___max__ tPtPtP iMiMiM


Power output (%)

CO

2eq

uiv

alen

t (k

g)

Partial load ratio (%)

Fu

el c

on

sum

pti

on (

kW

h)


3. Operational and Power Reserve Distribution Strategies

Strategy 1:

Power Reserve Covered by only MGT:

Classical strategy for passive PV

generators;

Strategy 2:

Power Reserve Covered by both MGT and

PV AG: the power reserve is distributed

into both MGT and PV AG.

Power reserve needs to be covered by MGTs and/or PV AG:

PR Dispatching in three MGT

PR Dispatching in three MGT + PV AG

8h 10h 13h 16h 19h 22h 1h 4h 7h0

5

10

15

20

Time (half hour)

Pow

er

Rese

rve

(kW

)

PR in MGT1

PR in MGT2

PR in MGT3

PR in AG

8h 10h 13h 16h 19h 22h 1h 4h 7h0

5

10

15

20

Time (half hour)

Pow

er

Rese

rve

(kW

)

PR in MGT1

PR in MGT2

PR in MGT3



4. Unit Commitment Problem with Dynamic Programming

Formulation of the UCP

)(),...,(),( =(t) _2_1_ tPtPtPx iMGTMGTMGT

)(),...,(),( =(t) 21 tttu i

8h 10h 13h 16h 19h 22h 1h 4h 7h0

15

30

PM

GT

1 (

kW

)

8h 10h 13h 16h 19h 22h 1h 4h 7h0

15

30

PM

GT

2 (

kW

)

8h 10h 13h 16h 19h 22h 1h 4h 7h0

30

60

PM

GT

3 (

kW

)

8h 10h 13h 16h 19h 22h 1h 4h 7h0

25

50

PA

G (

kW

)

Time (half hour)

Start

t=1

t=T

Stop

t=t-1

Optimization: calculate the minimum J(t) with

vectors x(t) and u(t), using recursive equation

F(t, x(t), u(t)).

Calculate F(T, x(T), u(T)) with x(T)

and u(T) to minimize J(T).

F1(1) F1(2) … F1(T-1) F1(T)

F2(1) F2(2)… F2(T-1) F2(T)

F3(1) F3(2)… F3(T-1) F3(T)

F4(1) F4(2)… F4(T-1) F4(T)

…. …. …. ….

F1(t) F1(T)

F2(t) F2(T)

F3(t) F3(T)

F4(t) F4(T)

…. ….

F1(T)

F2(T)

F3(T)

F4(T)

….

No

Yes

Dynamic Programming

Optimization Objectives:

1. Economic criteria: minimize total cost;

2. Environmental criteria: minimize CO2

emission;

3. Best compromise criteria: make a compromise.

Optimal Operational of a cluster of MGTs (PV power is prior source)


46/17

5 Simulation Results (1)In this case: rated load (110 kW), rated PV power (55 kW) and the power reserve (with 1 % of

LOLP) coming from the net demand uncertainty assessment (with second method).

Scenario Optimized criteria Cost (€) Pollution (kg) PR on AG (%) Ebattery_Max (kWh)

Without PV

power

None 219 1392 0 0

Environmental 212 1196 0 0

Economic 210 1263 0 0

Strategy 1:

PR from MGT

None 183 1156 0 80.2

Environmental 181 1067 0 80.2

Economic 178 1120 0 80.2

Strategy 2:

PR from MGT

and PV AG

None 182 1098 40 54.1

Environmental 179 991 40 54.1

Economic 177 1061 40 54.1

47%

14%

39%

Without PV power

MGT1

MGT2

MGT3

AG=071%

10%

19%

Strategy 1

MGT1

MGT2

MGT3

AG=0

30%

12%

18%

40%

Strategy 2

MGT1

MGT2

MGT3

AG

Power reserve dispatching, one day ahead 40% of PR is on AG !

Day-ahead Optimal Operational Planning of Generators


47/17

5. Simulation Results (2):

Daily PV energy and power reserve VS. percentage

of PV energy over Load.

8h 10h 13h 16h 19h 22h 1h 4h 7h0

25

50

PA

G (

kW

)

8h 10h 13h 16h 19h 22h 1h 4h 7h-20

0

20

PB

atte

ry (

kW

)

8h 10h 13h 16h 19h 22h 1h 4h 7h0

30

60

EB

atte

ry (

kW

h)

Time (half hour)

Increase the PV power Penetration Rate Battery states and power security

7h 10h 13h 16h 19h 22h 1h 4h 7h0

0.5

1

LO

LP

(%

) Strategy 1

7h 10h 13h 16h 19h 22h 1h 4h 7h0

1

2

Strategy 2

LO

LP

(%

)

Time (half hour)

0 20 40 60 80 100 120 140 1600

100

200

300

400

500

600

700

800

900

Rated PV Power (kW)

Dai

ly E

nerg

y (

kW

h)

Daily PV Energy

Energy Reserve

Max Battery Energy

Day-ahead Optimal Operational Planning of Generators


Conclusion and perspectives

49Conclusion and perspectives

Design optimization and supervision of multi-source energy systems is a

relevant issue and becomes more and more common.

An optimized and well managed system leads to better reliability and lower

costs.

Perspectives

It is possible to combine implicit and explicit optimization for design and

supervision of multi-source energy systems (for example : optimizing fuzzy

logic member functions with GA or SQP).

Adapting system models for optimization (simple and representative models

for fast simulation).

Improving convergence and rapidity of optimization algorithms.

Developing tools for optimization and supervision of complex energy systems.

References and publications

51References and publications

• Dhaker Abbes, André Martinez, Gérard Champenois, Jean Paul Gaubert, Étude d’un système hybride éolien photovoltaïque avec stockage : Dimensionnement et analyse du cycle de vie, European Journal of ElectricalEngineering (2012) DOI:10.3166/EJEE.15.479-497 © 2012 Lavoisier. (IF: 0,55).

• Dhaker Abbes, André Martinez, Gérard Champenois, Eco-Design Optimization of an Autonomous Hybrid Wind-PV System with Battery Storage, Revue IET: Renewable Power Generation, DOI: 10.1049/iet-rpg.2011.0204 ( IF : 2,54).

• Dhaker Abbes, André Martinez, Gérard Champenois, Life cycle cost, embodied energy and loss of power supply probability for the optimal design of hybrid power, J. Math. Comput. Simulat (Matcom), DOI: 10.1016/j.matcom.2013.05.004 (IF : 0,738).

• Dhaker Abbes, André Martinez, Gérard Champenois, Benoit Robyns, Real time supervision for a hybrid renewable power system emulator, Simulation Modelling Practice and Theory, 03/2014; 42:53–72. DOI:10.1016/j.simpat.2013.12.003. (IF: 1,16).

• Toufik Madani Layadi, Gérard Champenois, Mohammed Mostefei, Dhaker Abbes, Etude du vieillissement d’un banc de stockage plomb-acide dans un système hybride multi-sources, Symposium de Génie Electrique (SGE’14) : EF-EPF-MGE 2014, 8-10 JUILLET 2014, ENS Cachan, France.

• Petronela Pankovits; Dhaker Abbes; Christophe Saudemont; Othman Moumniabdou; Julien Pouget; Benoit Robyns, Energy Management Multi-Criteria Design for Hybrid Railway Power Substations, 11th International Conference on Modeling and Simulation of Electric Machines, Converters and Systems (ELECTRIMACS 2014), 19-22 May, Valencia, Spain.

• Dhaker Abbes, André Martinez, Emulation of a hybrid PV-Wind-Battery system, 14th International conference on Sciences and Techniques of Automatic control & computer engineering, Sousse, Tunisia; 12/2013.

• Dhaker Abbes, Gérard Champenois, André Martinez, Benoit Robyns, Modeling and simulation of a photovoltaic system: An advanced synthetic study, 3d International Conference on Systems and Control (ICSC13), 29 to October 31, 2013, in Algiers, Algeria.

• Petronela Pankovits, Maxime Ployard, Julien Pouget, Stephane Brisset, Dhaker Abbes, Benoit Robyns, Design and Operation Optimization of a Hybrid Railway Power Substation, EPE-ECCE Europe Conference, 3-5 September 2013, Lille, France.

52References and publications

• X. Yan, B. Francois, and D. Abbes, ”Solar radiation forecasting using artificial neural network for local power reserve,” in Electrical Sciences and Technologies in Maghreb (CISTEM), 2014 International Conference on, 2014, pp. 1-6.

• X. Yan, B. Francois, and D. Abbes, “Operating power reserve quantification through PV generation uncertainty analysis of a microgrid,” in PowerTech, 2015 IEEE Eindhoven, 2015, pp. 1-6.

• X. Yan, B. Francois, D. Abbes and Hassan BEVRANI, “Day-ahead Optimal Operational and Reserve Power Dispatching in a PV based Urban Microgrid,” EPE’2016 ECCE.

• Petronela Pankovits, Dhaker Abbes, Christophe Saudemont, Stephane Brisset, Julien Pouget, Benoit Robyns, “Multi-criteria fuzzy-logic optimized supervision for hybrid railway power substations”, Mathematics and Computers in Simulation, 2016, IF : 0,949.

Merci pour votre attention

Loads

Thank you for your attention!

www.da-engineering.com

Documents

Optimization and supervision of complex energy systems