Upload
others
View
0
Download
0
Embed Size (px)
Citation preview
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 :
7Types of optimization and supervision
* Explicit Optimization
- Direct (from an explicit mathematical function to optimize)
Multi-objective optimization :
8Types of optimization and supervision
* 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
9Types of optimization and supervision
• 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.).
10Types of optimization and supervision
• Supervision types :
11Types of optimization and supervision
• 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
14Examples and case studies
Case study 1 : Design and supervision of a hybrid wind-photovoltaic system with batteries
Example of explicit linear optimization:
15Examples and case studies
Case study 1 : Design and supervision of a hybrid wind-photovoltaic system with batteries
Example of explicit linear optimization:
Linear optimization using Excel (SIMPLEX)
16Examples and case studies
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
Case study 1 : Design and supervision of a hybrid wind-photovoltaic system with batteries
17Examples and case studies
Example of non-linear explicit optimization:
- Proposed method based on a dynamic system simulation associated with an
optimization algorithm:
Case study 1 : Design and supervision of a hybrid wind-photovoltaic system with batteries
18Examples and case studies
Example of non-linear explicit optimization:
A mono-objective formulation solved by the SQP algorithm(Newton-Wilson Algorithm)
Minimum LCC [€] or EE[MJ]
Case study 1 : Design and supervision of a hybrid wind-photovoltaic system with batteries
19Examples and case studies
Example of non-linear explicit optimization:
A mono-objective formulation solved by the SQP algorithm (Newton-Wilson Algorithm)
Case study 1 : Design and supervision of a hybrid wind-photovoltaic system with batteries
20Examples and case studies
Case study 1 : Design and supervision of a hybrid wind-photovoltaic system with batteries
Example of non-linear explicit optimization:
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.
21Examples and case studies
Case study 1 : Design and supervision of a hybrid wind-photovoltaic system with batteries
Example of non-linear explicit optimization:
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 [%]
22Examples and case studies
Case study 1 : Design and supervision of a hybrid wind-photovoltaic system with batteries
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
23Examples and case studies
Case study 1 : Design and supervision of a hybrid wind-photovoltaic system with batteries
Examples of extremal and intuitive optimization methods:
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
24Examples and case studies
Case study 1 : Design and supervision of a hybrid wind-photovoltaic system with batteries
Examples of extremal and intuitive optimization methods:
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
25Examples and case studies
Case study 1 : Design and supervision of a hybrid wind-photovoltaic system with batteries
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
Examples and case studies
27Examples and case studies
Case study 2 : Design and supervision of a hybrid railway power station
28Examples and case studies
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)
29Examples and case studies
Case study 2 : Design and supervision of a hybrid railway power station
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
30Examples and case studies
Case study 2 : Design and supervision of a hybrid railway power station
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
31Examples and case studies
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)
Case study 2 : Design and supervision of a hybrid railway power station
32Examples and case studies
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
33Examples and case studies
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)
Case study 2 : Design and supervision of a hybrid railway power station
34Examples and case studies
Methodology for HRPS energy management (implicit optimization)
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
Case study 2 : Design and supervision of a hybrid railway power station
35Examples and case studies
Methodology for HRPS energy management
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
36Examples and case studies
Methodology for HRPS energy management
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)
Case study 2 : Design and supervision of a hybrid railway power station
37Examples and case studies
Methodology for HRPS energy management (implicit optimization)
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).
Case study 2 : Design and supervision of a hybrid railway power station
38Examples and case studies
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)
39Examples and case studies
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
Case study 2 : Design and supervision of a hybrid railway power station
40Examples and case studies
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
Case study 2 : Design and supervision of a hybrid railway power station
41
Case study 3 : Day-ahead Optimal
Operational Planning of Generators
Examples and case studies
42Examples and case studies
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
43Examples and case studies
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
Case study 3 : Day-ahead Optimal Operational Planning of Generators
Power output (%)
CO
2eq
uiv
alen
t (k
g)
Partial load ratio (%)
Fu
el c
on
sum
pti
on (
kW
h)
44Examples and case studies
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
Case study 3 : Day-ahead Optimal Operational Planning of Generators
45Examples and case studies
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)
Case study 3 : Day-ahead Optimal Operational Planning of Generators
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
Examples and case studies
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
Examples and case studies
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