You are on page 1of 33

OPTIMIZATION in RENEWABLE ENERGY SYSTEMS

OPTIMIZATION in RENEWABLE
ENERGY SYSTEMS
Prof. Dr. Serdar PLK
Pamukkale University, Dept. of Electrical and Electronics Eng., Knkl Campus, 20040, Denizli
e-mail: iplikci@pau.edu.tr

web: www.pau.edu.tr/iplikci
phone: +90 (258) 2963197

16/01/2015

The 4th Renewable Energy Systems Winter School Prof. Dr. Serdar PLK

1/33

OPTIMIZATION in RENEWABLE ENERGY SYSTEMS

CONTENTS
1. Introduction
2. Optimization Techniques
1.

Introduction

2.

Heuristic Optimization

3.

Gradient-Based Optimization

3. Optimization Applications in the Renewable and


Sustainable Energy Systems
1.

Wind Power

2.

Solar Energy

3.

Geothermal Energy

4.

Bioenergy

5.

Hybrid Systems

4. Modeling and Prediction


5. Conclusions
16/01/2015

The 4th Renewable Energy Systems Winter School Prof. Dr. Serdar PLK

2/33

OPTIMIZATION in RENEWABLE ENERGY SYSTEMS

1) Introduction
1.1) Why Optimization?

Due to the increasing demand and limited sources worldwide, sustainability is of great
importance

It is necessary to develope efficient methods that allow us to produce and consume


the energy very effectively.

The extension of energy sources and the information structure allow a fine screening
of energy resources, but also require the development of tools for the analysis and
understanding of huge datasets about the energy grid.

Key technologies in future ecological, economical and reliable energy systems are

Energy prediction of renewable resources

Prediction and monitoring of energy consumption

Efficient planning and control strategies for network stability

To enable ecologically and financially feasible projects, optimization methods have


taken over a key role for planning, optimizing and forecasting sustainable systems.

16/01/2015

The 4th Renewable Energy Systems Winter School Prof. Dr. Serdar PLK

3/33

OPTIMIZATION in RENEWABLE ENERGY SYSTEMS

1) Introduction
1.2) Optimization in the Literature

Source: Web of Science

16/01/2015

The 4th Renewable Energy Systems Winter School Prof. Dr. Serdar PLK

4/33

OPTIMIZATION in RENEWABLE ENERGY SYSTEMS

2) Optimization Techniques
2.1) Introduction
2.1.1) General Format and Short Notation

Objective Function

min

Design Variables

= 1,2, ,

(1 , 2 , , )

1 ,2 ,,

s. t:

(1 , 2 , , ) = 0
(1 , 2 , , ) 0

1
2
=

min

()

s. t. : () = 0
() 0

16/01/2015

()

Local minimum
Global minimum

The 4th Renewable Energy Systems Winter School Prof. Dr. Serdar PLK

5/33

OPTIMIZATION in RENEWABLE ENERGY SYSTEMS

2) Optimization Techniques
2.2) Heuristic Optimization
2.2.1) Introduction

Heuristic methods are the methods that produce sufficient (even if not optimum)
solutions to the large scale problems very rapidly.
Meta-heuristics are generalizations of heuristics in the sense that they can be applied
to a wide set of problems.

Heuristic methods can be categorized as follows:

16/01/2015

Trajectory vs population
memory-based vs memoryless
nature-inspired vs non-nature-inspired

The 4th Renewable Energy Systems Winter School Prof. Dr. Serdar PLK

6/33

OPTIMIZATION in RENEWABLE ENERGY SYSTEMS

2) Optimization Techniques
2.2) Heuristic Optimization
2.2.2) Trajectory Methods

Trajectory methods start the search process with a single solution. During the search
process, the solution is updated iteratively and the thus the outcome is also a single
optimized solution. Most of them are extensions of simple iterative improvement
procedures that incorporate techniques that enable the algorithm to escape from local
optima. Some of the trajectory methods are:

Hill Climbling (HC)


Simulated Annealing (SA)
Tabu Search (TS)
Greedy Randomized Adaptive Search Procedures (GRASP)
Variable Neighborhood Search (VNS)

Some of the meta-heuristics trajectory methods are :


Iterated Local Search (ILS)
Pareto Archived Evolution Strategy (PAES)
Multi-Objective Simulated Annealing (MOSA)
16/01/2015

The 4th Renewable Energy Systems Winter School Prof. Dr. Serdar PLK

7/33

OPTIMIZATION in RENEWABLE ENERGY SYSTEMS

2) Optimization Techniques
2.2) Heuristic Optimization
2.2.3) Population-based Meta-Heuristics Methods

Population-based heuristics use a population of solutions which evolve during a given number of
iterations, also returning a population of solutions when the stop condition is fulfilled. The
main population-based heuristics include:
Genetic Algorithms (GA)
Evolutionary Algorithms (EA)
Scatter Search (SS)
Path Relinking (PR)
Memetic Algorithms (MA)
Ant Colony Optimization (ACO)
Particle Swarm Optimization (PSO)
Estimation of Distribution Algorithm (EDA)
Differential Evolution (DE)
Artificial Bee Colony Optimization (ABCO)
Population-based meta-heuristics are:
The Multiobjective Tabu Search (MOTS)
Non-dominated Sorting Genetic Algorithm (NSGA/NSGA-II)
Pareto Simulated Annealing (PSA)
Single Front Genetic Algorithm (SFGA)
Strength Pareto Evolutionary Algorithm (SPEA/SPEA-II)
Pareto Envelope-based Selection Algorithm (PESA/PESA-II).
16/01/2015

The 4th Renewable Energy Systems Winter School Prof. Dr. Serdar PLK

8/33

OPTIMIZATION in RENEWABLE ENERGY SYSTEMS

2) Optimization Techniques
2.2) Heuristic Optimization
2.2.4) Genetic Algorithm Cycle
GENETIC ALGORITHM CYCLE

(1 , 2 , , )

Population

fitness function
X1

N
number of population

X3

= 2

ith

roulette wheel

calculate F1 , F2 ,, FN
Sort from smallest to highest

XN

mutation

individual (solution)

original

1 0

0 1 1 1

0 1 0 0

1 0 1 1

0 1

mutated

1 0

0 0 1 1

0 0 0 1

1 1 1 1

0 0

= (1 , 2 , , )

crossover

fitness of the ith solution

= decimal2binary
binary counterpart of the
ith solution
16/01/2015

selection

Table of fitness values

X2

Parent 1

1 0

0 1 1 1

0 1 0 0

1 0 1 1

0 1

Parent 2

1 1

0 0 1 0

1 1 0 1

1 1 0 1

0 0

Offspring1

1 0

0 1 1 1

0 1 0 1

1 1 0 1

0 0

Offspring2

1 1

0 0 1 0

1 1 0 0

1 0 1 1

0 1

The 4th Renewable Energy Systems Winter School Prof. Dr. Serdar PLK

9/33

OPTIMIZATION in RENEWABLE ENERGY SYSTEMS

2) Optimization Techniques
2.3) Gradient-Based Optimization
2.3.1) Mathematical Basics

Gradient Vector

Hessian Matrix

Jacobian Matrix

16/01/2015

1 , 2 , ,

2 1 , 2 , ,

1 , 2 , ,

1 , 2 , ,
1
1 , 2 , ,
=
2

1 , 2 , ,

2 1 , 2 , ,
12
2 1 , 2 , ,
=
2 1

2
1 , 2 , ,
1

2 1 , 2 , ,
1 2
2 1 , 2 , ,
22

2
1 , 2 , ,
2

1 1 , 2 , ,
1
2 1 , 2 , ,
=
1

1 , 2 , ,
1

1 1 , 2 , ,
2
2 1 , 2 , ,
2

1 , 2 , ,
2

2 1 , 2 , ,
1
2 1 , 2 , ,
2

2
1 , 2 , ,
2

1 1 , 2 , ,

2 1 , 2 , ,

1 , 2 , ,

The 4th Renewable Energy Systems Winter School Prof. Dr. Serdar PLK

10/33

OPTIMIZATION in RENEWABLE ENERGY SYSTEMS

2) Optimization Techniques
2.3) Gradient-Based Optimization
2.3.2) Taylor Expansion, Descent Direction and Optimality Conditions

Taylor Expansion
+ = +

1
+ 2 + . .
2

Descent Direction

+ +

<

<0

Optimality Conditions
First-order Optimality Conditions: =
Second-order Optimality Conditions : 2 positive definite
Update Rule
If one can find a search direction that satisfies + <
+1 + kural ile gncelleme yaplr.
16/01/2015

The 4th Renewable Energy Systems Winter School Prof. Dr. Serdar PLK

11/33

OPTIMIZATION in RENEWABLE ENERGY SYSTEMS

2) Optimization Techniques
2.3) Gradient-Based Optimization
2.3.3) First- and Second-order Methods

First-order Methods
Steepest Descent:
Conjugate Gradient:

= + 1


1 1

Second-order Methods
Newton:

= 2

Modified Newton:

= 2 +

16/01/2015

The 4th Renewable Energy Systems Winter School Prof. Dr. Serdar PLK

12/33

OPTIMIZATION in RENEWABLE ENERGY SYSTEMS

2) Optimization Techniques
2.3) Gradient-Based Optimization
2.3.4) Quasi-Newton Methods and Second-order Approximate Methods

Quasi-Newton Methods
Davidon-Fletcher-Powell (DFP) :

= + ,
+1 = +

Broydon-Fletcher-Goldfarb-Shanno (BFGS) :
= +

, +1 = + +

= +1

= +1

Second-order Approximate Methods

Gauss-Newton (GN):

Levenberg-Marquardt (LM):

16/01/2015

The 4th Renewable Energy Systems Winter School Prof. Dr. Serdar PLK

13/33

OPTIMIZATION in RENEWABLE ENERGY SYSTEMS

3) Optimization App. in the Renewable and Sustainable Energy Systems


PROBLEM

ZM

Community-scale renewable energy systems planning is an


important problem consisting of justifying the allocation
patterns of energy resources and services, formulation of local
policies
regarding
energy
consumption,
economic
development and energy structure, and analysis of
interactions among economic cost, system reliability and
energy-supply security.

a long-term dynamic multi-objective planning model for


distribution network expansion along with distributed energy
options

Immune Genetic Algorithm (I-GA)

minimum cost expansion of power transmission networks


under carbon emission trading programs

annual peak load forecasting in an electrical power system


with the aim of minimizing the error associated with the
estimated model parameters

Particle Swarm Optimization (PSO)

new renewable energy sources penetration


and congestion management so that electricity supply and
demand are always evenly balanced

NelderMead Simplex (NMS) and PSO


Honey Bee Mating Optimization (HBMO)
Ant Colony Optimization (ACO), ANN, GA

Energy demand prediction

Yapay Sinir Alar (ANN)


Destek Vektr Makineleri (SVM)

16/01/2015

Interval Linear Programming (ILP)


Chance-Constrained Programming
Mixed Integer-Linear Programming (MILP)

Mixed-Integer Programming (MIP)


Genetic Algorithms (GA)
Simulated Annealing (SA )
Tabu Search (TS)

The 4th Renewable Energy Systems Winter School Prof. Dr. Serdar PLK

14/33

OPTIMIZATION in RENEWABLE ENERGY SYSTEMS

3) Optimization App. in the Renewable and Sustainable Energy Systems


3.1) Wind Power
3.3.1) Introduction

Wind is a periodical phenomenon for large geographical areas like Mexico.


the increasing sizes of turbines and the lower prices per installed production
capacity of electricity.

Wind energy systems may not be technically viable in all locations because of
low wind speeds and the fact that it is more unpredictable than solar energy
Areas where winds are stronger and more constant, such as offshore and high

altitude sites, are preferred locations for wind farms.


An accurate estimation of wind speed distributionthe site selection of wind
farms
Bayesian modelmodeling long-term wind speed distributions

16/01/2015

The 4th Renewable Energy Systems Winter School Prof. Dr. Serdar PLK

15/33

OPTIMIZATION in RENEWABLE ENERGY SYSTEMS

3) Optimization App. in the Renewable and Sustainable Energy Systems


3.1) Wind Power
3.3.1) Optimization in Wind Turbine Design

In particular, two important problems are often considered: wind turbine and wind farm layout
The power output of a turbine is a function of the:
density of the air
area swept out by the turbine blades
cube of the wind speed
Numerous metrics are used to measure the power quality of a wind turbine, such as
the power factor, reactive power,
Harmonic distortion
the optimization of the geometrical parameters of the rotor
configuration of stall-regulated horizontal-axis wind turbines with
the aim of achieving the best trade-off performance between the
total energy production per square meter of wind park and cost

Multi-Objective Evolutionary Algorithm


(MOEA)

the optimization of the ranges of gearbox ratios and power ratings


of multihybrid permanent-magnet wind generator systems

Genetic Algorithms (GA)

determining the optimum capacity taking into account uncertainties


arising from wind speed distribution and powerspeed
characteristics

Mixed-Integer Nonlinear Programming


(MINLP)

prediction of wind speed at a selected location based on the data


collected at the neighbouring locations

Fuzzy Logic Modelling

16/01/2015

The 4th Renewable Energy Systems Winter School Prof. Dr. Serdar PLK

16/33

OPTIMIZATION in RENEWABLE ENERGY SYSTEMS

3) Optimization App. in the Renewable and Sustainable Energy Systems


3.1) Wind Power
3.3.2) Optimization in Wind Farm Layout

Wind farm layout consists of determining the optimum positions of wind turbines
within the farm in order to maximize energy production.
optimal placement of wind turbines for maximum production capacity while
limiting the number of turbines installed and the acreage of land occupied by
each wind farm

Genetic Algorithms (GA)

optimum wind farm configuration problem which is driven by an integral


wind farm cost model based on the cumulative net cash flow value
throughout the wind farms lifespan

Evolutionary Algorithms (EA)

wind turbine placement based on wind distribution with the aim of both
maximizing the wind energy capture and minimizing an index that
determines constraint violations

Multi-Objective Evolutionary
Algorithm (MOEA)

determining the optimal type, number and placement of wind turbines


considering the given wind conditions and wind park area

Mixed-Integer Nonlinear
Programming (MINLP)

16/01/2015

The 4th Renewable Energy Systems Winter School Prof. Dr. Serdar PLK

17/33

OPTIMIZATION in RENEWABLE ENERGY SYSTEMS

3) Optimization App. in the Renewable and Sustainable Energy Systems


3.2) Solar Energy
3.2.1) Introduction

Solar energy is radiant energy that is produced by the sun. In many parts of the world,
direct solar radiation is considered to be one of the best prospective sources of energy.
The main ways to convert solar radiation into energy are active and passive solar
design.

Passive solar design is often based on the optimal design of buildings that capture the
suns energy in order to reduce the need for artificial light and heating. Regarding
passive solar systems, a primary interest for researchers in solar energy is related to
the design and optimization of solar energy homes.
Active solar design is based on water heating converting solar radiation into heat using
photovoltaic panels and solar cells to convert the solar radiation into energy.
In order to design both active and passive solar energy systems, radiation data are
needed for the studied location.

calculating solar radiation levels over complex mountain terrains using data from only
one radiometric station
16/01/2015

ANN, Neuro-Fuzzy

The 4th Renewable Energy Systems Winter School Prof. Dr. Serdar PLK

18/33

OPTIMIZATION in RENEWABLE ENERGY SYSTEMS

3) Optimization App. in the Renewable and Sustainable Energy Systems


3.2) Solar Energy
3.2.2) Active Solar Design
identifying the electrical parameters of photovoltaic solar cells and modules to
determine the corresponding maximum power point from the illuminated current
voltage characteristic

Genetic Algorithms
(GA)

maximizing the thermal performance of flat plate solar air heaters by considering
the different system and operating parameters

Genetic Algorithms
(GA)

determining the tilt angle of photovoltaic modules with the aim of maximizing the
electrical energy output of the modules

Particle Swarm
Optimization (PSO)

An interesting problem related to photovoltaic systems is the optimal determination of


their size. The sizing optimization of a stand-alone photovoltaic system is a complex
optimization problem which aims to obtain acceptable energy and economic cost for
the consumer, and a relatively correct energy supply quality.
Optimal sizing of photovoltaic systems

16/01/2015

ANN ve GA

The 4th Renewable Energy Systems Winter School Prof. Dr. Serdar PLK

19/33

OPTIMIZATION in RENEWABLE ENERGY SYSTEMS

3) Optimization App. in the Renewable and Sustainable Energy Systems


3.3) Bioenergy
3.3.1) Introduction

Bioenergy is renewable energy made available from materials derived from biological
sources. Biomass, a renewable energy source, is biological material from living, or
recently living organisms, including plants and animals.
Biomass is one of the most promising renewable energy sources, but more research is
required to prove that power generation from biomass is both technically and
economically viable
Biomass can be burned to produce steam for making electricity, or to provide heat to
industries and homes.
In addition biomass can be converted to other usable forms like methane gas, ethanol
fuel and biodiesel fuel.
Biomass power plants exist in over 50 countries around the world and supply a
growing share of electricity.
The sustainability of electricity generation from biomass must be assessed according
to the key indicators of price, efficiency, greenhouse gas emissions, availability,
limitations, land use, water use and social impacts.
Finding optimal location of biomass-fuelled systems for distributed power generation
with forest residues as biomass source
16/01/2015

binary PSO-based
method

The 4th Renewable Energy Systems Winter School Prof. Dr. Serdar PLK

20/33

OPTIMIZATION in RENEWABLE ENERGY SYSTEMS

3) Optimization App. in the Renewable and Sustainable Energy Systems


3.4) Geothermal Energy
3.4.1) Introduction

Geothermal energy is the energy contained as heat inside the Earth. Geothermal heat
pumps are a highly efficient, renewable energy technology for heating and cooling.
This technology relies on the fact that, at depth, the Earth has a relatively constant
temperature, warmer than the air in winter and cooler than the air in summer.

The main advantage of using geothermal energy is that this renewable energy source
can provide power 24 h a day due to it is constant, without intermittence problems
compared to other renewable resources such as wind or solar energy.
It is expensive to build a power station but operating costs are low, resulting in low
energy costs for suitable sites.
Geothermal power plants now exist in 19 countries, and new plants are commissioned
annually. However, only a small fraction of the geothermal potential has been
developed so far, and there is ample space for an accelerated use of geothermal energy
both for electricity generation and direct applications.
optimization of the exploitation system of a low enthalpy geothermal aquifer, with
the aim of determining the annual pumping cost of the required flow and the
amortization cost of the pipe network, which carries the hot water from the wells
to a central water tank, situated on the border of the geothermal field
16/01/2015

Genetic Algorithms (GA)

The 4th Renewable Energy Systems Winter School Prof. Dr. Serdar PLK

21/33

OPTIMIZATION in RENEWABLE ENERGY SYSTEMS

3) Optimization App. in the Renewable and Sustainable Energy Systems


3.5) Hybrid Systems
3.5.1) Introduction

In the last decade, there has been a spectacular increase the interest in optimizing the design and
control of stand-alone hybrid power generation systems in order to manage energy between the
maximum energy captured and consumed energy.
The aim of optimizing the mix of the renewable system is maximizing its contribution to the peak
load, while minimizing the combined intermittence at a minimum cost.
the economic environmental dispatching of a hybrid power system
including wind and solar thermal energies.

Multi-Objective Evolutionary
Algorithm (MOEA) and GA

the multi-objective design of isolated hybrid systems where the


objectives to minimize are the total cost throughout the useful life of
the installation and the pollutant emissions

Multi-Objective Evolutionary
Algorithm (MOEA) and Strength Pareto
Evolutionary Algorithm (SPEA)

Optimal sizing a hybrid solarwind-battery system with the aim of


minimizing the annualized cost system and the loss of power supply
probability

Genetic Algorithms (GA)

Solving the wind-photovoltaic capacity coordination for a time-of-use


rate industrial user with the aim of maximizing the economic benefits
of investing in a wind generation system and a photovoltaic
generation system

Particle Swarm Optimization (PSO)

Optimal sizing of stand-alone photovoltaic-wind generator systems,


which selects the optimal number and type of units to minimize the
cost subject to the constraint that the load energy requirements are
completely covered

Genetic Algorithms (GA)

16/01/2015

The 4th Renewable Energy Systems Winter School Prof. Dr. Serdar PLK

22/33

OPTIMIZATION in RENEWABLE ENERGY SYSTEMS

4) Modeling and Prediction


4.1) Introduction
4.1.1) The Concept of Modeling

Modeling is a method that is used when a real-world process cannot be described


analytically or mathematically. Here, the process can be all sorts of biological,
physical, chemical, electrical, mechanical, meteorological, social and financial
dynamical systems.
There are many such systems in the real world. For instance, dynamics of weather
conditions is so complex that it cannot be described mathematically, and thus it should
be modeled.similarly, stock market is a very complex system that incorporates many
variables and parameters. There some problems in the energy systems that need
modeling:
Electrical load prediction
Energy demand prediction
Therefore, when needed, we have to collect sufficient data from such energy systems
and then try to obtain a reliable model.

16/01/2015

The 4th Renewable Energy Systems Winter School Prof. Dr. Serdar PLK

23/33

OPTIMIZATION in RENEWABLE ENERGY SYSTEMS

4) Modeling and Prediction


4.1) Introduction
4.1.2) Data Types Input-Output Data
1

MIMO

1
2

Input Data

Output Data

...

...

11

12

...

11

12

...

21

22

...

21

22

...

...

16/01/2015

unknown system

...

The 4th Renewable Energy Systems Winter School Prof. Dr. Serdar PLK

24/33

OPTIMIZATION in RENEWABLE ENERGY SYSTEMS

4) Modeling and Prediction


4.1) Introduction
4.1.2) Data Types Time Series

TS

1
2

16/01/2015

Input Data

+1

Output Data

1
1
2

2
2
3

...
...
...

+1

+1

...

+1
+2

The 4th Renewable Energy Systems Winter School Prof. Dr. Serdar PLK

25/33

OPTIMIZATION in RENEWABLE ENERGY SYSTEMS

4) Modeling and Prediction


4.2) Modeling and Model Selection
4.2.1) Generalization, Approximation and Prediction Errors


MIMO

1
2

1
11
21
1


Input Data

2
...
12
...
22
...

2
...

1
2

MIMO = ,

Output Data

1 2 ...
11 12 ... 1
21 22 ... 2

1 2 ...

=1

Model Output

1 2 ...
11 12 ... 1
21 22 ... 2

1 2 ...

real model


predicted model

optimum model

HYPOTHESIS SPACE

TARGET SPACE

16/01/2015

The 4th Renewable Energy Systems Winter School Prof. Dr. Serdar PLK

26/33

OPTIMIZATION in RENEWABLE ENERGY SYSTEMS

4) Modeling and Prediction


4.2) Modeling and Model Selection
4.2.1) Emprical Error, Overfitting and Splitting Data
Under-learning

Over-learning

model complexity

Emprical error

= + +
training

test
validation

16/01/2015

The 4th Renewable Energy Systems Winter School Prof. Dr. Serdar PLK

27/33

OPTIMIZATION in RENEWABLE ENERGY SYSTEMS

4) Modeling and Prediction


4.2) Modeling and Model Selection
4.2.2) Objective Function

min

, ,
=1 TRA

, ,

=1 VAL

=
=1 TRA

Complexity
best model

16/01/2015

The 4th Renewable Energy Systems Winter School Prof. Dr. Serdar PLK

28/33

OPTIMIZATION in RENEWABLE ENERGY SYSTEMS

4) Modeling and Prediction


4.3) Example: Chaotic Time Series Predictyion by Artificial Neural Nets
4.3.1) Architecture of an Artificial Neural Network

Neural network

Single neuron

input signals

input weight

summing block

output weight

activation
function

1 bias weight


=
+

1,1

1
output signal

1,1

input weight

1,1

,1

1,1

1,

1,2

1,2

,2

1,
=

,
1

1,1

1,

+1 +
=
16/01/2015

The 4th Renewable Energy Systems Winter School Prof. Dr. Serdar PLK

29/33

OPTIMIZATION in RENEWABLE ENERGY SYSTEMS

4) Modeling and Prediction


4.3) Example: Chaotic Time Series Predictyion by Artificial Neural Nets
4.3.2) Data

+1 = 3.9 1 ,

Chaotic Logistic Map

0 = 0.2

0.9

0.8

0.7

0.6

0.5

0.4

0.3

0.2

0.1

16/01/2015

20

40

60

80

100
k

120

140

160

The 4th Renewable Energy Systems Winter School Prof. Dr. Serdar PLK

180

200

30/33

OPTIMIZATION in RENEWABLE ENERGY SYSTEMS

4) Modeling and Prediction


4.3) Example: Chaotic Time Series Predictyion by Artificial Neural Nets
4.3.3) Model Output
LM VERI:logistic EGITIM VERISI SAYISI=119 TEST VERISI SAYISI=80 NORON SAYISI=10 GURULTU GENLIGI=0.01
1

gercek yorunge
en iyi model cikisi
tahmin cikisi

0.8
0.6
0.4
0.2
0

50

100

150

200

250

0.8

egitim hatasi
0.2

0.6
0.4

0.15

hata

0.2
0

0.1

190

192

194

196

198

0.05

200

202

204

206

training error
test error

0
50

100

150

200

250

iterasyon

16/01/2015

The 4th Renewable Energy Systems Winter School Prof. Dr. Serdar PLK

31/33

OPTIMIZATION in RENEWABLE ENERGY SYSTEMS

5) Conclusions

The optimization methods that have beenn used for solving optimization problems in
renewable energy systems have been developed day-by-day. They are especially used
in hybrid systems.

Some of these methods are based on the traditional methods like mixed-integer,
interval linear-programming, Lagrangian relaxation, quadratic programming, and
NelderMead Simplex, while others are based on some heuristics methods such as GA
and PSO.

On the other hand, the multi-objective function problems in the energy systems have
been solved by Pareto-optimization techniques.

The problems in the renewable energy systems that can be solved by optimization
techniques are

16/01/2015

Planning
Management of supply-demand balance
Optimization of design parameters
Prediction of power curve
Configuration
Optimization of economical load distribution
Wind-photovoltaic capacity coordination
Modeling and prediction

The 4th Renewable Energy Systems Winter School Prof. Dr. Serdar PLK

32/33

OPTIMIZATION in RENEWABLE ENERGY SYSTEMS

THANKS...

16/01/2015

The 4th Renewable Energy Systems Winter School Prof. Dr. Serdar PLK

33/33