Power System Stabilizers Tuning using Bio-Inspired Algorithm

Power System Stabilizers Tuning using Bio-Inspired Algorithm

AUTHORS:

Wesley Peres – Edimar Oliveira

João A. Passos Filho – Diego Arcanjo – Ivo Chaves Junior

Federal University of Juiz de Fora – Brazil

June, 2013

Outline

1. Introduction;

2. Proposed Methodology;

3. Results;

4. Conclusion.

2

Introduction

3

• Frequency and amplitude problems; • Size and non-linearity; • Operating condition.

Power System Oscillation Damping has been the subject of many studies.

System Loads...

.

.

.

GH1

GH2

GHn

GT1 GT2

GTk

Hydroelectric Generators

Thermoelectric Generators

Introduction

Some Controllers include:

– Automatic Voltage Regulator (AVR);

– FACTS;

– HVDC Links.

Power System Stabilizers (PSS).

4

Introduction

Tuning techniques:

– Linear Techniques;

– Bio-Inspired Algorithm…

5

PSS(s) : in a feedback loop (Additional Control Signal) P(s) : open-loop transfer function

6

• The objective in this work is to evaluate a Bio-Inspired Algorithm for tuning multiple PSSs.

• Modified Cuckoo Search (MCS) technique to achieve this goal.

Objective

Problem Formulation

State Space Equations frequency domain

A damping ratio is associated with each eigenvalue:

PacDyn (CEPEL)

7

Problem Formulation

8

Determine the system stability:

j

Unstable Condition

Problem Formulation

9

j

Stable Condition

Problem Formulation

PSS

10

PSS Structure

11

Tw = 5 NB = 3 values from the literature for the New England Test System.

PSS Parameters

• Each solution is represented by a vector considering the set of PSS to be tuned.

• p stabilizers are tuned.

12

Lead Lag Stage Gain

Set of Operating Conditions

• For the tuning a set of closed-loop eigenvalues are obtained for over pre-specified operating conditions:

13

Fitness Function associated with the PSS tuning

Optimization Problem

14

The solution provides the best tuning for all PSSs.

Variables limits

Metaheuristic

Inspired

CUCKOO SEARCH TECHNIQUE

Reproduction Strategy

Nest Parasitism

Other birds nest

15

CUCKOO SEARCH TECHNIQUE

Consequencesof

Parasitism Egg is

discovered

Egg is not discovered

Nest will be abandoned or egg will

be destroyed

Egg will be hatch and

Cuckoo will survive

16 Cuckoos need to find a good host.

Modified Cuckoo Search (MCS)

• In this work: cuckoo = egg = nest = prob. solution.

• First Change: Cuckoos are divided into two groups

• Group of Top Eggs:

– Individuals (eggs) with the best fitness;

– A new egg of this group is generated by using two eggs from this group (information exchange).

• Group of eggs to be replaced:

– This group is composed of eggs that will be replaced.

17

Modified Cuckoo Search (MCS)

• Generating new solutions:

18

Encourage localized search improving the optimal solution.

Lévy distribution: the probability of returning to a

site previously visited.

t: generation number

Results • The New England test system is used:

– 39 AC-buses; – 10 Power generators; – 9 PSS should be tuned.

• Settings:

– Population size: 15 individuals – Convergence Criterion: 250 generations. – Top eggs group = 25% and 75% to the replaced group

• The results obtained by using MCS are with Genetic Algorithm;

• MCS and GA are initialized with the same solution.

19

Results

• Four operating conditions without PSS:

20

Results

21

Results

22

5 simulations for each methodology.

Results

• Minimum Damping Ratio(%)

23

MCS leads the system to a higher stability condition.

Results

24

• MCS uses only two parameters: size of population and percentage of top group;

• The results obtained with the proposed methodology are consistent with results from literature;

• The use of MCS presented promising results for tuning stabilizers and damping power system oscillations.

25

Conclusions

Acknowledgments

26