A NEW TECHNIQUE FOR MAXIMUM LOAD MARGIN ESTIMATION AND PREDICTION

This paper presents the application of Fast Artificial Immune System (FAIS) for maximum load margin estimation and hybrid Fast Artificial Immune Support Vector Machine (FAISVM) for maximum load margin prediction. The newly developed techniques are marked by its significant fast computation time. A n...

Full description

Saved in:
Bibliographic Details
Published in:ARPN journal of engineering and applied sciences 2015-12, Vol.10 (23), p.17566-17572
Main Authors: Ab Aziz, Nur Fadilah, Rahman, Titik Khawa Abdul, Yasin, Zuhaila Mat, Zakaria, Zuhaina
Format: Article
Language:English
Subjects:
Artificial intelligence
Cloning
Data buses
Evolutionary algorithms
Load margins
Mathematical analysis
Mutations
Support vector machines
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:This paper presents the application of Fast Artificial Immune System (FAIS) for maximum load margin estimation and hybrid Fast Artificial Immune Support Vector Machine (FAISVM) for maximum load margin prediction. The newly developed techniques are marked by its significant fast computation time. A new developed index, Voltage Stability Condition Indicator (VSCI) was used as the fitness function for FAIS and FAISVM in order to evaluate the stability condition of load bus in the system. In FAIS, various mechanisms techniques of AIS were investigated and intensive comparisons were made in order to obtain the best implementation of AIS for maximum load margin estimation. The mechanisms were investigated and compared on three main AIS principles; cloning, mutation and selection. In addition, FAISVM is another new hybrid technique developed for maximum load margin prediction that integrates the application of FAIS and Support Vector Machine (SVM). For validation, FAISVM was compared with Evolutionary Support Vector Machine (ESVM) that uses Evolutionary Programming (EP) as the search algorithm. Based on the results, it shows that FAISVM outperforms ESVM with a higher accuracy prediction value.
ISSN:1819-6608
1819-6608