Optimization of Cable Fault Recognition Using Particle Swarm Optimization

 




 

Chan, Hou Jing (2018) Optimization of Cable Fault Recognition Using Particle Swarm Optimization. Final Year Project (Bachelor), Tunku Abdul Rahman University College.

[img] Text
Chan Hou Jing.pdf
Restricted to Registered users only

Download (1MB)

Abstract

Partial Discharge(PD), a type of breakdown that due to the defect that occurs in the insulation material of the power system. During the occurrence of PD, it may lead to serious insulation damage and decrease the life span of the high voltage equipment. Hence, PD monitoring is required in order to minimize the chance of PD lead to serious impact to the power system In this final dissertation, a research work to develop a hybrid cable fault recognition system had been done. The classifiers include Artificial Neural Network (ANN), Support Vector Machine (SVM) and Adaptive Neuro Fuzzy Inference System (ANFIS) and being optimized by modified Chaotic Inertia Weight Particle Swarm Optimization with Time Varying Acceleration Coefficient(CIW-PSO-TVAC) to improve its performance under noise-free and noise-contaminated condition. The result obtained for optimized and unoptimized ANN, SVM and ANFIS under noise-free and noise contaminated condition were compared side by side. Based on the result shown, optimized ANN outperform compare to SVM and ANFIS for noise-free and noise-contaminated condition which show the highest classification accuracy and noise tolerance level. Hence, this method can be proposed as one of the method to improve the performance of classifiers for PD pattern classification under noise-free and noise-contaminated condition.

Item Type: Final Year Project
Subjects: Technology > Electrical engineering. Electronics engineering
Faculties: Faculty of Engineering and Technology > Bachelor of Engineering (Honours) Electrical and Electronics
Depositing User: Library Staff
Date Deposited: 09 Oct 2018 04:01
Last Modified: 03 Apr 2019 09:16
URI: https://eprints.tarc.edu.my/id/eprint/181