Fault Location and Classification Using CAD Software

Lau, Kim Hui (2021) Fault Location and Classification Using CAD Software. Final Year Project (Bachelor), Tunku Abdul Rahman University College.

Text Lau Kim Hui.pdf Restricted to Registered users only Download (1MB)

Abstract

Power distribution systems are the core of electrical power system and hence require a fast fault locating system to minimize the downtime due to occurrence of fault. IEEE 34 bus feeder system with slight modification was simulated in PSCAD software and the current data from the fault signal were collected and used to train Support Vector Machine algorithm for fault location and classification. DWT and PCA was used for dimension reduction and feature extraction of the data to reduce the computational burden and training time. Different fault resistance and fault type occurred at different point in the distribution system was simulated to examine the performance of the machine learning algorithm under different condition and 5-fold Cross validation was used to prevent overfiring of data. An accuracy of 99% was achieved when classify the 4 fault category namely single phase to ground fault, line to line fault, double line to ground fault and three phase line to ground fault. Besides that, an RMSE of 0.19 was achieved for fault distance estimation. The proposed machine learning algorithm is to be run using MATLAB power simulation software using Statistics and machine learning toolbox.