MONITORING HARMONICS IN UTP DISTRIBUTION NETWORK USING VIRTUAL HARMONIC ANALYSER AND FORECASTER(V-HAF) SYSTEM

In power distribution system across the world, the need to maintain its power quality remains crucial. As technology takes over various aspects of life, the amount of non-linear equipment has been increasing, forming distortion in the distribution network. This is mainly due to the harmonics current injected by non-linear loads into the distribution line, which then pollutes the power quality and poses risks and losses. In Universiti Teknologi PETRONAS’s (UTP’s) distribution network, similar issues being faced. Hence, this study aims to propose and demonstrate an intelligent system that not only enable harmonic monitoring in UTP, but also reduces instrument cost at data gathering phase. This thesis proposed implementation of Virtual Harmonic Analyser and Forecaster (V-HAF) system to monitor harmonic fluctuation in UTP distribution network. The proposed V-HAF system compromises two different intelligent tool, which are estimation (Tool#1) and time-series prediction (Tool#2) tools in order to serve the objective. Techniques for both tools proposed based on literature study done in respective fields. For the estimation tool, feed-forward neural network with Levenberg-Marquardt back-propagation training technique was proposed and tested against the common Gradient Descent training method