Smart metering data analytics for non-intrusive load monitoring

With the rise of advanced metering infrastructure, Non-Intrusive Load Monitoring (NILM) has been extensively used in many fields from industrial to residential. The aim of NILM is to successfully disaggregate the aggregated load into individual appliances profile. This project survey NILM methodology and outlines its basic principle framework. In the initial stage, data acquisition was conducted in Nanyang Technological University (NTU) Clean Energy Research Lab (CERL) and the targeted appliances were a standing fan and table lamp. The data acquired was then disaggregate with Artificial Neural Network (ANN). To further evaluate the NILM performance, a large public dataset, UK-Dale, was used and a comparison was done between ANN, Convolution Neural Network (CNN), and Long-Short Term Memory (LSTM) model. In the result section, both classification and regression approach would be evaluated. In this project, pre-processing and post-techniques were designed and employed to improve the performance of the prediction significantly.Bachelor of Engineering (Electrical and Electronic Engineering