Model Development Of The Energy Demand By Utilizing Artificial Intelligence (AI) Technique

• Main Author: Saifullah, Muhammad Nasih
• Format: Final Year Project
• Language: English
• Published: Universiti Teknologi PETRONAS 2014
• Subjects: Instrumentation and Control
• Online Access: http://utpedia.utp.edu.my/14409/1/Final%20Report_13732.pdf; http://utpedia.utp.edu.my/14409/
• Institution: Universiti Teknologi Petronas

Load forecasting has been one of the major researches in electrical engineering in the recent years. It plays a very important role in power system planning and operation. Through load forecasting, power generation can be balanced with load demand, which subsequently contributes to an efficient electricity management in power system. One way of forecasting load demand is by using Artificial Intelligence (AI) technique. There are two AI technique’s methods discussed in this project which are Artificial Neural Network (ANN) method and Fuzzy Logic (FL) method. Both approaches utilize MATLAB software. The accuracy of the forecast is based on the Mean Absolute Average Error (MAPE). Instead of using a year-long historical data, this project uses selective seasonal historical data, focusing on the autumn season. 1-hour ahead and 24-hour ahead load forecasts are developed for each approach. The first chapter of this report discusses the fundamental of each method and also statistical analysis of data. The following chapter describes how each method is developed by using MATLAB. Followed next is a chapter consisting of results of both models and followed by discussions of the results obtained. The last chapter is on conclusion as well as recommendations for any possible future continuation of the project.