Dynamic load forecasting for commercial power network

• Main Author: Alzalet, Abdusalam Rajb
• Format: Thesis
• Language: English; English; English
• Published: 2014
• Subjects: TK Electrical engineering. Electronics Nuclear engineering; TK1001-1841 Production of electric energy or power. Powerplants. Central stations
• Online Access: http://eprints.uthm.edu.my/1423/1/24p%20ABDUSALAM%20RAJB%20ALZALET.pdf; http://eprints.uthm.edu.my/1423/2/ABDUSALAM%20RAJB%20ALZALET%20COPYRIGHT%20DECLARATION.pdf; http://eprints.uthm.edu.my/1423/3/ABDUSALAM%20RAJB%20ALZALET%20WATERMARK.pdf; http://eprints.uthm.edu.my/1423/

Load forecasting is an important component for power system energy management system. The electrical load is the power that an electric utility needs to supply in order to meet the demands of its customers. It is therefore very important to the utilities to have advance knowledge of their electrical load, so that they can ensure the load is met and thus minimising any interruptions to their service. It also plays a key role in reducing the generation cost, and also essential to the reliability of power systems. The electric power demand in Universiti Tun Hussein Onn Malaysia (UTHM) has increased as the power system network is getting larger with more consumption is to be expected. This loading trend is certain to continue in the near future. The aim of this project is to forecast the medium term loading of UTHM Linear regressions and polynomial based methods as well as artificial neural networks (ANN) approach have been adapted in the load forecasting from 2006 to 2012. The results attained are validated with the real data obtained from the Tenaga Nasional Berhad (TNB) which represents the monthly load electric consumption in UTHM. By comparing the forecasted results with the real data, the most suitable method has been proposed. When the approaches are compared according to their highest prediction error, the highest error for linear regression and Polynomial equation approaches are very high compared to the ANN approach. Generally the ANN approach has produced better results.