Fuzzy approach performance of shortterm electricity load forecasting in Malaysia

• Main Authors: Mansor, Rosnalini, Zulkifli, Malina, Mat Yusof, Muhammad, Ismail, Mohd Isfahani, Ismail, Suzilah, Yip, Chee Yin
• Format: Monograph
• Language: English
• Published: Universiti Utara Malaysia 2014
• Subjects: QA75 Electronic computers. Computer science
• Online Access: http://repo.uum.edu.my/24771/1/12407.pdf; http://repo.uum.edu.my/24771/
• Institution: Universiti Utara Malaysia

Many activities (such as economic, education and etc.) would paralyse with limited supply of electricity but surplus contribute to high operating cost.Therefore electricity load forecasting is important in order to avoid shortage or excess.Many techniques have been employed in forecasting short term electricity load.They can be classifies either by statistical or artificial intelligent (AI) or hybrid of those two techniques; Statistical techniques and AI techniques. Electricity load demand is influenced by many factors, such as weather, economic, social activities and etc.The relation between load demand and the independent variables is complex and it is not always possible to fit the load curve using statistical models.The complexity and uncertainties of this problem appear suitable for fuzzy methodologies.Hence, the Fuzzy approach was used to forecast electricity load demand.Previous findings showed festive celebration has effect on shortterm electricity load forecasting.Being a multi culture country Malaysia has many major festive celebrations (EidulFitri, Chinese New Year, Deepavali and etc.) but they are moving holidays due to non-fixed dates on the Gregorian calendar.Therefore, the performance of fuzzy approach in forecasting electricity loads when considering the presence of moving holidays was studied.Autoregressive Distributed Lag (ARDL) model was estimated using simulated data by including model simplification concept (manual or automatic), day types (weekdays or weekend), public holidays and lags of electricity load.The result indicated that day types, public holidays and several lags of electricity load were significant in the model.Overall, model simplification improves fuzzy performance due to less variables and rules.