Forecasting peak load electricity demand using statistics and rule based approach

• Main Authors: Ismail, Zuhaimy, Yahaya, Azizi, Mahpol, K. A.
• Format: Article
• Language: English
• Published: Science Publications 2009
• Subjects: L Education (General)
• Online Access: http://eprints.utm.my/id/eprint/9714/1/ajas681618-1625.pdf; http://eprints.utm.my/id/eprint/9714/; http://www.scipub.org/fulltext/ajas/ajas681618-1625.pdf

Problem statement: Forecasting of electricity load demand is an essential activity and an important function in power system planning and development. It is a prerequisite to power system expansion planning as the world of electricity is dominated by substantial lead times between decision making and its implementation. The importance of demand forecasting needs to be emphasized at all level as the consequences of under or over forecasting the demand are serious and will affect all stakeholders in the electricity supply industry. Approach: If under estimated, the result is serious since plant installation cannot easily be advanced, this will affect the economy, business, loss of time and image. If over estimated, the financial penalty for excess capacity (i.e., over-estimated and wasting of resources). Therefore this study aimed to develop new forecasting model for forecasting electricity load demand which will minimize the error of forecasting. In this study, we explored the development of rule-based method for forecasting electricity peak load demand. The rule-based system synergized human reasoning style of fuzzy systems through the use of set of rules consisting of IF-THEN approximators with the learning and connectionist structure. Prior to the implementation of rule-based models, SARIMAT model and Regression time series were used. Results: Modification of the basic regression model and modeled it using Box-Jenkins auto regressive error had produced a satisfactory and adequate model with 2.41% forecasting error. With rule-based based forecasting, one can apply forecaster expertise and domain knowledge that is appropriate to the conditions of time series. Conclusion: This study showed a significant improvement in forecast accuracy when compared with the traditional time series model. Good domain knowledge of the experts had contributed to the increase in forecast accuracy. In general, the improvement will depend on the conditions of the data, the knowledge development and validation. The rule-based forecasting procedure offered many promises and we hoped this study can become a starting point for further research in this field.