Prediction Modeling for Future Electrical Energy Demand in Malaysia

• Main Author: Ahmad Khan, Imtiaz
• Format: Thesis
• Language: English; English
• Published: 2006
• Online Access: http://psasir.upm.edu.my/id/eprint/610/1/1600449.pdf; http://psasir.upm.edu.my/id/eprint/610/
• Institution: Universiti Putra Malaysia

Accurate forecasting of energy requirement for future development of the country is one of the most important factors of energy management. Adequacy of energy is the main factor for the development of a country. Electricity producing natural resources are depleting very fast, all over the world, since their quantity is limited and their use is increasing very rapidly. But the pace of development can not be compromised. Financial limitations do not permit to, increase the generation capacity to meet the peak demand and produce surplus electricity after peak hours, and obtain new technologies of electricity generation. For installation and maintenance of generation capacity, transmission and distribution infrastructure long term forecasting is very important. Energy requirement depends on number of variables, some of them which are cardinal to the energy consumption and addressed here are population, number of electricity consumers, per capita electricity consumption, peak electricity demand, gross domestic product and annual electricity consumption of the country. Data for these variables are available annually and have very firm relation with time. These data were analyzed in this work. Annual electricity consumption has been taken as dependent and rest as independent variables. All the variables have been evaluated for first, second and third order polynomial with time and mathematical relation was found. This mathematical relation was then extrapolated into future for next ten years, the forecast horizon. Out of these, evaluated values of independent variables having minimum standard deviation from the past data trend, were used in developing multi variable model. All the evaluation work was performed on MATLAB software. The chance of error is low in this model since it takes the variation of data into consideration and follows the previous trend by checking standard deviation. Once the data are keyed in the program it takes less than a minute in giving the forecasted values and its corresponding graph. The achievement of this work is that by just updating the data of the variables for recent year in the program the current updated forecast for next ten years can be obtained. This forecast may be of great use for energy managers. Since it is sensitive to six independent variables, it gives more reliable forecast. This program can be used for any country for the same forecast horizon with the assumption that the previous trend will persist.