Electric load demand forecasting for Aborlan-Narra-Quezon distribution grid in Palawan using multiple linear regression

摘要

Load forecasting is needed by an electric utility to anticipate the amount of power to be generated and distributed to the consumers. It will be a big help for them in power system planning, scheduling, and operation especially now that there is power crisis in the country. In this paper, the use of regression based approach will be used to forecast the load demand of Aborlan-Narra-Quezon distribution grid in Palawan for the next ten (10) years. The factors considered were the actual historical data from the local power utility company, the number of consumers for the past five (5) years, and the development plans (i.e. commercial, industrial) on the subject area for the next ten (10) years ahead as the input data. Results showed that the proposed method is satisfactory. The overall load forecasting error or the mean average percentage error (MAPE) is 2.26% which is very much feasible and acceptable. The use of multiple linear regression in MS Excel is implemented and very useful to evaluate the accuracy and significance of the forecasted data. The forecasted data determined the need for capacity addition of an existing generating plant and proposed another source of energy. If the proposed model is adopted, it can be applied to other distribution system in Palawan to forecast the entire electricity demand of the main distribution grid as well as the island grid.