Seasonal Short-Term Electricity Demand Forecasting under Tropical Condition using Fuzzy Approach Model

Abstract

Concern of this work is analysis and short-term electricity demand forecasting under tropical condition using fuzzy approach. Two different demand models are proposed for dry season and rainy season to forecast a total load demand in Makassar, Indonesia for 24 hours ahead in each season. Based on the typical characteristic of seasonal demand, three inputs (time of load, temperature, and type of day) are used for load model in dry season, and four inputs (time of load, temperature, type of day, and rainfall) for load model in rainy season. Meanwhile, output is estimated load in related seasons. Some forecasting error analyses are applied to models. Under tested cases, both seasonal models have good forecasting results with MAPE values smaller than 2.95%. Estimated demand values when holidays and non-holidays in each season which are relatively close to actual load have confirmed effectiveness of the fuzzy based models.