Analysis on anomalous short term load forecasting using two different approaches

Abstract

The problem of optimizing Short Term Load Forecasting (STLF) is a task in the management of electrical power systems. STLF problem solving using soft computing approach has been becoming the interest of researchers. This study aims to compare two approaches on anomalous short term load forecasting. These approaches are one method based on Soft Computing, which is Fuzzy Subtractive Clustering (FSC) algorithm and Holt-Winter exponential smoothing, which is included in Time Series Analysis. Load forecasting is defined by the base load and peak load for weekdays and weekend days. The experimental results verify that the influences range optimization from FSC algorithm provided a better accuracy of forecasting results affecting on the efficiency of generation cost. Meanwhile, the exponential smoothing method produces a good result although they tend to lag behind the observed values a little bit.