Load Forecasting using Fuzzy Methods

Today, it´s the need of developed and developing countries to consume electricity more efficiently. Though developed countries do not want to waste electricity and developing countries cannot waste electricity. Hence, the wise use of electricity is the need of hour. This leads to the concept - load forecasting. This paper is written for the short term load forecasting on daily basis. Though this can be extended to hourly or half-hourly or real time load forecasting. But as we move from daily to hourly basis of load forecasting the error of load forecasting increases. This paper is written on the practical analysis of previous year´s load data records of an Engineering College in India using the concept of fuzzy methods. The analysis has been done on Mamdani type membership functions. In order to reduce the error of load forecasting on daily basis fuzzy method has been used with artificial network (ANN) with some iteration processes. The error has been reduced to a considerable level in the range of 2-3%. Further studies are going on with fuzzy regression methods to reduce the error more.