APPLICATION OF FUZZY LOGIC TO LOAD FORECASTING

Abstract

The present work is done for 30 days load data taken for Nov.2003 for IITR campus for load forecasting and is tested for Fuzzy Time Series methodology, which is an integration of conventional and non-conventional approaches. To test the developed MATLAB program, load data for 30 days i.e. from 1st Nov. 2003 to 301h Nov. 2003 has been collected from the different substations located inside the campus of Indian Institute of Technology, Roorkee. These required data are taken from the substations located in Indian Institute of Technology, Roorkee. In campus, there are five substations, which are taking the input from 33KV main receiving substation under Uttaranchal Power Corporation supplying to the 11 kv substation. These substations are located in Cautley Bhawan, Welding Research Laboratory, High Voltage laboratory, Wind tunnel substation and the old substation near Physics Deptt. For the development of fuzzy time series load forecasting, the load data is collected from these five substations and the daily record of weather variables, average temperature and relative humidity has also been obtained from Hydrology department of I.I.T. Roorkee for the same period. The data is compiled in terms of various pre-decided variables for load forecasting using regression-based analysis. The proposed method is well suited for online load forecasting as the part of the program is deal with the historical data worked out for off line and only the adaptive part of it done on the online basis. This method can be constructed in simpler procedure than the conventional approaches.