APPLICATION OF FUZZY LOGIC FOR SHORT TERM LOAD DEMAND FORECASTING

This Final Year Project report presents the implementation of Fuzzy Logic techniques in
designing the short term load demand forecasting model for Gas District Cooling (GDC)
of Universiti Teknologi PETRONAS and New South Wales, Australia. The
implementation of Fuzzy Logic technique was carried out by FuzzyTECH software. The
performance of the forecast model is evaluated through the prediction for the days in the
year 2005. The fuzzy forecasted load values are compared with the actual load to get the
forecast accuracy. From the model error analysis, the Mean Absolute Percentage Error
(MAPE) is about I% to 6%. It shows that the forecast models still need an improvement
in order to reduce errors until reach the acceptable accuracy level. The meteorological
variables such as weather temperature, humidity is required in order to achieve high
precision of STLF model.