Application of Artificial Neural Network for Short Term Electricity Demand Forecasting

Many researchers around the world work on short term electricity demand forecasting (STLF) in order to establish an accurate power planning and generation system their countries. This research, with its focus short-term load forecasting, aims fill this gap by implementing two methodologies based Artificial Neural Network (ANN) Autoregressive Integrated Moving Average (ARIMA) applied a set of half hourly data six years, provided Ceylon Electricity Board Sri Lanka. The first five years (~70% dataset) were used train algorithms those last year (~30% for testing. effect historical patterns making prediction next 24 hours studied. Moreover, data, unlike most literature which forecasts only one value (either peak day or hour), entire (48 values each hour) is forecasted at once. predictions obtained application ANN compared ARIMA methodology benchmark comparing STLF. None applications deviated other can be predict half-hourly since was successful grasping periodic that exist series.