Neural-wavelet Methodology for Load Forecasting

Intelligent demand-side management represents a future trend of power system regulation. A key issue in intelligent demand-side management is accurate prediction of load within a local area grid (LAG), which is defined as a set of customers with an appropriate residential, commercial and industrial mix. Power consumption is deemed to be unpredictable in some sense due to the idiosyncratic behaviors of individual customers. However, the overall pattern of a group of consumers is possible to predict. The developed neural-wavelet approach is shown capable of handling the nonlinearities involved and provides a unique tool for intelligent demand-side management. The paper presents the neural-wavelet approach and its implementation to load identification and forecasting.