Forecasting short term electric load based on stationary output of artificial neural network considering sequential process of feature extraction methods

With the advent of deregulation in electric utilities, short-term load forecasting (STLF) becomes even more important especially to the system operators and market participants in which this may assist them towards organizing appropriate planning strategies of risk management and competitive energy trading. This is important to ensure the electric utilities are operating in an economic, reliable, secure and uninterrupted service to the customers. This paper presents the application of artificial neural network (ANN) that used to perform the STLF. The Malaysian hourly peak load in the year 2002 is used as a case study in the assessment of STLF using ANN. The proposed methodology comprises of ANN model incorporating with stationary output and sequential process of feature extraction methods. The multiple time lags of input data and principal component analysis (PCA) are performed in a sequential process of feature extraction methods so that this will reduce the size of significant input data for improving the performance of ANN in providing accurate result of STLF.