An improved RBF neural network for short-term load forecast in smart grids

Abstract

With the rapid development of smart grids, a large volume of smart meter data has been collected. The analysis of these big data can be leveraged to channel the data from individual meters into knowledge valuable to electric power utilities and end-consumers. Short-term load forecasting is essential for ensuring both efficient and reliable operations of smart grids. This paper proposes an improved Radial Basis Function neural network model (RBF-PCA-WFCM) for short-term load forecast, which use the weighted Fuzzy C-Means (FCM) clustering algorithm based on Principal Component Analysis (PCA) to determine the basis function centers, and use the gradient descent algorithm to train the output layer weights. The proposed model is implemented on real smart meter data and simulation results show that better forecasting accuracy could be achieved by using the proposed model comparing with the conventional RBF neural network model based on K-Means (RBF-KM).