Short-term load forecasting for microgrid energy management system using hybrid HHO-FNN model with best-basis stationary wavelet packet transform

Abstract

Accurate prediction of load has become one of the most crucial issue in the energy management system of the microgrid. Therefore, a precise load forecasting tool is necessary for efficient power management in the microgrid, which can lead to economic benefits for consumers and power industries. This paper proposes a hybrid approach for short-term forecasting of load demand in a typical microgrid, which is a combination of the best-basis stationary wavelet packet transform and Harris hawks optimization-based feed-forward neural network. The Harris hawks optimization is applied to the feed-forward neural network as an alternative training algorithm for optimizing the weight and basis of neurons. The proposed model is applied to predict load demand in the Queensland electric market and is compared with existing competitive models. Numerical results are obtained using MATLAB. These results demonstrate that the proposed approach reduces the average mean absolute percentage error by 33.30%, 49.54% and 60.76% as compared to the particle swarm optimization (PSO) based artificial neural network, PSO based least-square-support vector machine and back-propagation based neural network, respectively.