HHO-GCN-LSTM APPLY IN LOADPROFILE FORECASTING FOR MICROGRID

Abstract

Power load graph forecasting is of great interest in microgrid (MG) energy management. The need for
accurate short-term load charting is particularly important for efficient power management for MG. This
paper proposes a new method for short-term load forecasting (STLF). This method uses the time series of
temperature and load data provided to the model based on the Graph Convolutional Network (GCN) model
to combine the characterization of the input and output data given into the calculation for the
corresponding LSTM network to forecast the hourly load graph in the future. In order to evaluate the
accuracy of the prediction model, this study used the HHO optimization algorithm to calculate the GCNLSTM network. To compare the results of the model with other forecasting models, we work with the load
data set of an MG model belonging to the Ho Chi Minh City power grid. The forecast model is compared with
previous forecasting models. The results show that our proposed model has superior results compared to
other deep learning-based models in terms of root mean square error (RMSE) and mean absolute
percentage error (MAPE).