Hybrid Method for Photovoltaic System Forecasting Based on Wavelet Transform and LSTM

Photovoltaic (PV) forecasting is done to help in managing planning and knowing PV energy production. However, PV systems are always associated with fluctuating weather parameters, such as solar irradiance, temperature, wind speed, and humidity. In order, to forecast uncertain data, this object over here study proposed a combination method by using a concept of wavelet transform (WT) and the ability of long-short-term memory (LSTM). The WT is used to decompose the output power with the feature extraction process. Then it is used as input to the LSTM to forecast the futuristic value of output power. The forecasting results are shown in a curve and compared with actual data at several time horizons of 1 day, 15 days, 30 days, and 60 days. In addition, the results are also compared with several other ML methods such as Ridge Regression (RR), Linear Regression (LR), Least Absolute Shrinkage and Selection Operator (LASSO), and Elastic Net (ENET). Each of these treatments shows that the forecasting results using the proposed method are superior. This is also supported by evidence on the percent error values of Root mean squared error (RMSE), mean absolute error (MAE), mean absolute percentage error (MAPE), and coefficient of determination (R2)