Short time load forecasting for Urmia city using the novel CNN-LTSM deep learning structure

In the present time, electricity stands as one of the most fundamental needs within human societies. This is evident in the fact that all industrial activities and a significant portion of social, economic, agricultural, and other activities rely heavily on this energy source. As a result, both the quality and continuity of electricity hold immense importance. The primary objective of this study is to predict short-term changes in load consumption. These predictions are based on a range of factors that influence electric consumption, factors characterized by complex nonlinear relationships. Notably, these factors encompass climate shifts and fluctuations within daily consumption cycles. The proposed method for short-term load forecasting (STLF) involves a hybrid neural network utilizing deep learning techniques. Specifically, it combines the convolutional neural network (CNN) and long short-term memory (LSTM) architectures. The CNN architecture is leveraged for its proficiency in extracting patterns from data, while the LSTM architecture excels in predicting time series data. The suggested approach enables the anticipation of future consumption patterns by considering upcoming weather conditions and analyzing past electricity consumption trends. Numerical results underscore the enhanced forecast precision achieved through this method, which is about 1% better than the best previous results, as evidenced by improvements in metrics such as the root mean square error (RMSE) and the mean absolute percentage error (MAPE). These improvements outperform the best available methods presented in prior research. Thus, this paper not only contributes a novel approach but also serves as a comprehensive review of the latest developments in the realm of short-term load forecasting.