Short-Term Load Forecasting in Power System Using CNN-LSTM Neural Network

The accurate forecasting of short-term load plays a significant role in power systems operation and planning. This paper suggests a short-term load forecasting model combining Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM). The developed CNN-LSTM aims to capture both spatial and temporal dependencies within the load data, leveraging the strengths of both architectures. Simulations are performed using real-world power system load data. Comparative analyses are carried out against standalone CNN and LSTM models. The CNN-LSTM has significantly better forecasting accuracy than other models, showcasing its effectiveness in short-term load forecasting.