A Hybrid CNN-RNN based Energy Consumption Forecasting for Smart Grids in Industries

Energy generation and delivery provide fundamental problems to civilization. Renewable energy sources now account for a larger portion of the energy. In these conditions, small-scale network grid structures operating as fully functional energy systems in small areas are becoming appealing. Reliable energy forecasting by major users, such as manufacturers, is critical to the smart grid's success. This study's goal was to enhance energy consumption by predicting software for end-users. The software calculates energy consumption using real-world information from a manufacturing facility or device. This method provides end-customers with energy demand forecasting capabilities, allowing them to participate more efficiently in the power grid energy system. A hybrid CNN and RNN model was developed to predict energy demand. The model was created using data from an industry's energy consumption. The model has a strong forecasting ability and projected energy usage with an MAEV of less than 1%. Other deep learning models, such as LSTM, KNN, BPNN, RFR, and PLSR, were compared to the developed model.