Renewable Energy Forecasting AI-Driven Framework for Peer to Peer Trading

This research aims to predict future energy demand by utilizing a new dataset that captures energy consumption patterns from Education City Community Housing (ECCH). This research employs sophisticated deep learning algorithms to improve the efficiency of peer-to-peer (P2P) energy systems while minimizing congestion and losses. Three different learning algorithms are utilized in the study: Bidirectional Long-Short- Term Memory (Bi-LSTM), Random Forest (RF), and Gated Recurrent Unit (GRU), to predict future energy consumption values for the complex. The results are assessed using metrics such as Mean Absolute Error (MAE) and Root Mean Square Error (RMSE). The findings indicate a strong potential for accurately forecasting future energy demand and showcase the superior performance of the GRU and Bi-LSTM models, particularly when implemented as univariate models that strictly focus on energy consumption data without additional features.