Optimization and analysis of passive optical networks using deep learning-based energy management in renewable smart

Abstract Electric-power consumption (EPC) and renewable-energy generation (REG) exhibit high variability, impacting grid stability and power quality. To address these challenges, this paper proposes a deep learning driven real-time demand side management (DSM) controller for smart buildings integrated with renewable energy sources (RES) and energy storage systems (ESS). The system employs an Internet of things (IoT) enabled architecture to monitor, forecast, and optimize energy usage at high temporal resolutions of five minutes. A bidirectional long short-term memory (B-LSTM) network is utilized to forecast short-term EPC and photovoltaic (PV) generation, surpassing LSTM, GRU, and other baseline models in terms of RMSE, MAE, and R 2 . The DSM controller applies dynamic incentive and penalty schemes to shift curtailable loads from peak to off-peak periods, reducing peak-to-average ratio (PAR) and lowering electricity costs for both prosumers and utility operators. Real-time decisions are enabled through dynamic pricing integration and IoT-based ESS control. The proposed framework provides a scalable, intelligent, and sustainable solution for future smart grid energy management.