Smart Grid Evolution: Deep Reinforcement Learning for Carbon-Free Based AI-Driven Energy Management | ||||
| Engineering Research Journal (Shoubra) | ||||
| Articles in Press, Accepted Manuscript, Available Online from 22 July 2025 | ||||
| Document Type: Research articles | ||||
| DOI: 10.21608/erjsh.2025.394317.1430 | ||||
 View on SCiNiTO | ||||
| Author | ||||
Mohamed A.Wahab ALI  
| ||||
| Department of Electrical Engineering, Faculty of Engineering at Shoubra, Benha University, Cairo, Egypt. | ||||
| Abstract | ||||
| This paper presents a Deep Deterministic Policy Gradient (DDPG) framework for real-time optimization of smart grids with high renewable energy integration. The proposed model addresses the critical challenge of balancing intermittent generation and dynamic demand while minimizing carbon emissions and maintaining grid stability. By employing a multi-objective reward function, the system simultaneously optimizes environmental impact, operational efficiency, and power quality. The proposed framework is tested on the IEEE 33-bus system, the DDPG-based solution demonstrates superior performance, achieving a 32% reduction in power losses (120 kW) and 28% lower carbon emissions compared to conventional methods. The framework's key advantages include continuous control of energy storage systems, adaptive renewable power allocation, and computationally efficient implementation suitable for large-scale deployment. These results highlight the potential of deep reinforcement learning to enable more sustainable, resilient, and intelligent power systems, offering a practical solution for the energy transition. The approach significantly outperforms traditional optimization techniques while maintaining the flexibility required for real-world grid operations. | ||||
| Keywords | ||||
| Smart Grid Optimization; Deep Reinforcement Learning (DRL); Renewable Energy Integration; Carbon Emissions Reduction; Deep Deterministic Policy Gradient (DDPG) | ||||
|
Statistics Article View: 40 |
||||
