A Parallel Framework for Fast Charge/Discharge Scheduling of Battery Storage Systems in Microgrids DOI Creative Commons
Wei Huang, Wu‐Chun Chung, Chao‐Chin Wu

et al.

Energies, Journal Year: 2024, Volume and Issue: 17(24), P. 6371 - 6371

Published: Dec. 18, 2024

Fast charge/discharge scheduling of battery storage systems is essential in microgrids to effectively balance variable renewable energy sources, meet fluctuating demand, and maintain grid stability. To achieve this, parallel processing employed, allowing batteries respond instantly dynamic conditions. By managing the complexity, high data volume, rapid decision-making requirements real time, ensures that microgrid operates with stability, efficiency, safety. With application deep reinforcement learning (DRL) algorithm design, demand for computational power has further increased significantly. address this challenge, we propose a Ray-based framework accelerate development fast microgrids. We demonstrate how implement real-world problem framework. focused on minimizing losses reducing ramping rate net loads by leveraging Asynchronous Advantage Actor Critic (A3C) algorithms features Ray cluster real-time decision making. Multiple instances OpenDSS were executed concurrently, each instance simulating distinct environment efficiently input data. Additionally, Numba CUDA was utilized facilitate GPU acceleration shared memory, significantly enhancing performance computationally intensive reward function A3C. The proposed enhanced performance, enabling efficient management complex, environments.

Language: Английский

A Deep Reinforcement Learning Optimization Method Considering Network Node Failures DOI Creative Commons
Xueying Ding, Xiao Liao, Wei Cui

et al.

Energies, Journal Year: 2024, Volume and Issue: 17(17), P. 4471 - 4471

Published: Sept. 6, 2024

Nowadays, the microgrid system is characterized by a diversification of power factors and complex network structure. Existing studies on fault diagnosis troubleshooting mostly focus detection operation optimization single device. However, for increasingly systems, it becomes challenging to effectively contain faults within specific spatiotemporal range. This can lead spread faults, posing great harm safety microgrid. The topology based deep reinforcement learning proposed in this paper starts from overall grid aims minimize failure rate optimizing grid. approach limit internal small range, greatly improving reliability operation. method optimize node multi-node fault, reducing influence range 21% 58%, respectively.

Language: Английский

Citations

2

A Parallel Framework for Fast Charge/Discharge Scheduling of Battery Storage Systems in Microgrids DOI Creative Commons
Wei Huang, Wu‐Chun Chung, Chao‐Chin Wu

et al.

Energies, Journal Year: 2024, Volume and Issue: 17(24), P. 6371 - 6371

Published: Dec. 18, 2024

Fast charge/discharge scheduling of battery storage systems is essential in microgrids to effectively balance variable renewable energy sources, meet fluctuating demand, and maintain grid stability. To achieve this, parallel processing employed, allowing batteries respond instantly dynamic conditions. By managing the complexity, high data volume, rapid decision-making requirements real time, ensures that microgrid operates with stability, efficiency, safety. With application deep reinforcement learning (DRL) algorithm design, demand for computational power has further increased significantly. address this challenge, we propose a Ray-based framework accelerate development fast microgrids. We demonstrate how implement real-world problem framework. focused on minimizing losses reducing ramping rate net loads by leveraging Asynchronous Advantage Actor Critic (A3C) algorithms features Ray cluster real-time decision making. Multiple instances OpenDSS were executed concurrently, each instance simulating distinct environment efficiently input data. Additionally, Numba CUDA was utilized facilitate GPU acceleration shared memory, significantly enhancing performance computationally intensive reward function A3C. The proposed enhanced performance, enabling efficient management complex, environments.

Language: Английский

Citations

0