Scholarlink

<p>The coordinated operation of hybrid photovoltaic (PV) and Small Modular Reactor (SMR) microgrids represents a promising pathway to achieve resilient, low-carbon energy supply in modern power systems. However, effective management of such systems requires advanced optimization frameworks that simultaneously address cost minimization, carbon emission reduction, and operational resilience under multi-source uncertainties. This paper proposes a comprehensive scheduling framework for hybrid PV-SMR microgrids, integrating multi-scale energy storage–lithium-ion batteries for short-term balancing and hydrogen storage for long-term seasonal regulation–while explicitly incorporating demand response flexibility. The proposed framework adopts a multi-objective distributionally robust optimization (DRO) approach to capture uncertainties in solar generation and load fluctuations, ensuring robust yet cost-effective dispatch decisions. The mathematical model addresses the multi-timescale coordination between variable PV generation, slow-ramping nuclear power, and dynamic battery and hydrogen storage operations. Key constraints include power balance, SMR ramping limits, battery state-of-charge evolution, hydrogen production and consumption cycles, and resilience-driven critical load prioritization. Furthermore, a real-time reinforcement learning (RL)-assisted mechanism enhances the system’s adaptability to evolving operational states, enabling dynamic adjustment of storage and demand response strategies based on live system feedback. A comprehensive case study is conducted on a 100 MW hybrid microgrid, integrating 40 MW of PV, a 50 MW SMR, a 20 MWh battery storage system, and a 15-ton hydrogen storage facility, supplying industrial and residential loads under realistic uncertainty scenarios. Results demonstrate that the proposed optimization achieves a 17.5% reduction in operational cost and a 32.8% reduction in carbon emissions compared to conventional microgrid scheduling, while enhancing resilience by maintaining continuous supply for critical loads even under extreme weather stress. The integration of DRO and reinforcement learning provides a 28% improvement in flexibility under solar variability, confirming the importance of adaptive, uncertainty-aware optimization for future hybrid microgrids. This work contributes an advanced, scalable framework for multi-energy hybrid microgrid management, providing valuable insights for resilient and low-carbon community microgrid development in the renewable-dominated era.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Coordinated operation and multi-layered optimization of hybrid photovoltaic-small modular reactor microgrids

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年11月19日

查看全文
  • Yao Duan,
  • Chong Gao,
  • Ye Huang,
  • Qiang Luo,
  • Zhiheng Xu

摘要

The coordinated operation of hybrid photovoltaic (PV) and Small Modular Reactor (SMR) microgrids represents a promising pathway to achieve resilient, low-carbon energy supply in modern power systems. However, effective management of such systems requires advanced optimization frameworks that simultaneously address cost minimization, carbon emission reduction, and operational resilience under multi-source uncertainties. This paper proposes a comprehensive scheduling framework for hybrid PV-SMR microgrids, integrating multi-scale energy storage–lithium-ion batteries for short-term balancing and hydrogen storage for long-term seasonal regulation–while explicitly incorporating demand response flexibility. The proposed framework adopts a multi-objective distributionally robust optimization (DRO) approach to capture uncertainties in solar generation and load fluctuations, ensuring robust yet cost-effective dispatch decisions. The mathematical model addresses the multi-timescale coordination between variable PV generation, slow-ramping nuclear power, and dynamic battery and hydrogen storage operations. Key constraints include power balance, SMR ramping limits, battery state-of-charge evolution, hydrogen production and consumption cycles, and resilience-driven critical load prioritization. Furthermore, a real-time reinforcement learning (RL)-assisted mechanism enhances the system’s adaptability to evolving operational states, enabling dynamic adjustment of storage and demand response strategies based on live system feedback. A comprehensive case study is conducted on a 100 MW hybrid microgrid, integrating 40 MW of PV, a 50 MW SMR, a 20 MWh battery storage system, and a 15-ton hydrogen storage facility, supplying industrial and residential loads under realistic uncertainty scenarios. Results demonstrate that the proposed optimization achieves a 17.5% reduction in operational cost and a 32.8% reduction in carbon emissions compared to conventional microgrid scheduling, while enhancing resilience by maintaining continuous supply for critical loads even under extreme weather stress. The integration of DRO and reinforcement learning provides a 28% improvement in flexibility under solar variability, confirming the importance of adaptive, uncertainty-aware optimization for future hybrid microgrids. This work contributes an advanced, scalable framework for multi-energy hybrid microgrid management, providing valuable insights for resilient and low-carbon community microgrid development in the renewable-dominated era.