Scholarlink

<p>This paper proposes an integrated hierarchical coordination framework for planning and operations to address the decoupling between long-term capacity planning and short-term operational scheduling in renewable energy systems and the resulting economic and reliability losses. The core innovation lies in its “planning-operation” closed-loop feedback system. The upper-level planning adopts the non-dominated sorting genetic algorithm III (NSGA-III) to optimize the capacity mix of wind, solar, and energy storage, balancing investment economy and system reliability. The lower-level operation introduces an adaptive distributed model predictive control (DMPC) strategy empowered by deep deterministic policy gradient (DDPG), enabling efficient response to real-time uncertainty. Through a multi-stage elite feedback mechanism, high-fidelity annual performance indicators from the lower-level operation directly guide the iterative optimization of upper-level planning. Simulation results on the modified IEEE 30-bus system show that the closed-loop coordination mechanism guides planning to configure Energy Storage Systems with higher energy-to-power ratios (3.0&#xa0;h vs. 1.5&#xa0;h in conventional methods), fundamentally transforming energy storage from a “power buffer” into an efficient “energy time-shifter.” This planning-driven structural optimization reduces the average daily operating cost by 24.6% compared with traditional decoupled methods and decreases the annual comprehensive curtailment rate from 9.9% to 3.4%. Importantly, the framework quantifies the pure gains of planning-operation integration: even compared with decoupled schemes using the same advanced scheduling, the integrated framework further lowers operating cost by 20.4%. Furthermore, when benchmarked against state-of-the-art integrated methods such as two-stage stochastic programming and end-to-end deep reinforcement learning, the proposed framework demonstrates superior performance by reducing daily operating costs by 13.9% and 15.7% respectively, while ensuring system safety and constraint satisfaction. When facing high-frequency stochastic disturbances, the maximum system frequency deviation is strictly controlled within ± 0.05&#xa0;Hz. This study shows that the proposed closed-loop coordination is a core mechanism for unlocking the potential of advanced control strategies and shifting the economic role of flexibility resources, providing an effective and robust paradigm for the full-life-cycle optimization of large-scale renewable energy.</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

An integrated optimization framework unlocks energy storage economic value in renewable energy bases through planning operation coordination

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年12月2日

查看全文
  • Ling Ji,
  • Yanjun Jiang,
  • Shunguo Ji,
  • Kai Chen

摘要

This paper proposes an integrated hierarchical coordination framework for planning and operations to address the decoupling between long-term capacity planning and short-term operational scheduling in renewable energy systems and the resulting economic and reliability losses. The core innovation lies in its “planning-operation” closed-loop feedback system. The upper-level planning adopts the non-dominated sorting genetic algorithm III (NSGA-III) to optimize the capacity mix of wind, solar, and energy storage, balancing investment economy and system reliability. The lower-level operation introduces an adaptive distributed model predictive control (DMPC) strategy empowered by deep deterministic policy gradient (DDPG), enabling efficient response to real-time uncertainty. Through a multi-stage elite feedback mechanism, high-fidelity annual performance indicators from the lower-level operation directly guide the iterative optimization of upper-level planning. Simulation results on the modified IEEE 30-bus system show that the closed-loop coordination mechanism guides planning to configure Energy Storage Systems with higher energy-to-power ratios (3.0 h vs. 1.5 h in conventional methods), fundamentally transforming energy storage from a “power buffer” into an efficient “energy time-shifter.” This planning-driven structural optimization reduces the average daily operating cost by 24.6% compared with traditional decoupled methods and decreases the annual comprehensive curtailment rate from 9.9% to 3.4%. Importantly, the framework quantifies the pure gains of planning-operation integration: even compared with decoupled schemes using the same advanced scheduling, the integrated framework further lowers operating cost by 20.4%. Furthermore, when benchmarked against state-of-the-art integrated methods such as two-stage stochastic programming and end-to-end deep reinforcement learning, the proposed framework demonstrates superior performance by reducing daily operating costs by 13.9% and 15.7% respectively, while ensuring system safety and constraint satisfaction. When facing high-frequency stochastic disturbances, the maximum system frequency deviation is strictly controlled within ± 0.05 Hz. This study shows that the proposed closed-loop coordination is a core mechanism for unlocking the potential of advanced control strategies and shifting the economic role of flexibility resources, providing an effective and robust paradigm for the full-life-cycle optimization of large-scale renewable energy.