针对极端气象条件下农村配电网运行不确定性增强以及调控能力不足的问题,对含分布式光伏与多类调节资源的配电网有功-无功协同优化调度进行了研究。根据历史负荷和光伏数据,使用K-means++算法聚类构建典型与极端气象场景,建立基于条件...针对极端气象条件下农村配电网运行不确定性增强以及调控能力不足的问题,对含分布式光伏与多类调节资源的配电网有功-无功协同优化调度进行了研究。根据历史负荷和光伏数据,使用K-means++算法聚类构建典型与极端气象场景,建立基于条件风险价值(conditional value at risk,CVaR)的多场景协同优化模型以计及电压约束、设备运行约束和功率平衡关系,结合集中训练分散执行框架,构造嵌入安全屏障与物理投影机制的安全多智能体近端策略优化(safe multi-agent proximal policy optimization,SMAPPO)算法。以改进IEEE 33节点系统为例进行了验证。结果表明,将CVaR纳入多场景优化目标后,尾部风险成本(CVaR值)降低7.2%,且极端场景下约束违规完全消除,增强了配电网在极端气象条件下的运行韧性与风险可控性。展开更多
为应对规模化新能源接入对电氢耦合系统运行的影响,解决传统两阶段鲁棒优化在不确定集选择上的主观性及新能源概率预测信息利用不足的问题,提出了面向新能源电网的两阶段风险鲁棒调度方法。首先,建立整合动态制氢-储氢-用氢一体化的新...为应对规模化新能源接入对电氢耦合系统运行的影响,解决传统两阶段鲁棒优化在不确定集选择上的主观性及新能源概率预测信息利用不足的问题,提出了面向新能源电网的两阶段风险鲁棒调度方法。首先,建立整合动态制氢-储氢-用氢一体化的新能源电网系统模型,并提出耦合热力学和气泡动力学的电解槽多物理场模型,精细刻画电解槽动态特性,提升电解效率随环境、运行状态等的表征精度;然后,构建了结合鲁棒优化与条件风险价值(conditional value at risk,CVaR)的新能源电网两阶段风险鲁棒发电-备用协调调度方法,第一阶段依据新能源预测确定储能、可控机组及制氢系统功率基点与备用容量,量化弃风及切负荷风险,第二阶段基于新能源实际功率与预留备用再调度,满足可消纳区间内所有场景需求;最后,采用列约束生成(column-and-con⁃straint generation,C&CG)算法求解优化模型,通过两阶段模型协同求解确定最优不确定集。算例分析验证了所提模型的有效性。展开更多
High renewable penetration improves the low-carbon performance of integrated energy systems,but it also increases scheduling uncertainty and renewable curtailment risk.This paper proposes a CVaR-based optimal scheduli...High renewable penetration improves the low-carbon performance of integrated energy systems,but it also increases scheduling uncertainty and renewable curtailment risk.This paper proposes a CVaR-based optimal scheduling model for an electric-heat-hydrogen integrated energy system with battery energy storage,hydrogen storage,and demand response.The proposed model minimizes a weighted objective that combines expected operating cost and tailrisk cost,while considering electricity purchase and sale,gas consumption,carbon emissions,battery degradation,hydrogen conversion,demand response compensation,and renewable curtailment penalty.Wind power,photovoltaic generation,and electric load uncertainty are represented by multiple scenarios,and the same scenario set is used for all comparative cases to ensure fairness.Five operation schemes are studied,including no storage,battery energy storage only,hydrogen storage only,battery-hydrogen storage,and the proposed battery-hydrogen-demand response scheme.The numerical results show that the proposed scheme achieves the lowest weighted objective,expected cost,and CVaR cost.Compared with the no-storage case,the proposed scheme reduces the weighted objective from 12337.22 to 9677.39,increases renewable utilization from 92.3%to 99.8%,and reduces expected carbon emissions from 4771.88 to 2994.13.These results indicate that coordinated scheduling of battery storage,hydrogen storage,and demand response can improve economic performance,reduce operational risk,and enhance renewable energy accommodation in highrenewable integrated energy systems.展开更多
随着大规模分散且多样化的分布式资源接入,虚拟电厂(virtual power plant,VPP)技术已成为有效管理和优化需求侧资源的重要工具。为使VPP更好地满足新型电力系统的发展需要,提出了一种考虑不确定性风险的VPP参与绿证-碳联合交易的优化调...随着大规模分散且多样化的分布式资源接入,虚拟电厂(virtual power plant,VPP)技术已成为有效管理和优化需求侧资源的重要工具。为使VPP更好地满足新型电力系统的发展需要,提出了一种考虑不确定性风险的VPP参与绿证-碳联合交易的优化调度模型。首先,构建了由风电机组、光伏机组、燃气轮机组、储能设备和用户侧柔性负荷组成的VPP优化运行模型,该模型以VPP运行成本最小为目标,并考虑了电市场、绿证-碳联合交易机制以及激励型需求响应。其次,综合考虑VPP中的源、荷以及需求响应等多重不确定性因素,运用条件风险价值(conditional value-at-risk ,CVaR)理论对不确定因素风险进行量化处理。最后,通过算例分析,验证了所提模型的经济性和环保性,所考虑的CVaR也为VPP利润与风险的平衡提供了有力的决策依据。展开更多
This paper analyzes the generalization of minimax regret optimization(MRO)under distribution shift.A new learning framework is proposed by injecting the measure of con-ditional value at risk(CVaR)into MRO,and its gene...This paper analyzes the generalization of minimax regret optimization(MRO)under distribution shift.A new learning framework is proposed by injecting the measure of con-ditional value at risk(CVaR)into MRO,and its generalization error bound is established through the lens of uniform convergence analysis.The CVaR-based MRO can achieve the polynomial decay rate on the excess risk,which extends the generalization analysis associated with the expected risk to the risk-averse case.展开更多
摘要针对极端气象条件下农村配电网运行不确定性增强以及调控能力不足的问题,对含分布式光伏与多类调节资源的配电网有功-无功协同优化调度进行了研究。根据历史负荷和光伏数据,使用K-means++算法聚类构建典型与极端气象场景,建立基于条件风险价值(conditional value at risk,CVaR)的多场景协同优化模型以计及电压约束、设备运行约束和功率平衡关系,结合集中训练分散执行框架,构造嵌入安全屏障与物理投影机制的安全多智能体近端策略优化(safe multi-agent proximal policy optimization,SMAPPO)算法。以改进IEEE 33节点系统为例进行了验证。结果表明,将CVaR纳入多场景优化目标后,尾部风险成本(CVaR值)降低7.2%,且极端场景下约束违规完全消除,增强了配电网在极端气象条件下的运行韧性与风险可控性。
摘要为应对规模化新能源接入对电氢耦合系统运行的影响,解决传统两阶段鲁棒优化在不确定集选择上的主观性及新能源概率预测信息利用不足的问题,提出了面向新能源电网的两阶段风险鲁棒调度方法。首先,建立整合动态制氢-储氢-用氢一体化的新能源电网系统模型,并提出耦合热力学和气泡动力学的电解槽多物理场模型,精细刻画电解槽动态特性,提升电解效率随环境、运行状态等的表征精度;然后,构建了结合鲁棒优化与条件风险价值(conditional value at risk,CVaR)的新能源电网两阶段风险鲁棒发电-备用协调调度方法,第一阶段依据新能源预测确定储能、可控机组及制氢系统功率基点与备用容量,量化弃风及切负荷风险,第二阶段基于新能源实际功率与预留备用再调度,满足可消纳区间内所有场景需求;最后,采用列约束生成(column-and-con⁃straint generation,C&CG)算法求解优化模型,通过两阶段模型协同求解确定最优不确定集。算例分析验证了所提模型的有效性。
摘要High renewable penetration improves the low-carbon performance of integrated energy systems,but it also increases scheduling uncertainty and renewable curtailment risk.This paper proposes a CVaR-based optimal scheduling model for an electric-heat-hydrogen integrated energy system with battery energy storage,hydrogen storage,and demand response.The proposed model minimizes a weighted objective that combines expected operating cost and tailrisk cost,while considering electricity purchase and sale,gas consumption,carbon emissions,battery degradation,hydrogen conversion,demand response compensation,and renewable curtailment penalty.Wind power,photovoltaic generation,and electric load uncertainty are represented by multiple scenarios,and the same scenario set is used for all comparative cases to ensure fairness.Five operation schemes are studied,including no storage,battery energy storage only,hydrogen storage only,battery-hydrogen storage,and the proposed battery-hydrogen-demand response scheme.The numerical results show that the proposed scheme achieves the lowest weighted objective,expected cost,and CVaR cost.Compared with the no-storage case,the proposed scheme reduces the weighted objective from 12337.22 to 9677.39,increases renewable utilization from 92.3%to 99.8%,and reduces expected carbon emissions from 4771.88 to 2994.13.These results indicate that coordinated scheduling of battery storage,hydrogen storage,and demand response can improve economic performance,reduce operational risk,and enhance renewable energy accommodation in highrenewable integrated energy systems.
摘要随着大规模分散且多样化的分布式资源接入,虚拟电厂(virtual power plant,VPP)技术已成为有效管理和优化需求侧资源的重要工具。为使VPP更好地满足新型电力系统的发展需要,提出了一种考虑不确定性风险的VPP参与绿证-碳联合交易的优化调度模型。首先,构建了由风电机组、光伏机组、燃气轮机组、储能设备和用户侧柔性负荷组成的VPP优化运行模型,该模型以VPP运行成本最小为目标,并考虑了电市场、绿证-碳联合交易机制以及激励型需求响应。其次,综合考虑VPP中的源、荷以及需求响应等多重不确定性因素,运用条件风险价值(conditional value-at-risk ,CVaR)理论对不确定因素风险进行量化处理。最后,通过算例分析,验证了所提模型的经济性和环保性,所考虑的CVaR也为VPP利润与风险的平衡提供了有力的决策依据。
基金Supported by Education Science Planning Project of Hubei Province(2020GB198)Natural Science Foundation of Hubei Province(2023AFB523).
摘要This paper analyzes the generalization of minimax regret optimization(MRO)under distribution shift.A new learning framework is proposed by injecting the measure of con-ditional value at risk(CVaR)into MRO,and its generalization error bound is established through the lens of uniform convergence analysis.The CVaR-based MRO can achieve the polynomial decay rate on the excess risk,which extends the generalization analysis associated with the expected risk to the risk-averse case.