电工技术学报2024，Vol.39Issue(16)：5060-5074,15.DOI:10.19595/j.cnki.1000-6753.tces.230874
奖励机制与用户意愿结合的高峰期负荷博弈调度策略
Peak Load Game Scheduling Strategy Combining Reward Mechanism and User Willingness
摘要
Abstract
With the substantial growth in the number of cooling/heating loads,load spikes are easily to occure during the concentrated residential power consumption hours,especially in old neighborhoods with high load rate,which affects the safe and stable operation of low-voltage distribution networks.To supplement the supply-side regulation capability and enhance the flexibility of regulation,there is an urgent need to guide the user-side adjustable resources to participate in the interaction between grid supply and demand.This paper proposes a peak load game scheduling strategy,which combines the reward mechanism with the users'willingness to ensure the benefits of users and load aggregators,and complies with the users'regulatory willingness to achieve the ideal load fluctuation smoothing and peak shaving effect at the same time. Firstly,based on the users'willingness of equipment regulation,the peak-hour loads are divided dynamically,and a differentiated subsidy mechanism based on the transformer load rate-reward function is formulated.The proposed mechanism fully stimulates the users'response motivation and stabilizes the load rate within a reasonable interval.Secondly,the Stackelberg game theory is introduced to solve the energy-use decision-making problem between load groups,where the load group that needs to boost power during peak hours is regarded as the leader,and the load group with flexible cutting ability is regarded as the follower.The Stackelberg game model is established along with the uniqueness of the game equilibrium proved.Then,the intraday optimal scheduling method of peak loads under the Stackelberg game is proposed,which searches for the optimal energy-use strategies for the main subjects of the game while guaranteeing their benefits.Finally,the multi-channel mixture-of-experts(MCME)network is constructed to solve the regulatory willingness at the equipment level,and a joint control strategy of single/multi-power-stage loads based on the users'willingness is proposed. Simulation results show that by using the proposed load intraday optimization method,the standard deviation,peak-to-valley difference,and load peak all reach their minimum values,which are 33.13％,35.31％and 8.25％lower than the pre-optimization curves.Both the user-side benefit and the net gain of the load aggregator reach their highest values,which verifies the advantages of the proposed strategy in peak shaving,load fluctuation smoothing and participants'gain guaranteeing.When applying the MCME model to solve the equipment regulatory willingness,both the mean absolute error and root mean square error reach their minimum values,indicating that the MCME model has a high solving accuracy.The optimal power change obtained by the intraday optimization method is very close to that obtained by the real-time control strategy,which verifies the effectiveness of the joint control strategy of single/multi-power-stage loads. The simulation analysis draws the following conclusions:(1)The load intraday optimization method based on the reward mechanism can achieve a win-win situation for load aggregators and users,and achieve significant peak shaving and load fluctuation smoothing effects.(2)The MCME model can effectively learn the common and characteristic laws from operation data,and accurately solve the users'willingness to regulate the equipment.(3)The joint control strategy of single/multi-power-stage loads,which considers the users'willingness,can realize fine load regulation.关键词
主从博弈/高峰期负荷/用户意愿/负载率-奖励函数/混合专家网络Key words
Stackelberg game/peak load/user willingness/load rate-reward function/mixture-of-experts network分类
信息技术与安全科学引用本文复制引用
杨雪莹,祁琪,李启明,杨春萍,祁兵..奖励机制与用户意愿结合的高峰期负荷博弈调度策略[J].电工技术学报,2024,39(16):5060-5074,15.基金项目
国家电网有限公司科技项目资助(5108-202218280A-2-379-XG). （5108-202218280A-2-379-XG）
