虚拟电厂（VPP）在新型电力系统中的地位日益凸显，而可聚合分布式能源（DER）的VPP内部风电机组、光伏发电机组的随机性和波动性负荷等不确定因素给VPP的运营决策和收益带来风险。如何建立公平、合理、透明的利益分配机制是成功维护 VPP 内DER合作关系的关键。为了激励VPP内部各主体参与市场的积极性，各主体可根据自身特点与贡献度协商分配收益，构建计及风险偏好的包含风电机组、光伏发电、可控分布式电源、储能及柔性负荷的多主体VPP内部优化调度决策模型，参与电力市场，并建立Nash-Harsanyi讨价还价解的VPP利益分配模型。算例分析以VPP最大化运行效益为目标，指导系统运营策略和降低VPP运行风险水平，验证了所提利益分配法能有效平衡各方利益，量化各DER对VPP利润的实际贡献度，提高各成员参与市场竞争的积极性。采用Nash-Harsanyi解的分配方案在VPP收益分配合理性和适用性方面优于Shapley值法方案。

Virtual power plants （VPPs） are playing an increasingly prominent role in new power systems， while the operational strategies and revenues of a VPP with distributed energy resources（DERs）are affected by the fluctuated outputs of wind turbines and PV generators in the plant. A fair， reasonable and transparent benefit distribution mechanism is the key to maintaining the cooperation between different DERs in a VPP. To motive various entities in a VPP to participate in trades， the profit is allocated among different entities based on their own characteristics and contributions though consultations. An optimal scheduling and profit distribution strategy is proposed for the VPP integrating wind turbines， PV generators， controllable distributed power suppliers and flexible loads considering their risk preferences. The multi-agent VPP participates in the electricity market （EM），and its profit distribution model is built based on Nash-Hassanyi bargaining solution. A numerical analysis is carried out，aiming to maximize the operational efficiency， guide the operation and reduce the operation risk of the VPP. The results verify that the proposed benefit distribution method can effectively balance the interests of all parties， quantify the actual economic contribution of each DER to the VPP， and improve the willingness of each entity to participate in market competition. The benefit allocation mechanism using Nash-Hassanyi solution is superior to that using Shapley value method in terms of allocation rationality and applicability.

Distributed power supplies have gradually become a new trend in power supply development, but access to a large number of distributed energy sources has a certain impact on the stable operation of the power grid. A virtual power plant (VPP) can integrate a variety of distributed power sources for coordination and optimization; thus, it can effectively solve the difficulties faced by a distributed energy grid connection and promote the complementarity of energy sources. However, renewable energy often has a degree of volatility and randomness when distributed, which can bring certain risks to the operation of the VPP. In order to consider the risks brought by renewable energy, an optimal scheduling model of the VPP, based on an improved generative adversarial network (GAN) and the conditional value at risk (CVaR), was proposed to measure the relationship between the benefits and risks. Firstly, the uncertainty of new energy is analyzed, and wind power and photovoltaic scenarios are generated by the improved GAN; then, typical scenarios are generated by the k-medoids method. Finally, based on the CVaR, the optimal scheduling model of the VPP is established to study the effect of risk weight on VPP revenue. The results show that the model can effectively measure the relationship between the benefits and risks and can provide some references for the VPP to make reasonable operational decisions with different risk preferences.