基于集群风电虚拟机组的源端系统分层协调日前计划

Virtual Power Generator Based Day-ahead Hierarchically Coordinated Generation Schedule for Source Side System

在高风电渗透率系统中,单个风电场以"负负荷"形式参与日前计划的传统模式未能充分发挥风电的空间平滑性特性,可调度性较低,而且风电场数量众多也增加了日前计划模型复杂度。文中提出将集群风电作为虚拟机组参与分层协调日前计划的新思路,集群风电虚拟机组(WVPG)对调度中心以"准常规电源"形式参与日前计划,对集群内部则实现多风电场互济运行。首先介绍了基于WVPG的日前分层协调发电计划的总体框架;然后分别介绍了WVPG内部多风电场、WVPG与火电机组的日前协调发电计划优化模型。仿真算例分析表明,分层协调日前发电计划模型比单个风电场直接参与日前计划策略更佳,可提高含风电系统日前计划的安全性和经济性指标;与传统的简单区间和场景模型相比,考虑WVPG安全性和经济性综合风险的概率优化模型可降低风电日前计划的安全性和经济性风险。

In the source side power system with high wind power penetration, taking wind power as negative load in the conventional day-ahead generation schedule cannot make full use of wind power＇ s spatial smoothing effect and result in low dispatch capability. In addition, the number of wind farms also increases the complexity of day-ahead probabilistic optimization model. This paper proposes a new idea of wind farm cluster virtual power generator （WVPG） based hierarchically coordinated day-ahead generation schedule strategy. For the dispatch center, a WVPG participates in the day-ahead generation schedule as quasi-regular power source. For the member wind farms, WVPG achieves the coordinated operation of multiple wind farms. Firstly, hierarchically coordinated day-ahead generation schedule framework is proposed. Then, a generation schedule optimal model for multiple wind farms inside a WVPG is proposed and the objective for this model is to minimize the overall secure risk of these wind farms. The generation schedule optimal model for WVPGs and coal-fired units is proposed and the objective for this model is to minimize the comprehensive secure and economic risks of WVPGs and the cost for coal-fired units. Simulation results show that this method presents better performance than treating each wind farm in the day-ahead schedule separately, which improves the secure and economic indices for day ahead schedule. In addition, the strategy proposed here can reduce the comprehensive secure and economic risk of wind power comparing to the simple interval and scenario models.