分布式模型预测电压控制方法研究

针对集中式模型预测控制应用于大型电力系统长期电压稳定控制中存在滚动优化计算量大的问题,提出了一种分布式模型预测电压控制算法,即将一个复杂大规模系统分解为若干区,从而将求解全网大型在线滚动优化问题分解为求解各个区的在线滚动优化问题,并通过通信网络实现各个区之间信息交换,降低问题的规模和复杂性。分别采用集中式和分布式模型预测控制方法,对New England 10机39节点设计电压控制器,通过仿真和对比分析证实分布式模型预测是可行的且在计算速度方面具有优势。

Research on distributed model predictive voltage control method

When the centralized model predictive control method is applied to long-term voltage stability of large-scale power systems,the solution of receding optimization problem needs huge computational effort.Thus,this paper proposes a distributed model predictive voltage control method,namely,a complex large-scale systems is decomposed into several subareas,and solution of the whole on-line receding optimization problem is decomposed to solve the receding optimization problem in each subarea through communication network for information exchange among these subareas,hence the size and complexity of the problem are reduced obviously.Furthermore,centralized and distributed model predictive voltage controllers are used respectively on New England 10-machine 39-bus system.Results obtained from simulation and comparative analysis demonstrate the feasibility and superiority in computing speed of distributed model control method.