城市电网网架结构与分区方式的两层多目标联合规划

United Urban Power Grid Planning for Network Structure and Partition Scheme Based on Bi-level Multi-objective Optimization With Genetic Algorithm

提出城市电网网架结构与分区方式的联合规划方法。首先提出城市电网分区方法,进而构建联合规划的数学模型,并使用多目标遗传算法求解。设定模型求解的前提条件,将两层规划模型的求解转化为相应的单层多目标优化问题的求解。针对联合编码策略所引起的孤岛问题提出连通网架形成方法。通过适应度计算、精英保存、分层进化等环节的设计,使多目标遗传算法在求解单层多目标优化问题过程中逐渐形成接近分层多目标优化问题有效解集的一组解。对实际系统的应用结果表明该文方法直接、有效。

A united planning method for urban power network planning and partition scheme optimization was proposed. The partition method was proposed first, and the bi-level multi-objective united planning model of network and partition scheme was constructed, which was solved by multi-objective genetic algorithm （MOGA）. Precondition of the solution was assumed, and the bi-level multi-objective model was transferred to a corresponding single level multi-objective model. To solve the isolated island problem caused by united coding strategy, a connective network formulation method was proposed. Key steps of MOGA such as fitness function, elite saving and stratified evolution were designed, and the efficient solution set of bi-level multi-objective problem was formed with the solution of the corresponding single level problem. The model and algorithm are intensively tested in real power system, proving their potential in practical applications.