基于正交试验设计的配电网自动化终端优化配置

Optimal Configuration of Distribution Network Automation Terminals Based on Orthogonal Experimental Design

中文关键词:配电网  自动化终端  优化配置  正交试验设计

keywords:distribution network  automation terminals  configuration optimization  orthogonal experimental design

中文摘要:针对元启发式优化算法普遍存在计算耗时久、易陷入局部最优、计算结果稳定性差等缺陷，求解过程易受初始点设置不合理与迭代过程随机化双重影响的问题，提出一种基于正交试验设计（orthogonal experimental design，OED）的配网自动化终端优化配置方法。首先，以等年值综合费用最小为目标函数，综合考量终端初期投资、年度运行维护及故障停电损失3类费用，并将平均供电可用率作为约束条件。其次，OED 通过正交表筛选试验组合，利用因素分析确定最优组合方案；当最优组合不满足约束条件时，依托极差分析选取次优解，整个求解过程无需迭代，计算量小且结果稳定。最后，基于IEEE RBTS BUS5系统与某县域10 kV 配网馈线的算例进行仿真分析。仿真结果表明，所提方法在计算成功率与计算效率上均优于4种元启发式算法，且求解结果唯一、稳定性强，为配电网自动化终端的优化配置提供了一种新的有效方法。

Abstract:Aiming at the common shortcomings of metaheuristic optimization algorithms，such as long computation time，easy convergence to local optima and poor stability of calculation results，as well as the problem that the solution process is dually affected by unreasonable initial point settings and randomness in the iterative process，this paper proposes an optimal configuration method for distribution network automation terminals based on orthogonal experimental design （OED）. Firstly，with the minimum annual equivalent comprehensive cost as the objective function，three types of costs—including the initial terminal investment cost，annual operation and maintenance cost，and fault outage loss cost—are comprehensively considered，and the average service availability index （ASAI） is set as the constraint condition. Secondly，OED screens test combinations via orthogonal tables （OT） and determines the optimal combination scheme through factor analysis （FA）； if the optimal combination fails to satisfy the constraint conditions，the suboptimal solution is selected based on range analysis. The entire solution process involves no iteration，featuring a low computational load and stable results. Finally，simulation analyses are conducted based on the IEEE RBTS BUS5 system and a 10 kV distribution network feeder in a certain county. The simulation results show that the proposed method outperforms four typical metaheuristic algorithms in both calculation success rate and computational efficiency，yielding unique and highly stable solution results and thus providing a new and effective method for the optimal configuration of distribution network automation terminals.