基于增广Lagrange-Hopfield神经网络的分布式电源最优配置的研究

Research on Distributed Power Source Optimum Location Based on Augmented Lagrangian-hopfield Neural Network

为了解决配电网中分布式电源的最优配置问题,首先构建了以总体投资费用和购电费用为主体的目标函数,并且建立了配电网潮流、节点电压、电流和有功功率的约束方程,提出了一种基于增广拉格朗日松弛法作为其能量函数的连续型Hopfield神经网络来计算配电网中配置分布式电源的位置和容量。最后通过IEEE33配电网系统,分析了在目标函数中不同的投资费用和购电费用权重时,分布式电源的配置位置对系统网络损耗和有功损耗影响,验证该算法具有可行性、操作简单和计算速度快的特点。

In order to solve optimum location problem of distributed power source, an objective function combined total investment cost with power purchase cost was built firstly. Then established constraint equations of distribution system power flow, nodes voltage and real power, and proposed continuous hopfield neural network which used augmented lagrangian relaxation as its energy function to compute the location and capacity of distributed source in the distributed power network. In the end, making use of IEEE 33 standard distribution systems to analyze different weights of investment cost and power purchase cost, found out different locations had effects on power loss in the systems and real power loss. Verified the presented algorithm had the properties of teasibility, simplicity processing and fast computing speed.