基于CMAC神经网络的配网重构模型

为使配电网的有功功率损失最小化,提出了一种基于小脑模型关节控制器(cerebellar model articulation controller,CMAC)神经网络配电网重构模型.借助于CMAC神经网络输入和输出之间的非线性映射关系和泛化能力,来建立变化的负荷水平与最优化网络拓扑之间的对应关系,即网络重构.还将该模型与基于BP网络的配网重构模型进行比较.经算例表明,模型可以快速地给出重构的结果,适合大型配电网使用.

Model of distribution power network reconfiguration based on CMAC neural network

The purpose of network reconfiguration is to minimize the active power losses in distribution power networks. Based on a CMAC neural network, a network reconfiguration model was proposed. By mapping the complex nonlinear relationship between the inputs and outputs and using the generalization capability, the model establishes a relationship between the varying level of loads and the optimal network topology very fast, which is suitable for large-scale distribution power networks. The method is also compared with that of BP neural network.