基于粗糙集和IRBF的配电网故障诊断系统

摘要

In order to improve artificial intelligence methods in fault diagnosis system of the distribution network are given RBF neural network model structure based on rough set theory, and then use the trained neural network for the distribution network fault diagnosis. Using Visual C+ + language development tools, called Matlab neural network toolbox to establish a simplified fault diagnosis system, and through examples to verify the method. It has proved that not only improve the distribution network fault diagnosis and fault tolerance, fault diagnosis become more accurate and effective, but also reduce the neural network sample data, greatly reducing the time for fault diagnosis process.%为了改进人工智能方法在配电网故障诊断系统中的应用,给出了基于粗糙集理论的IRBF神经网络的模型结构,然后利用训练好的神经网络对配电网进行故障诊断;采用VC++语言开发工具,调用Matlab神经网络工具箱建立了一个简化的故障诊断系统,并通过配电网实例验证了方法的正确性;实践证明不但提高了配电网故障诊断的容错性,使故障诊断变得更加准确有效,而且减少了神经网络样本数据,大大地减少了故障诊断过程的时间.