一种基于ANFIS的配电网故障分类方法及其适应性分析

An ANFIS based fault classification in distribution network and its adaptability analysis

基于自适应神经模糊推理系统,利用配电网系统故障后产生的丰富的暂态信号,设计了一种用于谐振接地配电系统故障分类方法。在PSCAD/EMTDC中建立了仿真模型,分别仿真研究了该方法在不同信噪比、电弧故障、不同负荷水平以及不同系统等效阻抗四种工况下的适应性。通过仿真结果得出,在电弧故障和不同系统等效阻抗两种工况下,分类方法具有较强的适应性;而在信噪比较低以及负荷加重工况下,分类的准确性和过渡电阻的大小相关。提出应增加滤波器环节和增加重负荷工况下的训练样本以提高分类方法的准确性。

Based on Adaptive Neural Fuzzy Inference System（ANFIS）,the fault classification technique for neutral resonance grounding distribution network which utilizes abundant transient signals after fault occurs at distribution network is proposed.The simulation model is established in PSCAD/EMTDC environment.Through simulation,the adaptability of proposed technique to noises,arc faults,different load levels and different source equivalent impedances is studied respectively.From the simulation results, it can be concluded that the adaptability of the proposed technique to arc faults and different source equivalent impedances is good; but in the case of low signal noise ratio and heavy load working condition,the accuracy of classification relates with the magnitude of transition resistance.To enhance the accuracy of proposed technique,it is proposed to add filter and train samples in heavy load situation.