基于小波神经网络和数据融合的直流系统故障检测方法及实现

Implementation of DC system fault detection method based on wavelet neural network and data fusion

针对直流供电系统接地故障,提出了一种将多尺度神经网络和数据融合技术相结合的故障检测方法,利用小波神经网络对来自直流系统的采样信号进行滤波,再应用数据融合技术对滤波后的信号进行分析处理以判断是否存在接地故障。该方法弥补了传统检测方法的缺陷,并可以实现利用微机装置在线检测。文章采用“PC机+数据采集卡”的形式实现了故障检测的硬件设计,并通过仿真分析和相关实验,验证了该方法应用于直流系统故障检测的可行性。

This paper presents a new fault detection method for the grounding fault of DC electrical source, which based on multi-resolution neural network （MRNN） and data fusion. MRNN is utilized to filter the signal sampled from DC system, and data fusion is employed to dispose the signal in order to judge whether grounding fault occurs. This method overcomes the limitations of existing methods, moreover, it can timely realize detection using computer control. PC and data collection cards are used to design the hardware of the system, simulation results are represented to show the feasibility of the algorithm in grounding fault detection.