电力电子电路智能故障诊断技术探讨

Discussion on Intelligent Fault Diagnosis Technology of Power Electronic Circuit

在人们科技水平快速发展的背景下,市面上各种新型电力电子产品不断问世。现代人对相应系统的各类品质需求随之拓展增加,导致电力电子电路故障诊断的智能化发展成为相关领域亟待解决的重要课题。因此,以该问题为导向,通过Matlab高效仿真计算机软件构建仿真模型,获得输出型电压ud,并用傅里叶分析方式提取基波幅值、直流分量、三次和二次波幅值等。将这些数据信息进行一体化处理分析,输入于BP神经网络整体结构中,获得拥有编码性质特征的6个基本数字,继而确立故障具体发生部位以及故障基本发生点位。同时,运用这个方式,将三相桥型整流电路作为仿真实验案例,具体测试误差能够降低至10-4。最终,通过该实验充分证明,此方式和其他传统诊断方式相比较,有着更加高效的诊断效率,并且能够在全面提升电力电子电路故障诊断效率的同时,提升诊断质量,是一项诊断水平高、稳定性强、可靠性高、应用范围广阔以及未来发展前景良好的电力电子电路诊断技术。

In the background of the rapid development of human science and technology,various new power electronics factory products on the market come out continuously. Modern people’s demand for various quality of the corresponding system expands and increases,which makes the intelligent development of power electronic circuit fault diagnosis become an important subject to be solved urgently in related fields. In this paper,based on this problem,a simulation model is built by using Matlab high-efficiency software to obtain the output voltage ud,and the fundamental amplitude,dc component,cubic and quadratic amplitude are extracted by means of Fourier analysis. These data are processed and analyzed in an integrated way,and input into the overall structure of BP neural network. Six basic Numbers with coding characteristics are obtained,and then the specific location and basic location of fault are established. Meanwhile,by using the above methods,the three-phase bridge rectifier circuit is taken as a simulation experiment case,and the specific test error can be reduced to 10-4. Finally,fully proved by the experiment,this method compared with other traditional diagnostic methods,useful efficiency,more efficient diagnosis and to improve the efficiency of the power electronic circuit fault diagnosis at the same time,improving the quality of diagnosis,is a diagnostic level is high,strong stability,high reliability,wide in application and has good prospects for future development of power electronic circuit diagnosis technology.