基于BP网络的开关电源故障诊断方法研究  

Research of fault diagnosis method of switch power supply based on BP neural network

作　　者：谢香峰[1] 雷电[1] 孙承波[1] 

机构地区：[1]上海大学机电工程与自动化学院,上海200072

出　　处：《电子测量技术》2012年第8期11-16,共6页Electronic Measurement Technology

摘　　要：分析了开关电源故障诊断方法研究的现实意义,以BP神经网络、数据模糊化和功能测试理论为基础,针对开关电源电路特征提出了采用功能BP网络与子模块BP网络相结合并行处理方法,首先避免了建立庞大的BP网络,其次解决了大规模模拟电路功能模块串行处理效率低的问题。测试和仿真结果表明,所提出的方法能够有效地诊断模拟电路故障模块和故障元件。This paper analyzes the practical significance of fault diagnosis method researching of switching power supply. For switching power supply circuit characteristics, the combination parallel processing method of functional BP neural network and submodule BP network base on BP neural network data fuzzy and functional test theory is presented in this paper. Firstly, the establishment of large BP neural network is avoided in this method. Secondly, it is used to solve the low efficiency of serial processing function modules of large-scale analog circuit. The test and simulation results show that the proposed method can effectively diagnose faults of analog circuit functional modules and components.

关 键 词：开关电源 故障诊断 BP神经网络 数据模糊化 功能测试 

分 类 号：TN277[电子电信—物理电子学]