现代模拟电路智能故障诊断方法研究与发展  

Advance in Modern Analog Circuit Intelligent Fault Diagnosis Methods

作　　者：郭珂[1] 伞冶[1] 朱奕[1] 

机构地区：[1]哈尔滨工业大学,黑龙江哈尔滨150001

出　　处：《电子设计工程》2012年第2期177-180,共4页Electronic Design Engineering

基　　金：国家自然科学基金资助项目(61074127)

摘　　要：对系统可靠性和经济性要求的提高使得模拟电路故障诊断的重要性日益凸显。首先在介绍了模拟电路故障原因及分类的基础上,详细分析了模拟电路故障诊断的特点。针对传统诊断方法的不足之处,介绍了基于人工智能和现代信息信号处理的现代故障诊断方法,包括专家系统诊断方法、神经网络诊断方法、模糊诊断方法和基于核的诊断方法,同时系统地分析了每种方法的基本原理、优缺点、研究进展和典型应用。最后探讨了目前模拟电路故障诊断研究存在的问题和未来的发展方向。The reliability and economy requirements of electronic system are getting higher and higher,which make analog circuit fault diagnosis become more and more important.Based on the introduction of the cause and classification for analog circuit faults,the characteristics of analog circuit faults were analyzed in detail.According to the deficiency of traditional methods,modern diagnosis methods based on artificial intelligence and modern information processing were introduced,including expert system diagnosis methods,neural networks diagnosis methods,fuzzy fault diagnosis methods and kernel-based methods.The principle,advantages and disadvantages,research advance and representative application of each method are introduced systematically at the same time.The issues which modern diagnosis methods are confronted and the future research directions are discussed at last.

关 键 词：模拟电路 故障诊断 人工智能 机器学习 核方法 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]