【作者】 蒋周娜；

【导师】 张玲霞；

【作者基本信息】 西安电子科技大学 ， 测试计量技术及仪器， 2010， 硕士

【摘要】 电子设备的故障大多出自模拟电路。随着电子工业迅速发展,电子设备的结构日益复杂,对模拟电路故障诊断的要求不断提高,对模拟电路故障诊断方法的研究也更具现实意义。因其自身存在许多困难,传统的故障诊断方法很难达到预期的效果,具有非线性映射特性的神经网络为模拟电路故障诊断提供了一条有效途径,本文研究了将小波分析与神经网络结合的模拟电路故障诊断方法。本文首先在分析模拟电路故障诊断基本思想的基础上,阐述了神经网络故障诊断的原理,探讨了用于故障诊断的BP网络的设计方法。然后,研究了将小波分析与神经网络结合用于模拟电路故障诊断的方法,松散型的结合通过小波分解输出信号提取各频段的能量作为故障特征输入到神经网络训练,紧致型的结合则将神经网络中的激励函数用特性良好的小波函数代替,通过对带通滤波器电路诊断的实例证明了小波神经网络方法的有效性、优越性。最后将虚拟仪器技术引入到模拟电路故障诊断领域,以负反馈放大器电路为例,设计完成基于LabVIEW的故障诊断系统,可实现故障的诊断和故障类型的显示,界面直观。总之,基于小波神经网络的方法对解决模拟电路故障诊断问题是切实有效的,将虚拟仪器技术应用到模拟电路故障诊断系统中,有很好的工程应用前景。更多还原

【Abstract】 Most of the electronic equipment failure is caused by analog circuit fault. with the development of electronic industry, requirements on analog circuit fault diagnosis become higher, so research of methods of analog circuit fault diagnosis is significative. For itself with many difficulties existing, traditional methods of fault diagnosis are hard to achieve the desired results. Neural Networks,which have nonlinear mapping and generalizing ability, provide a powerful way to diagnose faults of analog circuits, morever, methods of combining Wavelet Transform and Neural Network can better deal with some critical problems in analog circuit fault diagnosis.Firstly, this paper illustrates the principle of Neural Network fault diagnosis after understanding basic analog circuit fault diagnosis theories, elaborates approaches of designing BP Network for fault diagnosis. Secondly, this paper researches methods of conbining Wavelet Transform with Neural Network for analog circuit fault diagnosis. for the loose method, using Wavelet decomposition to process the impulse response drastically reduce the number of input fed to the Neural Network, for the compactive method, Wavelet function is used as neural network’s inspirit function because of it’s good performance, and simulation result indicates that the two methods are feasible and effective by diagnosing a Band-pass filter. Finally, virtual instrument technology is introduced in the area of analog circuit fault diagnosis, LabVIEW procedures are designed to accomplish a fault diagnosis system for an example of negative feedback amplifier circuit, the system can diagnose fault and show fault types intuitively.In a word, the Wavelet Neural Network based approches for fault diagnosis of analog circuits are practical and effective, application of virtual instrumentation to the analog circuit fault diagnosis system, have a good engineering application prospect.更多还原