基于神经网络的某装备电路单元故障诊断研究

某装备是一种较为先进的防空武器系统.该系统结构和电路复杂,元器件种类与品种多,技术保障中其性能检测与故障诊断的难度大.为此,针对系统电路单元功能模块的组成情况,进行了单元电路模块化诊断方法研究.以该装备某单元电路板为例,介绍了利用神经网络和模式识别相结合,实现功能模块的故障判断并对故障元器件定位的方法.结果表明:这种方法能够大大降低测前仿真的工作量,提高故障诊断的精度和速度,增强该武器装备的维修保障能力.     

基于仿真的电子元器件故障模型研究

  故障知识获取是基于知识的电路故障诊断方法的瓶颈问题,严重限制了电子设备智能诊断系统的发展.提出了基于仿真获得故障知识的方法,介绍了数字/模拟电子元器件的故障模式仿真模型的建立方法.采用FMECA分析法选择仿真对象,有效地减少了故障注入的次数.建立故障仿真模型库,降低了编写仿真文件的难度.通过数模混合电路验证了故障模型的有效性.     

基于PSpice的电子设备故障自动注入平台研究

针对目前故障仿真注入过程中存在着效率低、自动化程度低等问题,设计了基于PSpice仿真环境的故障自动注入平台.在该平台中对常见的十几种故障模式在PSpice中进行了建模,分别封装在3类故障库中,为故障的注入提供了便捷清晰的路径.并深入研究了PSpice仿真器的结构和电路描述语言,采用面向对象的思想编写了可对PSpice模拟电路进行分析的解析模块,实现将指定故障注入到目标电路的故障自动注人过程,在一定程度上改进了故障注入繁琐和不智能问题.经过实验验证了该设计的正确性与有效性.     

基于遗传BP网络的模拟电路故障诊断

提出了一种用故障的典型特征作为训练样本、遗传算法与BP网络相结合的模拟电路故障诊断新方法.该法采用故障的典型特征作为样本训练网络,遗传算法来优化BP网络的结构与初始权值分布的策略.详述了其诊断原理及诊断步骤,并给出了仿真实例及工程应用.结果表明,该方法具有特征提取简便、测前工作量小、故障诊断速度快及故障定位准确率高等特点.     

基于斜率特征的容差模拟电路故障诊断方法

针对基于斜率特征的模拟电路软故障诊断字典法中存在的容差问题,提出了基于蒙特卡罗(Monte Carlo,MC)仿真的测前仿真故障字典法,将电路容差允许范围的电路特征以及各元件发生故障的斜率特征建立统一的故障字典,用于测后故障检测及故障元件定位,不仅能判定待测电路参数是否在容差范围内,而且能诊断容差电路的软故障.论文给出了基于MC仿真的容差模拟电路故障字典建立的详细步骤,该容差处理方法仿真简单,易于实现,有利于基于斜率特征的故障字典法的进一步实用化.论文最后通过仿真实例验证了提出方法的有效性.     

基于灵敏度预分类的BP神经网络故障诊断方法

故障特征信息的获取和处理对电路故障的可靠分类和准确诊断有很大的影响．在电路故障诊断时,对于不同的故障模式,存在信息混叠的现象,需要解决特征信息的有效提取和故障的可靠分类等问题．为此,本文提出了一种结合灵敏度特性分析的BP神经网络故障诊断方法．基本思想是通过灵敏度的计算,对电路故障样本作预分类,再根据电路灵敏度的计算结果分别提取相应特征信息,以此构造故障样本特征集,然后作为BP神经网络的输入对网络训练,并进行故障诊断．对滤波器的仿真结果表明,该方法能分类不同的元件故障,且对模拟电路故障诊断的平均正确率优于传统方法．     

基于Matlab的变频器开路故障诊断方法

本文针对三相变频器典型拓扑电路,通过理论分析和仿真实验,着重对功率管开路故障诊断问题进行了研究.论文首先将变频器故障分为不可控整流桥故障、三相可控逆变电路的故障,及其两者组合故障进行分析;建立了变频器故障的Matlab模型加以仿真验证;提出了基于输出端电感滤波前相电压波形的开路故障诊断法,验证了该诊断方法不受负载突变的影响,并探讨了该方案在变频器智能诊断方向的应用前景.     

一种求解模拟电路测试向量的故障仿真方法

本文提出了一种模拟电路测试向量故障仿真方法。与传统的故障开关法相比 ,使用该方法求解模拟电路各种故障状态下的测试向量时 ,所需要的计算量大大减少     

一种新的模拟电路故障模糊集确定算法

提出了一种在模拟电路故障诊断中故障类模糊集的确定算法.该算法基于测前仿真法,不但同时适用于线性电路和非线性电路,而且可以确定元件模糊组以外的故障类模糊集.算法利用电路仿真故障特征数据计算故障的类间差异矩阵,从而完成故障类模糊集的确定.计算示例验证了该算法的有效性.     

基于LASAR的数字电路可测试性设计仿真

随着故障诊断技术的发展,利用专业的仿真工具对实际电路进行可测试性分析仿真用的越来越普遍。LASAR(逻辑自动激励与响应)就是一套优秀的用于数字电路测试开发和逻辑分析的仿真软件系统。介绍了利用LASAR故障仿真进行数字电路可测试性分析的方法。通过对一个实际电路进行仿真,具体说明了该方法在实际工程当中的应用。     

Analog fault diagnosis of actual circuits using neural networks

We have developed a neural-network based analog fault diagnostic system for actual circuits. Our system uses a data acquisition board to excite a circuit with an impulse and sample its output to collect training data for the neural network. The collected data is preprocessed by wavelet decomposition, normalization, and principal component analysis (PCA) to generate optimal features for training the neural network. This ensures a simple architecture for the neural network and minimizes the size of the training set required for its proper training. Our studies indicate that features extracted from actual circuits lie closer to each other and exhibit more overlap across fault classes compared to SPICE simulations. This implies that the neural network architecture which can most reliably perform fault diagnosis of actual circuits is one whose outputs estimate the probabilities that input features belong to different fault classes. Our work also shows that SPICE simulations can be used to select appropriate features for training the neural network. Reliable diagnosis of faults in an actual circuit, however, requires training data from the circuit itself. Our fault diagnostic system, trained and tested using data obtained from real sample circuits, achieves 95% accuracy in classifying faulty components     

基于改进的经验模态分解的滚动轴承故障诊断研究

传统的时频分析方法不能有效地处理非平稳信号,经验模态分解(EMD)非常适合处理非平稳信号,但结果可能出现伪内禀模态函数(IMF)和不敏感内禀模态函数。针对EMD的不足,提出能量门限法和敏感IMF选择法相结合来识别真IMF和敏感IMF的方法,对同时满足这两个条件的IMF作频谱变换,频谱图上可以清晰呈现故障特征信息。将改进后的EMD应用到滚动轴承故障诊断实例,证明了此方法的可行性和准确性。     

A method for analysing the reliability of a transmission grid

The paper describes a probabilistic method for transmission grid security evaluation. Power system security is the ability of the power system to withstand sudden disturbances such as short circuits. The method presented here uses event and fault trees and combines them with power system dynamic simulations. Event trees model the substation protection and trip operations after line faults. Different event tree end states (fault duration, circuit breaker trips) are simulated with power system dynamic analysis program. The dynamic analysis results (power system post-fault states) are then classified into secure, alert, emergency and system breakdown. The probabilities, minimal cut sets and grid level importance measures (Fussell-Vesely, risk increase and decrease factors) are calculated for the total and partial system breakdown. In this way, the relative importance of the substation devices regarding to the system breakdown can be reached. Also the more and less likely contributing factors to system breakdown are received. With this method, an existing 400 kV transmission grid with its line fault and device failure statistics is analysed.     

Analog system-level fault diagnosis based on a symbolic method inthe frequency domain

Concurrently, a symbolic approach for analog system-level fault. Diagnosis and a systematic approach to maximize the fault location capability are proposed. This unified approach is realized as a result of combining the simulation before test (SBT) fault dictionary diagnosis method with a symbolic approach. The traditional SBT fault dictionary method is often costly aid inefficient because of a high number of simulations, but it can become very efficient when a symbolic approach is employed. This symbolic approach only requires one analysis for circuit topology to generate the network transfer function and a parameter substitution to obtain the frequency response (or time response) of the system. An efficient program is developed to deal with the frequency responses of the system to provide the optimum testing point set, and to automatically generate the fault dictionary. The “distance” between the measurement data and the frequency responses from the fault dictionary is evaluated to determine the diagnosis results. A practical example is presented in order to illustrate the main features of this proposed analog system fault diagnosis approach     

Oscillation-test methodology for low-cost testing of active analogfilters

The oscillation-test strategy is a low cost and robust test method for mixed-signal integrated circuits. Being a vectorless test method, it allows one to eliminate the analog test vector generator. Furthermore, as the oscillation frequency is considered to be digital, it can be precisely analyzed using pure digital circuitry and can be easily interfaced to test techniques dedicated to the digital part of the circuit under test (CUT). This paper describes the design for testability (DFT) of active analog filters based on oscillation-test methodology. Active filters are transformed to oscillators using very simple techniques. The tolerance band of the oscillation frequency is determined by a Monte Carlo analysis taking into account the nominal tolerance of all circuit under test components. Discrete practical realizations and extensive simulations based on CMOS 1.2 μm technology parameters affirm that the test technique presented for active analog filters ensures high fault coverage and requires a negligible area overhead. Finally, the DFT techniques investigated are very suitable for automatic testable filter synthesis and can be easily integrated in the tools dedicated to automatic filter design     

自动舵电路板故障测试系统研究

针对自动舵故障电路板需要返厂测试的问题,设计自动舵电路板便携式故障测试系统。在分析自动舵电路板故障测试需求的基础上,采用专用接口电缆和测试软件而复用数据采集系统的方法,仅用少数硬件就实现了多块电路板的测试;基于神经网络方法设计自动测试软件,能够定位电路板常见故障;采用Measurement Studio和Visual Studio联合编程的方法,设计了辅助信息丰富的故障诊断程序界面。实践表明:该系统具有成本低、体积小、功能强的特点,可在船上直接对电路板进行故障测试。     

火控系统故障测试仪研究

设计了一种针对火控系统性能和故障检测的测试仪,检测方式上分为全系统级检测、分系统检测和板级检测三种检测模式;基于全系统检测主要用于性能检测和故障检测,可将故障定位到分系统;基于分系统检测主要用于故障检测和分系统性能检测,可将故障定位到板级,板级检测主要对电路板性能和故障进行检测,可将故障定位到元器件级．硬件设计采用PC总线系统为平台,在此基础上构建测试仪的不同测试模式的接口模块,包括数字-轴角转换卡、轴角-数字转换卡和激励火控计算机系统．以C＋＋Builder为测试仪的软件开发环境,检测软件采用与硬件相适应的模块化结构,并根据硬件模块配置变化,自动配置软件环境．     

基于ISSA⁃SVM的露点测量系统电路故障诊断方法研究

针对高精度谐振式露点测量系统中电路故障诊断问题,提出了一种基于改进的麻雀搜索算法（Improved Sparrow Search Algorithm, ISSA）优化智能分类器参数的电路故障诊断模型,采用测前仿真故障诊断方法中的智能诊断方法,选择适用于小样本、非线性问题的支持向量机（Support Vector Machine, SVM）作为智能分类器,针对麻雀搜索算法中收敛速度慢、易陷入局部最优等问题进行改进,并将改进后的优化算法用于SVM参数寻优,构建ISSA?SVM故障诊断模型用于谐振电路故障诊断。实验结果显示,ISSA?SVM模型在建立的电路上能够达到88.9%的故障诊断率,可靠性较强,能够作为高精度谐振式露点传感器电路的故障诊断方法。     

利用小波尺度谱同步平均的时频脊故障特征提取

旋转机械的早期故障特征微弱,容易受到噪声的干扰,不容易准确识别。小波尺度谱存在受噪声干扰影响大、高频部分频率分辨率低等缺点, 对小波尺度谱进行重排可以提高其时频聚集性。为此,结合小波尺度谱同步平均和小波脊线分析的优点,提出了基于时频脊线特征提取方法。首先对多周期的振动信号进行小波连续变换,并重排小波尺度谱;再根据信号的周期性,对尺度谱进行同步平均;最后提取同步平均后的尺度谱小波脊线,计算信号的包络幅值并进行频谱分析,最终提取出弱故障特征。通过仿真和实例验证了本方法的有效性,为旋转机械的早期故障诊断提供了新方法。