Based on random set and conditional evidence theory, this paper uses a method which reflects the influences of prior knowledge to realize accurate positioning of the analog circuit fault. Firstly, the fault evidence probability assignments based on the fault eigenvalues which obtained by processing the measurement data is gotten, and the evidence is changed into the form of random set. Then the evidence is used to combine by the conditional evidence theory. Finally, the concordance between the measurement evidence and prior knowledge is calculated, and the decision based on diagnostic decision rules is made. This method makes full use of all kinds of information, so it can improve the reliability of fault diagnosis and have a relatively high rate of fault components positioning. The second order filter circuit with tolerance is taken as an example, and the specific diagnosis methods and procedures are given, the results prove the effectiveness of the method in analog circuit fault diagnosis.%为实现模拟电路故障的准确定位,采用一种基于随机集与条件证据的融合先验知识的模拟电路故障诊断方法；首先对测量数据进行处理获取故障电路特征值,得到故障证据可信度分配,并将其表示为证据的随机集形式,然后利用条件证据理论对证据进行组合,最后计算组合后的证据与先验信息之间的和谐度,并依据决策规则得到诊断决策；这种方法充分利用了各种信息,提高了故障诊断的可靠性,故障元器件定位率也比较高；以含容差的二阶滤波电路为例,对电路元件单软、硬故障进行诊断,给出了具体诊断方法及步骤,诊断正确率达到95％以上,结果证明了该诊断方法的有效性；另外,该方法的诊断步骤相对程式化,便于计算机实现,可以实现模拟电路的在线诊断.
展开▼
