摘要
为了更有效地对数字电路进行多故障诊断,提出了一种利用电路的故障真值表来表征电路在有故障和无故障状态下的特征信息来对电路进行故障诊断的新方法。该方法使用BP型神经网络,运用误差反向传播算法,把从电路中提取的有效特征信息作为样本对网络反复训练,从而实现用单故障对多故障进行诊断的目的。实验表明该方法可以有效、方便地实现电路的故障检测和定位,为多故障诊断的研究提供了一种新思路。
For diagnosing the multiple-fault more effectively, this paper provides a new method for digital circuit fault diagnosis by using the fault truth table to token the circuit information in the state of fault and free-fault. The method uses the model of BP network and error back propagation algorithm, using the valuable information extracted from the circuit as the sample to study the model, to achieve the intention of fault diagnosis from the single-fault to multiple-fault. We can draw a conclusion from the experiment that this method is not only effective but also convenient for the fault detection and location and provide a new thought for the study of multiple-fault diagnosis in the future.
出处
《电子器件》
EI
CAS
2006年第2期490-492,共3页
Chinese Journal of Electron Devices
关键词
BP网络
FTT
多故障诊断
back propagation algorithm
a fault truth table （FTT）
multiple-fault diagnosis
