基于集团法和自组织神经网络的容差电路故障诊断(英文)

Fault diagnosis of circuits with tolerance based on body and KNN method

提出了一种基于集团法和自组织神经网络的容差电路故障诊断方法。按照给定的撕裂原则将大规模电子电路撕裂成子电路后,运用集团法诊断出含有故障的子电路,并将它们的测量数据输入到自组织神经网络,神经网络的输出就是故障元件。文章详述了故障诊断的原理和步骤,并引实例结果论证其诊断方法。

A new method for fault diagnosis of large-scale analogue circuits with tolerance based on body algorithm and Kohonen neural networks(KNN) is presented in the paper. The large-scale circuits are decomposed to several sub-networks according to specific rules. The faulty blocks are detected based on body method. The faulty components are illustrated by applying the experimental data to KNNs. Diagnosis principles and steps are described. The validity of the method is confirmed by practical examples.