电视电路直流故障的BP网络快速诊断

Neural Network Approach for DC Fault Diagnosis in TV Circuits

测后诊断速度和诊断精度是电视直流故障诊断性能的重要衡量指标。文中将神经网络的自学习和分类技术应用于电视电路的直流故障诊断之中 ,并把反向传播BP网络训练成一个故障学习库。考虑到元件参数容差对诊断的影响 ,文中提出了优选训练样本的具体方法。此外 ,为实现快速诊断 ,重新定义了BP网络输出误差函数 ,提出了可变步长的快速选择方法。

For analog TV circuit DC fault diagnosis, the post-fault diagnostic speed and the diagnostic accuracy are the main of its performance. In the paper, the self-learning ability and the classification function of the artificial neural network have been used to TV circuit DC fault diagnosis, training a back propagation (BP)network as a fault learning-base. The specific method for optimally selecting the training patterns has been presented with the element parameter tolerances taken into account. Furthermore, for realizing to diagnose fault fast, the output error function of BP network has been redefined and the iteration steps are given . Simulation results demonstrate that the neural network approach is the method that realize to diagnose TV circuit DC fault fast.