基于OWPB和LS-SVM的电路板故障诊断系统

针对雷达电路板检修困难的问题,提出了基于最优小波包基和最小二乘支持向量机相结合的雷达电路板故障诊断方法。利用小波变换对采样数据进行去噪处理,通过小波包分解选择最优小波包基提取熵值作为故障特征向量,并作为基于最小二乘支持向量机的雷达故障诊断模型的输入向量,经诊断模型输出后,完成雷达电路板故障诊断。基于此方法设计了雷达电路板故障诊断系统,提高了雷达故障诊断的正确性和效率。

PCB fault diagnosis system based on OWPB and LS-SVM

According to the reality of maintaining the circuit board of radar,this paper proposed a new method in radar circuit board fault diagnosis,which combined optimal wavelet packet basis(OWPB) and least squares support vector machine(LS-SVM).First,the sampled data was denoised by wavelet transformation,then after wavelet packet decomposition,applied the entropies of the best wavelet packet basis selected as the input vectors of characteristic vectors of the fault,they were the input vectors of the fault diagnosis model of radar based on least squares support vector machine,and the outputs of the model were used to complete the categories of the typical faults in the radar circuit board.And based on this method,this paper designed fault diagnosis system for the circuit board of some kind of radar,it improves the accuracy and efficiency of the fault diagnosis for radar.