【摘要】 针对模拟电路故障难以识别等问题,提出一种基于小波包Tallis熵和多分类相关向量机(Relevance Vector Machine,RVM)的模拟电路故障诊断方法。该方法采用脉冲信号仿真模拟电路,应用小波包变换对采集到的故障响应信号进行分解,通过提取不同频带内的Tsallis熵作为故障特征值,利用相关向量机对各种状态下的特征向量进行分类决策,实现模拟电路的故障定位。实验结果表明,提出的故障诊断方法相较于现有的故障诊断方法能较好地提取故障特征,极大地提高模拟电路故障诊断的效率。更多还原

【Abstract】 In order to solve the problem that the faults of analog circuit are difficult to recognize, a new method based on wavelet packet Tallis entropy and relevance vector machine(RVM) is proposed. A pulse signal was selected as the circuits excitation source. Then, wavelet packet transform was employed to decompose the collected fault response signals. After computing the Tsallis entropy in different bands, multiclass RVM was conducted to classify the feature vectors under different conditions, and the fault location of analog circuit was realized. The results show that the proposed method can better extract fault features than the existing fault diagnosis methods, and it also greatly improve the efficiency of analog circuit fault diagnosis.更多还原