基于支持向量机的模拟电路故障诊断研究

Research on the Fault Diagnosis Based on Support Vector Machine for Analog Circuits

针对模拟电路的故障诊断问题,详细介绍了支持向量机算法,由于它在非线性映射、小样本学习方面的独特优势,故将它引用到模拟电路的故障诊断过程中。并提出了一种基于支持向量机的诊断方法,该算法能够对被测电路的故障进行有效并且精确地分类。以折线逼近平方曲线的近似测量电路为例,设计了基于支持向量机的模拟电路故障诊断系统。以实际测试数据作为训练样本进行学习训练后,对其它实际测量数据进行诊断,其结果正确,验证了算法的有效性。

Aiming at the issue of fault diagnosis of analog circuits, the support vector machine algorithm is introduced in detail. Because its unique superiority in non-linear mapping and small sample learning, it is adopted into fault diagnosis of analog circuits; thus the diagnostic method based on support vector machine is proposed. By using this algorithm, the faults of the measured circuit can be effectively classified. With the approximate effective value of the approximate square curve of broken line as example, the fault diagnosis system based on support vector machine is designed for analog circuits. Through learning and training with practical test data as training samples, diagnosis is conducted for other practical measured data, the resuh is correct; the effectiveness of the algorithm is verified.