作　　者： ; ; ;

机构地区： 电子科技大学自动化工程学院

出　　处： 《仪器仪表学报》 2011年第7期1564-1569,共6页

摘　　要： 针对非线性模拟电路故障诊断中Volterra级数(Volterra核)计算复杂的难题,基于Volterra核的频谱子带分析,提出了利用将Volterra核进行子带分解以降低计算复杂度的方法。首先将激励信号分别输入正常电路和故障电路得到两种电路的输出,然后分别计算正常电路和故障电路的Volterra子带核,通过求解两种电路Volterra子带核的相关系数实现故障元件的特征提取。实验结果表明,该方法不仅能提高故障诊断效率,还能有效提取灾难型故障(硬故障)和参数型故障(软故障)的故障特征。 Considering the problem of complex computation of Volterra series（Volterra kernel） in nonlinear analog circuit fault diagnosis,an approach based on subband decomposition of Volterra kernel for reducing computational complexity is proposed.Firstly,the output signals of fault-free and faulty circuits are obtained under the excitation of input signals.Then,the subband kernels of the fault-free and faulty circuits are computed.The correlation coefficients between the subband kernels of the fault-free and faulty circuits are used to extract the signatures of the faulty parts.Experiments show that the method proposed in this paper can not only improve fault diagnosis efficiency but also extract the fault signatures of catastrophic and parametric faults effectively.

关 键 词： 非线性电路 故障诊断 级数 子带分解 相关系数

领　　域： [电子电信]