基于并发模型的LAC在线故障检测

On-Line Fault Detection of LAC Based on Concurrent Model

针对线性模拟电路的在线故障检测进行了研究。首先,通过对并发测试问题的分析及建模要求,从线性模拟电路的网表描述得到可用于残差生成的扩展状态空间模型,这种扩展的状态空间模型定义了电路的变量/方程依赖矩阵即因果依赖矩阵,进而从因果依赖矩阵可得到电路系统的标称模型和未知输入及故障模型。然后,基于无差拍观测器方法采用扩展的状态模型来生成最优故障检测器的电路方程,从而实现故障的在线残差检测。仿真实验结果表明,与传统的线性模拟电路的离线和在线故障检测方法相比,所提出的方案具有更好的故障检测能力,适用于任何线性模拟电路。

On-line fault detection of linear analog circuits(LAC)widely used in various functional electronic circuits in nuclear electron systems is studied.Firstly,based on the analysis and modeling requirements of concurrent testing problem,the extended state space model which can be used to generate residuals is obtained from the netlist description of linear analog circuit.This extended state space model defines the variable/equation dependence matrix of the circuit,namely the causal dependence matrix,from which the nominal model and the unknown input and fault model of the circuit system can be obtained.Then,the extended state model is used to generate the circuit equations of the optimal fault detector based on the dead-beat observer method,so as to realize the on-line residual fault detection.Simulation results show that,compared with the traditional off-line and on-line fault detection methods of linear analog circuit,the proposed scheme has better fault detection capability and is suitable for any linear analog circuits.