基于相关性分析的3容水箱多参数故障诊断

A Multi-parameter Fault Diagnosis of Three-tank Based on Correlation Analysis

针对3容水箱的故障诊断,提出一种多参数故障诊断方法,包括故障检测、隔离与辨识。利用真实系统与标称系统输出的残差对于故障的敏感性,通过残差实现故障的判断。为消除系统不确定性造成的故障误判,引入相关系数法对真实系统与故障模拟系统分别与标称系统比较后构成的残差信号进行相关性分析;通过构建以相关系数为优化目标的适应度函数,借助差分进化算法(DE)对系统的故障参数进行优化求解,准确辨识故障的模式和参数。仿真结果表明,提出的故障诊断机制可实现多参数并发故障及时发现和准确识别,具有较好的应用价值。

A multi-parameter fault diagnosis method for three-tank is proposed.Diagnosis methods include fault detection,fault isolation and identification.Firstly,the sensitivity of the residuals between the real system and the nominal system to the fault is used to judge the fault.In order to eliminate the fault misjudgment caused by the system uncertainty,the correlation coefficient method is introduced to analyze the residual signals formed by the real system and the fault simulation system compared with the nominal system respectively.By constructing the fitness function with correlation coefficient as the optimization target,the fault parameters of the system are optimized by differential evolutionary algorithm(DE),so as to identify the fault patterns and parameters accurately.The simulation results show that the fault diagnosis mechanism proposed in this paper can realize timely detection and accurate identification of multi-parameter concurrent faults and has good application value.