Robust Data-Driven Design for Fault Diagnosis of Industrial Drives

Abstract

Due to the presence of actuator disturbances and sensor noise, increased false alarm rate and decreased fault detection rate in fault diagnosis systems have become major concerns. Various performance indexes are proposed to deal with such problems with certain limitations. This paper proposes a robust performance-index based fault diagnosis methodology using input–output data. That data is used to construct robust parity space using the subspace identification method and proposed performance index. Generated residual shows enhanced sensitivity towards faults and robustness against unknown disturbances simultaneously. The threshold for residual is designed using the Gaussian likelihood ratio, and the wavelet transformation is used for post-processing. The proposed performance index is further used to develop a fault isolation procedure. To specify the location of the fault, a modified fault isolation scheme based on perfect unknown input decoupling is proposed that makes actuator and sensor residuals robust against disturbances and noise. The proposed detection and isolation scheme is implemented on the induction motor in the experimental setup. The results have shown the percentage fault detection of 98.88%, which is superior among recent research.

Development of a fault diagnostics and tolerance system: An application to continuous stirred tank reactor

“…The fault detection rate of the proposed diagnostic scheme consistently performs higher as compared to both conventional diagnostic approaches. Quantitative comparison is provided in table 1, demonstrating the superiority of the proposed approach in terms of fault alarm rate and false alarm rate when compared to the conventional approaches [32,40]. Residual plots for sensors and actuators fault using conventional [32,40] and proposed approach.…”

“…Quantitative comparison is provided in table 1, demonstrating the superiority of the proposed approach in terms of fault alarm rate and false alarm rate when compared to the conventional approaches [32,40]. Residual plots for sensors and actuators fault using conventional [32,40] and proposed approach. It is noticed that the proposed approach stands out as a computationally efficient solution for nonlinear dynamic systems, with the strategic use of pre-computed parity vectors being a key contributing factor.…”

“…It is noteworthy that, for conventional approaches referenced in the literature [32,40], the entire dataset is utilized to compute a single parity vector. Figure 4(a) illustrates the residual generated under sensor fault conditions using the conventional approach proposed by Rashid et al [40].…”

“…It is noteworthy that, for conventional approaches referenced in the literature [32,40], the entire dataset is utilized to compute a single parity vector. Figure 4(a) illustrates the residual generated under sensor fault conditions using the conventional approach proposed by Rashid et al [40]. In figure 4(b), the residual generated using the conventional approach-2 proposed by Huo et al [32] is presented, while figure 4(c) demonstrates the residual plot using the proposed approach.…”

“…The state feedback gain is designed to guarantee the stability of the closed-loop system. The diagnostic observer designed using ( 38)- (40) gives estimates of z k , where z k = T d x k . Considering (G i dz , H i dz ) to be controllable, the feedback gain (F) is constructed by using conventional approach [24]…”