An active approach for detection of incipient faults

Process variables can be classified into three stages: normal operation, incipient fault, and significant fault stage. A two-step incipient fault detection strategy was proposed for monitoring the complex industrial process. The first step aims at the significant fault detection using the traditional multivariate statistical process monitoring methods. Then a method combining the wavelet analysis with the residual evaluation was carried out for monitoring the incipient fault. Wavelet analysis aims at extracting the incipient fault features from process noise. The residual generation is optimization based on the robustness and sensitivity index, which can be realized directly using the test data. An improved kernel density estimation based on signal to noise ratio is proposed to adaptively determine the detection threshold. The proposed incipient fault detection scheme is tested on a numerical example and the Tennessee Eastman process. Compared to other traditional fault detection methods, good monitoring performances, such as higher fault detection rate and lower false alarm rate, are obtained.

show abstract

“…] s s s s s s N (5) where s and sample number N are selected by users such that s ≥ n and N is large enough.…”

Section: Industrial and Engineering Chemistry Researchmentioning

confidence: 99%

Incipient Fault Detection Based on Fault Extraction and Residual Evaluation

Jing

Zhou

et al. 2015

Ind. Eng. Chem. Res.

View full text Add to dashboard Cite

show abstract

“…Next we consider the incipient fault algorithm of [22,1,23,24]. Suppose we have the model of the form [22]:ẋ…”

Section: Auxiliary Signal Design For Incipient Faultsmentioning

confidence: 99%

Auxiliary Signal Design for Fault Detection in Nonlinear Systems

Sweetingham¹,

∗²

2009

Control and Intelligent Systems

View full text Add to dashboard Cite

SWEETINGHAM, KELLY A. Auxiliary Signal Design for Fault Detection in Nonlinear Systems. (Under the direction of Professor S. L. Campbell). Recently, research has developed in the area of active fault detection and model identification algorithms for linear systems. These algorithms compute an auxiliary input signal which guarantees fault detection, assuming a bounded noise. This dissertation addresses the question of when the previous linear theory can be applied to nonlinear systems. Several case studies are presented to verify that linearizations can in fact produce results in the nonlinear case. Two results are proven about the use of linearizations. The first result gives a parameter β to scale the entire problem. Using this parameter, the scaled auxiliary signal, along with a scaled noise bound, will guarantee fault detection in the nonlinear problem. The second result shows how to compute the acceptable noise bound for the nonlinear problem using the exact auxiliary signal from the linearized problem. Also presented is a computational study to verify these two results. Two secondary projects are also presented in this dissertation. The first is a comparison of two different linear algorithms used to compute the optimal auxiliary signal. The second is a port of one of the existing pieces of fault detection software into Matlab. This new software is included, as well as a discuss of the complications of the port.

show abstract

“…An incipient fault detection and isolation scheme was proposed in [9, 10] by combining SMOs (sliding mode observers) with a Luenberger observer. In [11], linear quadratic cost function was used to design an auxiliary signal that could guarantee the separation of fault information from external disturbances.…”

Section: Introductionmentioning

confidence: 99%