Incipient fault detection and diagnosis of nonlinear industrial process with missing data

Mou, Miao; Zhao, Xiaoqiang

doi:10.1016/j.jtice.2021.10.015

Cited by 9 publications

(5 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Finite Gaussian Mixture Model (FGMM) [9] and Independent Component Analysis [10] excel in handling non-Gaussian process data. Additionally, various strategies for improving and combining existing MSPM methods have been proposed [11][12][13][14]. For instance, Amin et al [12] introduced a hybrid method based on PCA and Bayesian Network (BN), enhancing fault diagnosis by updating BN with multiple likelihood evidence to identify the root cause of process faults and their propagation paths.…”

Section: Introductionmentioning

confidence: 99%

Process monitoring via dependence description based on variable selection and vine copula

Bai,

Qiu,

Liu

et al. 2024

Meas. Sci. Technol.

View full text Add to dashboard Cite

Process monitoring is crucial to ensure the safety of industrial processes. Generally, the monitoring process involves all measured variables; however, large industrial processes contain many redundant variables. For a method based on describing the intrinsic correlation relationships among variables, vine copula-based dependence description (VCDD) method shows significant advantages for describing nonlinear and non-Gaussian processes. However, redundant and irrelevant variables adversely affect the correlation between variables containing the most important information, reducing model performance. The lack of research in this area may substantially weaken the advantages of VCDD for fault monitoring. Therefore, this article introduces a variable selection vine copula dependence description (VSVCDD) monitoring model. It utilizes known faults as validation data to select the relevant variables for constructing the VCDD model, specifically designed for monitoring known faults. Furthermore, to prevent information loss, the remaining unselected variables are also employed to create a separate VCDD model, dedicated to monitoring unknown faults. The performance of the proposed method is verified by a numerical example, the Tennessee-Eastman (TE) process and the Penicillin fermentation process (PFP).

show abstract

Section: Introductionmentioning

confidence: 99%

Process monitoring via dependence description based on variable selection and vine copula

Bai,

Qiu,

Liu

et al. 2024

Meas. Sci. Technol.

View full text Add to dashboard Cite

show abstract

“…Kong et al [12] proposed a fault reconstruction model based on kernel dynamic ICA, which introduced a residual reduction fault subspace (RRFS) and used RRFS for fault diagnosis. For the nonlinear characteristics of the process, Mou and Zhao [17] combined the Gaussian kernel function and polynomial kernel function into a mixed kernel function and proposed a fault diagnosis method based on the mixed kernel function NPE, which has a better fault diagnosis capability for nonlinear processes. However, when these multivariate statistical methods are used for fault diagnosis, it is necessary to manually select a threshold to detect the occurrence of a fault, and then further diagnose fault variables.…”

Section: Introductionmentioning

confidence: 99%

A latent representation dual manifold regularization broad learning system with incremental learning capability for fault diagnosis

Mou

Zhao

et al. 2023

Meas. Sci. Technol.

Self Cite

View full text Add to dashboard Cite

When new fault occurs, the parameters and structure of fault diagnosis model based on deep learning need to be adjusted for retraining, which is often very time-consuming. For the above problem, a latent representation dual manifold regularization broad learning system (LRDMR-BLS) with incremental learning capability is proposed for process fault diagnosis. This model embeds latent representation learning into feature selection and utilizes the link information between data to guide feature selection. Meanwhile, manifold regularization is introduced in the objective function to preserve the local manifold structure of the original data space. Further, a manifold regularization term is added to the objective function of the broad learning system to preserve the local structure of the features. Finally, the incremental learning capability of the proposed model is given, which enables the proposed model can be updated quickly when new fault occurs. The superiority of the proposed model is demonstrated by two Chemical processes.

show abstract

“…Then, according to the design requirements, a proper cost function can be designed to find the optimized frequency and amplitude of the auxiliary input signal, and the multi-objective optimization method is used. The design idea of the frequency is based on the thought of random resonance of minor fault detection [25][26][27][28] , but in our research method, the frequency is designed as a fixed value in connection with a certain type and size fault.…”

Section: Introductionmentioning

confidence: 99%

Active Fault Detection Based on Auxiliary Input Signal Design

Zhang

Wang

et al. 2023

IEEE Access

View full text Add to dashboard Cite

Active fault detection (AFD) is considered an effective incipient failure-detection method. In this paper, an auxiliary input resonance signal proposal that can detect different incipient faults quickly while keeping the system stable is proposed for AFD.To maintain the stability of the system, auxiliary input signals with a sinusoidal form, which should minimize and maximize the effect in fault-free cases and fault cases, are designed on the YJBK parameterization architecture. Subsequently, the fault detection signal is transferred from the traditional system output signal to the cumulative sum of the designed residual signal. Finally, some illustrative examples are given to indicate the effectiveness of the proposed unified operation architecture and some encourage results have been obtained, the simulation results show that the designed auxiliary input signal can not only ensure the normal operation of the system, but also quickly and effectively detect the occurrence of minor faults.

show abstract

Incipient fault detection and diagnosis of nonlinear industrial process with missing data

Cited by 9 publications

References 35 publications

Process monitoring via dependence description based on variable selection and vine copula

Process monitoring via dependence description based on variable selection and vine copula

A latent representation dual manifold regularization broad learning system with incremental learning capability for fault diagnosis

Active Fault Detection Based on Auxiliary Input Signal Design

Contact Info

Product

Resources

About