Covariance eigenpairs neighbour distance for fault detection in chemical processes

Shang, Jun; Chen, Maoyin; Zhou, Donghua

doi:10.1002/cjce.22920

Cited by 4 publications

(3 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To better ensure the operational safety of the industrial processes, process monitoring methods have been studied by many researchers in the last several decades. [1][2][3][4] According to different mechanisms of the method, process monitoring methods can be mainly divided into two types, model-based and data-based. [5] Obtaining effective and accurate process data has become easier and more accessible because of the fast development of modern sensors and computer technologies.…”

Section: Introductionmentioning

confidence: 99%

Survey on recursive statistical process monitoring methods

et al. 2022

View full text Add to dashboard Cite

Recursive statistical process monitoring (RSPM) methods have superior monitoring performance for industrial processes, especially those with time-varying characteristics, and have recently been studied by many researchers. However, there is no survey paper yet that summarizes and analyzes the existing RSPM methods. In this survey, approximately 60 papers related to RSPM methods are reviewed and categorized from different aspects. Existing RSPM methods are

show abstract

Section: Introductionmentioning

confidence: 99%

Survey on recursive statistical process monitoring methods

et al. 2022

View full text Add to dashboard Cite

show abstract

“…Hence, researchers devote themselves to this issue and fruitful results have been proposed. Venkatasubramanian et al conclude that the existing approaches can be divided into data‐driven methodologies, quantitative model‐based strategies, and qualitative model‐based strategies. Generally, fault detection and diagnosis consists of three essential tasks: fault detection, fault isolation, and fault estimation .…”

Section: Introductionmentioning

confidence: 99%

Interval state and sensor fault estimation based on unknown input observer and interval hull computation

et al. 2019

View full text Add to dashboard Cite

This paper investigates the problem of interval estimation of state and sensor fault simultaneously based on the unknown input decoupling technique and interval hull computation approach. To facilitate sensor fault estimation conveniently, it is first regarded as an auxiliary state, and the faulty system is transformed into an augmented descriptor system. Next, an unknown input observer is designed based on the P-radius method for this augmented descriptor system to decouple some of the unknown system disturbance. Then, the bounds of the augmented state set effected by the undecoupled system uncertainties are calculated via an interval hull approximation technique. Finally, the effectiveness and practicality of this work are illustrated by a numerical example and a quadruple-tank process system. This work demonstrates that by decoupling some unknown disturbances from the estimation error system, the zonotope's size, corresponding to state and sensor fault estimation, can be further reduced. K E Y W O R D Sinterval fault estimation, interval hull computation, unknown input observer

show abstract

“…A new data‐driven process monitoring method, termed as covariance eigenpairs neighbour distance, was presented in Shang et al The eigenpairs of sample covariance matrices in a one‐step sliding window are recursively calculated and then the neighbour distance for the eigenpairs is used as a fault detection index. The proposed method has been implemented in both CSTR and TE processes.…”

mentioning

confidence: 99%