Monitoring, fault diagnosis, fault-tolerant control and optimization: Data driven methods

MacGregor, John F.; Çınar, Ali

doi:10.1016/j.compchemeng.2012.06.017

Cited by 251 publications

(147 citation statements)

References 37 publications

Supporting

Mentioning

142

Contrasting

Unclassified

Order By: Relevance

“…PCA is based on the construction of a new reduced set of variables, the so-called principal components, which explain properly predominant trends in multivariate data. With PCA it is possible to reduce random measurement errors (MACGREGOR and CINAR, 2012). There are several algorithms to generate the principal components, being the singular value decomposition (SVD) a commonly used algorithm (TEÓFILO, 2007).…”

Section: Introductionmentioning

confidence: 99%

“…Some methods rely on modeling, while others use historical process data. The last ones, especially the multivariate, have gained industry attention given the accuracy and speed with which they warn about errors (MACGREGOR and CINAR, 2012). Principal Component Analysis (PCA) is a statistical method that performs data compression and makes visualization in multidimensional space easier.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Algorithm for Detection and Diagnosis of Industrial Furnace Faults Applying Singular Value Decomposition and Graphic Visualization

Fonseca

Teófilo

Soares

et al. 2017

JCEC

View full text Add to dashboard Cite

ABSTRACT:The development of an algorithm based on singular value decomposition with graphic visualization to aid in the fault detection and diagnosis was the aim of this work. In order to test the algorithm, real temperature data from an industrial furnace of a natural gas processing unit were used. Nine temperature detectors were monitored over time. In addition to a real fault that caused the shutdown of the equipment, five simulated faults in the furnace detectors were also tested. Results showed that the algorithm was effective and has great potential to assist in the monitoring of industrial processes. RESUMO: O desenvolvimento de um algoritmo baseado na decomposição dos valores singulares com visualização gráfica para auxiliar na detecção e no diagnóstico de falhas foi o objetivo deste trabalho. Para testar o algoritmo foram utilizados dados de temperatura reais de um forno industrial

show abstract

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Algorithm for Detection and Diagnosis of Industrial Furnace Faults Applying Singular Value Decomposition and Graphic Visualization

Fonseca

Teófilo

Soares

et al. 2017

JCEC

View full text Add to dashboard Cite

show abstract

“…Meanwhile, with the rapid progress in sensor technology, distributed process control and data acquisition systems, more and more process variables can be measured on a routine basis. As a result, data-driven FDD methods have attracted substantial attention in both academia and industry (Venkatasubramanian et al, 2003b;MacGregor & Cinar, 2012). Multivariate statistical process monitoring (MSPM) is a well-known datadriven FDD method (Qin, 2012;Yao & Gao, 2009).…”

Section: Introductionmentioning

confidence: 99%

Root cause analysis in multivariate statistical process monitoring: Integrating reconstruction-based multivariate contribution analysis with fuzzy-signed directed graphs

Chen

Yang

2014

Computers & Chemical Engineering

View full text Add to dashboard Cite

“…In most industrial process control applications, the inherent (gradual) time-varying nature of the plant dynamics diminishes the lifetime performance of model-based control systems (see MacGregor & Cinar, 2012;Qin, 2012;Yin, Ding, Haghani, Hao, & Zhang, 2012;Shardt et al, 2012 for recent reviews on performance monitoring and diagnosis). Changes in the plant dynamics over time increase the plant-model mismatch, which may eventually invalidate the model identified at the commissioning stage of a model-based control system.…”

Section: Introductionmentioning

confidence: 99%

Least costly closed-loop performance diagnosis and plant re-identification

Mesbah

Bombois

Forgione

et al. 2015

International Journal of Control

View full text Add to dashboard Cite

The inherent time-varying nature of dynamics in chemical processes often limits the lifetime performance of model-based control systems, as the plant and disturbance dynamics change over time. A critical step in the maintenance of model-based controllers is distinguishing control-relevant plant changes from variations in disturbance characteristics. In this paper, prediction error identification is used to evaluate a hypothesis test that detects if the performance drop arises from controlrelevant plant changes. The decision rule is assessed by verifying whether an identified model of the true plant lies outside the set of all plant models that lead to adequate closed-loop performance. A unified experiment design framework is presented in the least costly context (i.e., least intrusion of nominal plant operation) to address the problem of input signal design for performance diagnosis and plant re-identification when the performance drop is due to plant changes. The application of the presented performance diagnosis approach to a (nonlinear) chemical reactor demonstrates the effectiveness of the approach in detecting the cause of an observed closed-loop performance drop based on the designed least costly diagnosis experiment.

show abstract

Monitoring, fault diagnosis, fault-tolerant control and optimization: Data driven methods

Cited by 251 publications

References 37 publications

Algorithm for Detection and Diagnosis of Industrial Furnace Faults Applying Singular Value Decomposition and Graphic Visualization

Algorithm for Detection and Diagnosis of Industrial Furnace Faults Applying Singular Value Decomposition and Graphic Visualization

Root cause analysis in multivariate statistical process monitoring: Integrating reconstruction-based multivariate contribution analysis with fuzzy-signed directed graphs

Least costly closed-loop performance diagnosis and plant re-identification

Contact Info

Product

Resources

About