Comparison of deep neural network architectures for fault detection in Tennessee Eastman process

Chadha, Gavneet Singh; Schwung, Andreas

doi:10.1109/etfa.2017.8247619

Cited by 23 publications

(18 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…In a review by [ 25 ], several other applications of FDI to the TEP were also presented. Recently, machine learning (ML/AI) approaches have also been applied to the TEP [ 26 ], including deep learning approaches [ 27 ]. Hybrid techniques, such as those of [ 17 ], have also been successfully applied to the TEP.…”

Section: Methodsmentioning

confidence: 99%

Application of Exergy-Based Fault Detection in a Gas-To-Liquids Process Plant

Greyling

Marais

Schoor

et al. 2019

Entropy

View full text Add to dashboard Cite

Fault detection and isolation (FDI) within the petrochemical industries (PCIs) is largely dominated by statistical techniques. Although a signal-based technique centered on exergy flows within a process plant was proposed, it has only been applied to single process units. The exergy-based scheme has not yet been applied to process plants that feature at least a single recycle stream. The Tennessee Eastman process (TEP) is commonly used as an FDI benchmark process, but due to obfuscation, the TEP cannot be directly implemented in a commercial process simulator. Thus, application of FDI techniques to proprietary processes will require significant investment into the implementation of the FDI scheme. This is a key impediment to the wide-spread comparison of various FDI techniques to non-benchmark processes. In this paper, a gas-to-liquids (GTL) process model is developed in Aspen HYSYS®, and the model’s performance is validated. The exergy-based FDI technique is applied to the GTL process while the process is subjected to carefully selected faults. The selected faults aim to affect several process units, and specifically, the resultant recycle stream of the GTL process is considered. The results indicate that even though the exergy-based technique makes use of fixed thresholds, complete detection and isolation can be achieved for a list of common process faults. This is significant since it shows, for the first time, that the exergy-based FDI scheme can successfully be deployed in processes with recycle streams.

show abstract

Section: Methodsmentioning

confidence: 99%

Application of Exergy-Based Fault Detection in a Gas-To-Liquids Process Plant

Greyling

Marais

Schoor

et al. 2019

Entropy

View full text Add to dashboard Cite

show abstract

“…For a fair comparison between the methods, for studies where only non-incipient faults were considered the results were compared to fault detection results obtained from the first level of the hierarchical structure model whereas for studies where all the faults were considered, the comparisons were done for results obtained from second level of the hierarchical structure model. The fault detection rate (FDR) for all the faults is compared in Table 5 for the proposed method, PCA [35], DPCA [35], ICA [23], Convolutional NN (CNN) [45], Deep Stacked Network (DSN) [7], Stacked Autoencoder (SAE) [7], Generative Adversarial Network (GAN) [46] and One-Class SVM (OCSVM) [46]. The fault detection rates for all non-incipient faults and 4 and 5 respectively for different methodologies along with the results from the proposed method.…”

Section: Tablementioning

confidence: 99%

Hierarchical Deep Recurrent Neural Network based Method for Fault Detection and Diagnosis

Agarwal¹,

Gonzalez²,

Elkamel³

et al. 2020

Preprint

View full text Add to dashboard Cite

A Deep Neural Network (DNN) based algorithm is proposed for the detection and classification of faults in industrial plants. The proposed algorithm has the ability to classify faults, especially incipient faults that are difficult to detect and diagnose with traditional threshold based statistical methods or by conventional Artificial Neural Networks (ANNs). The algorithm is based on a Supervised Deep Recurrent Autoencoder Neural Network (Supervised DRAE-NN) that uses dynamic information of the process along the time horizon. Based on this network a hierarchical structure is formulated by grouping faults based on their similarity into subsets of faults for detection and diagnosis. Further, an external pseudo-random binary signal (PRBS) is designed and injected into the system to identify incipient faults. The hierarchical structure based strategy improves the detection and classification accuracy significantly for both incipient and non-incipient faults. The proposed approach is tested on the benchmark Tennessee Eastman Process resulting in significant improvements in classification as compared to both multivariate linear model-based strategies and non-hierarchical nonlinear model-based strategies.

show abstract

“…For Hotelling's T 2 , generally, the mean matrix and the covariance matrix change two essential factors representing the process change from the normal situation to some fault situation [27]. There are two portions of the result in each variable's contribution to the process change.…”

Section: B Fault Diagnosis Using the Directionmentioning

confidence: 99%

“…In order to identify new faults from the known ones, the F direction, which is based on partial F-values with the cumulative percent variation (CPV), is introduced. While the conventional approach based on partial F-values works well in detecting faults in most cases [27]- [30], it still suffers from irrelevant variables and low computation efficiency [31]- [36]. In this paper, the CPV, based on the equivalent variation of each variable, is proposed to determine candidate variables.…”

Section: Introductionmentioning

confidence: 99%

Fault Diagnosis in Chemical Processes Based on Class-Incremental FDA and PCA

et al. 2019

View full text Add to dashboard Cite

A class-incremental scheme of fisher discriminant analysis is proposed to improve the performance ofprocess fault diagnosis. Fisher discriminant analysis seeks directions which are efficient for discrimination and has excellent fault detection and diagnostic performance for the sample set with the tag. However, due to the property of the model, it has no detection and diagnostic capabilities for unseen faults. In order to address this issue, the F direction, which is based on a partial F-values with the principle component analysis, is proposed in this paper. After a new fault being detected and added into the known fault collection, a class-incremental scheme is used to update the fisher discriminant analysis model to enhance the model's ability for continuous fault identification. The proposed approach is validated by the Tennessee Eastman process for the fault diagnosis. The results demonstrate that the proposed class-incremental fisher discriminant analysis method outperforms other conventional fisher discriminant analysis methods. INDEX TERMS Fault diagnosis, class-incremental, fisher discriminant analysis, principle component analysis.

show abstract

Comparison of deep neural network architectures for fault detection in Tennessee Eastman process

Cited by 23 publications

References 24 publications

Application of Exergy-Based Fault Detection in a Gas-To-Liquids Process Plant

Application of Exergy-Based Fault Detection in a Gas-To-Liquids Process Plant

Hierarchical Deep Recurrent Neural Network based Method for Fault Detection and Diagnosis

Fault Diagnosis in Chemical Processes Based on Class-Incremental FDA and PCA

Contact Info

Product

Resources

About