Chemical process fault diagnosis based on a combined deep learning method

Bao, Yu; Bo, Wang; Guo, Pandeng; Wang, Jingtao

doi:10.1002/cjce.24153

Cited by 25 publications

(16 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To evaluate the effectiveness of the proposed ENM-LCIbased method, application in monitoring two classical nonlinear processes (a numerical nonlinear example [25] and the TE benchmark process [26][27][28] ) is considered, and comparisons with the aforementioned nonlinear process monitoring methods are provided as well.…”

Section: Illustration and Discussionmentioning

confidence: 99%

“…The benchmark datasets generated from the TE process provide a well-known platform for evaluating the monitoring performance achieved by different MSPM methods. [28][29][30] The TE benchmark process is derived from a real chemical plant, which requires the collaboration of a reactor, a condenser, a vapour/liquid separator, a recycle compressor, a stripper, and a plant-wide control system. A total of 33 continuous measured variables is usually monitored.…”

Section: The Te Benchmark Processmentioning

confidence: 99%

See 1 more Smart Citation

An explicit nonlinear mapping‐based locality constrained index for nonlinear statistical process monitoring

2022

View full text Add to dashboard Cite

Different from the mainstream nonlinear multivariate statistical process monitoring approaches, which usually implement offline feature extraction for a given dataset sampled from the normal operating condition, the proposed method analyzes the inconsistency inherited in each specific online monitored sample of current interest in a timely manner so that a novel monitoring statistic called locality constrained index (LCI) can be simultaneously calculated. Through taking advantage of an explicit nonlinear mapping (ENM), the sampled data is first transformed into a higher‐dimensional space so as to explicitly reflect the inherited nonlinearity between measured variables. With the involvement of the online monitored sample in designing the corresponding objective function, the calculation of LCI is then targeted to point out the deviation within the neighbourhood. According to the comparisons with the counterparts, the proposed ENM‐LCI‐based method demonstrated its salient superiority and consistent effectiveness in nonlinear process monitoring.

show abstract

Section: Illustration and Discussionmentioning

confidence: 99%

Section: The Te Benchmark Processmentioning

confidence: 99%

An explicit nonlinear mapping‐based locality constrained index for nonlinear statistical process monitoring

2022

View full text Add to dashboard Cite

show abstract

“…Step 3 Iteration

t = 1, 2, \dots T

Calculate the label weighting function:

q_{t} (i, y) = \frac{w_{i, y}^{t}}{W_{i}^{t}}, y \neq y_{i}

where

W_{i}^{t} = \sum_{y \neq y_{i}} w_{i, y}^{t}

, and

\sum_{y \neq y_{i}} q_{t} (i, y) = 1

. Calculate the sample weight:

D_{t} (i) = \frac{W_{i}^{t}}{\sum_{i = 1}^{n} W_{i}^{t}}

Create a decision tree classifier as the weak classifier

h_{t}

and pass

D_{t} (i)

and

q_{t} (i, y)

to it. Calculate the pseudo error

ε_{t}

h_{t}

according to Equation (). Set the weight‐updating coefficient: [ 31–33 ]

β_{t} = \frac{ε_{t}}{1 - ε_{t}}

Calculate the new weight:

w_{i, y}^{t ...}

…”

Section: Multi‐fault Classification Based On the Adaboostm2 Algorithmmentioning

confidence: 99%

“…(4) Calculate the pseudo error ε t of h t according to Equation ( 26). ( 5) Set the weight-updating coefficient: [31][32][33]…”

Section: Multi-fault Classification Based On the Adaboostm2 Algorithmmentioning

confidence: 99%

A multi‐fault diagnosis method based on improved SMOTE for class‐imbalanced data

Zhao

et al. 2022

Can J Chem Eng

View full text Add to dashboard Cite

With the development of industrial processes, how to effectively diagnose the faults in an increasingly complex production process has attracted widespread attention. It is worth noting that there may be multiple types of faults in the actual industrial process, and there is an extreme class imbalance between the normal samples and the fault samples. Therefore, it is of practical significance to carry out research on the multi-fault diagnosis method for class-imbalanced data. In this paper, a multi-fault diagnosis method based on improved synthetic minority sampling technology (SMOTE) is proposed. First, aiming at the class imbalance, an improved SMOTE algorithm based on Mahalanobis distance (Mahalanobis distance-based SMOTE [MSMOTE]) is proposed for over-

show abstract

“…Chen et al [24] presented using one-dimensional convolutional auto-encoder based feature learning for the fault diagnosis of multivariate processes. Bao et al [25] presented a combined deep learning approach for chemical process fault diagnosis.…”

Section: Introductionmentioning

confidence: 99%

An Intelligent Process Fault Diagnosis System Based on Neural Networks and Andrews Plot

Wang

Zhang

2021

Processes

View full text Add to dashboard Cite

This paper proposes a neural network-based process fault diagnosis system with Andrews plot for information pre-processing to enhance the performance of online process fault diagnosis. By using features extracted from Andrews plot as the inputs to a neural network, as a classifier, the diagnosis speed and reliability are improved. A method for determining the important features in the Andrews function is proposed. The proposed fault diagnosis system is applied to a simulated continuous stirred tank reactor process and is compared with two conventional neural network-based fault diagnosis systems: scheme B where the monitored measurements are directly fed to a neural network after scaling and scheme C where the monitored measurements are converted to qualitative trend data before feeding to a neural network. Of all the considered faults, the proposed fault diagnosis system diagnosed the abrupt faults on average 5.45 s and 2.66 s earlier than schemes B and C, respectively and diagnosed the incipient faults on average 13.82 s and 5.09 s earlier than schemes B and C, respectively. The results reveal that Andrews plot method utilized in online process monitoring has a great potential in industrial process monitoring.

show abstract

Chemical process fault diagnosis based on a combined deep learning method

Cited by 25 publications

References 39 publications

An explicit nonlinear mapping‐based locality constrained index for nonlinear statistical process monitoring

An explicit nonlinear mapping‐based locality constrained index for nonlinear statistical process monitoring

A multi‐fault diagnosis method based on improved SMOTE for class‐imbalanced data

An Intelligent Process Fault Diagnosis System Based on Neural Networks and Andrews Plot

Contact Info

Product

Resources

About