Parallel PCA–KPCA for nonlinear process monitoring

Jiang, Qingchao; Yan, Xuefeng

doi:10.1016/j.conengprac.2018.07.012

Cited by 181 publications

(85 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…where τ ∈ [−1,1], τ = 1 means the change of u 1 same with u 2 , τ = − 1 indicates that the change of u 1 is just the opposite of u 2 , and τ = 0 shows that it cannot judge whether there has relationship between the two variables. The details for Equations (12) and (13) can be found in Nelsen. 43 That is, the copula-correlation analysis is the copula function-based Kendall rank correlation coefficient analysis and Spearman rank correlation coefficient analysis.…”

Section: Two-dimensional Copula-correlation Analysismentioning

confidence: 99%

See 1 more Smart Citation

Decentralized monitoring for large‐scale process using copula‐correlation analysis and Bayesian inference–based multiblock principal component analysis

Tian

Peng

et al. 2019

Journal of Chemometrics

View full text Add to dashboard Cite

Due to the massive monitored samples and their features such as strong coupling and time delay in large-scale industrial process, the decentralized monitoring methods, which divide the variables into several blocks and perform local monitoring in each subblock, have been brought up to handle the complex characteristics. However, most of the existing decentralized methods utilize the correlation degree among variables for block division, while the correlation patterns among variables are ignored. The missing of correlation pattern will influence the monitoring performance. To address this problem, the copula-correlation analysis, which considers both the correlation degree and the correlation pattern, is adopted for block division; then the principal component analysis (PCA)-based method is used for subblock monitoring, and the Bayesian inference is introduced to achieve decision fusion of fault detection. The superiority and effectiveness of the proposed method are illustrated through comparison studies on a numerical example and the Tennessee Eastman (TE) benchmark process.considering that both linear and nonlinear relationships may exist among process variables, parallel PCA-KPCA for nonlinear process monitoring has been proposed 12 ; considering the effect of variable selection on PCA monitoring performance, fault-relevant variable selection-based PCA method has been addressed. 13 Other than aforementioned PCA method and its extensions, other methods have been also developed for monitoring under various industrial process applications in the past decade. Tian introduced average multivariate cumulative sum (MCUSUM) residual analysis to improve the independent component analysis (ICA), which achieved the fault detection and diagnosis for non-Gaussian processes with periodic disturbance. 14 Peng proposed online performance monitoring and modeling paradigm based on just-in-time learning (JITL) and extreme learning machine (ELM) for non-Gaussian chemical process. 15 Du combined lazy learning (LL), support vector data description (SVDD), and modified receptor density algorithm (MRDA) to monitor the nonlinear multiple modes process. 16 Peng proposed sparse modeling and dictionary learning-based multimode process monitoring and fault detection method. 17 Chen proposed improved canonical correlation analysis (CCA) to realize the detection of multiplicative faults in industrial processes. 18 Gajjar and Tian explore the visualization technique for process fault detection and diagnosis, respectively. 19,20 However, despite these techniques, considering that the process presents novel complex characteristics due to the expansion of mass industrial production, such as massive monitoring points, strong coupling, and time delay among different operating regions, usually, a unique model that covers all possible operating conditions and characteristics should be extracted correctly to warn when fault occurs in the large-scale process. [21][22][23] The commonly used method is the multiblock method, which provides effective descr...

show abstract

Section: Two-dimensional Copula-correlation Analysismentioning

confidence: 99%

“…XMEAS (4,20,21,22,29,31),XMV (4,5,9) 46 XMEAS (1,7,11,13,16,20,23,27,29,33,35,38),XMV(3,5,6) XMEAS (7,10,11,13,16,18,19,20,25,31,33),XMV (2,5,6,9) 48 XMEAS (12),XMV (5) 49 XMEAS (15,23),XMV (8) 50 XMEAS (7,10,11,13,16,18,19,20,23,25,29,…”

mentioning

confidence: 99%

Decentralized monitoring for large‐scale process using copula‐correlation analysis and Bayesian inference–based multiblock principal component analysis

Tian

Peng

et al. 2019

Journal of Chemometrics

View full text Add to dashboard Cite

show abstract

“…The batch process is an important part of modern industry, and the safety monitoring of the batch process is very important and meaningful . Theoretical research based on data‐driven modeling methods, including soft sensor modeling and process monitoring, has made significant progress and has become increasingly intelligent with industrial processes . For the batch process, multiway principal component analysis (MPCA) is one of the most basic and the most extensively used monitoring methods .…”

Section: Introductionmentioning

confidence: 99%

“…[1][2][3] Theoretical research based on data-driven modeling methods, including soft sensor modeling and process monitoring, has made significant progress and has become increasingly intelligent with industrial processes. [4][5][6][7][8][9][10][11][12] For the batch process, multiway principal component analysis (MPCA) is one of the most basic and the most extensively used monitoring methods. 13,14 The requirement for PCA-based monitoring methods is that the data must be subject to Gaussian distribution.…”

Section: Introductionmentioning

confidence: 99%

Just‐in‐time learning–multiple subspace support vector data description used for non‐Gaussian dynamic batch process monitoring

Yan

Jiang

et al. 2019

Journal of Chemometrics

Self Cite

View full text Add to dashboard Cite

Batch process data are time‐varying dynamic and non‐Gaussian distributed. In addition, for multivariate statistical process monitoring, their variability can be overwhelmed when considering local variability behavior. To address the abovementioned issues, an improved batch process monitoring approach is presented that integrates just‐in‐time learning and multiple subspace support vector data description (JITL‐MSSVDD). A new multiple subspace segmentation method is proposed that classifies a contribution array that is calculated on the mixing matrix of independent component analysis (ICA). Offline, the variable subspace segmentation rule can be obtained from the proposed method. The subspace monitoring models can reduce the risk of the variability being overwhelmed. Online, local modeling samples are collected through JITL, which can reduce the impact of the time‐varying dynamic on modeling accuracy. Then, in accordance with the variable subspace segmentation rule, which is obtained offline, MSSVDD models are constructed to solve non‐Gaussian problems. The advantage of the proposed JITL‐MSSVDD is demonstrated through the standard test model fed‐batch penicillin fermentation process.

show abstract

“…In past years, many researchers have applied data-driven methods to determine the reliability and safety of industrial processes, forming a variety of methods called the multivariate statistical process monitoring (MSPM) methods. [1][2][3][4][5][6][7][8][9][10][11][12] Among them, principal component analysis (PCA), [5][6][7] partial least squares (PLS), [8] and independent component analysis (ICA) [9] are the most commonly used methods. These classic algorithms assume that the process is static where the data in the current sample time is independent from the data in previous moment.…”

Section: Introductionmentioning

confidence: 99%

Tensor sequence component analysis for fault detection in dynamic process

et al. 2019

View full text Add to dashboard Cite

For dynamic processes, using sequence information to augment the data can improve fault detection performance. Traditional approaches transform raw data into augmented vectors, which leads to losses in structural information in the variables and increases the data dimension. This paper proposes a novel data dimension reduction algorithm called tensor sequence component analysis (TSCA) and applies it to dynamic process fault detection. The algorithm extends each sample into a matrix comprising current and past process data, and simultaneously reduces the dimensions of time delay and the variables for feature extraction, solving the problem of the curse of dimensionality. For the dimension reduction of time delay, in order to extract similar information from the samples, each sample is reconstructed with time neighbourhoods. For the dimension reduction of the variables, considering the information of different variables variance information of the latent variables is maximized for feature extraction. Finally, a numerical example and the Tennessee Eastman process are used to demonstrate the efficacy of the proposed method.

show abstract

Parallel PCA–KPCA for nonlinear process monitoring

Cited by 181 publications

References 29 publications

Decentralized monitoring for large‐scale process using copula‐correlation analysis and Bayesian inference–based multiblock principal component analysis

Decentralized monitoring for large‐scale process using copula‐correlation analysis and Bayesian inference–based multiblock principal component analysis

Just‐in‐time learning–multiple subspace support vector data description used for non‐Gaussian dynamic batch process monitoring

Tensor sequence component analysis for fault detection in dynamic process

Contact Info

Product

Resources

About