Batch process modeling by using temporal feature and Gaussian mixture model

Multi-model/multi-phase modeling algorithm has been widely used to monitor the product quality in complicated batch processes. Most multi-model/ multi-phase modeling methods hinge on the structure of a linearly separable space or a combination of different sub-spaces. However, it is impossible to accurately separate the overlapping region samples into different operating sub-spaces using unsupervised learning techniques. A Gaussian mixture model (GMM) using temporal features is proposed in the work. First, the number of sub-model is estimated by using the maximum interval process trend analysis algorithm. Then, the GMM parameters constrained with the temporal value are identified by using the expectation maximization (EM) algorithm, which minimizes confusion in overlapping regions of different Gaussian processes. A numerical example and a penicillin fermentation process demonstrate the effectiveness of the proposed algorithm.

[1]  Miao Aimin,et al.  Neighborhood preserving regression embedding based data regression and its applications on soft sensor modeling , 2015 .

[2]  Biao Huang,et al.  Robust identification of piecewise/switching autoregressive exogenous process , 2009 .

[3]  Chunhui Zhao,et al.  Regression modeling and quality prediction for multiphase batch processes with inner-phase analysis , 2014 .

[4]  Pierantonio Facco,et al.  Transfer of process monitoring models between plants: Batch systems , 2014 .

[5]  Zhiqiang Ge,et al.  Co-training partial least squares model for semi-supervised soft sensor development , 2015 .

[6]  Hiromasa Kaneko,et al.  Adaptive soft sensor model using online support vector regression with time variable and discussion of appropriate hyperparameter settings and window size , 2013, Comput. Chem. Eng..

[7]  Anil K. Jain,et al.  Unsupervised Learning of Finite Mixture Models , 2002, IEEE Trans. Pattern Anal. Mach. Intell..

[8]  Aude Billard,et al.  On Learning, Representing, and Generalizing a Task in a Humanoid Robot , 2007, IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics).

[9]  A. Yazdizadeh,et al.  Identification of Damavand tokamak using fractional order dynamic neural network , 2018, Trans. Inst. Meas. Control.

[10]  Zhiqiang Ge,et al.  Soft sensor model development in multiphase/multimode processes based on Gaussian mixture regression , 2014 .

[11]  Hazem Nounou,et al.  Kernel PLS-based GLRT method for fault detection of chemical processes , 2016 .

[12]  Zhihuan Song,et al.  Semisupervised learning for probabilistic partial least squares regression model and soft sensor application , 2018 .

[13]  Wenhai Qi,et al.  Soft sensor modelling of propylene conversion based on a Takagi-Sugeno fuzzy neural network optimized with independent component analysis and mutual information , 2019 .

[14]  M. Naraghi,et al.  Minimum variance lower bound estimation with Gaussian Process models , 2018, Trans. Inst. Meas. Control.

[15]  Damien Fay,et al.  Multiple adaptive mechanisms for data-driven soft sensors , 2017, Comput. Chem. Eng..

[16]  Barry Lennox,et al.  The development of an industrial-scale fed-batch fermentation simulation. , 2015, Journal of biotechnology.

[17]  Fang Wu,et al.  Step-wise support vector machines for classification of overlapping samples , 2015, Neurocomputing.

[18]  Biao Huang,et al.  A Bayesian approach to design of adaptive multi-model inferential sensors with application in oil sand industry , 2012 .

[19]  Gee Wah Ng,et al.  Classification for overlapping classes using optimized overlapping region detection and soft decision , 2010, 2010 13th International Conference on Information Fusion.

[20]  L. Luo,et al.  Quality prediction and quality-relevant monitoring with multilinear PLS for batch processes , 2016 .

[21]  Nitin Singh,et al.  Model order reduction using factor division algorithm and fuzzy c-means clustering technique , 2019, Trans. Inst. Meas. Control.

[22]  Raghunathan Rengaswamy,et al.  A Novel Interval-Halving Framework For Automated Identification of Process Trends , 2004 .

[23]  Jie Yu,et al.  Soft Sensor Model Maintenance: A Case Study in Industrial Processes∗ , 2015 .

[24]  Jie Yu,et al.  Quality relevant nonlinear batch process performance monitoring using a kernel based multiway non-Gaussian latent subspace projection approach , 2014 .

[25]  Yan-Lin He,et al.  Data driven soft sensor development for complex chemical processes using extreme learning machine , 2015 .

[26]  Shengrui Wang,et al.  Measuring the component overlapping in the Gaussian mixture model , 2011, Data Mining and Knowledge Discovery.

[27]  Xianchao Zhang,et al.  Novel density-based and hierarchical density-based clustering algorithms for uncertain data , 2017, Neural Networks.

[28]  Alberto Bemporad,et al.  A Greedy Approach to Identification of Piecewise Affine Models , 2003, HSCC.

[29]  Zhiqiang Ge,et al.  Semi-supervised PLVR models for process monitoring with unequal sample sizes of process variables and quality variables , 2015 .

[30]  Dražen Slišković,et al.  Adaptive soft sensor for online prediction and process monitoring based on a mixture of Gaussian process models , 2013, Comput. Chem. Eng..

[31]  Sarthak Tiwari,et al.  A deep learning based data driven soft sensor for bioprocesses , 2018, Biochemical Engineering Journal.

[32]  Young-Il Kim,et al.  A cluster validation index for GK cluster analysis based on relative degree of sharing , 2004, Inf. Sci..

[33]  Alberto Bemporad,et al.  Piecewise affine regression via recursive multiple least squares and multicategory discrimination , 2016, Autom..

[34]  A. Ferrer,et al.  PCA model building with missing data: New proposals and a comparative study , 2015 .

[35]  Huizhong Yang,et al.  A Multi-model Approach for Soft Sensor Development Based on Feature Extraction Using Weighted Kernel Fisher Criterion , 2014 .