Small moving window calibration models for soft sensing processes with limited history

Abstract While many soft-sensing strategies have been developed for monitoring chemical processes, most require extensive historical data to train the soft sensor calibration. Many of these strategies involve very complex models to account for drift and other effects in the process. When a process changes, perhaps because of some upset or a change in process parameters, it is often necessary to rebuild the soft-sensing model, a process that can take considerable time. This paper focuses on use of simple models that can be used to provide soft-sensing with minimal calibration data, and with minimal training of hyperparameters to permit process monitoring while a more complex model is rebuilt. The goal of this study was to explore simple soft sensing strategies for the case where a new process has been put online with limited history. Five soft sensor methodologies with two update conditions were compared on two experimentally-obtained datasets and one simulated dataset. The soft sensors investigated were moving window partial least squares regression (and a recursive variant), moving window random forest regression, the mean moving window of the property y, and a novel random forest partial least squares regression ensemble (RF-PLS), all of which can be used with small sample sizes so that they can be rapidly placed online. It was found that, on two of the datasets studied, small window sizes led to the lowest prediction errors for all of the moving window methods studied. On the majority of datasets studied, the RF-PLS calibration method offered the lowest one-step-ahead prediction errors compared to those of the other methods, and it demonstrated greater predictive stability at larger time delays than moving window PLS alone. Both the random forest and RF-PLS methods most adequately modeled datasets that did not feature purely monotonic increases in property values, but both methods performed more poorly than moving window PLS models on one dataset with purely monotonic property values. Other data dependent findings are presented and discussed.

[1]  Barry Lennox,et al.  Industrial application of neural networks — an investigation , 2001 .

[2]  S. Wold,et al.  PLS-regression: a basic tool of chemometrics , 2001 .

[3]  Steven D. Brown,et al.  A frequency‐localized recursive partial least squares ensemble for soft sensing , 2018 .

[4]  Soon Keat Tan,et al.  Localized, Adaptive Recursive Partial Least Squares Regression for Dynamic System Modeling , 2012 .

[5]  อนิรุธ สืบสิงห์,et al.  Data Mining Practical Machine Learning Tools and Techniques , 2014 .

[6]  Steven D. Brown,et al.  A localized adaptive soft sensor for dynamic system modeling , 2014 .

[7]  Bhupinder S. Dayal,et al.  Recursive exponentially weighted PLS and its applications to adaptive control and prediction , 1997 .

[8]  S.Joe Qin,et al.  Neural Networks for Intelligent Sensors and Control — Practical Issues and Some Solutions , 1997 .

[9]  S. Qin.  Recursive PLS algorithms for adaptive data modeling , 1998 .

[10]  Gülnur Birol,et al.  A modular simulation package for fed-batch fermentation: penicillin production , 2002 .

[11]  Leo Breiman,et al.  Classification and Regression Trees , 1984 .

[12]  Rui Araújo,et al.  Review of soft sensor methods for regression applications , 2016 .

[13]  Chaohong He,et al.  Moving window as a variable selection method in potentiometric titration multivariate calibration and its application to the simultaneous determination of ions in Raschig synthesis mixtures , 2009 .

[14]  K. Helland,et al.  Recursive algorithm for partial least squares regression , 1992 .

[15]  Rui Araújo,et al.  Mixture of partial least squares experts and application in prediction settings with multiple operating modes , 2014 .

[16]  Luigi Fortuna,et al.  Soft Sensors for Monitoring and Control of Industrial Processes (Advances in Industrial Control) , 2006 .

[17]  Rolf Sundberg.  Small sample and selection bias effects in calibration under latent factor regression models , 2007 .

[18]  Ian H. Witten,et al.  Data mining: practical machine learning tools and techniques, 3rd Edition , 1999 .

[19]  Michael J. Piovoso,et al.  Process data chemometrics , 1991 .

[20]  Ming-Wei Lee,et al.  Adaptive soft sensors using local partial least squares with moving window approach , 2012 .

[21]  Leo Breiman,et al.  Random Forests , 2001, Machine Learning.

[22]  B. Kowalski,et al.  Partial least-squares regression: a tutorial , 1986 .

[23]  Ping Wang,et al.  Online soft sensor design using local partial least squares models with adaptive process state partition , 2015 .

[24]  Bogdan Gabrys,et al.  Data-driven Soft Sensors in the process industry , 2009, Comput. Chem. Eng..

[25]  W. Harmon Ray,et al.  Dynamic PLS modelling for process control , 1993 .

[26]  Zhiqiang Ge,et al.  Nonlinear Soft Sensor Development Based on Relevance Vector Machine , 2010 .

[27]  Jiahui Wang,et al.  Modeling Financial Time Series with S-PLUS® , 2003 .

[28]  N. L. Ricker.  The use of biased least-squares estimators for parameters in discrete-time pulse-response models , 1988 .

[29]  Bogdan Gabrys,et al.  Local learning‐based adaptive soft sensor for catalyst activation prediction , 2011 .

[30]  Sirish L. Shah,et al.  Modeling and control of multivariable processes: Dynamic PLS approach , 1997 .

[31]  Willi Gujer,et al.  Data-driven modeling approaches to support wastewater treatment plant operation , 2012, Environ. Model. Softw..

[32]  Olivier Sigaud,et al.  Many regression algorithms, one unified model: A review , 2015, Neural Networks.

[33]  Wei-Yin Loh,et al.  Fifty Years of Classification and Regression Trees , 2014 .

[34]  Zhang Yi,et al.  Advances in Neural Networks - ISNN 2006, Third International Symposium on Neural Networks, Chengdu, China, May 28 - June 1, 2006, Proceedings, Part III , 2006, ISNN.

[35]  S. Qin,et al.  Multimode process monitoring with Bayesian inference‐based finite Gaussian mixture models , 2008 .

[36]  Leo Breiman,et al.  Bagging Predictors , 1996, Machine Learning.

[37]  Bogdan Gabrys,et al.  Review of adaptation mechanisms for data-driven soft sensors , 2011, Comput. Chem. Eng..