Abstract
Latent variable (LV) models provide explicit representations of underlying driving forces of process variations and retain the dominant information of process data. In this chapter, we develop a new soft sensor model called probabilistic slow feature regression (PSFR). Slow features as temporally correlated LVs are first derived using probabilistic slow feature analysis (PSFA). Probabilistic slow features that evolve in a state-space form effectively represent nominal variations of processes, some of which may be potentially correlated to quality variables and hence help improving the prediction performance of soft sensors when used as inputs. An efficient expectation maximization algorithm is proposed to estimate parameters of the PSFA model, and two alternative criteria are put forward to select quality-relevant SFs in the PSFR model. The validity and advantages of the proposed method are demonstrated via two case studies.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
For general LDS, it is quite difficult to define \(\mathbf {F}^{t_k - t_k^-}\) for an arbitrary matrix \(\mathbf {F}\). For PSFA, however, such generalization is unique because \(\mathbf {F}\) is diagonal, which particularly applies to this unideal case.
- 2.
Available at http://www.springer.com/gp/book/9781846284793.
References
MacGregor JF, Cinar A (2012) Monitoring, fault diagnosis, fault-tolerant control and optimization: data driven methods. Comput Chem Eng 47:111–120
Ge Z, Huang B, Song Z (2014) Mixture semisupervised principal component regression model and soft sensor application. AIChE J 60:533–545
Ge Z, Huang B, Song Z (2014) Nonlinear semisupervised principal component regression for soft sensor modeling and its mixture form. J Chemometr 28:793–804
Dayal BS, MacGregor JF (1997) Recursive exponentially weighted PLS and its applications to adaptive control and prediction. J Process Control 7:169–179
Zamprogna E, Barolo M, Seborg DE (2004) Estimating product composition profiles in batch distillation via partial least squares regression. Control Eng Practice 12:917–929
Bengio Y, Courville A, Vincent PP (2013) Representation learning: a review and new perspectives. IEEE Trans Pattern Anal Mach Intell 35:1798–1828
Qin SJ (2014) Process data analytics in the era of big data. AIChE J 60:3092–3100
Ku W, Storer RH, Georgakis C (1995) Disturbance detection and isolation by dynamic principal component analysis. Chemom Intell Lab Syst 30:179–196
Kano M, Miyazaki K, Hasebe S et al (2000) Inferential control system of distillation compositions using dynamic partial least squares regression. J Process Control 10:157–166
Negiz A, Cinar A (1997) Statistical monitoring of multivariable dynamic processes with state-space models. AIChE J 43:2002–2020
Tipping ME, Bishop CM (1999) Probabilistic principal component analysis. J R Stat Soc: Ser B 61:611–622
Bechman CF, Smith SM (2004) Probabilistic independent component analysis for functional magnetic resonance imaging. IEEE Trans Med Imaging 23:137–152
Koller D, Friedman N (2009) Probabilistic graphical models: principles and techniques. MIT Press, Cambridge
Turner R, Sahani M (2007) A maximum-likelihood interpretation for slow feature analysis. Neural Comput 19:1022–1038
Digalakis V, Rohlicek JR, Ostendorf M (1993) ML estimation of a stochastic linear system with the EM algorithm and its application to speech recognition. IEEE Trans Speech Audio Process 1:431–442
Ghahramani Z, Hinton GE (1996) Parameter estimation for linear dynamical systems. Technical Report CRG-TR-96-2, University of Toronto, Ontario
Lee TW, Lewicki MS, Girolami M et al (1999) Blind source separation of more sources than mixtures using overcomplete representations. IEEE Signal Process Lett 6:87–90
McLachlan G, Krishnan T (2007) The EM algorithm and extensions. Wiley, Hoboken
Jin X, Huang B (2010) Robust identification of piecewise/switching autoregressive exogenous process. AIChE J 56:1829–1844
Bishop CM (2006) Pattern recognition and machine learning. Springer, New York
Kalman RE (1960) A new approach to linear filtering and prediction problems. J Fluids Eng 82:35–45
Shang C, Yang F, Gao X et al (2015) Extracting latent dynamics from process data for quality prediction and performance assessment via slow feature regression. In: Proceedings of American Control Conference, Chicago, IL
Chen L, Tulsyan A, Huang B et al (2013) Multiple model approach to nonlinear system identification with an uncertain scheduling variable using EM algorithm. J Process Control 23:1480–1496
Keshavarz M, Huang B (2014) Bayesian and expectation maximization methods for multivariate change point detection. Comput Chem Eng 60:339–353
Keshavarz M, Huang B (2014) Expectation maximization method for multivariate change point detection in presence of unknown and changing covariance. Comput Chem Eng 69:128–146
Ge Z, Song Z (2011) Semisupervised Bayesian method for soft sensor modeling with unlabeled data samples. AIChE J 57:2109–2119
Shang C, Huang X, Suykens JAK et al (2015) Enhancing dynamic soft sensors based on DPLS: a temporal smoothness regularization approach. J Process Control 28:17–26
Shang C, Yang F, Gao X et al (2015) A comparative study on improved DPLS soft sensor models applied to a crude distillation unit. In: Preceedings of International Symposium on Advanced Control of Chemical Processes, Whistler, Canada
Fortuna L, Graziani S, Rizzo A et al (2007) Soft sensors for monitoring and control of industrial processes. Springer Science & Business Media, Berlin
Kadlec P, Grbić R, Gabrys B (2011) Review of adaptation mechanisms for data-driven soft sensors. Comput Chem Eng 35:1–24
Xie L, Zeng J, Gao C (2014) Novel just-in-time learning-based soft sensor utilizing non-Gaussian information. IEEE Trans Control Syst Technol 22:360–368
Shang C, Huang B, Yang F et al (2015) Probabilistic slow feature analysis-based representation learning from massive process data for soft sensor modeling. AIChE J 61:4126–4139
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
Copyright information
© 2018 Springer Nature Singapore Pte Ltd.
About this chapter
Cite this chapter
Shang, C. (2018). Probabilistic Slow Feature Regression for Dynamic Soft Sensing. In: Dynamic Modeling of Complex Industrial Processes: Data-driven Methods and Application Research. Springer Theses. Springer, Singapore. https://doi.org/10.1007/978-981-10-6677-1_5
Download citation
DOI: https://doi.org/10.1007/978-981-10-6677-1_5
Published:
Publisher Name: Springer, Singapore
Print ISBN: 978-981-10-6676-4
Online ISBN: 978-981-10-6677-1
eBook Packages: EngineeringEngineering (R0)