Integrated multiscale latent variable regression and application to distillation columns

Muddu Madakyaru, Mohamed N. Nounou, Hazem N. Nounou

Research output: Contribution to journalArticle

7 Citations (Scopus)

Abstract

Proper control of distillation columns requires estimating some key variables that are challenging to measure online (such as compositions), which are usually estimated using inferential models. Commonly used inferential models include latent variable regression (LVR) techniques, such as principal component regression (PCR), partial least squares (PLS), and regularized canonical correlation analysis (RCCA). Unfortunately, measured practical data are usually contaminated with errors, which degrade the prediction abilities of inferential models. Therefore, noisy measurements need to be filtered to enhance the prediction accuracy of these models. Multiscale filtering has been shown to be a powerful feature extraction tool. In this work, the advantages of multiscale filtering are utilized to enhance the prediction accuracy of LVR models by developing an integrated multiscale LVR (IMSLVR) modeling algorithm that integrates modeling and feature extraction. The idea behind the IMSLVR modeling algorithm is to filter the process data at different decomposition levels, model the filtered data from each level, and then select the LVR model that optimizes a model selection criterion. The performance of the developed IMSLVR algorithm is illustrated using three examples, one using synthetic data, one using simulated distillation column data, and one using experimental packed bed distillation column data. All examples clearly demonstrate the effectiveness of the IMSLVR algorithm over the conventional methods.

Original languageEnglish
Article number730456
JournalModelling and Simulation in Engineering
Volume2013
DOIs
Publication statusPublished - 2013

Fingerprint

Distillation
Distillation columns
Latent Variables
Regression
Latent Variable Models
Feature Extraction
Prediction
Regression Model
Filtering
Packed Bed
Principal Component Regression
Integrated Modeling
Canonical Correlation Analysis
Model Selection Criteria
Multiscale Modeling
Partial Least Squares
Feature extraction
Synthetic Data
Modeling
Model

All Science Journal Classification (ASJC) codes

  • Modelling and Simulation
  • Engineering(all)
  • Computer Science Applications

Cite this

@article{e9dbd8bd60194174af30ed8964979e86,
title = "Integrated multiscale latent variable regression and application to distillation columns",
abstract = "Proper control of distillation columns requires estimating some key variables that are challenging to measure online (such as compositions), which are usually estimated using inferential models. Commonly used inferential models include latent variable regression (LVR) techniques, such as principal component regression (PCR), partial least squares (PLS), and regularized canonical correlation analysis (RCCA). Unfortunately, measured practical data are usually contaminated with errors, which degrade the prediction abilities of inferential models. Therefore, noisy measurements need to be filtered to enhance the prediction accuracy of these models. Multiscale filtering has been shown to be a powerful feature extraction tool. In this work, the advantages of multiscale filtering are utilized to enhance the prediction accuracy of LVR models by developing an integrated multiscale LVR (IMSLVR) modeling algorithm that integrates modeling and feature extraction. The idea behind the IMSLVR modeling algorithm is to filter the process data at different decomposition levels, model the filtered data from each level, and then select the LVR model that optimizes a model selection criterion. The performance of the developed IMSLVR algorithm is illustrated using three examples, one using synthetic data, one using simulated distillation column data, and one using experimental packed bed distillation column data. All examples clearly demonstrate the effectiveness of the IMSLVR algorithm over the conventional methods.",
author = "Muddu Madakyaru and Nounou, {Mohamed N.} and Nounou, {Hazem N.}",
year = "2013",
doi = "10.1155/2013/730456",
language = "English",
volume = "2013",
journal = "Modelling and Simulation in Engineering",
issn = "1687-5591",
publisher = "Hindawi Publishing Corporation",

}

Integrated multiscale latent variable regression and application to distillation columns. / Madakyaru, Muddu; Nounou, Mohamed N.; Nounou, Hazem N.

In: Modelling and Simulation in Engineering, Vol. 2013, 730456, 2013.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Integrated multiscale latent variable regression and application to distillation columns

AU - Madakyaru, Muddu

AU - Nounou, Mohamed N.

AU - Nounou, Hazem N.

PY - 2013

Y1 - 2013

N2 - Proper control of distillation columns requires estimating some key variables that are challenging to measure online (such as compositions), which are usually estimated using inferential models. Commonly used inferential models include latent variable regression (LVR) techniques, such as principal component regression (PCR), partial least squares (PLS), and regularized canonical correlation analysis (RCCA). Unfortunately, measured practical data are usually contaminated with errors, which degrade the prediction abilities of inferential models. Therefore, noisy measurements need to be filtered to enhance the prediction accuracy of these models. Multiscale filtering has been shown to be a powerful feature extraction tool. In this work, the advantages of multiscale filtering are utilized to enhance the prediction accuracy of LVR models by developing an integrated multiscale LVR (IMSLVR) modeling algorithm that integrates modeling and feature extraction. The idea behind the IMSLVR modeling algorithm is to filter the process data at different decomposition levels, model the filtered data from each level, and then select the LVR model that optimizes a model selection criterion. The performance of the developed IMSLVR algorithm is illustrated using three examples, one using synthetic data, one using simulated distillation column data, and one using experimental packed bed distillation column data. All examples clearly demonstrate the effectiveness of the IMSLVR algorithm over the conventional methods.

AB - Proper control of distillation columns requires estimating some key variables that are challenging to measure online (such as compositions), which are usually estimated using inferential models. Commonly used inferential models include latent variable regression (LVR) techniques, such as principal component regression (PCR), partial least squares (PLS), and regularized canonical correlation analysis (RCCA). Unfortunately, measured practical data are usually contaminated with errors, which degrade the prediction abilities of inferential models. Therefore, noisy measurements need to be filtered to enhance the prediction accuracy of these models. Multiscale filtering has been shown to be a powerful feature extraction tool. In this work, the advantages of multiscale filtering are utilized to enhance the prediction accuracy of LVR models by developing an integrated multiscale LVR (IMSLVR) modeling algorithm that integrates modeling and feature extraction. The idea behind the IMSLVR modeling algorithm is to filter the process data at different decomposition levels, model the filtered data from each level, and then select the LVR model that optimizes a model selection criterion. The performance of the developed IMSLVR algorithm is illustrated using three examples, one using synthetic data, one using simulated distillation column data, and one using experimental packed bed distillation column data. All examples clearly demonstrate the effectiveness of the IMSLVR algorithm over the conventional methods.

UR - http://www.scopus.com/inward/record.url?scp=84877289228&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84877289228&partnerID=8YFLogxK

U2 - 10.1155/2013/730456

DO - 10.1155/2013/730456

M3 - Article

VL - 2013

JO - Modelling and Simulation in Engineering

JF - Modelling and Simulation in Engineering

SN - 1687-5591

M1 - 730456

ER -