DocumentCode :
637142
Title :
Enhanced modeling of distillation columns using integrated multiscale latent variable regression
Author :
Madakyaru, Muddu ; Nounou, Mohamed Numan ; Nounou, Hazem Numan
Author_Institution :
Chem. Eng. Program, Texas A&M Univ. at Qatar, Doha, Qatar
fYear :
2013
fDate :
16-19 April 2013
Firstpage :
73
Lastpage :
80
Abstract :
Operating distillation columns under control requires inferring the compositions of the distillate and bottom streams (which are challenging to measure) from other more easily measured variables, such as temperatures at different trays of the column. Models that can be used in this regard are called inferential models. Commonly used inferential models include latent variable regression (LVR) techniques, such as principal component regression (PCR), partial least square (PLS), and regularized canonical correlation analysis (RCCA). Unfortunately, measured practical data are usually contaminated with errors, which degrade the prediction accuracy of inferential models. Therefore, noisy measurements need to be filtered to enhance the prediction ability of these models. Wavelet-based multiscale filtering has been shown to be a powerful denoising 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 filtering. 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 two examples, one using synthetic data and the other using simulated distillation column data. Both examples clearly demonstrate the effectiveness of the IMSLVR algorithm.
Keywords :
distillation; distillation equipment; filtering theory; least squares approximations; principal component analysis; production engineering computing; regression analysis; signal denoising; IMSLVR modeling algorithm; LVR techniques; PLS; decomposition levels; denoising tool; distillation columns; integrated multiscale latent variable regression; latent variable regression; noisy measurements; partial least square; prediction accuracy; principal component regression; regularized canonical correlation analysis; simulated distillation column; synthetic data; wavelet-based multiscale filtering; Computational modeling; Data models; Load modeling; Mathematical model; Pollution measurement; Predictive models; Vectors;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Computational Intelligence in Control and Automation (CICA), 2013 IEEE Symposium on
Conference_Location :
Singapore
Type :
conf
DOI :
10.1109/CICA.2013.6611666
Filename :
6611666
Link To Document :
بازگشت