Entity

Time filter

Source Type

Ancaster, Canada

Duchesne C.,Laval University | Liu J.J.,Pukyong National University | MacGregor J.F.,McMaster University | MacGregor J.F.,ProSensus Inc.
Chemometrics and Intelligent Laboratory Systems | Year: 2012

This paper provides an overview of the history, methods and applications of multivariate image analysis methods as developed for use in the process industries. It presents a general framework for the methods and their applications and discusses them under image analysis methods that use only spectral features of the images, such as Multivariate Image Analysis (MIA), and Multivariate Image Regression (MIR), and those that use textural features and combinations of textural and spectral features. The emphasis is on applications such as monitoring, prediction and control, and on those aspects of the methods that make them suitable for these tasks. © 2012 Elsevier B.V.


MacGregor J.,ProSensus Inc. | MacGregor J.,McMaster University | Cinar A.,Illinois Institute of Technology
Computers and Chemical Engineering | Year: 2012

Historical data collected from processes are readily available. This paper looks at recent advances in the use of data-driven models built from such historical data for monitoring, fault diagnosis, optimization and control. Latent variable models are used because they provide reduced dimensional models for high dimensional processes. They also provide unique, interpretable and causal models, all of which are necessary for the diagnosis, control and optimization of any process. Multivariate latent variable monitoring and fault diagnosis methods are reviewed and contrasted with classical fault detection and diagnosis approaches. The integration of monitoring and diagnosis techniques by using an adaptive agent-based framework is outlined and its use for fault-tolerant control is compared with alternative fault-tolerant control frameworks. The concept of optimizing and controlling high dimensional systems by performing optimizations in the low dimensional latent variable spaces is presented and illustrated by means of several industrial examples. © 2012 Elsevier Ltd.


Golshan M.,McMaster University | MacGregor J.F.,McMaster University | MacGregor J.F.,ProSensus Inc. | Bruwer M.-J.,ProSensus Inc. | Mhaskar P.,McMaster University
Journal of Process Control | Year: 2010

Latent Variable Model Predictive Control (LV-MPC) algorithms are developed for trajectory tracking and disturbance rejection in batch processes. The algorithms are based on multi-phase PCA models developed using batch-wise unfolding of batch data arrays. Two LV-MPC formulations are presented, one based on optimization in the latent variable space and the other on direct optimization over a finite vector of future manipulated variables. In both cases prediction of the future trajectories is accomplished using statistical latent variable missing data imputation methods. The proposed LV-MPCs can handle constraints. Furthermore, due to the batch-wise unfolding approach selected in the modeling section, the nonlinear time-varying behavior of batch processes is captured by the linear LV models thereby yielding very simple and computationally fast nonlinear batch MPC. The methods are tested and compared on a simulated batch reactor case study. © 2010 Elsevier Ltd. All rights reserved.


Tzoc Torres J.M.G.,McMaster University | Nichols E.,ProSensus Inc. | Macgregor J.F.,ProSensus Inc. | Hoare T.,McMaster University
Polymer (United Kingdom) | Year: 2014

The design of stimulus-responsive materials, particularly those intended to respond to more than one stimulus, is an inherently challenging and typically trial-and-error process involving multiple synthesis/characterization iterations in the laboratory. In this work, latent variable models are applied to existing, "failed" polymer formulations and characterizations to facilitate the rational design of materials with specific, targeted properties and to predict responsive polymer properties before synthesizing the materials in the laboratory. The models are capable of simultaneously predicting three targeted polymer properties (cloud point, molecular weight, and % recovery of polymer mass) for poly(N-isopropylacrylamide)-based materials that can be reversibly photo-crosslinked. Model inversion and optimization are used to identify new polymer formulations that exhibit significantly improved properties relative to the formulations developed by chemical intuition based on available literature. This model-based design approach moves away from the traditional trial-and-error approach to save time, energy, and resources in the production of novel materials while at the same time generating responsive polymers with improved properties.


MacGregor J.F.,ProSensus Inc. | Liu Z.,ProSensus Inc. | Bruwer M.-J.,ProSensus Inc. | Polsky B.,Mondelez International | Visscher G.,Mondelez International
Chemometrics and Intelligent Laboratory Systems | Year: 2016

Establishing meaningful multivariate specification regions on the multiple properties of a single raw material has been presented and illustrated by Duchesne & MacGregor Duchesne and MacGregor (2004). However, the manufacture of most final products usually involves the use of many raw materials each with multiple measured properties and from different suppliers. Setting specifications separately on each of these materials is unreasonable since it is the simultaneous combination of the properties of all the materials that will affect final quality. This paper presents an approach to determining the acceptability of new lots of raw materials from multiple suppliers and of assessing the suitability of combining specific lots of materials currently in inventory that will minimize the risk of manufacturing a poor quality product. Multivariate statistical models based on PLS are used to determine the importance of all the properties of each of the materials and to develop the specification methodology. Use of the models for achieving improved control over the product quality is also discussed. © 2016 Elsevier B.V.

Discover hidden collaborations