Krishna Kant
CNS Division of Computer and Network Systems
CSE Directorate for Computer & Information Science & Engineering
Start Date:
January 1, 2006
Expires:
December 31, 2009 (Estimated)
Awarded Amount to Date:
$480000
Investigator(s):
James Rawlings rawlings@engr.wisc.edu (Principal Investigator)
Stephen Wright (Co-Principal Investigator) Michael Graham (Co-Principal Investigator) John Eaton (Co-Principal Investigator)
Sponsor:
University of Wisconsin-Madison
21 North Park Street
MADISON, WI 53715 608/262-3822
NSF Program(s):
ITR-DYNAMIC DATA DRIV APP SYS, FLUID DYNAMICS, PARTICULATE &MULTIPHASE PROCES, PROCESS & REACTION ENGINEERING
Field Application(s):
0000912 Computer Science
Program Reference Code(s):
HPCC, 9218, 7481, 1443, 1415, 1403
Program Element Code(s):
7581, 1443, 1415, 1403
ABSTRACT
This project will advance capabilities in two complex and previously unaddressed applications: (i) measuring, controlling and preventing the formation of turbulence in fluid flow to achieve drag reduction, and (ii) measuring size and shape distributions of populations of crystalline particles, and model-based feedback control of the manufacture of these particles in real time. Given recent developments in the underlying science and required technology, the project provides for the first time a realistic chance at addressing these two complex applications. In turbulence control, computational power now allows direct simulation of turbulent flows, as well as the promise of model-based control approaches. Second, the advent of MEMS technology makes it possible to envisage sensors and actuators that can work at the scale of turbulence-producing coherent structures, which can be on the order of 100-1000 um. Finally, a better fundamental understanding of these coherent structures has recently been achieved. In measuring and controlling particle populations, the project will integrate computational and measurement capability to analyze video microscopy images in real time to determine particle size and shape distributions. By manipulating environmental variables such as pH, impurity concentration, and temperature, we can influence the evolving particle shape, which can be used as a marker for crystal structure (enabling polymorph control) in pharmaceutical applications.
The applications chosen are ideal for developing DDDAS tools because of the following features: complex models with large numbers of degrees of freedom, high complexity measurements, significant sources of noise and uncertainty, and significant industrial and economic impact. The research conducted under this project will develop and demonstrate the state estimation method known as moving horizon estimation as the algorithm for assimilating in real time the data and dynamic, nonlinear model, and will develop and implement the autocovariance least squares method for identifying the disturbance structures from the measurement data and models. By identifying the disturbance structures from data, the derived models do not have to be perfect in order to represent and predict the data accurately and enable model-based feedback control. The project will develop the new optimization tools that are required to enable state estimation, disturbance identification, and model-based feedback control. Industrial collaborations and participation in industrial consortia provide ample opportunities for technology transfer of the state estimation and model predictive control (ExxonMobil, Eastman Chemical, Shell), video imaging (MettlerToledo), and crystal size and shape distribution control (Mitsubishi and GlaxoSmithKline).
PUBLICATIONS PRODUCED AS A RESULT OF THIS RESEARCH
Figueiredo, M., Nowak, R., and Wright, S. J.. "Gradient projection for sparse reconstruction: Application to compressed sensing and other inverse problems," IEEE Journal of Selected Topics in Signal Processing, v.1, 2007.
Larsen, P. A., D. B. Patience, and J. B. Rawlings. "Industrial
crystallization process control," IEEE Ctl. Sys. Mag., v.August, 2006, p. 2.
Larsen, P. A., J. B. Rawlings, and N. J. Ferrier. "An algorithm for
analyzing noisy, in situ images of high-aspect-aspect ratio crystals
to monitor particle size distribution," Chem. Eng. Sci., v.61, 2006, p. 5236.
Larsen, P. A., J. B. Rawlings, and N. J. Ferrier.. "Model-based object recognition to measure crystal size and shape distributions from in situ video images.," Chemical Engineering Science, v.62, 2007, p. 1430.
Mastny, EA; Haseltine, EL; Rawlings, JB. "Two classes of quasi-steady-state model reductions for stochastic kinetics," JOURNAL OF CHEMICAL PHYSICS, v.127, 2007.
Pannocchia, G., J. B. Rawlings, and S. J. Wright. "Fast, large-scale model predictive control by partial enumeration," Automatica, v.43, 2007, p. 852.
Pannocchia, G., J. B. Rawlings, and S. J. Wright. "A partial enumeration strategy for fast large-scale linear model predictive control," Proceedings of the Chemical Process Control-7, 2006, p. Paper 12.
Rajamani, M. R., J. B. Rawlings, and S. J. Qin. "Equivalence of MPC
disturbance models identified from data," Proceedings of
the Chemical Process Control-7, 2006, p. Paper 38.
Rawlings, J. B. and B. R. Bakshi. "Particle filtering and moving horizon estimation," Computers and Chemical Engineering, v.30, 2006, p. 1529.
Please report errors in award information by writing to: awardsearch@nsf.gov.
