Main menu >> Chemical Engineering >> Systems and Control
Chemical Engineering
Systems and Control
- Analysis and Control of Nonlinear Dynamical Systems
- Analysis of Uncertain Large-Scale Systems with Time Delays
- Control-relevant Modeling of Sheet and Film Processes
- Fault Detection and Diagnosis for Large-Scale Systems
- Modeling and Control of Pharmaceutical Crystallization
^ Analysis and Control of Nonlinear Dynamical Systems
R. D. Braatz,* D. L. Ma, E. Rusli
Computational Science and Engineering Program
Creating stability and performance analysis tools that are both computationally efficient and nonconservative for nonlinear dynamical systems is an important and challenging problem. This task is especially challenging for systems described by complex simulation codes in which a closed form representation for the system is unavailable (for example, the code may include if–then statements that implement discrete switches at times that are computed during the simulation). Algorithms are being investigated for the analysis and control of such complex nonlinear dynamical systems.
^ Analysis of Uncertain Large-Scale Systems with Time Delays
R. D. Braatz,* R. Gunawan
E. I. du Pont de Nemours & Co.
Time delays and model uncertainties are prevalent in a broad range of large-scale systems, including paper machines, manufacturing plants, and wireless networks. Existing algorithms for computing stability and performance margins do not adequately address the effect of time delay uncertainties and process constraints on the overall closed-loop system. An algorithm maps time delay uncertainties to equivalent finite-dimensional real parametric variations that can be analyzed using available techniques. These analysis techniques are being used to create polynomial-time algorithms for optimal controller design based on linear and bilinear matrix inequalities.
^ Control-relevant Modeling of Sheet and Film Processes
R. D. Braatz,* A. J. McHugh, J. C. Pirkle, Jr.
Procter & Gamble
Sheet and film processes, which include coating, papermaking, and polymer film extrusion processes, are of world-wide industrial importance. Existing identification and estimation techniques construct models for these processes based only on input–output data, which limits the model quality. The objective of this project is to exploit the fundamental physics for these processes to improve the accuracy of process models and the robustness of model and state estimates. This project includes dynamic simulations and experimental verification on a blown film extruder in the laboratory.
^ Fault Detection and Diagnosis for Large-Scale Systems
R. D. Braatz,* L. H. Chiang
International Paper Co.
Fault detection and diagnosis algorithms are being created that are computable for large data sets (gigabytes of data) collected from large-scale manufacturing plants. This project includes new residual-based statistics based on canonical variate analysis, a suite of fault diagnosis techniques based on Bayesian statistics, new data-driven methods for diagnosing unknown and multiple faults, and a detailed evaluation of these methods through simulations of a fault-prone chemical facility and application to an industrial paper machine.
^ Modeling and Control of Pharmaceutical Crystallization
R. D. Braatz, * D. L. Ma, T. Togkalidou, M. Fujiwara
Merck; Computational Science and Engineering Program
An approach is being created to control pharmaceutical crystal formation that incorporates first-principles simulation models, dynamic optimization, and state-of-the-art analysis. The approach includes simulating the nucleation and growth of crystals with multiple characteristic length scales, designing algorithms for constructing multidimensional crystal size distributions from in-situ optimal microscopy and laser-backscattering, designing analysis techniques for improved in-situ supersaturation measurement in dense crystal slurries using Fourier Transform Infrared Spectroscopy with an Attentuated Transmittance Reflectance probe, and incorporating dynamic optimization algorithms to control the properties of the product crystals for intermediate and end-use applications.
