Model Predictive Control using Neural Networks

Mr. Steve Piche
Pavilion Technologies
Austin, TX

Friday, Feb. 5, 1999, 11:00 AM, ENS 537

Abstract

Optimization and control of many chemical and polymer properties using traditional linear control technologies is difficult, tedious and often unrealizable. With the emergence of fast, large-scale nonlinear model predictive control techniques, control and optimization of these processes and properties becomes feasible. This talk will provide an overview of a neural network based control system. Inferential models used in the system are developed using neural networks while the actual control system is implemented by a model predictive control scheme based upon neural network models. Successful industrial application of this control system to two polymer reactors (polypropylene and polyethylene) will be presented.

Pavilion is looking to fill full-time/summer positions.

A list of Signal and Image Processing Seminars is available at from the ECE department Web pages under "Seminars". The Web address for the Signal and Image Processing Seminars is http://anchovy.ece.utexas.edu/seminars