Two-layer Real Time Optimization of the Drying of Pastes in a Spouted Bed: Experimental Implementation
| Authors: | Costa Carlos E. S., Sym Consulting on Industrial Process Control, Brazil
Freire Fábio B., University of British Columbia, Canada
Correa Nivaldo Ap., State University of São Paulo, Brazil
Freire José T., Federal University of São Carlos, Brazil
Correa Ronaldo G., Federal University of São Carlos, Brazil |
|---|
| Topic: | 6.1 Chemical Process Control |
|---|
| Session: | Process Optimization and Predictive Control |
|---|
| Keywords: | process control, chemical industry, predictive control, quadratic programming, optimization |
|---|
Abstract
This paper presents an experimental implementation of a real time optimization strategy applied to a spouted bed dryer where a paste of homogenized chicken eggs is drying. The regulatory and economical optimization problems are solved by a QP (Quadratic Programming) algorithm at the same frequency and in two steps (two-layer structure). The economical objective function was a quadratic function based on the energy supplied to an air heater. The best operational conditions are found and the use of energy is minimized. Future stationary outputs and stationary control actions are calculated from the inverse of the process gain matrix based on predictions upon a constraint QDMC (Quadratic Dynamic Matrix Control). The algorithm was able to keep the controlled variables close to the best operational conditions.
