Date of Award

Winter 1999

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Mechanical & Aerospace Engineering

Program/Concentration

Engineering Mechanics

Committee Director

Jen-Kuang Huang

Committee Member

Chuh Mei

Committee Member

Sebastian Bawab

Committee Member

Keith M. Williamson

Abstract

A Kraft recovery boiler in a pulp-paper mill provides a means for recovery of the heat energy in spent liquor and recovery of inorganic chemicals while controlling emissions. These processes are carried out in a combined chemical recovery unit and steam boiler that is fired with concentrated black liquor and which produces molten smelt. Since the recovery boiler is considered to be an essential part of the pulp-paper mill in terms of energy resources, the performance of the recovery boiler has to be controlled to achieve the highest efficiency under unexpected disturbances.

This dissertation presents a new approach for combining system identification technique with predictive control strategy. System identification is the process of building mathematical models of dynamical systems based on the available input and output data from the system. Predictive control is a strategy where the current control action is based upon a prediction of the system response at some number of time steps into the future. A new algorithm uses an i-step-ahead predictor integrated with the least-square technique to build the new control law. Based on the receding horizon predictive control approach, the tracking predictive control law is achieved and performs successfully on the recovery boiler of the pulp-paper mill. This predictive controller is designed from ARX coefficients that are computed directly from input and output data. The character of this controller is governed by two parameters. One parameter is the prediction horizon as in traditional predictive control and the other parameter is the order of the ARX model. A recursive version of the developed algorithm can be evolved for real-time implementation. It includes adaptive tuning of these two design parameters for optimal performance. The new predictive control is proven to be a significant improvement compared to a conventional PID controller, especially when the system is subjected to noise and disturbances.

DOI

10.25777/8hwp-nr65

ISBN

9780599285255

Share

COinS