MODELLING AND CONTROL OF THE COKE MAKING PROCESS

DAH-LAIN TANG, Purdue University

Abstract

The coke making process of the steel industry was modelled and a computer-based control system was developed to help improve the process operation. Due to the complexity of the nature of the coal, which is the raw material used to make coke, the operations of the coke making process are presently based on the judgment of experienced operators and the use of the results of regression analyses. Consequently, there is high variation in the required production time of the coke, including an over-long cooking time, to guarantee a well-coked product. The latter results in excessive energy costs. In this study, a coking thermal model and a flue combustion model were integrated into an overall, multi-oven, computer simulation package. The coking times, energy costs, and flue temperature patterns generated from the package have been validated by comparison with data available from industrial measurements. A new heating pattern strategy that would reduce the required energy costs of the operation is proposed and demonstrated. This simulation package was also used as the real-plant example in a computer simulation used to design the adaptive, self-tuning controller proposed here. The pole placement, self-tuning, adaptive control method was utilized to develop the controller which would be implemented by a computer in the plant. The simulation results obtained have demonstrated the effectiveness of this controller method in compensating for any uncertainties in the characteristics of the coal which was charged into the oven. Further applications of the simulation package to the industrial computer control of the coke plant were also investigated. Conclusions of the study and a list of recommendations for future research are included.

Degree

Ph.D.

Subject Area

Mechanical engineering

COinS