Internal Model Control of Coupled Distillation Columns

Abstract

Experimental studies in the control of coupled distillation columns have seen performed to satisfy two objectives: 1. To provide an experimental test of previously developed operability and resiliency measures for multivariable systems. 2. To demonstrate the design of Internal Model Control (IMC) controllers for such systems and make comparisons on the basis of performance and ease of tuning with conventional algorithms like PID controllers. To attain these goals, two pilot plant distillation columns were arranged in a coupled configuration and step tests were used to identify input-output relationships. Systems with different levels of operability and resilience were selected and corresponding IMC controllers were designed and tested. The experimental results show clearly that the operability analysis is sufficient to distinguish extremes, but for fine distinctions, a knowledge of desired output changes, disturbances and model uncertainties must be included in the decision process. In addition a method to combine the sometimes contradictory resilience criteria needs to be defined. IMC is shown to be simple to implement, easy to tune, and to produce superior control performance.