Analysis of performance and design of several non-linear controllers for a level process

Abstract

The level process is an integrating process that is considered to be one of the most common control problems in the process industry. The objective of this work is to analyze performance of a few optimally designed non-linear controllers for the level process, compare their performance with the performance of an optimally tuned PID controller, and provide conclusions about perspectives of use of these non-linear controllers for the considered application. The nonlinear controllers are a second-order sliding mode controller in particular the super-twisting controller and the continuous control version of a Variable Structure (VS) PID controller for a level process. To achieve the project objectives, the following tasks were defined and completed: • Optimal design and tuning of PI/PID controller using valid methods. • Optimal design of a continuous control version of the VS PI/PID controller. • Optimal design of the STW controller. • Development of a tuning algorithm for the STW controller for the specified level process. • Optimal tuning of the STW controller and provide the used tuning rules. • Development of the software required to achieve the listed tasks. • Simulation and experiments. The major conclusions of the research are: • STW tuning should be an optimization on minimizing the level amplitude, the controller amplitude and the IAE or ITAE caused by the disturbance. • From the experimental results, it is shown that it is difficult to outperform an optimally tuned PI controller whenever a huge amount of noise is present. The Performance of the VS PI controller is more or less similar to the PI controller. Therefore, since both of the nonlinear controllers did not show significantly better results, the extra complexity of the design of the VS PI is not justified whenever a large amount of noise is present. • Once the noise is reduced, the VSPI controller can potentially outperform the optimally tuned PI controller.

Date of Award	2012
Original language	American English
Supervisor	Igor Boiko (Supervisor)