Model Predictive Control of Compressor Systems

Abstract

The purpose of this thesis was to investigate the suitability of applying MPC for the performance control of compressor systems. Dynamic model of the Kårstø gas compression system was used as a process model in D-spice simulator. It consists of two stages with three parallel compressors on each and intercooling to reduce the temperature and total required compression work. The study has deepen into direction of optimization of operation for this system.

A plantwide control design procedure was applied for the systematic approach of the control problem. Economic potential for the optimization was examined for the different operational conditions and controlled variables were stated.

The main result of the study, a developed and tested MPC application for the optimization of operation of export gas compression system in Septic MPC. It was discovered significant nonlinear interactions in the plant during the modeling and simulation. Acquired step response models works properly only in a specific operational region where they taken.