Temperature Control of an External Cleaning Unit

Abstract: The thesis work was performed at Tetra Pak where a temperature control system for an external cleaning unit was designed. The external cleaning unit is responsible for the heating and supply of cleaning fluids for an aseptic packaging and filling machine. This thesis presents a least squares estimated model predictive controller that uses an ILC inspired algorithm to predict the process behavior. The control problem countered in this thesis is a long variable time delay combined with fast process dynamics and load disturbances. Development of the prediction model was done using Matlab and Simulink, with the ability to use logged process data from the process PLC system in Simulink for verification. The model equations describe the actual physical relations in the process and are easily modified giving the ability to model several different outputs and inputs to the process. These equations were used to predict the main process output in Simulink. In the control system, the equations were modified to calculate the correct control signal and acted as a feedforward. Since the processes varies between units, a least squares estimation program was used to calculate the matching feedforward model to that specific unit. To improve the performance of the feedforward, an ILC algorithm was used to predict the process behavior during large transients. The complete control system and least squares algorithm was implemented using structured text on a Rockwell Logix PLC controller.