An experimental study of spatial temperature profile control of a distributed parameter heating system using model predictive control

Abstract

The problem of profile control is important owing to its industrial significance. This problem is viewed as a multi-objective problem and solved using different approaches such as augmentation of objectives or prioritised solving of each objective separately. The focus in our work is to experimentally study the lexicographic optimisation approach for prioritised control of different objectives. An experimental rig has been designed for the purpose of priority driven spatial property control of a distributed parameter system (DPS). The setup consists of a thin metal plate with four temperature sensors and four electric heaters located axially. Through this experimental rig, we demonstrate the concept of controlling the spatial profile in a DPS and address the relevant issues. Specifically, when the desired spatial profile is unachievable, we may be interested in controlling different parts of the profile depending upon their importance. We show an MPC formulation to achieve such a spatial profile control with user defined priority using lexicographic optimisation approach in this work.