PWA Modelling and Co-ordinated Continuous and Logical Control of a Laboratory Scale Plant with Hybrid Dynamics

Abstract

Abstract Abstract Many process plants are nonlinear and together with this they include a combination of continuous valued and logical control inputs and subsystems. This paper attempts to explore the potential of hybrid model predictive control (MPC) to cope with both of these problems. It uses a laboratory scale plant that was designed for experiments with hybrid systems. This plant has both continuous and logical control inputs and it is considerably nonlinear. An approximate hybrid model of the plant in the form of a piecewise affine (PWA) system is developed and evaluated in the first part of the paper. After that a hybrid MPC based on PWA model is applied to the control of the plant. While designing hybrid MPC and evaluating its performance, there is a special focus on the following question. Logical and continuous control systems are usually designed separately. This may result in unforeseen interactions between logical and continuous control and in the deterioration of the control performance. However, hybrid MPC is based on hybrid model that captures both logical and continuous dynamics in one unified framework. Hence it can reasonably be expected that hybrid MPC can avoid undesirable interactions and possibly also make use of these interactions in a positive way (e.g. to speed up the control response using logical inputs). Control results obtained with hybrid MPC are indeed fairly good and they show clear improvement over the results achieved with separate design of logical and continuous control.