Model predictive control implementation on a programmable logic controller for DC motor speed control

Abstract

Model predictive control (MPC) has been widely applied in various industrial plants, which are relatively slow processes. MPC application for faster processes is usually limited with the availability of faster optimization algorithms or digital processor. While it is widely employed, programmable logic controller (PLC) is not usually utilized as a implementation target of MPC in industry due to its low computation performance. A special MPC processor module attached to the standard PLC is used instead. This paper presents an experiment of using Mitsubishi Electric MELSEC-Q PLC as an MPC for DC motor speed control. A very simple quadratic program (QP) algorithm that has been proposed in the literature is adopted here. The algorithm involves a simple computation with a large number of iterations and may be optimized for faster convergence speed. The MPC with this QP algorithm is programmed into the PLC by standard ladder program and an proportional-integral ant-windup (PI-AW) controller is also implemented in the PLC for comparative purpose. Before PLC implementation, some simulation runs are taken to verify and fine-tuned the PI-AW and MPC for a DC motor speed control. The results of MPC implementation in the PLC have shown that MPC with a simple QP algorithm may be implemented in a PLC with a comparable performance with PI-AW