Designing MPC algorithms for velocity control of brushed DC motor and verification with SIL tests

Abstract

The model predictive control (MPC) tries to find the best control output by optimizing among the predictions. In this context, determining the objective function of optimization is a critical process. This study discusses the velocity control of the mini brushed direct current (BDC) motor used in the actuator with MPC. Four different functions have been defined to design the most appropriate cost function for MPC by considering the parameters that may be effective in the motor dynamics such as velocity, current, power, and switching states. Software in the Loop (SIL), one of the modeling-based testing methods, was used within the scope of testing the algorithms. After the SIL tests, it has been observed that the mini BDC motor can be successfully controlled with simple adjustments to the cost function of the MPC. As a result of the analysis obtained the best results with the objective function consisting of velocity error, estimated current, and the difference between the two estimated velocity values. With this controller, the BDC motor can be controlled without overshoot and with a steady-state error of less than 2% under load.