Real-time constrained current control of permanent magnet synchronous machines for automotive applications

A challenging problem in automotive industry is the torque control of permanent magnet synchronous machines employed in automotive electrical traction drives. These applications are characterised by fast dynamics that are subject to physical and computational constraints. The goal of this study is to provide a controller synthesis method that can deal with these challenges. To this end, a one step ahead model predictive state-feedback control scheme based on flexible Lyapunov functions concept was designed. The existing constraints related to the system states and control inputs are approximated with polytopic constraints, which are linear constraints. Moreover, the rotor temperature behaviour, which is crucial in automotive electrical drives, is taken into account directly in the controller synthesis methodology. As a result, the derived model predictive control problem is reduced to a linear program that can be solved on-line at each sample instant. Then, the obtained control solution is implemented in a real electronic control unit and tested in real-time using an industrial hardware-in-the-loop test-bench. Real-time results are reported and analysed and show significant improvement compared with the classical proportonal-integral approach.