A Review of Virtual-Flux Model Predictive Control and Receding Horizon Estimation in Motor Drives

Model predictive control and receding horizon estimation are advanced control and estimation techniques for next-generation high performance motor drives. These methods benefit from low noise and accurate system models, which are challenging to realize with parameter-based motor models. The idea of virtual-flux, where the machine is modelled with flux instead of current, has been seen as a solution, as the parameters and their nonlinearities can be captured by a function that maps measured current onto the corresponding flux it generates. In this way, all system information can be encoded in a single static function, simplifying the stability and robustness analyses, as well as online computational requirements. Furthermore, virtual-flux modelling allows for many electric machines and even the electric grid to be described in a similar way. This review condenses the results in literature into a uniform virtual-flux framework and explores the applications and potential of model predictive control and receding horizon estimation. The combination of these concepts is shown to strongly benefit their respective problems, ranging from finite control set and convex control set MPC, full and reduced phase current sensor set flux estimation, and position and speed co-estimation.