Sensorless Predictive Multiscalar-Based Control of the Five-Phase IPMSM

This article proposes multi-scalar variables based predictive control of sensorless multiphase interior permanent magnet synchronous machine. Estimated parameters from adaptive observers are used to implement the proposed control scheme. The control approach is divided into two parts: for the fundamental plane, torque and its dual quantity from the multi-scalar model are directly predicted by the controller, and torque density is improved by injecting a third harmonic current in the second plane. The multi-scalar model of the 3rd harmonic plane is controlled by classical linear controllers. The analysis of the five-phase interior permanent magnet synchronous machine is done deeply in the stationary reference frame. Moreover, the proposed control scheme is compared with traditional predictive control-based field-oriented control for the fundamental plane and the field-oriented control (linear controllers-based) for the second plane in the
$(d-q)$
reference frame. Compared with the previous control strategy, the proposed control structure provides a fast-dynamic response, reduces the computation resources by eliminating the reference frame transformation to obtain control signals, and improves overall control dynamics. The performance of the proposed control scheme is formally validated by simulation and experimental results.