Analysis of phase-detection algorithms for back-EMF-based sensorless strategies through real-time simulations

The work developed in this paper presents a comparison between three different phase detection algorithms in order to study the benefits of using each of them when they are applied inside a sensorless rotor position estimation technique. In particular we will analyse the various aspects related to the use of each single algorithm when it is applied into a specific model-based sensorless strategy and when we also take into account the typical impairments affecting the implementation of the real setup, such as unwanted harmonic components and measurement offsets. As a first step a short overview on each phase-detection algorithm under consideration will be given and subsequently the proposed model-based estimation technique will be presented. As a second step the three algorithms are analyzed in detail highlighting their dynamical performances and their robustness properties against implementative knots, in particular through the use of software simulations and real-time simulations reproducing the specific test case of a five-phase motor.