Identification of PMSM Parameters With Time-Error Compensated Based on Contractile Factor Antipredator PSO

The identification error of permanent magnet synchronous motor (PMSM) parameters is influenced by the dead time of the voltage source inverter (VSI). In order to solve this problem, the contractile factor antipredator particle swarm optimization (CFAPSO) is proposed and applied to the identification of parameters of the PMSM. In the CFAPSO algorithm, the negative sequence current is injected into the
$d$
-axis, so the under-rank problem of the equation is solved, and the full-rank equation is constructed. After establishing the full-rank equation, a time error compensation method is proposed to compensate for the dead time of VSI; the real-time recognition model is more accurate, and the recognition error is reduced. In addition, the contractile factor is introduced to restrict the learning factor. In this way, the algorithm trapping in local optima is avoided, and the recognition accuracy is enhanced; furthermore, an antipredator particle swarm algorithm is added to exclude the worst solution position, more unknown regions are explored, and the optimality-seeking accuracy is improved. Finally, the simulated and experimental results are used to verify the correctness and effectiveness of the algorithm.