Steady and Transient State Performance Balancing Strategy of Adaptive Complex-Vector Filter in PMSM Sensorless Control

In sensorless control of permanent magnet synchronous motors (PMSMs), the complex-vector filter (CVF) attached behind the observer is effective at suppressing harmonics. However, a fixed gain parameter cannot match the requirements of different working conditions, i.e., it cannot balance harmonic elimination and dynamic performance. To deal with this issue, a gain adaptive law of CVF is studied based on the reconstructed objective function. First, the frequency-locked loop (FLL) is embedded in the CVF to increase its stability margin. Then, the reason why the CVF output signal cannot serve as the objective function is analyzed. Thus, an objective function is developed that can correlate the speed error with the CVF gain. Based on this objective function, the gain-adaptive law of CVF is obtained with the gradient ascent method. For the studied adaptive iterative algorithm, its influence on the stability and dynamics of the position estimator system is investigated. The effectiveness of the studied adaptive law is tested by the experiments on a 2.3-kW PMSM platform.