Adaptive parameter identification based on dead-time compensation for Permanent Magnet Synchronous Machines for the 2011 11th international conference on control, automation and systems (ICCAS 2011)

The objective of this paper is to investigate an accurate on-line parameters identification method for PMSM. Because a majority of controllers are based on motor models, the parameters of stator resistance and dq-axis inductances play important roles for achieving high-performance control. However, voltage drop between the reference and output voltage in a pulse width-modulation (PWM) voltage source inverter affects the accuracy of identification. In this paper a model-reference adaptive identification method is proposed with a simple method for compensating the voltage drop of inverter in the driven system. The global stability of adaptive laws is proved by the Popov's hyper-stability theory. Simulation results proved that the estimated stator resistance and dq-axis inductances converge to the constant after the rotor speed becomes stable. The compensation method introduced in the MRAS is easy to implement without any additional circuits. Experimental results demonstrate the effectiveness of the proposed identified method and the compensation method.