Computation-Efficient Position Estimation Algorithm for Permanent Magnet Synchronous Motor Drives Under Distorted Conditions
Dianxun Xiao, Shamsuddeen Nalakath, Zekun Xia, Gaoliang Fang, Yingguang Sun, Jason Wiseman, Ali Emadi
Abstract
This article proposes a computationally efficient position-sensorless control scheme for medium- and high-speed permanent magnet synchronous motor (PMSM) drives under distorted conditions. An adaptive comb observer (ACO) is designed for accurate position estimation by eliminating primary distortions simultaneously, including dc offsets from current measurement, odd harmonics caused by inverter nonlinearity, and even harmonics due to the modulation. A frequency-locked loop based on the Lyapunov stability theory is embedded in the ACO to ensure the observer's stability and fast convergence of the estimated variables. Compared with conventional adaptive position observers that usually require high-order filtering networks, the proposed scheme only consists of four delay operators and does not require complicated discretization. Therefore, it can simplify digital implementation and reduce the computational burden of the sensorless algorithm. To validate the effectiveness of the proposed observer, comparative experiment tests of the proposed ACO and conventional methods are performed on a 15-kW PMSM setup.