Encoderless Deadbeat Predictive Current Control for SPMSM Drives With Online Parameter Identification

The encoderless deadbeat predictive current control (DBPCC) has been widely applied in surface-mounted permanent magnet synchronous motor (SPMSM). However, the DBPCC suffers from poor performance against motor parameters mismatches. An improved encoderless DBPCC method with online parameter identification is proposed to enhance parameter robustness in this article. First, a modified rotor position estimator is directly designed in the discrete-time domain. No observer is required to estimate back electromotive force (EMF). Hence, the proposed EMF estimator features simple structure and easy implementation. Then, parameter sensitivity of the encoderless DBPCC method is analyzed. A simple demodulation method is proposed to identify motor parameters based on the injected high-frequency (HF) current signals in the d-axis. The proposed online parameter identification algorithm would not introduce the extra torque harmonics in SPMSM. Meanwhile, theoretical analysis shows that the proposed parameter identification method presents high precision even with the severe initial motor parameters' mismatches. Therefore, the mismatched parameters used in the encoderless DBPCC method can be replaced by the identified motor parameters to precisely correct the rotor position estimation errors. Finally, the experiments were tested in a two-level inverter-fed SPMSM drives to validate the effectiveness of the proposed encoderless DBPCC algorithm with online parameter identification.