Predictive Current Control and Field-Weakening Operation of SPMSM Drives Without Motor Parameters and DC Voltage

The prior model-parameter-free predictive current control (MFPCC) applies a 1-D lookup table (1-D LUT) to store current variation, where the index input is the number of the basic voltage vector. However, current variation is not only related to the input voltage but also depends on the rotor position. As the influence of rotor position is not considered in the prior MFPCC, high current spikes may occur during normal operation. Aiming to improve the steady-state performance, the proposed MFPCC utilizes a 2-D LUT to backup current variations, where the second-dimensional input is the discretized rotor position. The proposed method keeps good robustness and eliminates current spikes in the prior MFPCC. Moreover, a field-weakening (FW) method is developed to extend the speed range. As no modulator is employed in MFPCC and only discrete voltage vectors are available, the conventional voltage magnitude-based FW control scheme cannot be directly applied. To address such issue, a complex-coefficient filter is employed to calculate the modulation index of the fundamental voltage, which is then fed to the FW controller to generate demagnetizing current. The proposed MFPCC can achieve satisfactory performance without motor parameters and the measurement of dc voltage. Experimental results confirm the effectiveness of the proposed method.