Sampling-Delay Error Compensation for Low-Speed Sensorless Control With Single-Current Sensor Based on Multiple-Branch Sampling

This article proposes a sampling-delay error compensation for high-frequency (HF) injection sensorless control based on multiple-branch sampling at zero and low-speed domains. In existing single-current sensor sampling schemes, sampling-delay errors are inherently present due to two samplings occurring at different moments. These errors deteriorate the accuracy of HF current reconstruction and constrain the injection frequency, thereby limiting system bandwidth. To address this issue, a model-based compensation strategy for sampling-delay errors is proposed. First, a low-speed motor model is designed to predict the current change rates within half a switching period, which avoids cumulative errors from position estimation and reduces computational complexity. Second, considering that dead-time effects cause discrepancies between actual and calculated vector durations, vector duration is compensated based on current states and reference voltage vector sectors. Finally, since instantaneous currents required for current change rate prediction cannot be directly sampled, this article employs a stepwise prediction method to calculate half-cycle current change rates at the asynchronous sampling point, realizing effective error compensation. Experimental results demonstrate that the proposed compensation strategy can reconstruct HF currents accurately, enabling sensorless control with higher injection frequencies and consequently improving system bandwidth.