2026/05/25 by Xiaozhuo Xu, L S Wang, Liangjie Wang +2
Engineering · #Advanced DC-DC Converters #Multilevel Inverters and Converters #Sensorless Control of Electric Motors
paper · doi:10.1109/tpel.2026.3696650
openalex publication_date 2026/05/25 · openalex created_date 2026/05/26 · openalex updated_date 2026/07/30
Deadbeat Predictive Current Control (DPCC) attains enhanced robustness via model-free predictive schemes with an ultra-local model. However, the bandwidth limitations of the observer and the controller gain's dependence on inductance parameters impair control performance. To address these challenges, this paper proposes a Novel Model-Free Deadbeat Predictive Current Control (NMFPCC) strategy for PMSM. The primary contributions are twofold. First, a novel Sliding Mode Observer (SMO) is designed to replace the traditional Extended State Observer (ESO). By employing a variable exponential reaching law, the proposed SMO avoid the linear bandwidth constraints of ESOs, enabling fast disturbance estimation while effectively mitigating high-frequency chattering. Second, to resolve the gain mismatch issue, a novel gain identification algorithm is developed. This algorithm estimates the reciprocal of the inductance in real-time and adaptively updates the controller gain, thereby decoupling the control system from parameter mismatches. Experimental results validate that the proposed NMFPCC strategy solely relies on input-output data and achieves superior robustness and dynamic response compared to existing methods, maintaining high control accuracy even under significant parameter variations.