关键词: Vienna 整流器, 快速排序, 自适应选择控制策略, 无模型预测

Abstract: An adaptive weight-free predictive control strategy is proposed for Vienna rectifiers, aiming to suppress neutral-point voltage imbalance and input current zero-crossing distortion while avoiding the difficulty of weighting-factor tuning. An ultra-local model is constructed to replace the conventional mathematical model of the rectifier,thereby reducing the dependence on model accuracy and enhancing robustness. To alleviate the computational burden caused by weightingfactor selection, a weight-free parallel cascaded cost function is proposed, in which the conventional cost function is decomposed into multiple sub-terms,and the optimal switching vectors for current-axis and neutral-point voltage constraints
are obtained separately via a fast sorting algorithm. An adaptive selection mechanism is designed to identify the intersection of optimal switching vectors,which serves as the constraint input for axis currents to suppress zero-crossing distortion and enhance both steady-state and dynamic performance. Experimental results validate that the proposed strategy outperforms conventional model predictive control across all evaluated metrics. Specifically,the total harmonic distortion of the current is reduced to 3. 16%,and the DC-bus voltage recovery times under step-change tests are shortened to 161 ms and 194 ms. Furthermore,the voltage fluctuation is reduced by 0. 9 V under load disturbances,while the current fluctuation decreases by 0. 14 A under inductance parameter mismatches.

Key words: vienna rectifier, quick sort, adaptive switching control strategy, model-free prediction