新能源变流器在电网故障恢复阶段易反复进出低电压穿越（low voltage ride through, LVRT）控制模式，由此引发的切换型控制诱导电压振荡问题日益突出。该现象具有典型的离散与非线性特性，导致系统多参数可行域难以通过传统解析方法准确刻画。为此，本文提出一种基于非迭代核学习的多参数可行域计算方法，旨在解决反复LVRT诱导的电压振荡问题。首先，建立计及LVRT控制的变流器并网系统模型，在此基础上对比分析切换型控制主导的电压振荡与小干扰稳定问题在机理上的本质差异；其次，采用点二列相关系数法量化系统参数与反复LVRT行为之间的关联程度，识别影响系统稳定性的关键参数；最后，构建非迭代核学习分类模型，实现对反复LVRT现象的准确判别与关键参数可行域的高效刻画。时域仿真对比结果表明，所提方法在可行域计算中具有较高的准确性与有效性，为新能源变流器反复LVRT问题的分析与抑制提供了新的解决途径。

During the grid fault recovery stage, renewable energy converters may repeatedly enter and exit the low voltage ride through (LVRT) control mode, leading to increasingly prominent voltage oscillations induced by switching-type control. This phenomenon exhibits distinct discrete and nonlinear characteristics, making it difficult to accurately characterize the multi-parameter feasible region using conventional analytical methods. To address this issue, this paper proposes a multi-parameter feasible region calculation method based on non-iterative kernel learning, aiming to resolve the voltage oscillation problem induced by repeated LVRT transitions. First, a grid-connected converter system model incorporating LVRT control is established. On this basis, a comparative analysis is conducted to reveal the essential differences in mechanisms between voltage oscillations dominated by switching-type control and small-disturbance stability issues. Then, the point-biserial correlation coefficient method is employed to quantify the relationship between system parameters and repeated LVRT behavior, thereby identifying the key parameters affecting system stability. Finally, a non-iterative kernel learning-based classification model is constructed to achieve accurate identification of repeated LVRT phenomena and efficient characterization of the feasible region of key parameters. Time-domain simulation results demonstrate that the proposed method achieves high accuracy and effectiveness in feasible region computation, providing a new solution for analyzing and mitigating the repeated LVRT problem of renewable energy converters.