针对尺寸参数摄动作用下滑阀副零位泄漏量的不确定性，建立了考虑参数摄动的滑阀副零位泄漏量数学模型，获得了尺寸参数摄动下的滑阀副零位泄漏量分布规律。进一步地，建立了工艺参数与零位泄漏量分布特征参数之间映射关系的Kriging代理模型，并基于代理模型利用Sobol法开展了零位泄漏量分布特性对工艺参数的全局灵敏度分析，最终为减小内泄漏量并提高一致性提供了理论依据。研究结果表明，尺寸参数摄动下的零位泄漏量服从正态分布规律，工艺参数之间的交互作用对滑阀副零位泄漏量分布特性的影响可以忽略不计。同时，相比轴向尺寸工艺参数，滑阀副零位泄漏量分布特性对径向尺寸工艺参数和节流边圆角半径工艺参数的变化更为敏感。滑阀副径向间隙最小值对滑阀副零位泄漏量平均值影响最大，而阀芯的节流边圆角半径最大值对滑阀副零位泄漏量的一致性影响最大。滑阀副内泄漏量实验结果表明零位泄漏量呈现出一定的波动性，与理论模型总体预测趋势相一致。

Aiming at the uncertainty of internal leakage for spool valve couple at zero position effected by the size parameters perturbation， a mathematical model of internal leakage at zero position considering parameter perturbation is established. Then， the distribution characteristics of internal leakage at zero position under the perturbation of size parameters are obtained. Furthermore， a Kriging surrogate model of the mapping relationship between process parameters and distribution characteristic parameters of internal leakage at zero position is established. Based on the Kriging surrogate model， the global sensitivity analysis of distribution characteristic of internal leakage at zero position to process parameters is carried out by Sobol method， which can provide a theoretical support for reducing internal leakage and improving consistency. The results show that the internal leakage at zero position under the size parameters perturbation obeys normal distribution law. Influence of the interaction among process parameters on the distribution characteristics can be ignored. Compared with the axial size process parameters， distribution characteristics are more sensitive to the variation of radial size process parameters and throttling edge corner radius process parameters. Minimum radial clearance of spool valve couple has the greatest influence on the average value of internal leakage at zero position， while the maximum corner radius of throttle edge of spool has the greatest influence on the consistency of internal leakage at zero position. Experimental values of the internal leakage at zero position of spool valve couple present a certain fluctuation， which is consistent with the overall prediction trend of the theoretical model.