Abstract: The calculation accuracy of traditional Taguchi method is not high enough to support assembly tolerance analysis and assembly success rate estimation， which could not adapt to the demand of high-precision， high-efficiency， short-cycle， multi-assembly dimensions of tolerance design in the manufacturing process of aero-engine. Aiming at this problem， a kind of assembly success rate calculation model based on improved Taguchi method is proposed. Combined with Pearson distribution theory， this calculation model is set up by using the improved Taguchi method to calculate the first four moments of statistics. On the basis of theoretical analysis， the proposed method is applied to calculate a turbine rotor’s assembly success rate. The results show that compared with Monte Carlo simulation results， the improved Taguchi method reduces the relative error from the original 19.38% to 0.21% when estimating the kurtosis of the assembly gap， and the relative error of the assembly success rate obtained by the proposed method is 0.02%， and the effectiveness of the proposed method is verified.

Key words: Aero-engine；Tolerance analysis；Taguchi method；Pearson distribution family；Assembly success rate

摘要： 考虑到运用传统的Taguchi方法在进行公差分析和装配成功率计算时，在计算精度上的不足，无法满足航空发动机研制过程中公差设计高精度、高效率、短周期、多组成环等要求。针对该问题，文中提出一种基于改进的Taguchi方法的装配成功率计算模型。利用改进的Taguchi方法对统计数据的前四阶矩进行计算，并结合Pearson分布理论建立装配成功率的计算模型。在理论分析的基础上，以某型航空发动机涡轮转子的装配成功率为例进行计算。结果表明:与蒙特卡洛仿真结果对比，改进的Taguchi方法在对装配间隙的峰度进行估计时使相对误差由原来的19.38%降低到0.21%；用本文所提方法得出的装配成功率与蒙特卡洛仿真结果的相对误差仅为0.02%，计算结果验证了该方法的有效性。

关键词: 航空发动机；公差分析；Taguchi方法；Pearson分布族；装配成功率