An improved AK-MCS-based reliability analysis method for gear systems

Abstract

The reliability of gear transmission systems critically depends on their weakest components, particularly the maximum contact stress locations in gear teeth. It is difficult to precisely and accurately identify the weak meshing position of the gear transmission system and its maximum stress moment, which is due to the presence of assembly and manufacturing errors having a visible influence on the weak links of the gear system. This will increase the complexity as the interaction occurs during the meshing transmission process. In order to find out the weak links in the meshing of gear system efficiently and accurately, this paper proposes a reliability analysis methodology. The approach integrates Active Kriging Monte Carlo Simulation (AK-MCS) with a parallel sampling strategy (Pseudo-Kriging method) to develop an active learning framework with parallel adaptive sampling. This innovative methodology leverages the unique capability of Kriging models to quantify prediction uncertainty through variance estimation, thereby establishing a systematic approach for analyzing critical meshing zones in gear transmission systems based on Kriging prediction variance. Analyze the planetary gear transmission system of a certain aviation helicopter as the object. Comparative analysis with uniform sampling methods demonstrates that the approach not only accurately identifies critical meshing zones during planetary gear system operation but also achieves significant computational efficiency improvements.