Rapid inverse design of mechanical seal groove based on kriging surrogate model

Abstract Aiming at the high requirement for personnel experience, time-consuming calculation, and the difficulty in providing many alternative combinations of geometric parameters in the simulation design of mechanical seal grooves, a rapid inverse design method was established in this paper. This method uses the universal Kriging surrogate model to improve the computing efficiency of mechanical seal grooves’ performance, and combines the elitist genetic algorithm and parallel computing to realize the rapid inverse design of mechanical seal grooves. After the input of the target performance, the corresponding geometric parameters can be quickly given, and the specific recommendation sequence can be given at the same time. At the case test point, the average relative error of each performance of the top 8 groups of recommended geometric parameters is 2.70%.