Research On Numerical Simulation Technology Of Micro Contact Theory And Application In Sealing

Sealing is an important mechanical instrument in spacecraft,sealing performance has a great impact on spacecraft life and mission.In order to ensure the success of space missions,the study of sealing mechanism is very necessary.In this paper,the simulation of surface morphology,multi-scale contact mechanics,leakage percolation threshold and leakage factors were carried out.The main contents and contributions of this paper are as follows:(1)Basis of Sealing Contact Simulation Researchsurface morephology research is the basis of contact mechanics simulation,three kinds of common fractal surface simulation methods are introduced,and several sets of fractal rough surfaces with different parameters are generated by using Random Midpoint Displacement method.The fractal characteristics are tested,which laid a foundation for the follow-up simulation of multi-scale contact mechanics.(2)Research on Sealing Contact Simulation AlgorithmIn the study of multi-scale contact mechanics simulation,this paper proposes a sliding window-based Gauss-Seidel algorithm based on the study of the distribution of the influence coefficients,which to cope with the problem that traditional algorithms are slow and hard to converge.The proposed algorithm reduces the dimensiona of the linear equations,gets a good precision and a nice efficiency.Finally,the contrastive experiments are carried out based on the Persson contact mechanics theory to verify the effectiveness of our algorithm.(3)Research on Percolation Algoithm of Random Rough Surface Leakage ProblemIn percolation threshold study,percolation discriminant algorithm is an important research point,the traditional method is through the depth-first search(DFS)algorithm to determine,and when the data is gettinglarger,DFS loses its efficiency.We redifine the percolation discriminant problem as a two-labels classification problem of two-values image,propose a new algorithm based onConvolutional Neural Network model to solve the problem.Through a series of simulation experiments with different settings,the optimal model hyperparameters and training sample numbers are obtained.The results show that the proposed model is superior to the traditional one.(4)Research on Influence Factors of Leakage between Random Rough SurfacesIn the study of the influence factors of leakage,the leakage rate and the size of the leakage channel are analyzed by Persson model.The results show that the leakage rate and leakage channel average height will decrease rapidly with the compressive force decreasing,but they can not reduce to 0.And the surface roughness smaller,the reduction more rapidly,this conclusion is consistent with the actual phenomenon.