Study Of Finite Element Model Updating Method For Spacecraft Based On Response Surface Method

Before launch,the satellite needs to be tested in a mechanical environment to ensure its normal operation in orbit.However,the physical satellite test has the disadvantages of long manufacturing cycle,complicated operation and high cost,and the finite element software is used for simulation.The prediction method of analysis can improve the efficiency of analysis and save development costs.However,due to the numerous uncertainties in the analysis and the multiple assumptions introduced in the modeling,the results of FEM simulations are often There are certain discrepancies with the real test data.Therefore,in order to reduce the bias predicted by the dynamics,the initial finite element model needs to be modified based on the experimental measurements to a more accurate finite element analysis(FEA)model is obtained to replace the actual structure for analysis and response prediction.In this paper,a parametric correction method based on the quadratic polynomial response surface model is proposed,taking into account the uncertainty factors in the actual project.An interval correction method based on the kriging model is proposed.The details of the study are as follows.Firstly,the basic process of model correction and the preparation work before correction are introduced.This includes parametric batch modeling simulation analysis based on ABAQUS and the correlation analysis between the test data and simulation results.In the model correction section,the response surface method is proposed to correct the model.The basic theory of the response surface method and the correction steps are introduced,including the experimental design method,the response surface form,the evaluation of the effectiveness of the response surface,and the analysis of the model correction.The objective function and optimization of the solution links are presented.A correction method using orthogonal test design and selection of quadratic polynomial response surface is proposed.The feasibility of the correction method is verified by simple arithmetic examples.Secondly,based on the quadratic polynomial response surface model,the finite element model correction of the satellite ontology model is carried out.The finite element model of the satellite is established,and the complex structure is reasonably simplified or equivalently processed.The modal analysis is carried out to obtain the natural frequency and vibration type;correlation analysis is carried out on the data of test and simulation results,and the initial selection of parameters is carried out.ANOVA was performed to screen the highly significant factors as the final parameters to be corrected,and the quadratic polynomial response surface method was used for model correction.The correction effect is evaluated,and the results show that therelative error of the first 6 orders of natural frequencies is reduced from 8%to less than1%after the correction,which proves that the the feasibility of the correction algorithm is investigated.Finally,a study of the uncertainty model correction method is carried out considering the uncertainty problems that arise in practical projects.The kriging response surface model is proposed to solve the response surface model uncertainty problem caused by the lack of sample points and the probabilistic correction and interval correction methods to solve the parameter uncertainty problem.And take the satellite star ontology model as an example,considering the response surface form uncertainty and parameter uncertainty,and adopt the method based on the the kriging model is corrected by the interval correction method.The correction results of the two correction methods are also compared.The results show that the correction effect of kriging model is better than quadratic polynomial response surface model.In addition,the correction method which takes into account the uncertainty of parameters is more practical and has a good prediction effect.