Study On Process Parameters Of Slug Rivet Interference-fit And Structural Optimization Of Rivet Die

Advanced assembly and connection technology are the key technique problems needed to be urgently solved in China's aircraft manufacturing.Slug rivet interference-fit can realize the homogeneous deformation along the rivet shank,which can meet the quality requirements of long fatigue life,high reliability and good sealing performance.Slug rivet interference-fit technology has become the major measure to improve the riveting quality and fatigue performance of a riveted lap joint.But the study of slug rivet connection starts late in China and the application level is very low.The related studies,which lack of theoretical support and methodological guidance,have a strong dependence on foreign technology.With the rapid development of the design and manufacture technologies of large aircraft in China,it is particularly urgent to improve the riveting quality and fatigue performance of a riveted lap joint.This thesis studies the slug rivet connection technology using theoretical analysis,numerical simulation,intelligent optimization and experiment methods.The experiments were conducted based on the automatic drilling and riveting machine developed by Zhejiang University.Theoretical model is built to make a pre-phase analyses.Then FE models are established to analyze the effects of process parameters and riveting die structure on riveting quality.The purpose of this thesis is to find out the appropriate parameters and riveting die structure to enhance the interference value as well as its distribution condition.The results of this thesis will be helpful to offer a theoretical support and methodological guidance to improve the panel riveting quality.This thesis describes the research background and significance at first.The research status and application condition of slug rivet interference-fit used in aircraft manufacturing are analyzed.The deficiencies existed in the study are summarized as well.And the study content and total frame of this thesis are proposed based on the project background.According to the study content,this thesis is divided into three parts:riveting modeling,parameter analysis and parameter optimization.Riveting modeling is the foundation of this thesis,which contains two targets.The theoretical model based on countersunk hole structure is formulated to map the relationship between squeezing force and riveting quality.Meanwhile,the FE models are built to simulate the riveting process.To provide a appropriate model for further study,the relationship of model size,operating efficiency and analysis precision between different FE models are analyzed.Similarly,parameter analysis contains two study targets as well.Firstly,to provide an appropriate parameter selection range,this thesis,considering the key parameters during the riveting process,focuses on the structural parameter(such as rivet geometry)and process parameter(such as squeezing force)which would have significant influences on the panel riveting quality.The results show that the interference condition should be considered as the main quality assessment criteria.Only in some particular situations,the driven head dimension can be regarded as the main quality assessment criteria.Secondly,this thesis studies the radial force of the driven head during the riveting process.To find out the appropriate riveting die structure,the association relationship between different types of riveting dies and the corresponding parameter combinations are analyzed.Experimental study is conducted to verify the validity of the numerical simulation.Besides,to increase the degree of modeling automation and intelligence,a parameterized FE model is established by using ABAQUS software as second development platform.The riveting quality is significantly improved using the appropriate riveting dies.The results further indicate that it is necessary to use the interference condition as the main quality assessment criteria.Then,this thesis proposes the optimization objective and begins to optimize the structure parameters of riveting die using machine learning method.A regression model is used to predict interference value.The genetic algorithm is developed to find out the appropriate structure parameters.And the structure parameters are substituted into the regression model to predict the corresponding interference value.The results point out that,for the optimization problem with multi-factor combined effects,it would be better to train the optimization model based on the data obtained from multivariate analysis.To avoid the uncertainty of regression model,this thesis combines the parameterized FE model and genetic algorithm to further optimize the structure parameters of riveting die.The parameterized FE model is applied to replace the regression model to directly calculate the interference value.The interference value after optimization reaches a 49.31%increase compared with the data before optimization under the same condition.The interference distribution condition after optimization is satisfactory as well.Experiment study verify the validity of optimization results.Finally,the main content of this thesis is summarized,and the orientation of further study is put forward.