Research On Anti-loosening Performance Prediction Method Of Screw Connection Structures For Automobile

Screw connection is widely used in assembly of automobile parts,and its performance has an important influence on the safety and reliability of the whole vehicle structure.Looseness is one of the common failure forms of screw connection in service,which leads to leakage,abnormal sound,fracture and other faults,and even becomes the main cause of overall structure damage and failure,leading to product failure and even major safety accidents.Therefore,it is of great significance and engineering application value to study the prediction method of anti loosening performance(critical loosening condition)of screw connection structures.However,the key screw connection structures of automobile is subjected to composite dynamic loads under complex working conditions,and there are different clamping,loading forms and slip States,which brings great challenges to accurately and efficiently predict and verify the anti loosening performance of screw connection.Based on this,this paper combines mechanics(finite element analysis),support vector machine and explicit surrogate model method,and uses the advantages of the three methods to improve the prediction accuracy and calculation efficiency of anti loosening performance.The main research work is as follows.(1)Structural characteristics and mechanical analysis of automobile screw connection based on dynamic load acquisition.Aiming at the problem of high difficulty and large error in obtaining the dynamic loads of screw connection structures under vehicle driving condition,the established vehicle dynamic model is calibrated through load calibration test and road test to simulate vehicle driving condition,so as to ensure accurate dynamic loads.On this basis,the characteristics and mechanical behavior of automobile screw connection structures under dynamic loads are analyzed,which provides data and theoretical basis for the subsequent research on screw connection performance prediction method.(2)The numerical calculation model of screw connection performance for engineering practice is established.In engineering practice,Due to the complexity of dynamic composite load and influencing factors,it is difficult to accurately predict and verify the performance of screw connection,the mechanical model of the screw connection structures is constructed by comprehensively considering the combined action of dynamic loads,the interface slip condition of the connected parts,the micro stress state of the screw surface,and the influence of eccentric clamping and eccentric loading on the elastic springback and load coefficient of the connection structures.The composite stress of bolt can be accurately calculated by measuring the application coefficient of shear stress,and the numerical calculation model between structural parameters,working loads and joint performance is established,which expand the application scope of the numerical calculation model.Finally,the accuracy and reliability of the model in engineering practice are verified by the test methods of critical rotation loosening residual preload and yield/fatigue failure-non failure assembly preload.(3)Based on sample fusion feature weighted and multiple kernel function weighted support vector machine,a loosening failure detection method is established.Aiming at the problem of low prediction accuracy of standard support vector machine for anti loosening performance,the sample features affecting anti loosening performance are fused and dimensionality reduced based on the numerical calculation model,the weight of feature vector is calculated based on the combination of prior knowledge and information gain considering the influence of feature weight on mixed kernel function,the principle of feature distance minimum-maximum is used to establish the deterministic calculation method of mixed kernel function weight,in order to improve the accuracy of weight calculation and feature selection.The accuracy and superiority of the new support vector machine are verified by the sample database,which provides a new idea to solve the problem of anti loosening performance prediction.(4)The prediction method of thread connection anti loosening performance based on fusion model is constructed.In order to further improve the prediction accuracy of anti loosening performance,the mechanical model,finite element analysis and support vector machine are integrated,and the advantages of the three methods are effectively used to achieve accurate prediction of anti loosening performance.Through the mechanical model and finite element simulation,the quantitative relationship among structural parameters,loads,friction moment and rotation torque is studied,and the friction moment and rotation torque are taken as the eigenvectors of support vector machine,so as to solve the problems of support vector machine with many dimensions and relatively low prediction accuracy.At the same time,support vector machine is used to reduce the difference between mechanical model,finite element analysis and actual output results in order to improve the prediction accuracy of anti loosening performance,a new method is proposed to solve the problem of anti loosening performance prediction.(5)An explicit proxy model prediction method with high computational efficiency is established.In order to improve the efficiency of structural failure prediction and optimization in engineering practice,the key sample features are used as the input variables of the surrogate model.Taking advantage of the high prediction accuracy of the fusion model,the critical looseness preload(as output response)of the screw connection structures is calculated based on the interpolation method and the fusion model,and the explicit surrogate model between the key features and the critical looseness preload is constructed.The method of failure mode prediction,performance optimization and cost optimization is proposed.The forward design and development of screw connection structures is realized,which provides intuitive and efficient design basis and technical support for engineering application.