Font Size: a A A

Study On Failure Of Carbon Fiber Composite Bolted Connection Structure

Posted on:2023-11-13Degree:MasterType:Thesis
Country:ChinaCandidate:S N ZouFull Text:PDF
GTID:2531307097476284Subject:Engineering Mechanics
Abstract/Summary:PDF Full Text Request
Carbon fiber composite materials have become one of the main materials used in aircraft fuselage with excellent properties.However,the composite material connection structure on the aircraft fuselage is prone to aging and damage,which will affect the strength and stability of the aircraft fuselage structure,and seriously endanger the service life of the aircraft and the personal safety of passengers.Improving the strength and stability of composite connecting structures is one of the important contents of aircraft fuselage structure design.Bolt connection has the advantages of good stability,high strength,eas y disassembly and maintenance,etc.,and is one of the commonly used connection methods for aircraft fuselage.However,the bolted connection is prone to stress concentration,damage and rupture.The factors affecting the failure are wide,and there are ma ny kinds of damage forms.Different damage forms often cause nonlinear coupling results,and the mechanism of action is complex.These factors cause great difficulties in the study of bolted connection structures.Therefore,it is of great engineering sign ificance to carry out in-depth research on the failure of the composite bolted connection structure.In this paper,the influence characteristics and laws of the main influencing factors on the strength of composite bolted connections are studied,and a pr ediction model of connection failure strength is successfully established by combining the deep learning method.The main work contents are as follows:(1)The connection models of laminates and metal plates and laminates and laminates are established,and the influence of connection parameters on these two different models is studied.The controlled variable method is used to study the influence of parameters such as lap type(single lap,double lap),bolt type(bore head bolt and countersunk head bolt)on the connection strength.The results show that compared with the single-lap connection,the double-lap connection can greatly improve the strength of the connection structure;Compared with countersunk head bolts,raised head bolts can significantly improve the conn ection strength,and the countersunk head bolt holes are prone to damage,which is the area with the most delamination failures;The strength of the connection structure is determined by the laminate and has no significant relationship with the metal plate.(2)Aiming at the double lap connection form with the best connection performance,a model of different parameter combinations is established to study the influence of typical parameters on the structural strength of the double lap joint with composite studs.Typical parameters include geometric parameters(board width-toaperture ratio,end-to-aperture ratio),number of bolts,and layering methods(layering sequence,unidirectional layering ratio).The results show that the connection strength increases with the increase of the plate width-to-aperture ratio and decreases with the increase of the end-to-aperture ratio;Compared with the single-nail connection,the strength of the double-nail connection structure is significantly improved,while the strength of the three-nail connection structure is close to that of the double-nail connection;When the 90° layup is placed on the outermost layer,the load-bearing strength of the laminate can be improved.(3)Using the deep learning method,the influence weight of each parameter on the connection strength was successfully quantified,and a deep learning failure strength prediction model was established.The model has the advantages of fast calculation speed,high accuracy,and wide application range.It can prov ide reference and guidance for the design of composite bolted connection structures.
Keywords/Search Tags:Carbon fiber composites, Bolt connection, Failure strength, Deep learning, Neural network
PDF Full Text Request
Related items