No abstract

Reported in this paper is the development and verification of a finite element model with the fewest solid elements that can predict the non-linear stiffness characteristics of adhesively bonded single lap-joints in vehicle bodies. This work was driven by the need to significantly reduce computing hardware resources and run times for whole body analyses, and to achieve this goal the lap-joint needs to be modelled by a 'small' number of shell elements. It is wellknown that the deformation of bonded lap-joints is dependent on seven key parameters, and that it is impractical to have a comprehensive characterisation of these by physical testing alone. To gain further understanding of the influence of these parameters on joint stiffness, over a wider range of variables than could be practically achieved in the laboratory, the ANSYS finite element code is used to simulate the highly non-linear (geometric and material) response of bonded joints. The authors use the laboratory results for joint stiffness from nineteen batches of laboratory specimens to aid the development, and verification, of numerical derived stiffness curves from a solid element model. Initially, a very refined mesh is used. This model is developed to have a 'coarse' solid element mesh that minimises run times without the calculated joint stiffness deviating by more than 10% from the batch mean * Corresponding author: E-mail I.T.Pearson@warwick.ac.uk Page 2 of 10 measurements from laboratory testing. Numerical results from many simulations using the 'coarse' solid mesh model are used to show that, for the shell modelling methodology (reported in Part 2 on this work) to be successful, a representative model must account for the four key parameters of: adherend stress-strain relationship, adherend thickness, bond line thickness, and the over lap length. The ANSYS results also confirm that stiffness is directly proportional to joint width.

No abstract