The interest of functionally graded adhesives (FGA) is growing as it is a mean to increase a bonded joint strength without any modification of the initial design of the adherends. The behaviour of bonded joints with variable adhesive properties along the overlap can be predicted with a potentially time-costly Finite Element (FE) analysis. Dedicated numerical procedures and design tools for FGA bonded joints would increase. The objective of this paper is to offer a mesh-free method for the analysis of functionally graded joints. The technique is based on the macro-element (ME) method and Taylor expansion in power series (TEPS) are used to approach the shape functions of the ME. The method has been developed so far for 1D-bar and 1D-beam kinematics frameworks. This mesh-free_method and a Finite-Element analysis give similar results.

International audiencePurpose - Large structures (e.g. plane, bridge, etc.) often include several hundreds of assembly points. Structural computations often use over-simplistic approximations for these points; among others, they do not take into account the thermo-mechanical history due to the assembling process. Running computations with each assembly point modelled completely would require too much time to achieve a simulation. There is thus a need to create equivalent elements for assembly points in order to: take into account the mechanical state of the assembly point in the design stage - while reducing the computational time cost at the same time. This paper aims to discuss these issues. Design/methodology/approach - This paper introduces an innovative strategy based on a coupling procedure between a finite element tool for modelling the assembly process in order to access to the mechanical state of the assembly point and an optimisation algorithm, in order to identify the equivalent element parameters. Findings - The strategy has proven to be successful. A connector model easier to use and much faster than the complete model, has been obtained. Results obtained with this element are in good agreement with experimental tests in the case of multipoint assemblies and with the simulation results of the complete numerical model. Finally the connector model appears to be easier to use and much faster than the complete model, more difficult to model properly. Originality/value - The main innovative aspects of this strategy lie in the fact that the creation of this equivalent element is based on a complete numerical approach. The thermo-mechanical history due to the assembly process is considered - the element parameters are identified thanks to an evolution strategy based on the coupling between a finite element model and a zero-order minimisation algorithm

No abstract