Font Size: 

As part of the project on modelling delaminations at the yarn-matrix interface of woven SiC/S iCcomposite, a RPIM meshfree frictionless contact code was built from scratch. At the first stage of the software design a 2D and 3D model for elastic small-strain single body problems was created. In the verification and convergence process, examples of a cube under uniform compression and a cantilever beam under a bending force were used to optimise the main parameters involved. The radial basis function method used for this study was the multi-quadric and its inverse, which contain two shape parameters of α_c and q to optimise. In addition to α_c and q, the use of polynomial terms for computing
the shape functions, the number of the field nodes in the support domain and the total number of field nodes in the problem domain were studied. It was found that using the first order polynomial terms in the shape function calculations leads to an overall error reduction but it also changes the terms of the optimisation process. Because while the optimised values of α_c and q for the beam bending example matched with and without polynomial terms, but for the cube under compression a different optimisation was obtained, which meant there was no generalised optimum value for all the applications. In order to investigate this further few other examples such as pressure inside a thick pipe, sphere under compressive force and few more are being studied so a more generalised
understanding of the optimisation can be reached. This optimisation will further be tested in the context of frictionless contact, which can then be applied to modelling the woven composite.

radial point interpolation, mesh free, multi-quadric, optimisation