Comparative evaluation of meshfree data transfer methods in hydromechanical problems

Abstract

The investigation of the combined effect of diverse physical phenomena in engineering problems through multiphysics numerical analyses is a trend in many challenging engineering problems. To increase the efficiency of the numerical procedures, multiphysics analyses can be carried out under uncoupled or pseudo-coupled approaches, which frequently require the transfer of data between two non-matching spatial discretizations. Hydromechanical analyses are often executed through pseudo-coupling schemes once each phenomenon requires a different level of discretization or even different numerical approaches. Using appropriate interpolation methods that guarantee the accuracy and conservation of the results is challenging. Methods based on mesh-free interpolation have been broadly used since they show good accuracy and conservation and are topology independent. This work studies the performance and influence of three mesh-free interpolation methods on the response in pseudo-coupled hydromechanical simulations with different meshes. The Moving Least Squares reconstruction, the Spline Interpolation method, and the Inverse Distance Weighting method are assessed regarding accuracy and conservation. An auto-search of the compact support radius scheme based on a minimum number of source points inside the compact support is proposed. The numerical results show the high dependency of the transfer quality on the compact support size. The proposed support size adaptation is shown to be advantageous for the moving least squares method. The inverse distance average method is a promising method for hydromechanical problems.