Project
Numerical simulation of weld pools
Project DescriptionIn the project proposed here, the possibility of simulating reasonably complete mathematical models is to be exploited. These simulations are characterized by complex physical phenomena such as surface tension driven convection, electromagnetic forces, heat and mass transfer, solidification and melting, etc. The surface tension must be modelled realistically, including the solution of convective diffusion equations restricted to the melt - gas interface, for surface concentrations of surface active substances. The Reynolds numbers that are encountered are such that the melt flow is between a rapid laminar flow with thin boundary layers, up to transitional unsteady or even turbulent flow. There are also different timescales in this problem, since the timescale for the melt flow is much shorter than the timescale for melting/solidification. These simulations would typically be either time dependent axisymmetric or steady three-dimensional. The main effort in this project will be to improve the modelling to make predictions of weldability for realistic steels. The aim is to make this sufficiently reliable to be used as a tool for predicting welding properties of new materials. The knowledge that is generated should eventually be made available to industry as improved expert systems or small simulations that can be done online. Apart from development of the physical models, a significant development of the numerical techniques must be carried out in order to deal with the somewhat nonstandard mathematical models in an efficient way. There will be a close coupling between simulations made at KTH and corresponding experiments at AvestaSheffield.
Publications
related to the project
Internal Reports
related to the project
|
|
