Efficient multiple time scale method for modeling compressible vapor plume dynamics inside transient keyhole during fiber laser welding

Efficient coupling modeling of multiple time scale interactions between keyhole, weld pool and compressible vapor plume during laser welding has long been limited. To address this problem, we present a highly efficient multiple time scale method combining a novel dual-time stepping and Ghost Fluid interpolation strategy with incompressible and compressible fluid solvers, which allows us predicting the compressible plume dynamics inside transient keyhole in fiber laser welding for the first time. In our method, the compressible dynamic vapor inside the transient keyhole is solved with a Roe scheme based algorithm and the incompressible molten liquid of weld pool is calculated by a Projection method. A novel temperature dependent boundary condition of vapor plume is also proposed for the consideration of the dynamic evaporation phenomena on the transient keyhole wall. It is found that the time dependent distributions of vapor plume characteristics, including temperature, pressure, velocity, density and Mach number distributions inside the transient keyhole induced by laser welding can be reasonably predicted by comparing to experimental and literature data. It is also shown that the proposed multiple time scale method is around 60 times faster than the vapor plume modeling method using a single nanosecond scale time step. For the vapor plume in a typical fiber laser welding process, the results indicate that the peak pressure can be greater than 2.0 atmospheric pressures; the average density is around 0.15-0.3 kg/m(3) which is much smaller than the air density; and the local Mach number can be greater than 0.8 or even 1.0 Mach which demonstrates the necessity to treat the vapor plume as a compressible fluid. (C) 2015 Elsevier Ltd. All rights reserved.