CFD Simulation Approach for Quenching and Thermal Treatment Optimization

Tuesday, May 12, 2015: 10:30 AM
Room 202A (Long Beach Convention and Entertainment Center)
Dr. David Greif , AVL-AST d.o.o., Maribor, Slovenia
The present paper summarizes the numerical approach to simulate heat transfer characteristics during the immersion quenching process. The methodology can be applied to any cast aluminum parts, e.g. landing gears. The functionality is available in a commercial Computational Fluid Dynamics (CFD) code and it efficiently accounts for the boiling mass exchange within the quenchant (water) using the Eulerian multi-fluid model. Boiling occurs due to the heat released from a heated cast part when submerged into the sub-cooled liquid. The applied finite volume method treats solid and liquid domains simultaneously using the so-called multi-material approach. The procedure allows for conjugate heat transfer during quenching. Capturing different boiling regimes is the key for accurate cooling rate prediction. During the film boiling phase the vapor acts as an insulator and the heat removal is slower until the Leidenfrost limit. At this temperature the transition to nucleate boiling occurs and the rate of cooling is accelerated. Below the saturation temperature the solid undergoes single phase, convective cooling. The boiling phenomenon is covered within the liquid domain. The same solver is used to calculate the cooling of the solid, where only the enthalpy equation is solved. Spatial and time dependent temperature fields within the structure are thereafter applied as the input for the Finite Element Analysis (FEA) of thermal stresses and deformations to detect the most favorable quenching configuration. The objective is to minimize the residual stresses resulting from the production process. Numerical prediction of the cooling shows very good agreement with the measured data for two complexity levels of test-type geometries as well as for a realistic cylinder head type component. Application of the CFD results as an input for the FEA analysis is demonstrated.