Advances in Numerical Investigation of Immersion Quenching at Different Pool Temperatures

摘要

This paper outlines an improved computational methodology to simulate the immersion quenching heat transfer characteristics. Main applicability of the presented method lays in virtual experimental investigation of the heat treatment of cast aluminum parts, above all cylinder heads of internal combustion engines. The boiling phase change process between the heated part and a sub-cooled liquid domain is handled by using the Eulerian multi-fluid modeling approach, which is implemented within the commercial Computational Fluid Dynamics (CFD) code AVL FIRE. Solid and liquid domains are treated simultaneously. While for the fluid domain mass, momentum and energy equations are solved in the context of multi-fluid modeling approach, only the energy equation is solved to predict the thermal field in the solid region. For the presented quenching simulation, the solid and fluid parts are contained in a single domain. This approach is known as AVL FIRE Multi-Material approach, where the surface temperature and local heat transfer coefficients are exchanged after each iteration and no longer after each time step as in the previously utilized ACCI (AVL Code Coupling Interface) method [1, 2, 3, 4, 5]. The applied heat transfer model utilizes an empirically correlated heat transfer coefficient, which is changing within different boiling regimes (from film to transition boiling), controlled by the variable Leidenfrost temperature. Preliminary results of the variable Leidenfrost temperature are presented, with additional interfacial forces such as lift and wall lubrication force, which were added in the context of momentum interfacial exchange terms. In the present research, the objective was to compare the simulation results with experimental data for different pool temperatures. Solid side temperature measurements along the height of the so-called step plate test piece, featuring different thicknesses along its length, were performed at different positions. The temperature histories predicted by the presented model correlate very well with the provided experimental data.