Experimental and transient computational fluid dynamic analysis of vehicle underhood in heat soak
Simulation-based analyses of underhood compartments are proving to be crucially important in a vehicle development program, reducing test work and time-to-market. While Computational Fluid Dynamics (CFD) simulations of steady forced flows have demonstrated reliable, studies of transient natural convective flows in engine compartments under thermal soak are not yet carried out due the high computing demands and lack of validated work. The present work assesses the practical feasibility of applying the CFD tool at the initial stage of a vehicle development programme for investigating the thermally-driven flow in an engine bay. A typical vehicle underhood was reproduced in half-scale for laboratory investigations. Surface temperatures of components, airflow patterns induced by the buoyant forces as well as the spatial distribution of the air temperature were measured under both steady and transient thermal conditions. Temperature mappings were obtained with thermocouples whereas airflow magnitudes and directions were determined with Particle Image Velocimetry (PIV) instrumentation. The detailed measurements were used as reference for validating the corresponding CFD simulations carried out with the software VECTIS. Experimental and numerical data correlated well in steady state, both quantitatively and qualitatively. A computation procedure that enables pseudo time-marching simulations to be performed with significantly reduced CPU time usage, in comparison to traditional fully-conservative transient simulations, was also developed. The methodology used a unique combination of CFD solver parameters to overcome the computationally challenging problem of solving for momentum transport in time-marching mode and for a long period of physical time. The procedure was successful in providing a detailed and time-accurate flow and thermal simulation of the underhood model during transient cooling. Such simulation would not have been practically feasible with a standard transient simulation. A reduction in CPU processing time in excess of 90% was achieved with good correlation between the CFD predictions and the experimental data.