NIGRO Norberto Marcelo
Assessment of a zero-dimensional model of tumble in four-valve high performance engine
International Journal for Numerical Methods in Heat and Fluid Flow
Emerald Group Publishing Limited
Año: 2007 vol. 17 p. 770 - 787
A predictive 0D-model to estimate tumble, one of the main in-cylinder flow patterns in four-valve pentroof engines, was studied by comparing with CFD simulations using ANSYS CFX 5.7 software. The 0D-model formulation was firstly proposed by Benjamin (Benjamin,1993) in order to compute the tumble and turbulence levels during the compression stroke in disk combustion chambers. The model was later extended by Achuth and Metha (Achuth, 2001) to include the tumble and turbulence generation from the intake period. The 0D-model is based on a mean flow analysis of tumble motion in conjunction with a k-e model to estimate the turbulence. In this model, the main flow phenomena, including shear flow inside the mean vortex, and frictional losses on cylinder walls, are considered. To validate the 0D-model, the intake and compression strokes of a single-cylinder engine with two different intake-port configurations, one of them a conventional intake port and the other a modified intake port in such a way as to promote tumble, were analyzed by CFD. Simulations were performed at 1500 and 2500 rpm considering two different charge levels and both port configurations.   Tumble results from CFD have been shown to be very sensitive to parameters such as the incoming angular momentum through the intake port. On the other hand, 0D-model have not accused significant differences to predict the tumble levels from the two intake port configurations. Besides, 0D-model results were not in agreement with CFD-simulation ones along intake and compression strokes in neither of the two configurations. As regards the intake stroke, the 0D-model overestimates the entrance-flow tumble momentum for the conventional port but underestimates it for the tumble one. Disagreements in tumble estimation during the intake period are due to a poor description of the entering flow direction through the valves. For this reason, a correction of the velocity through the intake valves was added to the 0D-model in order to match the CFD mass flow rate distribution over the valve curtain. With this velocity field correction, only the second port showed an improvement. This analysis allows to conclude that the unique 0D-model presented so far should be applied only in those cases where the port design promotes high tumble and only if extra information about the orientation of the incoming mass flow rate is incorporated.