The Effect of Intake Flow Modelling on Flame Front Shape and Its Displacement in Cylindrical Combustion Chamber

In this paper some results concerning the effect of intake flow modeling obtained by 3D numerical modeling of combustion on flame front shape and its displacement through combustion chamber geometry of s.i .engine consisting of flat head with two ver tical valves and cylindrical bowl were presented. These results obtained with KIVA3V code were compared with results of flame front shape and its displacement obtained for the case with no valves i.e. without intake flow modeling .In order to alleviate th e application of KIVA3V code and to enhance its flexibility two additional computer codes were applied as well, i.e. AVL TYCON code intended for cam design calculations and calculations of the dynamic behavior of timing drives and gear transmission units and AVL BOOST code intended for engine cycle calculations, including 1.5D fluid flow calculations through pipelines. The first code was used for the calculation of valve lift curve while the other was used for the calculation of relevant data set in valve regions. It was found that for particular combustion chamber shapes considered the entirely different flame front shapes and propagation velocities were encountered for these two cases ensuing primarily from the entirely different fluid flow patterns in the vicinity of TDC. 1. INTRODUCTORY REMARKS It is known for a long time that various types of organized flows in combustion chamber of i.c engines are of predominant importance for combustion particularly with regards to flame front shape and its propagation [ 1, 2 ]. Some results related to synergic effect of squish and swirl on flame front shape and its propagation through various combustion chamber layouts are already analyzed [ 3, 9 ] but the isolated or combined effect of the third type of organized flow i.e. the effect of tumble or “tumble-like” intake flow on flame front shape and its displacement is not sufficiently clarified ensuing partly from the ambiguity concerning the exact definition of tumble flow. Namely, in spite of the fact that tumble flow is inherent to multi-valve engines some two valve engines exhibit some characteristics similar to tumble flow [ 4, 8 ] and therefore in the rest of the paper , in lieu of tumble or tumble-like flow we are operating more general intake flow. In addition, one of the reason is also CFD with all the difficulties ensuing from 3D grid generation of real multi-cylinder engine and experimental verification of numerical results. For that reason in this paper, as a part of broader research concerning the effect of intake port/valve geometry and maximum valve lift variations on flame parameters the combined effect of squish and intake flow on flame front shape and its propagation was analyzed. These results were compared with results of flame front shape and its displacement for the case with no intake flow modeling (no valves) indicating the isolated effect of squish thereafter. 2. COMBUSTION MODELING The analysis of this type is inherent to multidimensional numerical modeling of reactive flows in complicated geometry therefore it was quite logical that such a technique (KIVA3V code [ 5 ] ) was applied for the analysis of the effect of intake flow modeling on flame parameters particularly due to fact that it is the only technique that encompasses the valve/port geometry in an explicit manner. Taking into account various options contained in aforementioned code the following assumptions were adopted: -flame propagation is controlled by turbulent diffusion modeled via k-ε model of turbulence; -chemistry was modeled with quasi global irreversible reaction of fuel oxidation (C8H18) followed by two group of reactions, those which proceed kinetically and those which reduce the chemical heat release; -the ignition was not modeled but rather bypassed as well as the entire ignition delay period by energy deposition in spark cells (artificial rise of temperature in spark cells); -the commencement and the length of energy deposition was adjusted with reference to the location of peak cylinder pressure determined with BOOST code [ 7 ]; -in the case with valves the relevant set of data on open boundaries were calculated with BOOST code as well; -valve lift curves and valve timing were calculated through series calculations with TYCON code [ 6 ].