J. Phys. IV France 10 (2000) O EDP Sciences, Les Ulis Computational modeling of high pressure combustion mechanism in scram accelerator J.-Y. Choi, B.-J. Lee* and I.-S. Jeung** Pusan National Universify, Pusan 609-735, Korea * Agency for Defense Development, Taejon 305-600, Korea ** Department of Aerospace Engineering, Seoul Nafional Universify, Seoul 151-742, Korea Abstract: A computational study was carried out to analyze a high-pressure combustion in scram accelerator. Fluid dynamic modeling was based on RANS equations for reactive flows, which were solved in a fully coupled manner using a fully implicit-upwind TVD scheme. For the accurate simulation of high-pressure combustion in ram accelerator, 9-species, 25-step fully detailed reaction mechanism was incorporated with the existing CFD code previously used for the ram accelerator studies. The mechanism is based on GRI-Mech. 2.11 that includes pressure-dependent reaction rate formulations indispensable for the correct prediction of induction time in high-pressure environment. A real gas equation of state was also included to account for molecular interactions and real gas effects of high-pressure gases. The present combustion modeling is compared with previous 8-step and 19-step mechanisms with ideal gas assumption. The result shows that mixture ignition characteristics are very sensitive to the combustion mechanisms, and different mechanism results in different reactive flow-field characteristics that have a significant relevance to the operation mode and the performance of scram accelerator. 1. INTRODUCTION In scram accelerator, explosive gas mixture filled typically at 20-50 bar is compressed by shock waves and generates thrust force by high-speed combustion mechanisms such as shock-induced combustion or oblique detonation wave. Previous studies[l,2,3] have shown that ignition is presumed to occur in the boundary layer behind an impinging point of oblique shock wave, and the combustion characteristics has a great relevance to the operation mode and performance. Therefore, a combustion mechanism must faithfully describe the induction time to capture the ignition point and the rate of energy release to predict the combustion process including the generation of detonation wave. In previous scram accelerator studies, a 7-species and 8-step chemistry mechanism developed by Evans and Schexnayder[4] has been used for its simplicity,[l,2,5] and more detailed 9-species, 19-step Jachimowski mechanism[6] has been used for low pressure cases where the mechanism is reliable.[3,7] However, the mechanisms used in the previous studies were developed under atmospheric condition and loose their accuracy at operating condition of ram accelerator. Because, these reaction mechanisms were based mostly on low pressure and high temperature data, whereas typical fill pressure of ram accelerators is about 20-100atm and its mixture ignition temperature is less than 1,40OK[8]. As an effort to overcome these limits, Petersen et a1.[8,9] developed an 9-species, 18-step reduced kinetics mechanism for the hydrogen based ram accelerator combustion. This model is nearly a subset of a 47-species, 279-step mechanism (RAMEC)[9] based on GRI-Mech. 1.2.[10] Since GRI-Mech. 2.11 comprised of 51-species, 277-reactions is available for the present,[ll] a fully detailed, 9 species 25-step hydrogenlair reaction mechanism based on GRI-Mech. 2.11 was used in this research. This includes pressure dependent rate formulation that is indispensable for predicting induction time correctly in high-pressure region. On the other hand, the ideal gas assumption loses its validity for ram accelerator flows since the combustion pressure in ram accelerator is extremely high. An assumption of ideal gas neglects intermolecular interactions and is generally valid for low pressure and/or high temperature systems. However, the combustion pressure in ram accelerator is order of thousand atms because of the high fill-pressure and the burned/unburned pressure ratio of 20-40. At this highly elevated pressure, it is Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jp4:20001114