REST HF-10 Test Case: Numerical Simulation of a Single Coaxial LOX-CH4 Injector with Forced Mass Flow Oscillations Using the DLR TAU-Code

Abstract

Combustion instabilities are high-pressure acoustic oscillations that often develop spontaneously and may lead to catastrophic failure of a liquid rocket engine. In order to investigate this phenomenon, the Rocket Engine Stability initiative (REST) was founded by CNES, ONERA, DLR and CNRS. Supporting current and future engine developments, the REST community has worked towards a better understanding of highfrequency instability phenomena. The advancement of numerical simulation methods created new opportunities to simulate combustion instabilities and allowed for a deeper understanding of processes in rocket engines that are difficult to observe experimentally. Numerical simulations were used to investigate the coupling between different internal processes like pressure oscillations and heat release fluctuations. With Prometheus as the engine of a future European launcher vehicle, predicting instabilities and flame dynamics in methane-oxygen (CH4-O2) combustion became an important step in developing reliable, efficient and reusable rocket engines. In order to support these developments numerically, the REST community proposed a representative single injector test case designed for fundamental research. The test case consists of a hexagonal combustion chamber with periodic boundary conditions and a representative injector geometry. The injector was designed within the frame of the "Sonderforschungsbereich Transregio 40", a research program founded by the German Research Foundation (DFG), and is intended to be suitable both for LOx-CH4 and LOx-H2 rocket engines. The chamber pressure is set to be 100 bar and instabilities are introduced by modulating the inlet mass flows for methane and the oxygen injector up to +-10 % of their nominal values at different frequencies. The main goal of this test case is to compare the different numerical codes and modeling approaches (different turbulence modeling, different combustion modeling) between the members of the REST community and various codes. This paper presents the current status of the DLR contribution to the test case. All simulations are conducted with the DLR in-house Code TAU. The combustion is modeled using a real-gas flamelet model whereas the turbulence is modeled using a 2-layer k-epsilon RANS model. In addition to the URANS simulations, Detached-Eddy simulation (DES) results for all test case load points are presented and compared to the URANS results. It is shown that the URANS simulations greatly overestimates the dense LOx core lengths while the DES results give more reasonable values. Investigations of the flame response to the longitudinal mass flux oscillations showed only a small effect for an excitation of the O2 inflow at 5 kHz while there is a stronger flame response for the other excited cases. We also present results for numerical grid resolution sensor and a grid convergence study indicating that the chosen mesh resolution allows for grid-converged results for this test case.