Unsteady RANS simulations with uncertainty quantification of a spray combustor under liquid rocket engine relevant conditions

We present the numerical characterization under uncertainty of a single-element liquid-rocket-engine fueled with liquid oxygen (LOX) and gaseous hydrogen (GH2) operating at subcritical pressure, which complies with the MASCOTTE A-10 test case. The proposed numerical investigation relies on the Eulerian-Lagrangian unsteady Reynolds-averaged Navier-Stokes (uRANS) framework. The harsh thermophysical conditions encountered within the combustion chambers of such devices, as well as the large number of sub-models embedded into the uRANS approach, require the knowledge of a multitude of model parameters. The uncertainty quantification (UQ) framework can reduce this lack of knowledge, estimating the influence of model and parameter uncertainties on the numerical solution accuracy. In the present work, the uncertain parameter of choice is the most probable diameter, d0, of the Rosin-Rammler liquid droplets’ distribution characterizing LOX spray injection. The associated uncertainty is propagated employing a polynomial chaos expansion (PCE) representation of the random variables (RVs) obtained through a limited set of uRANS simulations. It allows for evaluating the sensitivity of the predicted mean temperature and OH fields to the uncertain input parameter. Lastly, the performance and reliability of the model are assessed through a detailed comparison with the available experimental database.