Uncertainty quantification in RANS prediction of LOX cross-flow injection in methane

This work presents the numerical characterization under uncertainty of a pintle-injector liquid rocket engine thrust chamber, fueled with LOX-CH4 and operated at subcritical pressure. Being the design optimization the ultimate goal of this effort, the numerical characterization is carried out employing a Eulerian-Lagrangian Reynolds-averaged Navier Stokes equations approach. The numerical model of choice, as well as the rich variety of physical phenomena taking place in such a device, require the knowledge of a large number of model parameters, many of which are challenging to be calibrated under the severe thermophysical conditions of interest. A possible way to overcome this lack of knowledge is to resort to the Uncertainty Quantification (UQ) framework to estimate the effects of model and parameter uncertainties on the solution accuracy. In particular, this research aims at propagating the uncertainty associated with the most probable diameter which characterizes the injection Rosin-Rammler distribution for the liquid droplets, employing a Polynomial Chaos Expansion (PCE) representation of the uncertainty. The pintle configuration consists of a horizontal gaseous methane inflow and a vertical LOX spray injection. A set of RANS are conducted to generate the PCEs surrogate model for the estimation of the probability distribution of the quantities of interest, as well as the visualization of their credibility intervals. Lastly, to assess whether the uncertainty on the droplet diameter can overshadow the sensitivity to the pintle design, the same uncertainty quantification analysis is performed for two geometries, which differ in the distance between the annulus final section and the fuel-oxidizer impingement location.