Low-order acoustic network model for the modeling of high-frequency combustion instabilities in rocket engines

Abstract

The prediction of high-frequency combustion instabilities is of relevance for liquid propellant rocket engines, to prevent risks or damages caused by overpressure and increased heat transfer to the walls. The main prediction method in rocket engines is through CFD simulations, which are computationally expensive, and require a significant set up process. This document presents a low-order tool capable of providing predictions in significantly less time than CFD simulations, with little to no setup time required by the user. The tool has the capability to predict stability, as well as computing the acoustic resonant frequencies of the system with an accuracy comparable to higher order models. The tool employs what is known as a "network model" and linear acoustics. The model includes new developments that improve its prediction capabilities with respect to the existing network model approaches present in literature. This includes acoustic elements capable of describing any isentropic flow, as well as distributed flame models that asses the stability of the combustion chambers of rocket engines. The document describes these new models in detail as well as their derivations. The new models are validated using experimental data from various rocket engines and load points tested at DLR Lampoldshausen. The validation shows that the acoustic resonant frequencies are predicted with an average error less than 5%. Furthermore, no single validation case presents an error over 10%. The tool is capable of predicting stability based on two parameters, the time lag and gain between perturbations in pressure and the flame response. Given these parameters the tool adequately predicts stability of the engines used for validation.