System Identification for Control Applications of Pump Fed Rocket Engines

Abstract

Motivated by control applications, linear system identification is used on data produced by EcosimPro simulations of a gas generator subsystem and a full gas generator engine. When limiting the operating range of valve actuators from 40 % to 60 %, the linear models are able to produce predictions with mean errors of less than 3.5 % for mixture ratios and chamber pressure. Full range identification of the gas generator engine fails to produce linear models that could be used for controller design. The toolchain developed is thereby validated to be used on LUMEN, an LOX/LNG engine in development at DLR Lampoldshausen, Institute of Space Propulsion.