Telemetry Data Prediction of Launch Vehicle Attitude Control Engine Using LSTM

Abstract

Abstract The attitude control engine used by the launch vehicle is mainly composed of a pressurized gas cylinder, a pressure reducer, a tank, a duct, a pitch extension, a roll extension, a yaw extension, and a sinker extension. In this paper, a long short-term memory network is used to train the telemetry data transmitted from the temperature and pressure sensor points distributed in various positions of the attitude control engine, such that the state of the rocket attitude control engine can be deduced and predicted in real time through the data-driven digital twinning system. It is possible to detect potential data anomalies in advance and locate the faults, so that the control system can change the attitude control strategy in sufficient time to alleviate the fuel loss issue as well as positional deviation caused by minor faults, so as to better ensure the success of the mission.