Failure detection algorithms for hybrid rocket engines

Kurzfassung

Hybrid rocket propulsion has the potential to enable cheaper and safer space launch systems than current propulsion technologies. The use of a solid fuel and liquid oxidizer means that a hybrid rocket motor is safer and less mechanically complex than liquid rocket motors. At the same time, the oxidizer mass flow provides thrust control in contrast to solid rocket motors. This feature can also enable future FDIR-routines: depending on the failure type, the motor could be throttled to a more conservative operating range or shut down entirely, provided the mission planning accommodates this. Especially for potential future reusable systems, a reliable FDIR capability will be an important safety mechanism. To provide the first step in a FDIR procedure, this work investigates several failure detection algorithms. The algorithms considered are a simple high-pass filter with a threshold, a density-based approach, and a neural-network-based approach. The algorithms are applied to measurement data from ground test of the VISERION hybrid rocket engine developed by DLR. During the test, a failure of the thermal-protection system of the post-combustion chamber occurred and the failure events can be identified using video recordings. Thus, this experiment serves as a benchmark for the different failure-detection approaches, which are applied to measurement data from temperature and pressure sensors at several positions in the motor, as well as thrust measurements. The findings will guide the design of FDIR strategies for future hybrid rocket systems.