Near Real-Time Impact Point Prediction for the Flight Termination System of the DART XL Sounding Rocket

Abstract

A flight termination system (FTS) is one of the most important safety features on a launch vehicle, as it allows for the vehicle's flight to be terminated in a controlled way when predetermined safety criteria or boundaries are exceeded. Currently, T-Minus Engineering B.V. aims to develop a custom flight termination system for the Dart payload stage of their DART XL sounding rocket. A sounding rocket is by definition sub-orbital, so an important safety constraint is its impact location. Based on this impact location, an advice can be formed for a range safety officer (RSO) to base flight termination decisions on. Therefore, the objective of this thesis is creating an algorithm to predict the instantaneous impact point (IIP) of the DART XL's ballistic payload stage in near real-time during a launch, considering impact probability area boundaries and requirements for the accuracy and update rate.

To meet this objective, two separate models were created. The first was a pre-flight nominal trajectory model aimed at evaluating the nominal vehicle trajectory and thereby producing inputs and verification tools to assist the development of the second model. The second model, the in-flight IIP prediction model, is intended to run during a DART XL launch to continuously predict the ballistic trajectory and IIP of the Dart stage in the post-separation phase. An important element of developing both models is determining the model definitions. The final environment model selection consists of a US76 atmospheric model, a central gravity model including the J2 effect, a local wind model and a rotating, ellipsoidal Earth model. Furthermore, using 3-degrees of freedom (3-DoF) for the IIP prediction model resulted in IIP predictions close to those predicted using a nominal trajectory model of more degrees of freedom.
Additionally, an Euler integrator with a step-size of 0.1 s was used for the IIP prediction model. With this step-size, the IIP prediction model successfully met the computation time requirement of an update rate below 1 s. For the accuracy requirement, the IIP prediction model shall only include parameters that influence the prediction in the order of 1 km. This requirement was successfully met using the step-size of 0.1 s, with a distance to a verification IIP of less than 600 m.
Lastly, the nominal impact point distribution was determined by performing Monte Carlo simulations of the pre-flight nominal trajectory model. The resulting IIP distribution was used to determine the 3-σ impact probability area, which serves as a measure of the uncertainty of the nominal impact point.

It should be noted that uncertainties remain in this nominal prediction larger than the desired 1 km accuracy radius. However, it is unlikely that the uncertainty of the nominal IIP prediction can be decreased below a radius of 1 km due to the inherent unpredictability of launch conditions. The overall conclusion is therefore that the thesis work has successfully developed a model for the near real-time instantaneous impact point prediction of the DART XL sounding rocket's Dart payload stage.