The Application of Machine Learning in Micrometeoroid and Orbital Debris Impact Protection and Risk Assessment for Spacecraft

Current spacecraft micrometeoroid and orbital debris impact risk assessments utilize semi-empirical equations to describe the protection afforded by a spacecraft component (e.g., pressure hull, critical component, etc.). These equations demand fundamentally limiting assumptions, for example of projectile shape and material, to reduce the complexity of the mechanics and material response under such extreme conditions. Machine learning approaches, however, are well suited to such high dimensionality problems and have previously been shown to provide comparable classification accuracy to state-of-the-art empirical techniques in this domain. We demonstrate that such models can readily incorporate additional complexity beyond that currently achievable with semi-empirical models, such as the effect of thermal insulation blankets, non-aluminium projectiles, and non-spherical projectiles on failure thresholds. To provide additional insight into the behaviour of these quasi black-box machine learning models we demonstrate feature importance ranking, learning bias, and the inclusion of physicsbased laws for improved model predictions (particularly in sparse data regimes)