Abstract:
Dynamic rollovers represent a major hazard for helicopters during near-ground operations, often resulting in significant aircraft damage and passenger injuries. To improve safety in operations, recent studies have focused on developing a Helicopter Flight Data Monitoring framework to provide data-driven insights on operational safety. This work contributes to that effort by proposing an approach to identify precursors to dynamic rollovers. According to NTSB reports, approximately 60% of such incidents occur during in-flight phases like hover, hover-taxi, or landing. To capture the complex non-linear dynamics of helicopters, physics-based simulations were conducted to estimate a first hitting time metric, defined as the time until blade-ground contact, across a wide range of initial conditions for an inflight initial state of the helicopter. Eight parameters were identified as driving the first hitting time, and a probabilistic model was created to predict the distribution of that metric for different values of those parameters. Based on the predicted distributions, a risk-based metric was derived to robustly assess the risk of dynamic rollover and identify safer operational boundaries.