Parachute-Smart Machine Learning Algorithm for Inflation Time Assessment From Test Data

A novel “parachute-smart” neural network algorithm is proposed as a tool for finding hard-to-get algebraic relationships tying performance and canopy design data collected during test drops, wind tunnel investigations or high-order computer simulations. Herein, inflation performance data of the USAF C-9 personnel recovery parachute and NASA’s Orion three-cluster landing ringsail parachutes are used to computationally establish, in each case, novel functional relationships between their inflation times (non-dimensional), inflated canopy volume, initial skirt area and overall canopy size scale. Interesting analysis byproducts are produced as well, thereby offering new insights on unmeasurable dynamical variables such as the average through-skirt flux and the role of the sails in ringsail designs acting as “air scoops” during the inflation process.