A comprehensive approach for the prediction and the analysis of the behaviour of jettisoned external loads

The present work describes a comprehensive approach to the problem of external load jettison from a heli- copter and the associated tool developed. The focus is on trajectory identification and aerodynamic parameters reconstruction from flight test data, simulation and collision detection. Particular attention is paid to the separa- tion problem in case of auto rotation when loads are jettisoned to reduce the overall weight or to drop apart from the helicopter potentially dangerous items. The first part of the work deals with the simulation of the separation and drop of an external suspended load. A 6-DOFs dynamics rigid body model with quaternion is used for the load. A simple kinematic model, with assigned linear and angular velocities, accounts for the position and the attitude of the helicopter and allows the release of the body in constant rate manoeuvres like turns. Depending on the type of body released, two aerodynamic models are used. The first is a static aerodynamics axis symmetric model while the second is a full static aerodynamic model. The velocity field around the helicopter and the eventual ejection mechanism are taken into account during the simulation. An important aspect of the simulation is the collision avoidance. By loading both the shape of the helicopter and the shape of the load by points, the minimum distance between the two bodies is evaluated at each simulation step. In this way it is also possible to monitor the clearance between load and helicopter during the entire manoeuvre. The second part of the work describes a methodology to reconstruct the trajectory and the attitude of the jettisoned load from video frames. This is done by manually selecting three points visible in two video frames taken at the same time by video cameras mounted on the helicopter. A simple camera calibration technique is also implemented. The last parts of the work assess the possibility of identification of the aerodynamics properties of the re- leased body from a previously reconstructed trajectory. Different aerodynamic models are used depending on the geometry of the external load. The reconstruction of aerodynamic parameters is performed by minimizing a cost function based on the error between the identified trajectory and a simulated trajectory with tentative pa- rameters. The minimization is performed by an Evolutionary Algorithm. This approach is somewhat slower than a classical gradient-based one, but it is not affected by discontinuity in the search space, it is much less sensitive to local minima and performs well even on non-convex solutions spaces. The performance of different algorithms is compared on a test case. Finally the results obtained with the proposed methodology are compared with reference cases obtained with different codes or experimental data available in literature.