BAYESIAN CONTROL OF A HELICOPTER MAIN GEARBOX SEMI-ACTIVE SUSPENSION SYSTEM - EXPERIMENTS ON A QUARTER-SUSPENSION PROTOTYPE

Jonathan Rodriguez (1) , François Malburet (2, 3)

1 LSIS - Laboratoire des Sciences de l'Information et des Systèmes
2 Arts et Métiers Paristech ENSAM Aix-en-Provence
3 LISPEN - Laboratoire d’Ingénierie des Systèmes Physiques et Numériques

This short paper considers the control of a helicopter gearbox semi-active suspension. As the future generation of helicopters will include variable engine RPM during flight, it is interesting to consider implementing control on their suspension systems in order to always optimally filter the main disturbance frequency. Here, a semi-active suspension based on the DAVI principle is developed, simulated and tested with its control algorithm based on Bayesian optimization. This control method based on the Bayes theorem is a trial/error algorithm allows to significantly reduce the number of evaluations of the real objective function for a given set of parameters. Thus the system is capable to fastly determine its own optimal set of parameters to maximize the objective function. The objective is to prove experimentally the ability of the Bayesian optimization to lead the learning behavior of a semi-active resonant suspension.

Domaines: Sciences de l'ingénieur [physics]

Jonathan Rodriguez, François Malburet. BAYESIAN CONTROL OF A HELICOPTER MAIN GEARBOX SEMI-ACTIVE SUSPENSION SYSTEM - EXPERIMENTS ON A QUARTER-SUSPENSION PROTOTYPE. MEDYNA 2017, Apr 2017, Sevilla, Spain. ⟨hal-02301446⟩