Progetto, modellazione e controllo attivo di un prototipo per un esperimento a bordo della stazione spaziale internazionale

In cooperation with the Aerospace Engineering Department, the Information and Electronics Department of Politecnico University in Milan and Carlo Gavazzi Space S.p.A, it has been carried out a study to adapt the Body Rotating Device (BRD) to the requirements of the International Space Station. The aim is the implementation of a rotating device (SOAR: Station Off Axis Rotator) for the module Columbus of the International Space Station in order to be able to perform trials similar to the ones already completed on board of Neurolab. The necessity to develop and validate control systems for the suppression of forces discharged on ground by structure, leads to design and build a scaled prototype of SOAR on which testing control strategies.
It has been used the Froude method which allows to reproduce the main dynamical characteristics of the true device to correctly design the scaled model. A rigorous scaled model of all the chair components isn’t necessary in the phase of control system developement, therefore it was decided to realize a simplified scaled model that was fully representative of the principal characteristic of SOAR, like: mass, centre of mass position and first mode frequency. Starting from this design it has been developed a multibody model of the prototype. Since it was necessary to reproduce the deformability of the structure, it was decided to model the basement and the rotating beam like flexible parts integrating FEM analysis of this components in the multibody model.
The flexible parts are scaled trying to respect the main bending frequencies of the structure, which are the most interesting for the development of the control law.
Operatively it has been designed an experimental device which respect, as much as possible, the design imposed by Froud scaling method, provided the availability on market of all the components. The scale factor called S was based on the total weight of the structure, in order to maintain a criterion to determinate the requirements for the restraining forces and the modal frequencies.
The multibody model results to be of fundamental interest for the validation of control law in microgravity field, since we can reproduce the condition of the ISS environment only by numerical simulation. This multibody model presents a detailed modelling of the motor and actuators groups. The actuators of control masses are modelled introducing the stepper motor characteristics and its behaviour. The motor group has been modelled reproducing the new design, and the electromechanical behaviour of brushless motor has been modelled using the transfer function between torque and supplied Voltage in relation with back EMF (Electro Motive Force) torque and viscous damping of motor.
Starting from this mutibody model it has been developed an active control law. The aim of the control system for SOAR is to attenuate the effect of the forces and moments generated by the rotating part of the device on the forces exchanged between SOAR and its rack.
The available actuators are the shifting masses located in the rotating frame of the device, while the available measurements consist of a subset of the aforementioned forces namely the one measured by the load cell. The vibratory forces which have to be attenuated are characterised by a spectrum which is entirely concentrated at the rotor angular frequency. For this reason, the preliminary design of the control algorithms for SOAR has been based on periodic disturbance attenuation techniques.
The multibody model of SOAR for the ISS resulting from this studies is an useful instrument on which making simulations. Simulations carried out on this multibody model, with two shifting masses, show that simultaneous attenuation of vertical and in plane forces cannot be achieved, however, satisfactory results in terms of attenuation of the vibratory loads have been achieved for each of the considered sub-problems. Preliminary simulation results, carried out on a modified multibody model of SOAR with three shifting masses, show that even better performance can be obtained at the cost of a slightly more complex control architecture.