Development of an experimental test bench for the validation of prognostic algorithms for electromechanical actuators

The thesis work presented includes a first part dedicated to the development of a test bench that reproduces the architecture of an electromechanical actuator (EMA) for secondary flight controls. In this section, it is analyzed the installation of the hardware and the realization of the connections between the different components. The test bench under examination consists of two modules: the actuation module and the braking module. The first module deals with the system drive and includes an sinusoidal brushless motor, the control electronics, as the Control Unit, the Inverter and the Microbox PC, manufactured by Siemens and a planetary gearbox, made with the FDM (Fused Deposition Modelling) technique at the Politecnico di Torino, in a previous study. The second module has the function to simulate the presence of an external load acting on the motor and includes a shaft placed parallel to the motor shaft and supported by two self-aligning bearings. On the brake shaft are installed a sprocket, necessary to support the chain that connects the actuation module with the brake module, the brake disc and a rectangular plate, interposed between two self-aligning bearings. The latter is a support for the brake caliper and servomotor and transfers the braking load to the load cell, connected to an Arduino board. The two modules described are connected to each other by a chain transmission system, consisting of a roller chain supported by two sprockets, one installed on the motor shaft and the other on the brake shaft. At the end of the assembly of the test bench, tests have been carried out in order to obtain the necessary parameters to complete the characterization of some of the components, previously described. The second part of the work is dedicated to the validation process of a "high-fidelity" numerical model, implemented in Matlab-Simulink environment. In particular, a genetic algorithm was used in order to optimize a set of parameters and tests were performed in the time and frequency domain in order to compare the responses of the numerical model and those of the test bench. Since the trends obtained from these tests are the same between the test bench and the numerical model, it was possible to validate the latter. The numerical model "high-fidelity" is then used together with a numerical model "low-fidelity" to implement prognostic algorithms.