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

From 2000 to today, air traffic due to civil aviation has seen an increase of 2.4 times, and with a forecast of doubling further in 15 years. The growth of the aerospace industry leads to raise up standards and considers cost effective solution. Electromechanical actuators in the last century have a high growth rate. These parts will become essential for primary and secondary flight controls, this transition leads to higher quality standards and the development of algorithms that can predict their operation. Through prognostics algorithms it will be possible to evaluate in real time the behavior of the aircraft and predict failures. An important problem is maintenance, with the use of these algorithms it will be possible to schedule and optimize maintenance using real-time data and models to reproduce actuators behavior. The initial aim of the thesis in the development of an experimental test bench is to develop and create a test bench to validate the prognostics models. The development of the test bench consisted of several stages; first an analysis and design of major parts, then the assembly of the test bench. The last step to complete the test bench is to characterize the components and to achieve the desired results. The test bench consists of an electric brushless motor, a planetary gearbox, a braking system, a motor actuation and control system, an encoder for closed loop and an Arduino board for brake control. In this thesis, the electromechanical actuator test bench is adapted to the requirements. In particular, after the characterizations carried out in previous works, it was possibly make tests on different ways: -??Variable backlash with zero load: backlash was inserted on the user. -??Load variation and backlash variation: application of load on the motor shaft with variable backlash. The development of prognostication algorithms (HF model) was carried out in the MatLab/Simulink environment. The model used takes into account the main physical effects present in the test bench. Using Simulink, the electromagnetic model of the motor, the model of the inverter controlling the electric motor, the control electronics and the planetary transmission were modelled. Simulink model takes into account also: -??Friction effects by Borello’s friction model. -??Include a short-circuit on each phase of the electric motor. -??Regulation of the noise. -??Clarke-Park's transformation. -??Some hysteresis effects. The validation of the prognostics algorithms was carried out by evaluating the data obtained with the bench test with those of the Simulink model. The comparison and validation of prognostics algorithms made it necessary to calibrate the Simulink model in order to better simulate the behavior of the test bench; this work made it possible to study and validate the backlash model implemented in the Simulink model.