COMBINED OPTIMISATION AND DESIGN OF EXPERIMENT PROCEDURES IN ORDER TO ESTIMATE MLGD INTERFACE LOADS UNDER UNSTEADY AERODYNAMIC VIBRATIONS

摘要

The Main Landing Gear Doors, as many other aircraft components, are affected by unsteady vibration loads during flight. The main source of this vibration is aerodynamic in nature and generally it is very difficult to characterise experimentally this unsteady aerodynamic excitation. This characterisation is further hindered if unsteady pressure sensors are not available and the instrumentation is limited to a set of accelerometers. In the present work, a method is proposed to characterise the shape of the unknown gust input load excitation, leading to the possibility of estimating the loads at the Main Landing Gear Door (MLGD) interfaces. To achieve this, a Flight Test (FT) has been completed on the Airbus A320 aircraft. The left MLGD was instrumented with a limited set of acceleration sensors and all interface loads were instrumented with strain gauges. Two steps have been followed. In the first step, acceleration data was considered as the only available source of experimental data. In order to estimate the loads at the MLGD interfaces, a combined NASTRAN -MATLAB optimisation procedure, based on the least square method, is applied. The purpose is to define the shape of the gust load able to fit the experimental data at all the experimental sensor points. Hence the gust load and the reactions at interfaces will be functions of the flight parameters, such as aircraft speed, angle of attack, sideslip and high lift configuration. The Design of Experiment technique is used in order to build-up a predictive model of the reactions as function of any combination of flight parameters. During the second step the FT instrumented interface loads were considered. This allowed these experimental interface loads to be compared with the predicted loads determined from the NASTRAN - MATLAB optimisation procedure. The accuracy and the reliability of the procedure are analysed and discussed. From the results it can be seen that the shape of the unknown gust load can be efficiently optimised in order to fit the experimental sensor responses and the loads at interfaces can be efficiently estimated.