No abstract

No abstract

This work presents an aerodynamic model capable of representing lift, drag and pitch moment coefficients at high angles of attack (AOA), accounting for compressibility effects and aerodynamic hysteresis during stall. The modeling was peformed using flight test data collected during accelerated stalls of a fighter trainer jet, which consisted in wind-up turn (WUT) maneuvers from Mach 0.40 up to 0.73. A nonlinear model structure, based on Kirchhoff 's theory of separated flow, was used to deal with high AOA effects. However, as the aircraft reaches high speeds, the aerodynamic coefficients become a function of Mach number. In order to deal with compressibility effects, the flight data were partitioned based on Mach number intervals, so that a different set of coefficients is applicable to each range of Mach. The values obtained for each coefficient as a function of Mach were mostly compatible with those expected from theory. During the validation tests, using a different set of data, it was clear that a single set of coefficients for a single Mach number was not able to reproduce the flight data in the entire envelope of Mach and AOA tested.

This work presents a new approach adopted in the application of flight path reconstruction (FPR) techniques to overcome limitations in the flight test instrumentation (FTI), even when one or more flight test parameters are unreliable or corrupted. The main motivation arose from the necessity of measurement of true airspeed during spin tests, for modeling purposes. However, airspeed measurements become unreliable during spin maneuvers, due to Pitot tube ineffectiveness when subjected to angles of attack and sideslip in excess of 40 • . At first, it is not possible to use this parameter for flight path reconstruction of the spin maneuver, because the unrealistic measurements during the developed spin would affect the entire estimation process. However, the measurements before and after the spin are valid and useful for estimation. The new approach proposed is used not only to separate the useful part of the flight data, but also to obtain reliable estimates of the corrupted parameters during the time intervals at which they were considered unreliable.

The main objective of this study was to experimentally obtain the most representative linear dynamic model of a two-stroke piston engine for small unmanned aerial vehicle applications with low-cost sensors and a friendly interface based on laboratory virtual instrument engineering workbench (LabVIEW®). The engine was mounted on a test bed equipped to measure the thrust and rotational speed. The throttle lever was actuated by a standard hobby servo motor, which controlled the carburettor’s valve opening. Input command and data acquisition were performed in a two-layer approach: low-cost hardware, where a micro-controller unit managed the sensor’s readings, servo input, and external communications through serial protocol; and LabVIEW for command signal generation, serial port write/read, data processing, and other high-level tasks. The motor dynamics, represented by its transfer function, was obtained by minimizing the output error between the experimental responses to various types of input signal and that obtained by the simulation of a fixed-topology, fixed-order reference model that included the servo and engine models in series.