Aircraft dynamic analysis methods using flight tests

Abstract

The dynamic response of an aircraft is an important aspect to be analyzed during its development program. When the airplane is disturbed from its steady state due to a wind gust or commands, the way by which it returns to the equilibrium is crucial for the pilots opinion of the aircrafts handling qualities. The characteristics of the aircraft response can be quantified by parameters such as the natural frequency, damping coefficient and time constant values of each of the well-known dynamic modes: Phugoid, Short-Period, Dutch Roll, Spiral and Roll subsidence. These modes are related to aerodynamic, inertial and propulsive properties of the aircraft, which are initially obtained from semi-empirical formulations and refined from wind tunnel tests data. However, for the validation of these parameters and for experimental analysis, flight test must be performed, in which the actual aircraft behavior can be observed. In this work we present techniques commonly used for flight tests execution, i.e., procedures to excite each of the aircrafts dynamical modes in flight. Initially, a classical methodology is used for data reduction. Such methodology is based on graphical analysis, where simplified dynamic models for the aircraft are used to estimate its characteristics. Such techniques depend on high data quality with low noise level and on the personal interpretation of the flight test engineer, which can generate inaccurate results. Therefore, other different methods are used for analysis, which are based on system identification techniques where different models of the flight mechanics can be used during analysis. To compare the different methods presented, two distinct aircrafts (VFW-Fokker 614 and ACS-100 Sora) are evaluated. The maneuvers are executed more than once over the same flight conditions to obtain different datasets, for the estimation and validation. Despite the many error sources, the classical methodologies has shown a great proximity when compared with the other applied methods. The parameters obtained using the system identification techniques have a higher reliability due to the statistical properties of the methods. Indeed, the results obtained by these methods show a greater consistency with the measured data, where a lower degree of dispersion is observed when comparing the results for the different system identification techniques. Thus, the system identification methods are more suitable for the evaluation of the flight dynamics.