Automatic Operational Modal Analysis for Aeroelastic Applications

Kurzfassung

The development of new aircraft requires the evaluation of the aeroelastic stability to avoid the phenomenon of flutter, a self-excited oscillation of the airframe. Since the rational analysis of the flutter stability comprises coupled simulations using numerical structural models and unsteady aerodynamic loads, the accomplishment is complex and the implementations must be checked for their validity by comparison of analytical and experimental results. In the so-called Ground Vibration Test (GVT) the natural modes, eigenfrequencies and damping ratios of the prototype aircraft are identified using classical Experimental Modal Analysis (EMA) methods. Depending on the complexity of the new design, conducting such a test requires a time slot of several days shortly before the first flight. Consequently, there is an ongoing need to reduce the testing time to improve the availability of the aircraft prototype. This paper addresses the application of Operational Modal Analysis (OMA) methods during the GVT of an aircraft, which might cut down the efforts in time and labour. An automatically running fast implementation of the Stochastic Subspace Identification method (SSI) is introduced, which analyses the output acceleration response of the airframe randomly excited by modal shakers. The identification process is specified in detail for a glider aircraft, where acceleration time series must be evaluated to generate the stabilization diagram. To isolate the physical mode shapes from the mathematical poles, a pole-weighted Modal Assurance Criteria (MAC) is evaluated for several model orders to clean the stabilization diagram. Since the process needs no further operator interaction, it is suitable for monitoring airframe vibrations of the aircraft in flight, provided that the changes in flight conditions are significantly slower than the duration of the vibration periods considered. For the sucess of the methods, the OMA requirements should be fulfilled, i.e. the excitation of aircraft should be non-deterministic with broad-band spectra. Such conditions are provided by atmospheric turbulence excitation and/or pilot control inputs. The presented autonomous process is applied to a simulated Flight Vibration Test (FVT) of an research aircraft with real-time modal identification where changes of eigenfrequencies and damping ratios are tracked with changes in flight conditions.