Authors: Potito Cordisco, Senior Project Manager, Vicoter
Mauro Terraneo, Chief Technical Officer, Vicoter
Aeroelasticity is often represented by the Collar triangle, whose three vertices correspond to aerodynamic forces, elastic forces, and inertial forces, with the sides illustrating the mutual interactions among them.
Although in real phenomena each of the three terms typically acts in interaction with one or both of the others, the identification and validation of each contribution are generally carried out separately. From an experimental standpoint, this means performing wind tunnel tests to assess aerodynamic effects, measuring masses and their distribution to characterize inertia, and applying known loads while recording deformations to determine structural stiffness.
In practice, when the investigated problem involves the interaction among all three types of forces—as in the case of flutter instability or gust response—it is common not to separate the effects of elasticity and inertia. These two contributions are evaluated jointly to ensure a correct understanding of their interaction mechanisms and to reduce uncertainty. When only these two are considered, with aerodynamics introduced at a later stage, the assessment takes the form of a classical dynamics problem and can be analyzed using techniques of standard structural dynamics.
From an experimental standpoint, this means performing modal analyses—referred to in the aerospace field as Ground Vibration Tests (GVTs)—to identify the modal frequencies, mode shapes, damping ratios, and modal masses. These results are then used either directly or to validate numerical or analytical models.
X-DIA is an aeroelastic 1:10 scale wind tunnel model developed by Politecnico di Milano to investigate the dynamic behavior of advanced aircraft configurations. It features a highly flexible structure and modular aerodynamic surfaces, enabling the study of aeroelastic phenomena such as flutter, control surface effectiveness, and gust response. Designed for both open- and closed-loop testing, X-DIA has served as a benchmark platform for experimental aeroelasticity for several years. The model has seen a constant evolution during the time:
-
- 2005–2008: Under the EU-funded Active Aeroelastic Aircraft Structures (3AS/FP5) project, X-DIA was used to test active flutter suppression (AFS) systems, featuring movable canards and distributed wing controls that significantly dampened bending and torsion modes.
- 2017–2019: During the AFS collaboration with the University of Washington, the model underwent a complete redesign, maintaining, de facto, only the original name. It was tested at the wind tunnel of Politecnico di Milano using open- and closed-loop configurations to validate numerical models and gather lessons on control strategies.
- 2019: A sliding-tip safety device was integrated to raise the flutter onset speed and enable automatic anti-flutter activation during testing. New flutter tests were performed.
- 2021–2023: The model evolved into the so-called F-X-DIA with installation on a pylon mount—allowing free pitch and roll motion— from the original free-free A redesign of the vertical tail, for LCO (Limit Cycle Oscillations) studies, introduced updated control surfaces and optimized actuators. Wind tunnel tests explored interactions between rigid-body stabilization and flutter suppression at speeds beyond the open-loop flutter point.
- 2025: The fuselage and pylon, and the interface between them, were modified as were some particulars of the tail. New tests of were made in the region of LCO with freeplay in the control Vicoter is a company specialized in performing this type of testing and was commissioned to carry out Ground Vibration Tests on the latest version of the X-DIA model.
During GVT tests carried out in the wind tunnel of the Department of Aerospace Science and Technology of Politecnico di Milano, Milan, Italy, at the end of June 2025, the aircraft was installed on the new pylon in the test chamber to reproduce correctly the stiffness of the support system.
During the GVTs four configurations of aircraft were tested: the clamping system between the aircraft and the pylon was released to permit to test the structure free movement in the pitch and roll directions. In such cases an elastic rope connected to a soft spring was used to constrain the rigid displacements while not introducing change in the dynamic behaviour of the device under test.
The setup of the accelerometer was used to visualize the first out of plane bending and torsional modes of the lifting surfaces: wings, horizontal tail, vertical tail. Control surfaces and pods, such as the in-plane movement of the T-tail, were also investigated .
68 locations on the aircraft and the supporting system were measured by single- (mostly PCB 333B32 ) and tri-axial (mostly PCB 356A32 and PCB 356B08) accelerometers for a total of 97 channels read contemporaneously.
In each of the four configurations, the FRFs (Frequency Response Functions) needed for the modal identification were acquired by two frontend SCADAS 316 data acquisition systems connected in series.
An impact fixed accelerometers MIMO technique was preferred. The accelerometers were glued to the structure while excitation was given to the aircraft by an instrumented hammer, changing step by step the excitation point. Ten impact points, spanning from the tip of the wing in vertical direction to the nose of the fuselage in lateral one or the end of horizontal tail plane in longitudinal direction were used as references to have signal with a high S/N ratio and avoid the nodes of the modes.
FRFs were acquired with a high sampling frequency of 512 Hz, since it was essentially cost-neutral, and with a frequency resolution of 0.03125 Hz due to the foreseen presence of low frequency resonances. The observation time of 32 s has been employed to give to the structure the time to decay autonomously and avoid the use of windows to contain the leakage. Each point was hammered five times and the data averaged in order to reduce measuring noise. FRFs were estimated by the Hv method, the best choice for modal analysis. As usual in these cases, for every excitation signal, the following functions have been acquired: FRFs, Autopowers, Crosspower and Coherence.
Coherence analyses of acquired signals showed values from very close to one up to more than 100 Hz so the excitation method was considered suitable for test purposes as the band of interest was from 5 Hz to 35 Hz.
Identification of natural frequencies, shapes, damping and modal masses was carried out with the LMS-Test.Lab software using the algorithm Polymax followed, at the occurrence, from an MLMM optimization procedure. Twenty-five modes were found in the band up to 35 Hz. In particular, the ones concerning the pitch, roll and yaw movement around the tip of the pylon, which are very difficult to recover in the finite element model due to the geometrical complexity of the part, were clearly visible.
Particularly important, especially in project involving certification, is the validation of the obtained modes that gives a numerical quantification of the quality of the results. It is performed in various ways starting from the comparison of the acquired FRFs with the ones synthetized from the identified modal base, passing through the AutoMAC matrix and finishing with the complexity checks (MOV, Mass sensitivity, MPC, MPD, Scatter).
Even if the modes were identified as complex, they were transformed into real modes at the end of the procedure by a quadrature algorithm upon request of the client. In such way it was possible to use them directly in software for flutter analyses (Nastran, NeoCass, …) or control without the need to correlate a numerical model.
Vicoter is grateful to Francesco Toffol, Donato Grassi and Elena Roncolini for their help during the setup of the test.
