[home] [Personal Program] [Help]

---

Session: Ground vibration testing
Session starts: Thursday 20 June, 16:00
Presentation starts: 17:00
Room: Room 1.6

---

Anders Ellmo (Saab AB)

Abstract:
Using complex composite structures in a wind tunnel model means that small differences in layup, thickness and other manufacturing outcomes will result in significant consequences for the structural dynamic properties of the model, and subsequently its aeroelastic behaviour. The KTH-NASA wind tunnel model has interchangeable wings, and it is shown that finite element representations of each individual wing are needed to make relevant comparisons to experimental data, and accurate predictions of flutter behaviour. The KTH-NASA wind tunnel model is constructed for acquiring flutter data in transonic conditions. It is specifically designed for the conditions at the NASA Transonic Dynamics Tunnel, and was tested there in 2016. That test ended pre- maturely since the wings were damaged in a flutter incident after recording two flutter points. The damage to the wings was an opportunity to re-design the wings to allow flutter without external stores mounted. The strategy chosen was to replace a plain weave in the glass fiber composite wing shells with a unidirectional material, and to vary the layup angle until desirable flutter speed was reached. Three pairs of wings were then manufactured using the new design. Experimental modal analysis of the three wing pairs shows differences in fre- quencies and mode shapes. Each wing was suspended in free-free conditions and excited with a force hammer. The analysis showed first elastic modes in the range 21.0 - 23.2 Hz, second elastic modes in the range of 43.1 - 45.8 Hz, and third elastic modes in the range 89.8 - 93.2 Hz. The experimental results forms the goal for a set of model update procedures. Structural optimization is used to reach a target level of correlation between each wing and its finite element representation. The updated finite element models are used to calculate flutter speeds for each set of wings. The span between the results represents a manufacturing related uncertainty on the flutter speed. Finally, all possible wing pairings are compared to each other in terms of the resulting flutter speed.