IFASD2024 Paper Submission & Registration
17 – 21 June 2024, The Hague, The Netherlands

Powered by
© Fyper VOF.
Conference Websites
13:30   Data-driven methods 2
Chair: alexis laporte
30 mins
Stall flutter suppression with active camber morphing based on reinforcement learning
Jinying Li, Yuting Dai, Chao Yang
Abstract: The stall flutter has been facing the difficulty of high-dimension, nonlinearity, and unsteadiness for a long time, which makes it hard to predict and control. In recent years, the rise of data-driven methods has brought an inspiring perspective on certain topic. Among the thriving data-driven methods, reinforcement learning (RL) shows its outstanding ability in complex model prediction, directness, and generalization ability. This study investigates the adaptation of RL into stall flutter suppression. The geometric model is an NACA0012 airfoil with active trailing edge morphing. Firstly, an offline, rapid responsive stall flutter environment is constructed with differential equations, where the aerodynamic force is predicted with reduced order models. A double-Q-network (DQN) algorithm is adapted to train the controlling agent with proposed offline environment. The agent has 5 optional actions: large downward morph, small downward morph, stay still, small upward morph and large upward morph. The reward function is designed with a linear combined punishment of pitching angle and angular velocity, a large bonus reward on complete suppression and a large punishment on over-limit morphing. The trained agent shows a rapid and complete stall flutter suppression performance in offline environment simulation. Test is further conducted in online, high-fidelity, fluid-structure-interacted computation, where the trained agent also performs significant suppression effect. Additionally, a broad generalization ability of the trained agent is also observed in high-fidelity tests, which effectively suppresses stall flutter under various ranges of inflow airspeeds and computation timestep size.
30 mins
A Procedure for Flutter Analysis with Nonlinear Experimental Modal Parameters
Martin Tang, Marc Böswald
Abstract: GVT are conducted to identify structural modal models which are used afterwards either to validate an existing numerical model or to establish a mathematical substitute model for further analysis, e.g. as input for flutter analysis. The use of experimental substitute models is not based on modelling assumptions. However, the system identification methods applied to the test data are based on the assumptions of linear and time-invariant systems (LTI). Most often nonlinear effects are observed with aerospace structures in a GVT. This would also affect the results of flutter analysis and critical flutter speeds. However, considering nonlinear effects in the analysis model and subsequent nonlinear simulations might not be feasible shortly before the first flight. This work proposes a procedure to perform flutter analysis in time domain with a nonlinear mathematical substitute model, which can be derived from experimental data. A numerical model of a wing section is considered in this work to assess the presented approach. It is shown that LCOs with a fair agreement in amplitude and flight speed of LCO onset in comparison to the nonlinear simulation can be obtained.
30 mins
Post-flight system identification and aeroservoelastic model updating for prediction and validation of the onset of flutter
Özge Süelözgen, Gertjan Looye, Thiemo Kier, Matthias Wüstenhagen, Ramesh Konatala, Keith Soal, Nicolas Guérin, Bálint Vanek
Abstract: A validated mathematical aircraft model allows, among other things, the extensive study of system performance and characteristics, the verification of wind-tunnel and analytical predictions, the support of flight envelope expansion during prototype testing, and the design of flight control laws. The aeroservoelastic (ASE) stability analysis is another crucial component of the configuration optimization and certification process over the intended operational envelope. In this context, the flutter phenomenon is a well-known example of a selfexcited aeroelastic instability resulting from the interaction between unsteady aerodynamic forces and structural vibrations. Significant amplitudes of vibration can be induced, eventually resulting in the structure's catastrophic failure. With the guidance of accurate aeroservoelastic models, a reliable prediction of an aircraft's susceptibility to flutter across its intended flight envelope is possible. Through developments in control systems theory and hardware, as well as the development of high-bandwidth actuators, it became feasible to suppress aircraft flutter instabilities through the actively controlled closed-loop action of control surfaces. Active flutter suppression (AFS) presents a robust and effective solution when passive approaches, such as modifying mass distribution or structural stiffening, are impracticable for eliminating flutter. The EU-funded projects FLEXOP and FliPASED investigated this topic and made technological advancements to the point where the AFS control laws were successfully demonstrated by the P-FLEX UAV during flight tests. Using data from flight tests of the fixed-wing P-FLEX UAV with a 6m wing span, this paper will demonstrate post-flight system identification and, by extension, ASE model updating outcomes. The predictions provided by the updated model regarding flutter boundary will be thoroughly assessed. An additional significant topic is the post-flight verification of the openloop flutter speed obtained from Operational Modal Analysis during flight flutter testing. In this context, the onset of the flutter during flight testing is demonstrated theoretically through the flight modal identification from the simulated flight flutter test. This is achieved through a qualitative comparison of the stability diagrams of the system's poles, which were gathered from flight test data and simulated flight flutter testing. Finally, flutter boundary expansion enabled by the AFS controller of the closed-loop system will be verified via post-flight modal identification.
30 mins
Experimental study on transonic buffeting of launch vehicles with large-diameter fairing using elastic models
Chen Ji
Abstract: The launch vehicle may experience transonic buffeting during atmospheric ascent. Usually, fluctuation pressure wind tunnel testing with rigid model is performed to assess the buffeting loads. However, for the shapes prone to buffeting, such as the large-diameter fairing with a hammerhead nose shape, it is recommended by NASA SP8001 to use an elastic model to study their buffeting response and evaluate potential hazards. In this paper, a set of elastic models with same structural dynamic characteristics and different diameters of fairing were investigated for theirs transonic buffeting behaviours. These configurations with different fairing-core diameter ratios, which are 1.55, 1.60, and 1.73, respectively. The aeroelastic damping and buffeting load response characteristics of different diameter-ratio configurations were obtained by conducting aeroelastic damping tests and buffeting load response tests. From the aeroelastic damping test results, it is shown that the model with a diameter ratio of 1.60 exhibits negative aerodynamic damping at certain Mach numbers and Angle-of-Attack. Thus, solely from the perspective of aerodynamic damping, it is not possible to conclude that larger diameter ratios lead to instability or increased buffeting. However, from the buffeting load response test results, it is evident that with an increase in diameter ratio, the structural dynamic load response, in terms of both response amplitude and the Mach number range of significant amplitude, increases, especially for the 1.73 diameter ratio case. This implies that with larger diameter ratios, the buffeting response amplitude becomes more severe, and the duration of significant response increases. Therefore, for the transonic buffeting issue of launch vehicles with large-diameter fairings, evaluation should be conducted from both the perspectives of aeroelastic damping stability and buffeting response dynamic loads.

end %-->