Experimental and numerical investigations of a 2D aeroelastic arifoil encountering a gust in transonic conditions Fabien Huvelin, Arnaud Lepage, Sylvie Dequand To cite this version: Fabien Huvelin, Arnaud Lepage, Sylvie Dequand. Experimental and numerical investigations of a 2D aeroelastic arifoil encountering a gust in transonic conditions. CEAS Aeronautical Journal, Springer, In press, 10.1007/s13272-018-00358-x. hal-02027968 HAL Id: hal-02027968 https://hal.archives-ouvertes.fr/hal-02027968 Submitted on 22 Feb 2019 HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés. 1 EXPERIMENTAL AND NUMERICAL INVESTIGATIONS OF A 2D 2 AEROELASTIC AIRFOIL ENCOUNTERING A GUST IN TRANSONIC 3 CONDITIONS 4 Fabien HUVELIN1, Arnaud LEPAGE, Sylvie DEQUAND 5 6 ONERA-The French Aerospace Lab 7 Aerodynamics, Aeroelasticity, Acoustics Department 8 Châtillon, 92320 FRANCE 9 10 KEYWORDS: Aeroelasticity, Wind Tunnel Test, CFD, Gust response, Unsteady coupled 11 simulations 12 ABSTRACT: 13 In order to make substantial progress in reducing the environmental impact of aircraft, a key 14 technology is the reduction of aircraft weight. This challenge requires the development and 15 the assessment of new technologies and methodologies of load prediction and control. To 16 achieve the investigation of the specific case of gust load, ONERA defined a dedicated 17 research program based on both wind tunnel test campaigns and high fidelity simulations. 18 To reach the experimental objectives, a set-up was designed, manufactured and implemented 19 within the ONERA S3Ch transonic wind tunnel facility. The first component, called gust 20 generator, consists of two oscillating airfoils installed upstream of the wind tunnel test section 21 and allows to produce air flow deflections. The second component, the test model, is a two 22 degrees-of-freedom aeroelastic model of a supercritical airfoil. A test campaign has been 23 performed leading to the generation of databases for high fidelity tools validation. 24 These databases have been used in order to assess the capabilities of the elsA code (ONERA- 25 Airbus-Safran property) using its aeroelastic module and a gust model based on the Field 26 Velocity Method. A validation process has been defined in order to move from experimental 27 results obtained in the wind tunnel with wall boundaries to industrial modeling computed with 28 farfield boundaries. The full process was applied to a transonic case with sine gust excitation 29 signals. 30 ACKNOWLEDGEMENT 31 A part of the research leading to these results has received funding from the European 32 Union’s Seventh Framework Program (FP7/2007-2013) for the Clean Sky Joint Technology 33 Initiative under grant agreement CSJU-GAM-SFWA-2008-001. 34 The numerical studies presented in this paper have been partially funded by Airbus, Safran, 35 and ONERA which are co-owners of the software elsA. 36 1 INTRODUCTION 37 The aircraft load analysis leads to perform a large number of aerodynamic and aeroelastic 38 simulations. Especially, the aircraft has to be designed to withstand loads resulting from gusts. 39 Unsteady gust analyses rely usually on linear techniques in frequency domain, based on 40 simple Doublet Lattice Methods (DLM) for the aerodynamic flow prediction [1], [2]. These 41 techniques are valid for subsonic flows, but could sometimes be not accurate enough to get 1 Corresponding author : [email protected], Tel : +33 1 46 73 46 35 1 1 realistic responses in the transonic regime, characterized by strong nonlinearities such as 2 shocks and flow separation. Different approaches have been developed in order to simulate 3 the gust response with Computational Fluid Dynamics (CFD) tools [3], [4]. ONERA has 4 implemented in the elsA software the capability to compute high fidelity aeroelastic gust 5 responses, directly in time-domain, for different discrete gust shapes [5]. The objective is to 6 complement the classical aircraft manufacturer’s tools during the critical response 7 investigation phase by computing steady and unsteady aerodynamics with high fidelity tools 8 (CFD) for a reduced number of critical configurations. 9 It is therefore required to have relevant experimental data to validate the models and tools for 10 load prediction that will serve to design innovative approaches and to prove the efficiency of 11 load alleviation methodologies. In the academic or industrial research area, there are few 12 works conducted on the experimental investigation of gust in wind tunnel, especially for the 13 transonic regime. Studies have been performed in low speed regime based on various 14 mechanisms for creating an oscillating flow in a wind tunnel, as for example a rotating slotted 15 cylinder [6] or [7] and oscillating airfoils [8]. In the transonic speed range, very few wind 16 tunnel experiments can be identified: concepts based on oscillating vanes [9] or using a single 17 pitching airfoil [10] have been recently built in transonic wind tunnel environments. But 18 considering these studies, there is still a lot of work to perform in order to improve the 19 understanding of the physical phenomena involved in fluid structure interaction with a gust 20 perturbation and therefore to improve the definition of gust load alleviation strategies. 21 To achieve the experimental investigation, ONERA defined a research tool for gust load 22 investigation in Wind Tunnel Test (WTT) conditions within the key objectives: 23 • to improve the understanding of a large scope of unsteady physical phenomena 24 induced by gust with sensitivity tests relative to the Mach number (from subsonic at 25 0.3 to transonic at 0.73), the gust frequency (from 20 Hz to 80 Hz), the gust amplitude 26 (up to +/-1°) for aerodynamic and aeroelastic configurations; 27 • to generate a comprehensive and relevant experimental databases for the validation of 28 methodologies and numerical simulation such as CFD tools. 29 This research tool is developed using high fidelity numerical computations for a mutual 30 benefit. On one hand, the elsA CFD code (ONERA-Airbus-Safran property) using its 31 aeroelastic module assesses different gust generator concepts. On the other hand, the 32 experimental results are used for gust response validation of the elsA code. 33 2 NUMERICAL TOOLS 34 2.1 CFD/CSM numerical tool 35 The high fidelity simulation tool developed at ONERA for aeroelastic applications is based on 36 the elsA CFD solver (ONERA-Airbus-Safran property) for the flow computation. A general 37 framework has been developed in the aeroelastic module of elsA over the last decade, giving 38 access in a unified formulation of several types of aeroelastic simulations, while minimizing 39 the impact on the flow solver. The motivation of these developments, detailed in [11], [12], 40 [13] is to provide a numerical tool for the prediction of various aeroelastic phenomena such as 41 flutter or limit cycle oscillations and aerodynamic phenomena involving complex flows 42 including shocks, vortex flow, and flow separation. 43 2.2 Fluid-structure coupling with a modal approach 44 Governing equations of the aerodynamics 2 1 A full detailed overview of the flow solver development can be found in [14], [15]. For an 2 aeroelastic problem, an Arbitrary Lagrangian Euler (ALE) formulation is used. The 3 formulation adds a grid velocity ( ) in the standard Eulerian formulation allowing to 4 switch from the Eulerian formulation (fixed grid, =0) to the lagrangian formulation 5 (moving grid, = ). : + = 0 (1) � � � ⋅ Ω ∂Ω (2) : + = ( + ) � � � ⋅ � ∇ ⋅ Ω ∂Ω Ω (3) : + = ( ( ) + ) � � � ⋅ � ∇ ⋅ ⋅ ⋅ 6 where, t is the time, Ω the reference∂Ω coordinate, Ω an arbitrary volume with a surface 7 boundary , the density, the convective velocity, the normal to the boundary surface, 8 the momentum, the specific total energy, theΩ Cauchy tensor, the material velocity 9 and the ∂specificΩ body force vector. The convective velocity is expressed with the material 10 velocity and the grid velocity: ( , ) = ( , ) ( , ) (4) 11 The unsteady aeroelastic simulation s are performed − using the dual time-stepping technique, 12 allowing large physical time-steps and mesh and grid velocity updates from the structural 13 solver during the iterative process. 14 Gust modelling in flow solver 15 There are several possibilities in order to implement a gust perturbation in CFD codes. One 16 way consists in introducing the gust velocity as a disturbance on the farfield boundary 17 conditions of the computational domain [16]. The gust fluctuation is then propagated into the 18 domain by solving the Navier-Stockes equations. The main advantage of this approach is to 19 take into account the effect of the gust on the aircraft, but also the reverse effect of the aircraft 20 on the gust. The main disadvantage of such modeling is the required grid refinement in order 21 to propagate the gust from the upstream boundary to the structural model without any 22 numerical dissipation schemes. 23 An alternative approach is to use the so-called Field Velocity Method (FVM) suggested by 24 Sitaraman et al [17] which has been implemented in elsA [5], [18]. A prescribed gust velocity 25 field ( ), depending on both space and time, is added to the grid velocity in each cell of 26 the aerodynamic grid: ( , ) ( , ) + ( , ) (5) 27 This modeling avoids the numerical → dissipation due to the grid refinement but does not take 28 into account the reverse effect of the aircraft on the gust [16], [19]. The absence of the aircraft 29 disturbance on the gust compromise the flow physics for small gust length (less than 5 chord) 30 and has a minor impact on large gust length (more than 30 chords) comparing to a split 31 velocity method [20].