Numerical Database Of Stall Flutter Gust Response - AGREE Project

NUMERICAL DATABASE OF STALL FLUTTER GUST RESPONSE AGREE Project – Deliverable #10 Authors: Konstantinos Rekoumis National Technical University of Athens The research project is implemented in the framework of H.F.R.I call “Basic research Financing (Horizontal support of all Sciences)” under the National Recovery and Resilience Plan “Greece 2.0” funded by the European Union – NextGenerationEU (H.F.R.I. Project Number:016749) CONTENTS K. Rekoumis Contents 1 Introduction 2 2 Examined cases 2 3 Geometrical Configuration 4 4 Numerical Setup 5 4.1 CFDSoftware ............................................. 5 4.2 ComputationalGrid ......................................... 5 4.2.1 Gustgenerationcase .................................... 6 4.2.2 Gustresponsecase...................................... 6 5 Data Organization 6 5.1 GustGenerationCase ........................................ 7 5.2 GustResponseCase ......................................... 7 1 Introduction This document contains information for the numerical database created to characterize gust-induced airfoil stall flutter. The aim of the present report is to present the data in a clear and concise manner for others to use. The present work investigates the numerical modeling of gust-induced stall flutter oscillations of a NACA 64-418 airfoil. The computational framework aims to numerically model the experimental apparatus installed in the NTUA Wind Tunnel , which combines a vane type Gust Generator (GG) and an elastically supported airfoil model [1]. In section 2, the examined cases will be described. In section 3 the geometrical arrangement and in section 4 the numerical setup employed will be discussed respectively. Finally, in section 5 the data organization protocol will be presented. 2 Examined cases This numerical investigation on the gust-induced stall flutter is based on experiments performed in the NTUA Wind Tunnel [2], where a vane type gust generator creates a gust inflow, which in turn sets an elastically mounted airfoil on a stall flutter motion. To perform the present study a two-step approach was followed. First, the gust was studied in an empty Wind Tunnel, see Figure 1. The generated gust inflow consisted of a series of four individual gusts, separated by 0.8 sec between them, see Figure 2. To generate each individual gust, the gust generation foils (hereafter called vanes) followed a “1-cos” profile (see eq. 1) with an Amplitude A= 34 deg and an oscillation frequency f= 7 Hz. The free-stream velocity U∞= 17.94 m/sis kept constant throughout this study. θ=(A 2[1 −cos(2πfδt)] , t ∈[t0, t0+1 f] 0, else (1) ,where δt =t−t0,f[Hz]the excitation frequency, and A[deg]the target vane angle. The velocity field was measured in several stations along the span of the computational domain at the same xy coordinate . The x coordinate is 1295 mm downstream of the gust generation vanes’ TE, and the y is +140 mm off the Centerline of the Tunnel. The z coordinates distribution is given in the table 1. To get better context of the arrangement, see figure 1. 2 K. Rekoumis Figure 1: Monitor point location on the xy plane and spanwise arrangement Table 1: Spanwise arrangement of the monitor points Monitor Point Z [m] A 0.1 B 0.2 C 0.3 D 0.4 E 0.5 F 0.6 Second, the gust response case is examined, where the airfoil is coupled with a torsional springdampener system and is able to freely oscillate about its elastic axis while having all its other degrees of freedom constrained. The details of the airfoil are given in table 2. It is noted that, the airfoil’s elastic axis is parallel to the Global Z axis and consequently no gravitational effects are considered. Initially, the airfoil was stabilized in the flow and then subjected to four gusts according to the timeseries in the figure 2. Table 2: Details of the elastically supported airfoil model Foil Profile NACA 64-418 Foil chord 500 mm Elastic Axis - x 35 % chord Elastic Axis - y 1.6 % chord Moment of Inertia 0.7282 kg m2 Damping ratio 0.21 Spring Coefficient 254 Nm/rad 3 K. Rekoumis 0.0 0.5 1.0 1.5 2.0 2.5 time - t 0 [sec] 0 5 10 15 20 25 30 35 vane [deg] Figure 2: The angle timeseries of the gust generation vanes The timeseries are given in offset form, where the time is offset by t0, where t0is the actuation time instance [s]. 3 Geometrical Configuration The Wind Tunnel test section has a height of 1.4 m and width 1.8 m (see table 3). The gust generation vanes and the airfoil model spanned the wind tunnel vertically. The distances between the gust generation vanes and the airfoil model are given in Figure 3. The former had a NACA 0015 profile and a chord of 200 mm while the latter had a NACA 64418 profile and a chord of 500 mm. Details about the test section, the elastically mounted airfoil model and the gust generation are given in table 2 and table 3 respectively. Table 3: Wind Tunnel & Gust Generator dimensions Wind Tunnel Working Length 3.3 m Geometry Width 1.8 m Height 1.4 m Gust Generation Number 4 Vanes Vane chord 200 mm Vane Profile NACA 0015 Center of Rotation 25 % chord 4 K. Rekoumis Figure 3: Top view of the computational domain with dimensions for the gust response case. All dimensions are in mm. 4 Numerical Setup 4.1 CFD Software In the present work, all numerical simulations were done using the in-house URANS code MaPFlow [3], [4]. In both gust generation and aero-elastic scenarios MaPFlow resolved the incompressible unsteady Reynolds Averaged Navier-Stokes equations, employing an artificial compressibility scheme. In both cases, to model turbulence Menter’s K-omega SST model [5] was used. Also, the boundary layer’s transition was modeled using the γ−Reθ transition model [6]. Finally, MaPFlow can resolve FluidStructure Interaction problems internally, via its strongly coupled Rigid Body Dynamics solver [4]. 4.2 Computational Grid For each of the two examined cases, both a 2D and a 3D computational grid were constructed. To construct each 3D grid, the respective 2D grid was extruded by a certain amount of steps in the Z-axis, exploiting the xy plane symmetry of the Wind Tunnel Section. Each grid has a total length of 10 m in the stream-wise direction, where 5 m in the middle are densely meshed (dx=dy=10 mm). The width of the computational grid is equal to the Wind Tunnel’s width (for details see table 3). In figure 4, the different boundary elements of the computational domain are color coded and presented in the included table. The bounding walls elements are modeled as Inviscid walls and the inflow & outflow elements with Inflow and Outflow Boundary Conditions respectively. Both the gust generation vanes and the airfoil elements are modeled with viscous wall boundary conditions and the y+ value is kept ≤1. 5 K. Rekoumis Figure 4: Geometrical Configuration of the Aero-elastic setup Each different element of the setup is color - coded differently. Notice that each solid boundary (gust generation vanes & airfoil) is treated as a separate entity, hinting at the versatility of our setup. 4.2.1 Gust generation case Both 2D and 3D simulations were considered for this case with practically identical results. For completeness the 3D grid and results are provided in this database. To generate the 3D grid, the 2D grid was extruded in the Z direction up to z=700 mm in 70 steps. The 2D and 3D mesh had 173k and 12.1M cells, respectively. 4.2.2 Gust response case Both 2D and 3D simulations were considered for this case, too. To generate the 3D grid, the 2D grid was extruded in the Z direction up to z=700 mm in 27 steps. The 2D and 3D mesh had 206k and 5.57M cells, respectively. It is noted that unlike the gust generation case, in the gust response case the 2D and 3D simulations lead to distinctively different results. Specifically, in the 2D case the airfoil response to the specific gust inflow is to start a limit cycle oscillation (LCO), whereas in the 3D case the oscillation is damped out. 5 Data Organization In the present section, the data naming and structure conventions for using the provided data will be described. All numerical data provided are in a .csv (comma-separated values) text file for ease of use. 6 5.1 Gust Generation Case K. Rekoumis The first row will contain the name of the respective column at each cell. In the subsections that follow, information is provided on how to read the uploaded data as well as the context of the numerical values. Finally, the computational grids for each simulation are present inside the simulation’s directory. The computational grids are stored in a .cgns file making them accessible with a wide range of applications. 5.1 Gust Generation Case In the present case, 6 data files and one computational grid are uploaded in the gust_generation directory. First, the comp_grid.cgns contains the computational grid. Then, the monitor_point_X.csv (X takes values from the ”Monitor Point” column of table 1) files contain the velocity vector at the respective point in space for each time instance. The mapping of header names to z coordinate follows the table 4. Table 4: Mapping of header names to physical context for monitor point files 1 time 7−→ Simulation Time [s] 2 velx 7−→ Flow Velocity x component [m/s] 3 vely 7−→ Flow Velocity y component [m/s] 4 velz 7−→ Flow Velocity z component [m/s] 5.2 Gust Response Case The files for the gust response case can be found under the gust_response directory. There, two subdirectories exist: the gust_response/2d_case for the 2-D simulation and the gust_response/3d_case for the 3-D simulation. To provide insight in the movement performed by the airfoil, files containing the movement and loading timeseries are uploaded. Inside the directory of each aero-elastic case you find the following files: 1. hist_rbd.csv : The file containing the motion history of the airfoil as it is resolved by the RBD solver. 2. hist_loads.csv : The file containing the raw loads and their non-dimensional counterparts developed on the global orthogonal axes. These loads are the direct result of integrating the pressure field and the shear stresses on the solid boundary. 3. comp_grid.cgns : The computational grid for the respective case. To ensure that the information of the files is relayed properly, in tables 5 and 6 an explanation about each columns data is presented. Table 5: Columns’ header explanation for hist_rbd.csv file 1 time 7−→ Simulation Time [s] 2 aoa 7−→ Angle of Attack [deg] 3 rot_speed 7−→ Rotation Speed [rad/s] 4 rot_accel 7−→ Rotational Acceleration [rad/s2] 7 REFERENCES K. Rekoumis Table 6: Columns’ header explanation for hist_loads.csv file 1 time 7−→ Simulation Time [s] 2 Fx 7−→ Force along the Global x-axis [N] 3 Fy 7−→ Force along the Global y-axis [N] 4 Mz 7−→ Moment about the Global z-axis [Nm] 5 Cx 7−→ Coefficient of Force along the Global x-axis [-] 6 Cy 7−→ Coefficient of Force along the Global y-axis [-] 7 Cm 7−→ Coefficient of Moment about the Global z-axis [-] References [1] M. Manolesos, C. Ampatis, D. Gkiolas, N. Papakonstantinou, A. Alexandris-Galanopoulos, K. Rekoumis, and G. Papadakis, Design of a wind tunnel set up to measure airfoil aeroelastic gust response, Zenodo, Jul. 2025. doi: 10.5281/zenodo.17405080. [2] D. Gkiolas and M. Manolesos, Experimental database of gust induced stall flutter, Zenodo, Nov. 2025. doi: 10.5281/zenodo.17664469. [3] G. Papadakis, “Development of a hybrid compressible vortex particle method and application to external problems including helicopter flows,” Ph.D. dissertation, National Technical University of Athens School of Mechanical Engineering, 2014. [4] D. Ntouras, “Adaptation of the artificial compressibility formulation for free surface flows with applications in ship & marine hydrodynamics,” Ph.D. dissertation, National Technical University of Athens School of Naval Architecture & Marine Engineering, 2023. [5] F. R. Menter, “Two-equation eddy-viscosity turbulence models for engineering applications,” AIAA Journal, vol. 32, no. 8, pp. 1598–1605, 1994. doi: 10.2514/3.12149. [6] F. R. Menter, P. E. Smirnov, T. Liu, and R. Avancha, “A One-Equation Local Correlation-Based Transition Model,” Flow, Turbulence and Combustion, vol. 95, no. 4, pp. 583–619, Dec. 2015. doi: 10.1007/ s10494-015-9622-4. 8