Andrea Da Ronch

Adaptive design of experiments for efficient and accurate estimation of aerodynamic loads

Purpose The purpose of this paper is to document an efficient and accurate approach to generate aerodynamic tables using computational fluid dynamics. This is demonstrated in the context of a concept transport aircraft model. Design/methodology/approach Two designs of experiment algorithms in combination with surrogate modelling are investigated. An adaptive algorithm is compared to an industry-standard algorithm used as a benchmark. Numerical experiments are obtained solving the Reynolds-averaged Navier–Stokes equations on a large computational grid. Findings This study demonstrates that a surrogate model built upon an adaptive design of experiments strategy achieves a higher prediction capability than that built upon a traditional strategy. This is quantified in terms of the sum of the squared error between the surrogate model predictions and the computational fluid dynamics results. The error metric is reduced by about one order of magnitude compared to the traditional approach. Practical implications This work lays the ground to obtain more realistic aerodynamic predictions earlier in the aircraft design process at manageable costs, improving the design solution and reducing risks. This may be equally applied in the analysis of other complex and non-linear engineering phenomena. Originality/value This work explores the potential benefits of an adaptive design of experiment algorithm within a prototype working environment, whereby the maximum number of experiments is limited and a large parameter space is investigated.

On Uncertainty Quantification of the Flow Predictions around the NATO STO AVT-251 Unmanned Combat Aerial Vehicle

Turbulence models based on Reynolds-averaged Navier-Stokes (RANS) equations remain the workhorse in the computation of high Reynolds-number wall-bounded flows. While these methods have been deployed to design the configuration developed within the NATO STO AVT-251 Task Group, their deficiencies in modelling complex flows are well-documented. However, an understanding of the sources of errors and uncertainties in RANS solvers, arising for example from different numerical schemes and flow modelling techniques, is missing to date. The aim of this work is to establish and quantify the impact that epistemic uncertainties within RANS solvers have on the flow predictions (shock wave locations, vortex breakdown, etc.). This will produce a range of all possible values of interest due to the inherent uncertainty of RANS solvers, which is expected to be highly dependent on the flow conditions and geometry configuration. This information, in turn, will be used to establish the robustness of the AVT-251 design and its performance metrics considering uncertain predictions of the dominant flow features. The benefits of this work will also extend to the structural design, whereby appropriate factors of safety can be integrated in the process.

Data-driven Optimisation of Closure Coefficients of a Turbulence Model

The solution of the Reynolds-averaged Navier-Stokes equations employs an appropriate set of equations for the turbulence modelling. The closure coefficients of the turbulence model were calibrated using empiricism and arguments of dimensional analysis. These coefficients are considered universal, but there is no guarantee this property applies to test cases other than those used in the calibration process. This work aims at revisiting the universality of the closure coefficients of the original Spalart-Allmaras turbulence model using machine learning, adaptive design of experiments and accessing a high-performance computing facility. The automated calibration procedure is carried out once for a transonic, wall-bounded flow around the RAE 2822 aerofoil. It was found that: a) an optimal set of closure coefficients exists that minimises numerical deviations from experimental data; b) the improved prediction accuracy of the calibrated turbulence model is consistent across different flow solvers; and c) the calibrated turbulence model outperforms slightly the standard model in analysing complex flow features around the ONERA M6 wing. A by-product of this study is a fully calibrated turbulence model that leverages on current state-of-the-art computational techniques, overcoming inherent limitations of the manual fine-tuning process.

Computational-Fluid-Dynamics-Based Aeroservoelastic Analysis for Gust Load Alleviation

Gust load alleviation using computational fluid dynamics as source of the aerodynamic predictions is carried out in the time domain. To this goal, an aeroservoelastic reduced-order model is generat...

Aerodynamic Shape Optimisation of Benchmark Problems Using SU2

In this paper, results are presented from the application of an open–source code, SU2, to a suite of benchmark cases defined by the AIAA Aerodynamic Design Optimisation Discussion Group. Two geometry parameterisation methods are employed, which are Hicks– Henne bump functions and Free–Form Deformation. Mesh deformation is achieved by solving linear elasticity equations. The adjoint solver within SU2 enables efficient sensitivity analysis, and gradient–based optimisation is performed using the SLSQP optimiser. The first optimisation problem studied is the drag minimisation of the NACA 0012 aerofoil in inviscid, transonic flow subject to a minimum thickness constraint. The shock wave is greatly weakened and moved downstream, achieving as much as 83% drag reduction. The second is the lift–constrained drag minimisation of the RAE 2822 aerofoil in transonic, viscous flow. The shock is eliminated, reducing drag by 38%. The NACA 0012 optimisation shows sensitivity to several numerical settings in the parameterisation approaches, whereas the RAE 2822 optimisation is insensitive to those parameter settings. The same pattern is also observed in the design variable dimensionality study. Moreover, for both two–dimensional optimisation problems, non–unique flow solutions exist on the optimised aerofoil. The third problem is the twist optimisation of a rectangular wing to minimise the induced drag at fixed lift in subsonic, inviscid flow. A nearly elliptical lift distribution is obtained using Free–Form Deformation twist parameterisation. The drag is reduced by approximately 1 count and an improved span efficiency is achieved.

An Efficient Implementation of Transonic Aeroelastic Tailoring based on a Reduced-Order Model using Structural Dynamic Reanalysis Method

Due to fuel efficiency, advanced aerodynamic and structural concern, more and more composite materials used in aircraft desgin. In composite structure aeroelastic tailoring process, an accurate and efficient method to evaluate the aeroelastic stability is very required. The traditional CFD-based POD/ROM has been shown its accuracy and efficiency for transonic aeroelastic analysis at fixed system. In aeroelastic tailoring process, in order to meet the requirements of aeroelastic performance, the parameters of the composite structure need to be modified repeatedly and the aerodynamic model have to be reconstructed. However, these reconstruction procedures take a considerable time, and greatly increasing the time cost of the aircraft design. To develop a more efficient composite structure aeroelastic tailoring method, starting with improving the efficiency of aeroelastic performance evaluation, this paper propose an approximate aeroelastic characteristics evaluation method based CFD-based POD/ROM by introducing the structural dynamic reanalysis method. The improved AGARD 445.6 composite wing was employed to verify the accuracy and efficiency of the proposed method. The results show that the proposed evaluation method can not only accurately predict the aeroelastic response of the structure, but also greatly improving the efficiency of transonic composite structure aeroelastic tailoring.

Dataset for Efficient Infinite-swept Wing Solver for Steady and Unsteady Compressible Flows

Supporting material: Franciolini, M. et al (2017). Efficient infinite–swept wing solver for steady and unsteady compressible flows. Aerospace Science and Technology, 1-25.The database contains data in support of the above manuscript. In particular, grid files and solution files that were used for the generation of the figures are made available. A toolbox that implements in Matlab a Discrete Fourier Transform algorithm is also shared. Each subfolder contains a README.dat file.