2014•Deep Blue (University of Michigan)Open access

High-Fidelity Aerodynamic Design Optimization of Aircraft Configurations.

Zhoujie Lu

Open full text 4 citations

Abstract

With increasing fidelity and efficiency of numerical simulations, it becomes possible to rely on computational simulations and optimization to achieve a better aircraft design. One of the most computationally intensive disciplines is the aircraft external aerodynamic design. Computational fluid dynamics based on Reynold-averaged Navier--Stokes equations is necessary to accurately resolve the flow field in order to achieve a practical design. High-fidelity CFD poses difficulties to numerical optimization due to its high computational cost, especially when large number of shape design variables are used. This thesis presents an approach to compute the gradients of Reynold-averaged Navier--Stokes equation equations with a Spalart--Allmaras turbulence model using a combination of the adjoint method and automatic differentiation algorithms, for use in gradient-based aerodynamic shape optimization. The resulting gradients are accurate, robust, and efficient. With this state-of-the-art Reynolds-averaged Navier--Stokes adjoint and aerodynamic shape optimization framework, we performed three high-fidelity aerodynamic design optimization studies in this thesis. The wing of a Boeing 777-sized aircraft is optimized for single and multiple flight conditions. The drag coefficient is minimized with respect to 720 shape design variables, subject to lift, pitching moment, and geometric constraints, using grids with up to 28.8M cells. Drag coefficient of the optimized design was reduced by 8.5% relative to the initial design. The second application is to optimize the aerodynamics of a near-term aircraft retrofit modification: a wing with morphing trailing edge. A drag reduction in the order of 1% is achieved for on-design conditions, and reductions up to 5% were achieved for off-design conditions. Finally, we extend the aerodynamic shape optimization studies to design an unconventional configuration, the blended-wing-body aircraft. The best compromise between performance and stability was achieved by enforcing a small static margin that can be tolerated in a commercial airplane (1%) and including the center of gravity position as a design variable. This resulted in a trimmed configuration that exhibits a nearly elliptical lift distribution and the lowest drag among the trimmed stable designs. This was achieved by a combination of optimized washout and reflex airfoils.

About this research paper

What this paper is about

With increasing fidelity and efficiency of numerical simulations, it becomes possible to rely on computational simulations and optimization to achieve a better aircraft design. One of the most computationally intensive disciplines is the aircraft external aerodynamic design. Computational fluid dynamics based on Reynold-averaged Navier--Stokes equations is necessary to accurately resolve the flow field in order to achieve a practical design. High-fidelity CFD poses difficulties to numerical optimization due to its high computational cost, especially when large number of shape design variables are used. This thesis presents an approach to compute the gradients of Reynold-averaged Navier--Stokes equation equations with a Spalart--Allmaras turbulence model using a combination of the adjoint method and automatic differentiation algorithms, for use in gradient-based aerodynamic shape optimization. The resulting gradients are accurate, robust, and efficient. With this state-of-the-art Reynolds-averaged Navier--Stokes adjoint and aerodynamic shape optimization framework, we performed three high-fidelity aerodynamic design optimization studies in this thesis. The wing of a Boeing 777-sized aircraft is optimized for single and multiple flight conditions. The drag coefficient is minimized with respect to 720 shape design variables, subject to lift, pitching moment, and geometric constraints, using grids with up to 28.8M cells. Drag coefficient of the optimized design was reduced by 8.5% relative to the initial design. The second application is to optimize the aerodynamics of a near-term aircraft retrofit modification: a wing with morphing trailing edge. A drag reduction in the order of 1% is achieved for on-design conditions, and reductions up to 5% were achieved for off-design conditions. Finally, we extend the aerodynamic shape optimization studies to design an unconventional configuration, the blended-wing-body aircraft. The best compromise between performance and stability was achieved by enforcing a small static margin that can be tolerated in a commercial airplane (1%) and including the center of gravity position as a design variable. This resulted in a trimmed configuration that exhibits a nearly elliptical lift distribution and the lowest drag among the trimmed stable designs. This was achieved by a combination of optimized washout and reflex airfoils.

Why it matters

OpenAlex reports 4 citations for this work. Citation counts describe recorded attention and do not establish research quality.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

With increasing fidelity and efficiency of numerical simulations, it becomes possible to rely on computational simulations and optimization to achieve a better aircraft design. One of the most computationally intensive disciplines is the aircraft external aerodynamic design. Computational fluid dynamics based on Reynold-averaged Navier--Stokes equations is necessary to accurately resolve the flow field in order to achieve a practical design. High-fidelity CFD poses difficulties to numerical optimization due to its high computational cost, especially when large number of shape design variables are used. This thesis presents an approach to compute the gradients of Reynold-averaged Navier--Stokes equation equations with a Spalart--Allmaras turbulence model using a combination of the adjoint method and automatic differentiation algorithms, for use in gradient-based aerodynamic shape optimization. The resulting gradients are accurate, robust, and efficient. With this state-of-the-art Reynolds-averaged Navier--Stokes adjoint and aerodynamic shape optimization framework, we performed three high-fidelity aerodynamic design optimization studies in this thesis. The wing of a Boeing 777-sized aircraft is optimized for single and multiple flight conditions. The drag coefficient is minimized with respect to 720 shape design variables, subject to lift, pitching moment, and geometric constraints, using grids with up to 28.8M cells. Drag coefficient of the optimized design was reduced by 8.5% relative to the initial design. The second application is to optimize the aerodynamics of a near-term aircraft retrofit modification: a wing with morphing trailing edge. A drag reduction in the order of 1% is achieved for on-design conditions, and reductions up to 5% were achieved for off-design conditions. Finally, we extend the aerodynamic shape optimization studies to design an unconventional configuration, the blended-wing-body aircraft. The best compromise between performance and stability was achieved by enforcing a small static margin that can be tolerated in a commercial airplane (1%) and including the center of gravity position as a design variable. This resulted in a trimmed configuration that exhibits a nearly elliptical lift distribution and the lowest drag among the trimmed stable designs. This was achieved by a combination of optimized washout and reflex airfoils.

Key concepts: Aerodynamics, High fidelity, Fidelity, Aerospace engineering, Computer science, Aeronautics, Engineering, Electrical engineering

Related papers

Back to paper searchBrowse research topicsOriginal source
High-Fidelity Aerodynamic Design Optimization of Aircraft Configurations. — Research Paper | ScholarLens