Aerion is on a mission to enable a step-change in travel by introducing a new era of supersonic flight – one that is more sustainable and efficient than the era of the Concorde. Technological breakthroughs, including simulation-driven design and composite materials, have allowed us to solve many of the challenges of the past. But simply solving the challenges of the past isn’t enough. Standards for noise and emissions have become increasingly strict over the years, and rightly so. At Aerion, we’re designing a supersonic jet that meets modern day community noise standards and emission requirements and gets people from A to B sustainably.

The design of such an aircraft requires the application of some of the most sophisticated optimization methods – and Aerion’s engineers are experts at just that. The aerodynamic engineers in Reno, and our software team in Palo Alto, have come together to develop design optimization capabilities more commonly found in university labs or research groups. With Aerion’s refined optimization capabilities, we have analyzed millions of design points – all to find that sweet spot that satisfies performance requirements, design constraints, and government regulations.

Let’s get into the weeds a bit and explain how we’ve constructed our design process. The aerodynamic design of the AS2 focuses on two major components: the airframe and the engine inlet. Although both systems influence one another, they exhibit distinct aerodynamic characteristics which require different design approaches. The design of the airframe is focused on optimizing the shape of the wing, the fuselage, and the empennage to minimize the aerodynamic drag on the aircraft when in flight while still allowing enough lift to perform necessary maneuvers. The design of the engine inlet focuses on ensuring that the air ingested by GE’s Affinity engine is at the right velocity and pressure across a huge range of flight conditions – take-off, climb, transonic cruise, supersonic cruise, and landing.

Let’s explain how we optimize the airframe. The biggest challenge with the airframe is reducing the wave drag (the drag caused by shockwaves that occur during supersonic flight). That wave drag can be significantly reduced by using thin wings, a slender fuselage, and a long, tapered nose – but these changes are usually at odds with constraints such as cabin volume, fuel tank size, structural stiffness, pilot visibility, and more. The way to handle these conflicts is to conduct a robust and intelligent search of the design space for a combination of design variables that perform well while satisfying the constraints – this is optimization in a nutshell. In order to develop an effective design optimization process, our engineers work with a high fidelity solver that can predict the aerodynamics of the aircraft accurately, a geometry modeler that can construct the outer mold line (OML) of the aircraft with high fidelity, and optimization software that can control an enormous number of design variables while accounting for constraints. The modeler, the solver, and the optimizer work together like the many parts of an orchestra with the design engineer as the maestro – constantly adjusting settings to make the search effective.

To many in the industry, design optimization is by no means a new concept. But here’s the problem – most of the methods and processes that have been widely adopted are either computationally expensive (meaning they take a long time to complete) or limited by how many design parameters they can tweak at once. Those weaknesses can be caused, among other things, by geometry modelers that don’t permit robust parametric variations, by solvers that are slow or difficult to automate, or by optimizers that are unable to efficiently search more than a handful of variables. By carefully finding, and in some cases building, the best tools to address each of those weaknesses, Aerion engineers were able to take their optimization game to the next level.

Any aircraft configuration determined by the optimization process also must be CAD (Computer-Aided Design) friendly, so engineers and other members of the Aerion team can use these solutions further along in the building process. Therefore, conventional “mesh-based” deformation methods would not work for our engineers’ purposes. So, our engineers developed a scriptable CAD framework called Pascale, which supports the automated CAD model generation and is used inside our optimization framework for geometry deformation. This “CAD in-the-loop” innovation has allowed the high-fidelity surface of the airframe to be obtained directly from the optimization process. It’s also enabled true scalability, which is one of the most challenging problems to overcome in simulation-driven design.

We evaluated several solvers and found that Cart3D, coupled with its Adjoint Design Framework, was the perfect answer for many of our design challenges. Cart3D’s CFD (Computational Fluid Dynamics) package provides a robust, automated meshing capability, meaning it can compute thousands of simulations for different designs without an engineer adjusting the settings. The Adjoint Design Framework is essentially an optimization process that couples with the flow solver; the name is derived from a mathematical construct called an “adjoint equation.” Adjoint methods belong to a class of optimization problems called gradient-based methods. Gradient-based methods search and find optimal points quickly for “well-behaved problems,” as opposed to gradient-free methods like simplex or genetic algorithms. Selecting an adjoint method provides one critical advantage: the time needed to compute the gradients (i.e. how will a small change in each variable impact the performance) is independent of the number of design variables. So now, instead of being limited to a dozen variables as we might have been with other methods, our engineers can control upwards of 120 simultaneously.

Again, the design of the engine inlet is also critical for meeting performance objectives. When the aircraft is cruising at supersonic speeds, the air that is entering the engine must decelerate to low subsonic speeds. In order to achieve that, the inlet needs to create strong shocks (these slow the air), but they also produce pressure losses. An optimal inlet design will slow down the airflow with minimal pressure losses. Inlet aerodynamics show significant non-linear variations, depending on the location of the shock, which makes it challenging for most optimizers to search. To get around the problems caused by such a non-linear design space, the engineers at Aerion have built out a surrogate model of the space using response surface methods (RSM). This model approximates the aerodynamic performance trends across the entire inlet design space using a subset of evaluations and allows for a much more effective search.

We’ve skimmed the surface on just a couple of ways that Aerion is innovating, bringing cutting edge design technologies to bear on difficult engineering challenges, but the culture of innovation is prevalent throughout our organization. Building a supersonic airplane is certainly an impressive feat no matter the circumstance, but our goals of doing it sustainably are ambitious, and even audacious. Sustainable global mobility will help bring people closer by enabling travel, saving time, and preserving the environment. It’s going to take a full-court press to rise to the challenge – lucky for us we brought our A game.

04 01 2020