Fluid flow is a critical part of defense and aerospace design, whether the product is a rocket’s fuel valve or a tank’s turret-mounted electronics pod. Military and aerospace systems must function reliably under the most adverse conditions, and commonly, that involves fluid forces — internal or external. The fluid itself may be aerodynamic airflow or a volatile liquid, each with its own characteristics. But the discipline of fluid dynamics encompasses them all.

Figure 1: Conventional (“traditional”) CFD. Designs must exit the CAD flow for each round of CFD analysis.
The study of fluid flow requires the most complex sort of math; in fact, the underlying Navier-Stokes equations are said to be insolvable by human hands within reasonable time. But a solution does exist in the form of Computational Fluid Dynamics (CFD), an analysis method based on numerical approximations.

Not surprisingly, CFD analysis today is the province of dedicated applications running on desktop or mainframe computers. But even with a computer’s help, the process has always required the expertise of specialists who can minister to CFD’s technical demands. Most mechanical engineers are experts with their MCAD tools, but CFD analysis has historically been outside their realm.

What’s so difficult about CFD? Imagine analyzing a manifold that conducts atomized fuel. First, the MCAD data must be ported to the CFD tool. Before the analysis can begin, additional details about the manifold’s dimensions and topology must be fed into the CFD application. Where the fluid flows through hollow areas in the design, these must be separately defined and modeled as solids. Where the engineer sees smooth curves and angles, the CFD tool needs to see a “mesh” of discretized features. This complex calculating grid must be devised and boundary conditions must be defined.

The MCAD and CFD domains have always been disparate worlds. Until recently, there was no expedient way for an MCAD application to deliver its knowledge of the design to a flow analysis tool. Complex transfers between the two environments add time and increase the chance of miscommunication between the engineers, if not the tools.

Conventional vs. Concurrent

Figure 2: Concurrent CFD.
Figure 1 illustrates the flow of a conventional CAD engineering process involving several CFD steps. Each CFD step “interrupts” the design process and takes time doing so. As the design advances through the flow, the designer must interact with the external CFD process repeatedly, accruing delays in the export, the import, and the CFD analysis itself. Moreover, multiple prototype iterations (not shown) require the same CFD loop as the preceding milestones.

In practice, the time spent on external CFD processes can limit the number of attempts to perfect one’s design. And in the world of defense and aerospace applications, designs need to be as perfect as humanly possible.

Think about a typical MCAD toolset. It automates many labor-intensive calculations and record-keeping operations. It imposes design rules and checks the user’s work. It draws and dimensions the emerging device on-screen while recording details about solids, surfaces, apertures, outlets, and chambers.

This is exactly the information the CFD process needs. A new technology known as Concurrent CFD is able to use this data directly, bringing MCAD/CFD integration to mechanical designers for the first time. Solutions such as Mentor Graphics FloEFD permit engineers to perform CFD analysis without leaving their familiar MCAD tool environment. Sophisticated fluid dynamics analysis is just another choice on the MCAD menu.

Figure 2 underscores the contiguous nature of a Concurrent CFD flow. There are no detours into costly external processes. The mechanical engineer maintains full control and visibility of the design throughout the process.

Concurrent CFD automates tedious/ intimidating steps such as configuring solid representations of hollow spaces, designing meshes, and more. Notably, the mesh is a variable-resolution type whose cells adapt in size to deliver optimal detail everywhere in the device under analysis. In the conventional CFD world, that step would require a seasoned practitioner.

The Concurrent CFD approach has been shown to produce accurate results in as little as 25% of the time required for conventional CFD methods.

Concurrent CFD Goes to Work on Defense Projects

Concurrent CFD is finding its way into defense and aerospace projects around the world. Designers are using the tools to check fluid flow characteristics early and often, and to validate evolving virtual prototypes through two iterations or twenty.

Image of an array addressed with 16 x 8 boxes.
A case in point is a project at Shaw Aero Devices (now part of Parker Aerospace Fluid Systems) in Naples, FL. The company was asked to develop a new solenoid fuel control valve for use in unmanned aerial vehicles (UAV). If engineers could reduce the pressure drop compared to the existing valve, then the UAV could use a smaller fuel pump and carry a larger payload.

The task involved redesigning the valve to eliminate constrictions and reduce the pressure drop as much as possible. Normally, such a redesign would require at least three prototypes at a cost of about $3,000, and one month each.

Using a Concurrent CFD toolset, engineers were able to build their first virtual prototype in just one day. With the embedded CFD tool, there was no need to translate and retranslate data as it moved from CAD to CFD and back repeatedly.

By modeling a similar existing valve, designers created a baseline that enabled them to make informed decisions. In the first design iteration, a cut plot of the pressure gradient (similar to Figure 3) revealed that an angled wing seat was a key source of the pressure drop. Designers modeled a new valve with a larger opening and eliminated the angled valve seat.

Ultimately the new valve design reduced the pressure drop from 6.09 psi (at a flow rate of 4.45 gal/min) to 0.71 psi — an 88% improvement that met the customer’s exacting specifications. And with days instead of months spent on prototyping, the project stayed on schedule.

In another instance, Bell Helicopter engineers used Concurrent CFD modeling to redesign the nitrogen-fill flow in a combat helicopter fuel tank. Nitrogen is used to supplant air in the fuel tank with the intent of preventing an explosion if the tank is pierced.

The simulation homed in on several areas in the tank that were not being adequately vented. Had this issue remained undetected until the physical prototype was tested, it would have been far more expensive to correct the problem and moreover, might have delayed the project.

Conclusion

Concurrent CFD solutions can transform the mechanical engineer’s job where fluid flow analysis is concerned. These embedded CFD tools speed design validation steps and encourage experimentation. Concurrent CFD moves fluid flow analysis into the designer’s everyday kit of MCAD tools, saving time and money in defense and aerospace product development.

This article was written by Dr. Keith Hanna, Product Line Director, Mechanical Analysis Division, at Mentor Graphics Corp., Wilsonville, OR. For more information, Click Here .