Case Studies Events Recorded Webinars Partners Blog White Papers Sign up for News Plot Gallery Forum

Taming fuel slosh—Tecplot 360 helps researchers at Embry-Riddle Aeronautical University model fuel slosh behavior in spacecraft tanks

April 2009

Contributed by Brandon Marsell
Embry-Riddle Aeronautical University | http://www.erau.edu/

A rocket blasts off and begins its escape from the earth’s atmosphere. Most objects are securely fastened in place, making it possible to predict and manage the dynamics of the rocket’s parts and payload under increasing speed and pressure. But fuel is a different matter—its movement can’t easily be restricted. As fuel is consumed, the space it occupied is filled by air, allowing it to shake, slosh and create a reaction force. If the frequency of the reaction force is the same as that of another part of the rocket, the two frequencies can couple, sometimes catastrophically.

This Tecplot 360 animation illustrates the characteristics of fluid in a spherical tank as a function of time. The upper left portion shows the free surface as it sloshes, the lower left displays the fuel-to-air ratio, while the lower right shows the velocity vector field. The experimental setup used by ERAU researchers is pictured in the upper right.

In the 1950s and 1960s, when the space program was in its infancy, fuel slosh issues provided plenty of visual and dramatic rocket failures. Scientists and engineers have since addressed the problem by developing sophisticated fuel tanks with complex systems of baffles, diaphragms, and bladders to better control the motion of fuel. While these advanced containment systems can’t eliminate fuel slosh completely, they keep it within a tolerable margin of safety for today’s typical space missions. But better solutions are needed to support long-term space exploration, and researchers continue to work the problem. At Embry-Riddle Aeronautical University (ERAU) in Daytona Beach, Florida, Professor Sathya Gangadharan and graduate student Brandon Marsell are two such researchers.

"If you’re launching a satellite to Earth orbit, you burn all the fuel in the first ten minutes and then there’s no more fuel slosh problem. But on longer flights there’s a greater chance it’ll be an issue. If you need to get back from Mars, you will have to keep some fuel and you will need to be especially precise," says Marsell. "We need to understand how the fuel and air will behave in a different atmosphere and under different gravity. We’ll also need to send sizable tanks of liquids such as water and liquid oxygen for life support, which will have slosh issues of their own."

"Right now, they take a fuel tank, fill it with water, shake it, and get their numbers ... It’s expensive and time-consuming—a real hassle. What we’re trying to do is come up with a way of getting these numbers on the computer, to make it faster, simpler, and cheaper."

Brandon Marsell, Embry-Riddle Aeronautical University

Compounding these challenges, the development of safer fuel tanks still revolves around physical testing like that being conducted at ERAU’s state-of-the-art fuel slosh research facility. The ERAU lab has produced interesting results and a fairly reliable estimation process that can characterize the parameters of a mechanical pendulum analog model for predicting fuel behavior in a tank. Simulation and analysis may offer better results, but computers have not been powerful enough to handle the massive amounts of data needed to develop accurate, useful Computational Fluid Dynamics (CFD) fuel tank models.

This may change soon, however, thanks to the work being carried out by Gangadharan and Marsell. Under a grant from the NASA Launch Services Program at the Kennedy Space Center, the duo is developing CFD methods that may help reduce the amount of laboratory testing required.

NASA’s Launch Services group is responsible for getting every rocket ready for launch — if someone wants to launch a satellite, for example, Launch Services determines what is needed to put it up. This requires an accurate mathematical model of every rocket they use and parameters for each component of the models they describe, including the fuel. "Right now, they take a fuel tank, fill it with water, shake it, and get their numbers," says Marsell. "They need different numbers for each tank. It’s expensive and time-consuming—a real hassle. What we’re trying to do is come up with a way of getting these numbers on the computer, to make it faster, simpler, and cheaper."

Marsell and his group are currently working on a CFD model for free surface sloshing of liquid in spherical tanks. They begin by generating experimental data in ERAU’s fuel slosh research laboratory, then using that data to create a CFD model. The model is then tested for accuracy in further experiments so that refinements can be made.

In one experiment, a spherical fuel tank was suspended from a frame by cables and attached to a linear actuator. A force transducer was placed at the interface between the linear actuator and the fuel tank to measure the forces induced by the fuel slosh, and the data obtained was transmitted to a computer for analysis (figure 1). Using the Fluent 3D solver software to manage the data and Tecplot 360 to visualize it, the group then ran several test scenarios on the computer. In this case, they compared the computer results to those from their physical tests, and determined that the CFD method was accurate in predicting the damping rates and natural frequencies of the system, ultimately yielding valuable information about the sloshing fluid.

Figure 1: ERAU’s fuel slosh research laboratory is equipped with a state-of-the-art linear actuator and data acquisition system. In one case, researchers generated their experimental data using an eight inch in diameter spherical fuel tank suspended from a frame by cables and attached to a linear actuator.

"What we’re doing for NASA with these experiments is creating reliable parameters for each of their fuel tanks," says Marsell. "Eventually, they’ll be able to run their tests on the computer rather than in the lab—saving millions of dollars."

Possibly more exciting, says Marsell, is the prospect of being able to run tests on the computer that you couldn’t carry out safely in the lab.

" ... in a computer simulation, you can put hydrazine in a tank and shake it all you want. If it explodes, you just reboot the computer."

Brandon Marsell, Embry-Riddle Aeronautical University

"Propellants like hydrazine are toxic and explosive, so we run physical experiments using water instead, adjusting for the differences in the physical characteristics of the fuel afterward," he says. "But in a computer simulation, you can put hydrazine in a tank and shake it all you want. If it explodes, you just reboot the computer."

Computer modeling also allows scientists to test things that can’t be tested on Earth, such as how a liquid might behave in a Martian atmosphere or under different gravity.

"These CFD models are in their infancy. We’re just now getting the computer power to do it—and it’s really exciting," he adds.

One of the last steps in the process is to visualize the results, and this is where Tecplot 360 comes in. Marsell has been using Tecplot 360 for over a year now and praises its ease of use.

Figure 2: The frequency at which the slosh oscillations occur is useful information that can be used to prevent frequency coupling and inducing resonance in a fuel tank system. The XY plot above, generated by Tecplot 360, is an example of how experimental CFD results can be visualized for easy analysis.

"It takes inputs from multiple sources, can plot tons and tons of things, and displays large amounts of information at one time very clearly," Marsell says. "I use Tecplot 360 to visualize the flow field. It’s easy to move the image around and look at different things or isolate an item, take a slice, and view how velocity is changing. But the animation feature is the best when you’re looking at slosh, you can actually see a movie of how it sloshes. You can watch it and say ‘hey, look at that little corner where the fuel is pooling.’ It’s really important to see the entire domain like that."

There are plenty of other potential applications for studying fuel slosh with CFD methods, says Marsell. The manufacture of anything that must hold large volumes of liquids can benefit from it: cars, gas trucks, planes, ships, and oil tankers, to name a few.

"We’re talking about anything that has a tank. Period," says Marsell. "Now that we’re finally approaching the computational power we need, I expect we’ll start seeing many practical commercial applications within the next year or so."

Already convinced? Request a quote or try it for free.