Boeing Engineers Lend Simulation Smarts to Giant Magellan Telescope
Ian Wright posted on January 12, 2018 |
GMTO and Boeing staff discuss the challenges of designing the telescope’s protective dome.
3D rendering of the Giant Magellan Telescope. (Image courtesy of GMTO.)

3D rendering of the Giant Magellan Telescope. (Image courtesy of Giant Magellan Telescope - GMTO Corporation.)

The Giant Magellan Telescope (GMT) is still years away from its first light, at which time it will be the largest optical observatory in the world. To put it in perspective, although the GMT is a ground-based telescope, when it’s completed it will have a better resolution than the Hubble Space Telescope.

Let that sink in for moment.

It goes without saying that building such a powerful scientific tool is no mean feat of engineering, but what might surprise you is how much engineering is involved at every level of the GMT’s construction.

For example, all the sophisticated optics that are going into the GMT need to be protected from the elements, which means they’ll need to be protected by a dome. But designing a dome to do that requires a surprising amount of computational heavy lifting—so much so that the Giant Magellan Telescope Organization (GMTO) turned to the experts at Boeing for help. had the opportunity to discuss this collaboration with GMTO Vice President for Operations and External Relations, Patrick McCarthy, as well as Boeing Research & Technology Engineering Manager, Abdollah ‘Abdi’ Khodadoust, and Senior Manager – Aerosciences, Bill Norby.

What does the dome actually protect the telescope optics from?

McCarthy: One thing it protects the telescope from—which might sound silly—is the sun. You have this big machine full of glass and mirrors, and if the sun hits it, the sunlight will bounce around and if you’re on the receiving end of that you’re in big trouble.

So, we have to keep the telescope out of the sun during the day, and even though we’re putting it on a mountain that has fantastic weather, every now and then you get snow or high winds, fog or blowing dust. The dome has to protect the telescope from all those elements when the weather is bad.

When the weather is good, you want something that can protect you from the direct wind or perhaps shade you from the moonlight. So, the dome has to have a lot of different functions and that’s part of what makes it such a difficult thing to engineer.

The GMT is a ground-based telescope and Boeing is an aerospace company; what brought your organizations together on this project?

Norby: It started about two years ago. We had previously participated in a NASA study called “CFD 2030” and as a result of that, and through some collaborations with other agencies, someone from the GMT project suggested they reach out to us. We received enquiries from some of the folks on Pat’s team about this project and it certainly seemed very interesting and applicable to us.

Abdi [Abdollah Kohdadoust] and I are both from Boeing Research & Technology, which is the portion of the company that sits between the commercial airplane and defense sides to work research and development for the benefit of the entire enterprise, as well as outside agencies and organizations. This project seemed like a good synergistic fit with the kind of work we do on our team, which is fluid-dynamics-related.

So, we got started with a smaller project, but it’s grown since then. Principally, we’ve been providing the GMTO folks with aerodynamic information to help them make the best decisions they can for the design elements of the telescope enclosure.

How does the CFD analysis for the GMT dome compare to a typical aircraft analysis?

Norby: Probably the most important thing is the velocity of the flow field we’re studying. In a commercial airliner or defense product, the velocities we’re concerned about are everywhere from very low speeds to hyper-sonic speeds. The GMTO application is predominantly in that low-speed regime, but it’s one that we’re very interested in for our work.

The other difference is that we’re talking about something that’s fixed to the terrain, so an important part of the modelling is the shape of the mountain and the terrain surrounding it, as well as the telescope and the enclosure itself. We received data from GMTO that included the geometry of all of those features, so we could include them in our modelling. Typically, we’re just modelling the geometry in question for an aircraft or a missile or a launch vehicle, so we’re not usually so concerned about the ground. In this case, we are. It’s about 2km2, so a fairly large area.

You tested the dome’s aerodynamics using both simulation and a 3D-printed model. Can you contrast the insights you gained from these two approaches?

Khodadoust: With CFD, you do a certain amount of simulation, but you really need confidence in those simulations. How do they compare against real, measured data? There’s a certain amount of data available through the measurements that have been taken at the GMT site, and that’s been really useful.

Boeing's 3D-printed model of the Giant Magellan Telescope. (Image courtesy of Boeing Company.)
Boeing's 3D-printed model of the Giant Magellan Telescope. (Image courtesy of The Boeing Company.)
However, you need to simulate the environment in a more detailed fashion in order to gain confidence in your computational tools. That’s really where the testing comes in. We used a 3D-printed, subscale replication of the enclosure, the telescope and a portion of the terrain. We placed that replica in a water tunnel with dye traces to observe the flow of stream lines around and through the enclosure.

We then compared those traces to the computed traces that we got through the CFD simulation, which gave us confidence that, yes, our simulation is generally predicting the appropriate velocity fields to the right level.

Did you 3D print a model of that entire 2km2 area?

Norby: No. For the water tunnel test, we zeroed into a region around the dome that included the terrain just around the summit, the top of the summit and the enclosure. Our CFD model includes that larger area, but the problem with testing that in the water tunnel is that the enclosure would be so small it would be difficult to take any useful measurements.

Close-up of the 3D-printed GMT. (Image courtesy of The Boeing Company.)
Close-up of the 3D-printed GMT. (Image courtesy of The Boeing Company.)
We did do some CFD calculations on the zoomed-in region to match-up with the geometry in the water tunnel, so we could make a direct, one-to-one comparison.

Do the materials used to make the dome make a significant difference here?

Norby: As far as the CFD analysis goes, we’re agnostic about that. We’re concerned specifically with the shape, and not so much the materials.

McCarthy: But when the Boeing team gets to working on the thermal challenges, the materials make a big difference.

Norby: Absolutely.

Aerial photo of the GMT site taken February 2017. (Image courtesy of Giant Magellan Telescope - GMTO Corporation.)
Aerial photo of the GMT construction site taken February 2017. (Image courtesy of Giant Magellan Telescope - GMTO Corporation.)
McCarthy: We need to ensure that we’re using materials that can come to equilibrium with the outside air quickly, but you have to balance that with the fact that we’re building the structure out of steel for structural integrity. So, when we’re talking about the panels on the outside of the dome, we look at their thermal properties very carefully. They sit out in the sun all day, so when you open the dome at night, you have to ensure it comes to the same temperature as the outside air very quickly, and doesn’t take all night to radiate that heat you stored all day. We need to minimize the amount of heat absorbed during the day and make it very radiantly efficient for when it’s time to cool off.

Is the mountaintop site itself subject to wide temperature variations?

McCarthy: No, and that’s one of the things that makes the location such a remarkably good site: the difference between the daytime and nighttime temperatures is really quite small—about ten degrees Centigrade—probably because it’s not that far from the coast, so we’ve got that large body of water to regulate the temperature flow and keep things very homogenous.

How did you set up the simulation to model the turbulent airflow?

Norby: Most CFD simulations start with a geometry as the starting point and then use that to define the boundary of the computational volume. We then impose a computational mesh on the surfaces as well as the volume, which can include—and I think in GMTO’s case, we’re getting close to it—a hundred million points within that volume.

A lot of those points are concentrated close to physical surfaces in the model—the ground, the enclosure, the telescope—and the reason for that is that a boundary layer forms whenever a flow is moving over a solid surface. This means the velocity gradients are going to be more pronounced in that boundary layer, which is why we include a lot of computational points within that region. Once you get farther away from the wall, you can spread out your point spacing a bit more.

(Image courtesy of Giant Magellan Telescope - GMTO Corporation.)
3D rendering of the completed GMT. (Image courtesy of Giant Magellan Telescope - GMTO Corporation.)
The computational mesh, once completed, will be entered into a flow solver, along with the appropriate conditions for the analysis—wind speed, direction, ambient temperature, pressure, etc.—and then the calculation will proceed on our high-performance computing systems, usually for a number of days.

Once it’s done, then it’s a matter of interrogating the solution with a number of different tools to evaluate the flow field and its impact on the structure and the telescope.


This sounds very computationally intensive. Were Boeing’s resources helpful in that regard?

Norby: Yes, certainly. We are using computational resources within the Boeing company, but we’re also looking to a future where there might be some opportunities to partner with national labs and use even greater computing horsepower, depending on where the GMT analysis needs to go.

Is there anything else you’d like to share with our audience about this collaboration?

McCarthy: To build the world’s largest telescope, we’ve had to call on many experts from different disciplines, including those you might not expect. When we started this project 15 years ago, no one said, “You know, we really should call up Boeing and get those guys to help.” But now it makes perfect sense.

This is a ground-based telescope, but an airplane lives in the atmosphere and so does a telescope. There’s a natural connection: turbulence is bad for all of us and there are similar physical problems. What we’ve done in building this telescope is reach out to the engineering experts, rather than trying to do everything ourselves.

3D rendering of the completed GMT. (Image courtesy of Giant Magellan Telescope - GMTO Corporation.)
3D rendering of the completed GMT. (Image courtesy of Giant Magellan Telescope - GMTO Corporation.)
Norby: From the Boeing standpoint, this is exciting because it’s of interest to all of humanity, but we also look at it from the standpoint that it’s going to help us exercise our muscles by applying our methods and tools to different sorts of problems. This is a flow field problem that’s of relevance to the Boeing Company and, even though it seems well outside our normal class of work, it really does have a nice synergistic value and gives us the opportunity to learn on a playing field that’s different from the usual one.

There are certainly opportunities to take what we’ve learned here and apply it to other Boeing projects. We’re concerned about airplanes and crosswinds on the ground, we look at the performance of the ventilation within our paint hangers, for example. These are very low speed conditions that we need to have expertise in.

For more information, visit the GMTO and Boeing websites.

Recommended For You