Meet Liberty Lifter

The U-boat sinkings of Liberty Ships, the vessels rushed into service during World War II, so disgusted American shipbuilder and industrialist Henry J. Kaiser that he got together with Howard Hughes to design the H-4 Hercules, better known as the Spruce Goose. The idea was to make a massive seaplane that would fly low to utilize the ground effect, a cushion of air between the surface and its wings. The Spruce Goose was flown just once, long after the war was over, and the concept was abandoned. Between 1966 and 1980, the Soviet Union flew research versions of a ground effect troop carrier, but only one operational aircraft, the Ekranoplan or Caspian Sea Monster, was flown, and it was limited to flying over calm waters.

Now it’s the U.S.’s turn again: DARPA and its contractor Aurora Flight Sciences are drawing up the preliminary design for the demonstrator, while in parallel preparing to make structural test articles of components. Managing the program is not a plane designer but a maritime expert, Christopher Kent, who holds a Ph.D. in naval architecture and marine engineering.

We spoke by phone between London and Arlington, Virginia.

Paul Marks: What does DARPA want to accomplish with Liberty Lifter?

Christopher Kent: We’re trying to support maritime logistics at a fundamental level, but there is a need for a layer of flexibility and speed and agility on top of that. I call the concept “maritime airlift.” Today, if you need to move a million pounds of stuff, ships are hard to beat for efficiency. If you need it to be there tomorrow, airplanes are hard to beat, but they cost a lot more per pound. Transport efficiencies are way worse. With Liberty Lifter, the aim is to occupy the space between conventional airlift and conventional maritime transport — there’s quite a large gap in there. The idea is to transport troops and materiel at helicopter speeds but with efficiencies that are much closer to that of a ship. On top of that, we’re trying to do it at a lower cost. The other thing Liberty Lifter can support is moving unmanned systems, placing them over a much larger region much more quickly: The Defense Department has unmanned air vehicles, unmanned surface vessels and unmanned underwater vessels. Liberty Lifter will provide an opportunity to transport groups of those over hundreds of miles instead of tens of miles, which is what you’d be able to do with a ship at 14 to 18 knots. We’re looking at speeds of around 180 knots, similar to a helicopter. We’re building a demonstrator at a smaller scale of about a C-130 size, with numbers very similar to a C-130’s performance capabilities.

When we originally visualized the program, we were building a larger demonstrator, but we’ve learned some hard lessons about the availability of the props and gearboxes that can support the kind of thrust we need for takeoff. Those are not available, so we’ve sized down. But the aim for the future production vehicle is capacity closer to that of a C-17: a 90-ton cargo capacity, so that’s 180,000 pounds of men and materiel, not including a Liberty Lifter’s crew or its fuel. It wouldn’t be quite a C-17 size because it doesn’t have the requirement to carry tanks, so the cargo bay is narrower.

Kent clarified this point in a follow-up email. “We’ve optimized Liberty Lifter down from approximately the size of a C-17 to a C-130 — we didn’t need to build the bigger plane to test our problem set. The Liberty Lifter is expected to have a wingspan of ~200 feet [60 meters], which is wider than either the C-17 (~170 feet) or the C-130 (~132 feet). So when we say ‘C-130-sized’ that refers to payload, not physical dimensions.”

DARPA’s description of the program mentions transforming “fast logistics missions for the DOD and commerce.” What’s the commerce angle?

I grew up on a small island, so I see an opportunity here for providing fast inter-island transport for cargo. Because our transport efficiency will be better than a regular aircraft, it’ll become a tool for UPS-type businesses, giving them an opportunity to get cargo delivered in a day, whereas at ship speed it’d take weeks. The other thing that it can do is Coast Guard rescue for large ship casualties. We have the opportunity to extract people at helicopter-type speeds at ship scale: rescuing hundreds of people from a disabled cruise ship, for instance, versus six, seven or eight, which is what you can take out right now. We also see a very critical need for humanitarian relief: Getting food into places would be very easy because Liberty Lifter is designed to go into an austere port. You’d just need a beach or landing area.

As a naval architect by training, are you approaching Liberty Lifter as a boat that can fly or a plane that can float?

In this case, we’re technically a flying boat, as seaplanes are sometimes amphibious and we are not planning to do that.

A Cessna Caravan Amphibian, for example, can take off from a runway and land in the water or vice versa. Liberty Lifter will only take off from the water and will need water to land. — PM

Fundamentally, we’re still an aircraft, and we still have aircraft requirements and aircraft safety standards to meet. The reason they hired me to run this program was because I had a good understanding and brought a viewpoint and a healthy skepticism to the takeoff and landing problems. We want to be able to operate in high sea states, up to Sea State 4 for takeoff and landing, and Sea State 5 for flight in ground effect. My background in hydrodynamics and systems engineering blends well with the understanding of the problem and the planning of the hull loads. At a fundamental level, the program is a systems engineering problem with a lot of hard science thrown on top. There was a lot of work done at DARPA, ahead of this program, analyzing the feasibility of filling the gaps and the opportunity space that a craft like Liberty Lifter could fill. We stand convinced that it’s possible. A large seaplane design is a trade between the same “iron triangle” that everybody else has: cost, performance and schedule. But we’re trying to get the whole program to push the boundaries on cost and performance, for a craft of this type.

Sea State 4 means waves up to 2.5 meters — would those have defeated the Soviet Union’s Caspian Sea Monster?

One-hundred percent, yes. The Ekranoplan [and its predecessors] were severely limited by sea state. They were operating in what we call the 2D ground effect, or deep ground effect, operating where the ground effect is driven by the ratio of the chord of the wing to the height off the water.

Kent explains ground effect like this: “The downwash from the wings and the downward component of the wingtip vortices create a higher than normal pressure area below the wings because the air hits the ground and can’t ‘escape,’ increasing lift. Induced drag is reduced because wingtip vortices (causing drag) can’t fully develop close to the ground.” Also, the chord is the distance from the leading edge of the wing to the trailing edge. — PM

With wings low on their structure, they were operating really close to the free water surface, right up on the wave. That offers advantages because it’s extremely efficient — you can go much faster than we are planning — and it’s naturally stable, which is great when you have flat water. But you’re coupled to that, so when you start getting waves, you start [experiencing] high [vertical] accelerations.

Kent elaborated by email: “In ground effect, you’ll experience vertical accelerations that vary proportionally to the height and frequency or wavelength of the swells you encounter. In layman’s terms, the ride quality is analogous to racing a pickup truck over a washed-out dirt road; it’s bone-jarring and from a human factors perspective not sustainable for more than a very short time. This phenomenon limited Ekranoplan operations to relatively calm water and precluded their use in open ocean conditions.”

DARPA and others have spent quite a bit of time studying that [acceleration phenomenon] and we’re pretty sure we understand how it worked. And Howard Hughes’ planned Spruce Goose seaplane never exited the 2D ground effect into the regime that we actually want to use: the 3D ground effect, which involves vortices being shed off the wingtips and interacting with the ground. So you have two vortices that shed off the wingtips, they interact with the ground, and they are slowed. That slightly increases your lift and slightly decreases your drag. It’s not as strong as the 2D ground effect, but it’s also not as coupled with the surface, so you can operate in higher sea states in that kind of flight. The shed vorticity actually gives the same effect of that little lift you feel just before you land in an aircraft. Part of that is flare, but part of that is also that the wake vortices are interacting with the ground and you get a little cushion just before you land. That wake vortex is interacting with the ground, slowly processing downwards. When it hits the ground, that slows it, and that shed vorticity is what gives you drag. So you’re putting energy into that vortex so if you slow that vortex — if you think about it just from a conservation of energy perspective — you’ve reduced your drag, and it actually also slightly increases your lift.

How do you differentiate between 2D and 3D ground effect?

2D ground effect is a height over chord length of about a half, meaning the plane is flying at an altitude only half the width of the wing. The Ekranoplans were flying at 400 knots [740 kph] at about 4 to 8 feet off the water. So it was really low. What we are using in 3D ground effect is less than a wingspan, so it is a much larger number.

Liberty Lifter’s planned wingspan of 200 feet and cruising altitude of 100 feet is advantageous for this 3D ground effect?

I’m not going to give you the final numbers, but we’re cruising at around half wingspan-ish or less. The deeper you go into it, the more that effect strengthens, until you get into 2D ground effect, and then you get another lift. That’s why the Russians went to that regime with Ekranoplan. But the negative of that is that you can’t realistically operate a craft like that in an open ocean; you need something that you can decouple from the waves. That’s why they used them on the [calm waters of the] Caspian Sea — because they couldn’t realistically use them anywhere else.

What makes the wing-in-ground effect a “DARPA hard” problem?

In some of the earliest photos of the design by Aurora Flight Sciences, you’ll see that they have inverted winglets, basically downward ones. So they’re not steering [the wingtip vortices] down, they’re pushing them down by using a winglet. As to why this is technically DARPA-hard, we’re going after both cost and operation regions here at a fundamental level, and technically we have three challenges: the takeoff and landing in high sea states, second driving costs out, and control. Why control? A significant negative of the 3D ground effect is it’s not naturally stable; the aircraft doesn’t want to stay there, so you have to use active control. What we’re really leveraging in the program right now is the world’s best state-of-the-art active control that is being used for other aircraft and leveraging that into the program to enable flight in that regime. So unlike a pilot steering a landing, we are going to automate the ability to stay in that 3D ground effect region for long periods. And we are targeting costs of one-half to one-third of that of equivalent military aircraft like C-17s, C-130s. We’re getting to that by using maritime-style components and construction techniques, using composites in an aircraft and getting an aircraft flight certification for that from DARPA. A maritime-type composite epoxy will be about $30 per pound, and the aircraft one will be something like $140, although it’s a very similar chemical epoxy. Maritime components generally do not have the same tolerances and do not have the same strength-to-weight ratios as those in aircraft, but we’re buying those back on by operating in ground effects. We’re using the additional lift and drag that we can support to get our transport efficiency still higher, regardless of the fact that our aircraft at a fundamental level could be built more lightly. One of the things that I am personally interested in, and the program is interested in, is bonded joints. Aircraft have epoxy-bonded joints, but manufacturers then spend all the manpower to still rivet it. So bonded-joints actually cost more, because they still have as much labor as a non-bonded joint held together with just rivets. That’s because from a flight safety certification, the certification process has not pushed that boundary and has not gotten around except for some very specific cases, the acceptance of joints of that type. So we’re pushing that again, trying to see if we can move there.

Are there success stories from the maritime world that give you confidence?

Some 20 to 30 years ago in the Volvo Ocean Race and the America’s Cup, boats were breaking all the time, failing catastrophically on the course, rudders would snap — that was bonded joint failure, and that doesn’t really happen anymore. The naval architecture world, and the boat world, has moved on and figured out how to design those joints well. This also has the potential to open up a whole new industrial base for being able to build aircraft components in small boat yards spread all over the world, all over the U.S. And you have the opportunity to build aircraft subcomponents in such places, which we’re going to test that out here. Large portions of the structure will be built at a shipyard.

What are the advantages of Aurora’s proposed monohull design?

The monohull is a much more conventional aircraft. It has some drag advantages for takeoff, but its cargo configuration is a little more challenging as you naturally carry almost twice as much stuff in the twinhull [the design proposed by the second contractor, General Atomic Aeronautical Systems]. But you’re much more stable in a high seaway, as catamarans are very stable in big waves. So you fundamentally had more stability for that and had the ability to unload cargo twice as fast in a twinhull design when it’s landed. It also gets you a big, fairly long chord length in between the two hulls, which gives you a lot of structural strength in the beam. On a conventional aircraft, the structural joint arrangement to the main fuselage can be challenging, but it’s also something that’s well understood. Fundamentally, they’re both executable designs; it’s just we’ve had to downselect to one performer.

Kent followed up by email: “Efficiency is baked into the DARPA model, which maximizes our opportunity to create transformational change. For Liberty Lifter, when we reached the point where we realized only one performer was meeting our aggressive schedule and technical goals we streamlined the program to continue to deliver innovation ASAP.”

Once we get through preliminary design review in the next year or so, we’ll decide if we’re moving out on doing detailed design and construction on the monohull aircraft.

Regarding propulsion, are electric motors a possibility for Liberty Lifter?

Right now, there’s just nothing in the 5,000 horsepower-ish range that we need for our takeoff that can consistently deliver the kind of power we need. We need eight 5,000 horsepower engines, and we are also adding in a very corrosive saltwater environment. So our technical challenges don’t focus on marinizing big, high-power, aircraft-grade electric motors at this point. Do we see an opportunity space in the future as those continue on the track at which they are going? I do, absolutely; I would not be surprised to see a commercial version of this. What we would really need is a hydrogen-powered commercial version of this because we’d have a lot of volume to fill with hydrogen. But that’s just not executable [for the demonstrator], and it doesn’t really drive out our need. It also is not necessarily very focused on the military need, and our name does start with “defense.”

Tell me more about the control avionics and automation that will be involved in keeping Liberty Lifter in that all-important 3D ground-effect region.

We’re looking at wind gust alleviation regimes and approaches that have been used for commercial aircraft and trying to apply them to this platform. Thus far, the analyses that have been done have showed that there seems to be a lot of promise for smoothing out the ride for Liberty Lifter. There’s also some interesting behaviors that you have to address. When you’re flying in ground effect, for example, a bank turn is no longer a bank turn — because if you just banked, one wingtip would touch the water. So you have to do an active lift control turn, where you lift yourself up and sort of bank around the wingtip, which involves a lot of use of active lift control.

Lay out the big milestones going forward.

We will decide about moving into detailed design and construction in about a year’s time, plus or minus a few months. When we will start construction depends on your definition of “construction.” We will start purchasing long-lead item stuff very shortly thereafter, but my current expectation is that construction will begin in 2026 and last about a year. We’re targeting floating Liberty Lifter sometime in late calendar 2027 — and potentially first flight in late calendar year 2027.

About Paul Marks

Paul is an award-winning journalist in London focused on technology, cybersecurity, aviation and spaceflight. A regular contributor to the BBC, New Scientist and The Economist, his current interests include electric aviation and innovation in new space.

Spruce Goose flight
Howard Hughes’ massive Spruce Goose made its sole flight off the coast of Long Beach, California, in 1947. The aircraft reached an altitude of 25 feet. Credit: FAA
C-17 and C-130 aircraft
Performance goals: DARPA wants Liberty Lifter to have the payload capacity of a C-130 (right), but believes that if one or more of the military services were to contract an operational version, an aircraft with C-17 capacity could be possible. Credit: U.S. Air Force/Staff Sgt. Mitch Fuqua; U.S. National Guard/Staff Sgt. Jon Alderman

Meet Liberty Lifter