Commercial aircraft designs that depart from today's tube-and-wing shape demand equally innovative propulsion systems if they are to stand any chance of breaking the mold and providing the benefits anticipated by their creators.

With its engines flush-mounted beneath a pi-shaped tail, the Massachusetts Institute of Technology's (MIT) D8 concept with its “double-bubble” lifting fuselage is no exception. Sized to replace aircraft in the Airbus A320/Boeing 737 category around 2035—a time frame NASA calls “N+3 generation”—the D8 promises fuel burn more than 70% lower than the 737-800's. And while the wide, twin-aisle fuselage has more drag than the 737's, its overall shape enables a lighter wing and landing gear and a smaller tail.

The make-or-break aspect of the D8 design, however, is whether the embedded engine location is feasible, and whether smaller, lighter turbofans will be able to operate in the challenging flow conditions over the aft fuselage. To answer these questions, research is underway at NASA, MIT, Aurora Flight Sciences, Pratt & Whitney and the United Technologies Research Center (UTRC) into how closely the airframe and propulsion system can be integrated and what fuel benefits are possible from ingesting the fuselage boundary layer.

MIT has just completed four weeks of tests of the D8 in a 14 X 22-ft. wind tunnel at NASA Langley Research Center, Va., aimed a quantifying the benefit of boundary-layer ingestion (BLI) through back-to-back comparison of the same 1:11-scale, 13.4 ft.-span model with embedded and conventional podded engines. Initial results show a benefit close to that predicted, with a measured 5-8% reduction in the electrical power required to drive the 6-in.-dia. fans in the embedded engines at the same cruise condition, says Alejandra Uranga, technology lead for MIT.

BLI is not new a new idea; it is already being used in torpedo and ship propeller design. When a podded propulsor is in freestream airflow, like a turbofan in a wing-mounted nacelle, the excess kinetic energy in the jet is wasted. But when the propeller is immersed in the slower-moving boundary-layer flow there is no excess kinetic energy, and less energy needs to be added to achieve the same thrust. The benefit comes from the propulsor reenergizing the wake and reducing drag.

The question regarding the D8 is whether the same principle can be applied to an aircraft-engine fan operating under higher propulsive loads, and whether that fan can withstand such a turbulent environment without its efficiency being overly compromised.

“The primary question is, 'Can you get a fan that can operate in that distortion?' Secondly, by ingesting the boundary layer, are we reducing drag, as most people believe we are? It is drag versus efficiency and whether the summation of the two is still positive,” says NASA Fixed Wing project head Ruben del Rosario.

“Based on what we know today, nobody would develop a BLI system, but if we advance our knowledge, then people will consider taking the risk. That's why we see BLI in the N+3 time frame. We don't see it in the next 10 to 15 years; it is more like 20 years away,” he says.

There are two main parts to NASA's research. One is testing of the D8 configuration to validate the BLI benefit and characterize the flow into the aft-mounted engines. The other is work by UTRC to design and test a distortion-tolerant fan for embedded engines. A third piece is the study by Pratt & Whitney of novel architectures for small-core engines that could power a 2035-time-frame D8-configuration airliner.

MIT developed the D8 for NASA under an N+3 Phase 1 study completed in 2010. “A key feature is the rear flush-mounted engines and BLI, but it is one of the highest risk,” says Mark Drela, professor of aeronautics and astronautics at MIT and the design's creator. “Phase 2 is focused on evaluating its feasibility and performance, so the wind- tunnel tests are focused on the rear of the aircraft.”

While the double-bubble fuselage and slender low-sweep wing were defined during Phase 1, the tail was not. During Phase 2, MIT will “improve of the aerodynamic design of the tail and how the engines blend in,” notes Uranga. The same 1:11 model will be back in the Langley wind tunnel in January with a refined tail.

The goal of the Langley tests was to make a fair comparison between embedded and podded engines. “It's about measuring the right thing,” says Ed Greitzer, professor of aeronautics and astronautics at MIT and the principal investigator. “So we have swappable tails with the same propulsors, the same fans and motors, to rule out that variability.” The BLI benefit will be determined by comparing the electrical power required to drive the fan motors in the cruise condition, defined as zero net force on the model in the tunnel, says Uranga.

“In addition to obtaining force and moments data, we did extensive flow surveys at the inlet and exit of the nacelles to examine the total-pressure profiles into and out of the propulsors. The ability of the fans to handle the total-pressure variation at the inlet is a key requirement for effective BLI,” says Greitzer. The pressure profiles will be used to estimate thrust and drag and provide a second method of calculating the BLI benefit. “Ideally, the two measurements will be the same, or close.”

The drag benefit from BLI is potentially large. On the D8, the engines are positioned to ingest most of the flow over the top of the fuselage. “About 40% of the entire fuselage boundary layer is ingested, and typically the fuselage is 25 to 30% of the total aircraft drag,” explains Uranga.

“The preliminary data shows an electrical power-saving with the integrated configuration of roughly 5 to 8 percent. There are several physical origins to this saving, such as reduced overall wake plus jet kinetic energy losses and reduced nacelle wetted area,” says Greitzer. “However, BLI is just a fraction of the potential fuel-saving that the D8 configuration could actually provide, via secondary benefits in gross-weight reduction, tail-size reduction and other areas. We feel that the double-bubble concept has a big overall potential that is well worth investigating.”

Confirming the BLI benefit assumed during Phase 1 design is crucial to the D8. “If BLI does not work, the whole concept is suspect,” admits Drela. A fan that can operate while ingesting boundary-layer flow is critical also. “At NASA Glenn [Research Center, Ohio], we are looking at how a fan operating in a distorted flow can be designed or improved so you don't lose a lot of efficiency,” says del Rosario.

To that end, NASA worked with MIT to put pressure rakes at the fan face on the D8 model in the 14 X 22-ft. wind tunnel. “At cruise, inflow conditions looked pretty much as expected, but not so at high angle of attack and off-design,” says Greitzer. Work with UTRC on a distortion-tolerant fan will lead to tests of the fan in an 8 X 6-ft. wind tunnel in 2015. “We are having preliminary design reviews before we go into production of the fan itself,” says del Rosario.

Michael Hathaway, technical lead for propulsion within NASA's Fixed Wing project, says UTRC “went through a vehicle systems study with different configurations and numbers of propulsors and fans ingesting the boundary layer. They optimized around five designs that showed the best potential for BLI benefit and fuel-burn reduction.” The final candidate will be scale tested on a 22-in.-dia. fan rig at NASA Glenn, previously used in early development of the General Electric GE90 and Pratt & Whitney PW1000G.

“We are trying to keep the efficiency loss below 2%, and our current aerodynamic analysis is telling us we can stay within 0.5% with the design we have. That would be amazing,” says del Rosario. The initial rig-testing will include a “false floor” to simulate the airframe in front of the inlet (see diagram). “It cannot be tilted that much [for angle-of-attack simulation], but by raising different floors we can simulate different shapes of boundary layer,” says Hathaway, who adds that the rig is able to test effectively for cruise efficiency. “We have different options for doing a lot of cross-flow or angle-of-attack tests,” he says.

As well as overall efficiency, another focus is on the aero-mechanical impact of “the fan [blades] being buffeted in and out of the clean and distorted flow,” Hathaway says. “Forward sweep might be less, but it will be operating in a lower Mach-number flow, so it won't look radically different from a conventional fan. We may also alleviate some of the distortion of the flow by modifying or tailoring the inlet and stationary structure. That's what we hope to get out of the tests at UTRC.”

MIT and NASA, meanwhile, plan three separate wind-tunnel campaigns for the D8, with progressively larger and more sophisticated models. These will culminate with the third model in 18 months. “About that time, we will be finishing the fan data, so we will have both external aerodynamic and internal fan data,” explains Del Rosario. The airframe and propulsion system will be integrated to see if the potential benefit of BLI is still there, he adds. Overall, this could result in an 8-10% fuel-burn reduction.

For the next tunnel entry, early next year, the 1:11 model will be fitted not only with a refined tail, but also custom-designed fans for the embedded propulsors. The first tests used off-the-shelf model-aircraft electric ducted fans. “They were the best we could find, but it's a question how efficient they were,” says Uranga. “The comparison of two fans will also help assess the role of fan design in this non-uniform flow. We will also use higher tunnel speeds for greater measurement accuracy,” Greitzer says.

Tunnel speed will be increased to 100 mph from 70 mph for the first tests. Low-speed tests are adequate for assessing BLI because “boundary-layer evolution is very weakly affected by Mach number,” says Drela. “And high-speed wind-tunnel tests cost a lot more money.” Greitzer praises the high flow quality in the 14 X 22-ft. tunnel and the “exemplary” force-and-moment balance provided by NASA, which contributed to “exceptional repeatability and confidence in the measured data,” he notes.

Beyond the BLI aspect, the NASA/MIT/Pratt team is also investigating advances in engine design that would be required to gain the most benefit from the D8 configuration. The substantial drag reduction promised by the D8 means significantly less thrust would be required, calling for less powerful engines than on today's single-aisle airliners. “The core of those engines is going to be extremely small because you are going to have very high bypass,” notes del Rosario. “Gas generators will have to get smaller, so to get the energy to drive the fan, you have to raise the overall pressure ratio, which raises the temperature aft of the compressor,” adds Hathaway.

As a result, researchers are looking to identify the efficiency limits of small compressors and whether they scale with size. “If the aft end of the high-pressure compressor is running at temperatures over 1,500F, can the materials still handle that?” Hathaway asks.

In addition, the combination of low thrust and high pressure ratio for an ultra-high-bypass D8 engine points to a decrease in compressor-exit corrected flow, which implies very small high-pressure compressor blade heights. “They could be 0.5 inches in height or smaller, and that will mean it will be hard to hold clearances,” Hathaway says. “We have research ongoing into how to mitigate those losses. We have to start looking at novel architectures when we look at small blades.”

But Uranga points out the D8 configuration could provide substantial fuel-burn reductions through boundary-layer ingestion even with existing engine technology.

“Small cores are not on the critical path,” says Greitzer. “The BLI benefit is completely decoupled from the core, and only the fan is the issue. But we have done some work on the behavior of turbomachinery in that flow regime, and Pratt & Whitney has conceptual designs for engines that have different architectures and thus different mechanical parameters in small sizes.”