Propulsion integration holds the key to low-noise, low-boom supersonic transports
The biggest barriers to development of economically viable and environmentally acceptable supersonic transports—sonic boom and airport noise—may be about to be breached. Through development of design tools allowing aircraft to be shaped to produce quieter booms, and noise-reducing nozzle concepts, and industry are growing confident that routine supersonic overland flight is within reach.
Wind-tunnel tests in 2011 ofand concepts for small supersonic airliners showed aircraft can now be designed that combine both low boom and low drag. Traditionally low boom and good performance have been mutually exclusive, but “we have broken that paradox,” says Peter Coen, Supersonic Fixed Wing program manager. “We achieved low boom with good supersonic-cruise lift-to-drag ratio.”
Coen attributes the breakthrough to “better CFD [computational fluid dynamics] capabilities to more accurately model the flowfield, geometry manipulation tools to explore complex three-dimensional perturbations, and optimizers that can quickly identify realistic configurations with promise.”
While progress is being made, the rear of the aircraft—and particularly the propulsion system—remains the most challenging to shape for low boom. This is because of the highly three-dimensional flowfield and complex interaction between the fuselage, wing, empennage and nacelles. Shocks caused by the inlet flow and exhaust stream have to be controlled while meeting engine operability and accessibility requirements.
Gulfstream is coming to grips with the problem as it refines the design of a supersonic business jet and its potential precursor, the still-unfunded X-54 Low Boom Experimental Vehicle. The X-54 would prove that a complete aircraft can be shaped for low sonic boom and be flown to gather the data on public acceptance of shaped booms needed to convince regulators to change the rules barring supersonic flight over land—an essential step for market viability.
Designing a low-boom nacelle is difficult. Engine flow requirements set the inlet and nozzle size; the diameter of the engine and its external accessories determine the nacelle's girth. These set the intake cowl slope and nozzle boat-tail angle, which govern the strength of the forebody compression field and aft-body expansion zone in supersonic flight. These elements can be reduced by stretching the nacelle, but that increases weight and losses.
Working with theTay turbofan for the X-54 demonstrator, Gulfstream hit a problem: The engine's externally mounted accessory gearbox caused a large nacelle bulge that adversely effected sonic boom. Its answer was the high-flow nacelle bypass concept. This encloses the powerplant in an aeroshell with a sharp intake lip and smooth internal flowpath that routes much of the captured air around the engine and gearbox with low losses. Curved channels overcome the problem of the gearbox blocking the flow.
The concept helps in other ways, allowing Gulfstream to design an axisymmetric inlet that minimizes both drag and boom. Inlet spillage, caused when intake supply exceeds engine demand and air overflows the cowl lip, increases supersonic drag and sonic boom. To avoid spillage, the inlet must capture the centerbody tip shock, but with a conventional external-compression inlet, zero spillage is almost impossible to achieve. With the high-flow nacelle, the engine cowl is inside the aeroshell, allowing the outer intake to capture and bypass spillage from the inner inlet.
“Classic inlet designs have good pressure recovery and low flow distortion, but produce a larger sonic-boom contribution due to high nacelle profile blockage and off-design spillage,” says Tim Conners, principal engineer for supersonic propulsion. “So we've come up with a hybrid-compression design which has a large secondary flowpath for bypassing flow around the engine. This produces a cowl-lip region within the outer mold line, but doesn't need variable geometry.”
Bypass drag is significant, but nacelle drag is reduced, while engine performance increases and interference drag is reduced. Gulfstream has combined the concept with relaxed isentropic external compression, which moves the initial shock to defocus the compression field at the inlet lip and tailors the terminal shock to reduce the cowling slope and increase core-stream pressure recovery, lowering drag and sonic boom.
“We get high inlet performance consistent with a mixed-compression design, but with the shock stability of external compression,” Conners says. Relaxed compression also weakens the cowl shock, for an 80% reduction in near-field overpressure, according to CFD analysis. Performance is improved, with 50% less cowl drag and 9.9% lower installed specific fuel consumption.
The relaxed-compression inlet was tested at NASA Glenn Research Center, at up to Mach 2, with good results for flow quality and shock stability. “This allowed us to move to the next step, but we knew the shaping we'd tested was not sufficient to solve supersonic boom,” says Conners. Gulfstream and Rolls then produced a notional flight demonstrator configuration aimed at showing nacelle bypass can enable additional low-boom shaping.
Flight-scale ground testing of the low-boom supersonic nacelle, including inlet and nozzle, was conducted on a Tay engine installed on a Gulfstream IV testbed in late 2009, “demonstrating stable inlet flow and excellent performance at static conditions,” he says.
This was followed by supersonic tunnel tests at Glenn in 2010. A single-flowpath model investigated relaxed compression and use of micro-ramps on the centerbody to manage the boundary layer and avoid inlet bleed. The second, dual-stream model evaluated the performance and stability of high-flow nacelle bypass at Mach 1.7. Scaled from a flight nacelle sized for a Tay, this bypassed about 40% of the captured flow.
The tests showed “we should have a low-boom inlet with excellent pressure recovery and very stable angle of attack from minus three to plus five degrees,” says Conners, adding that the concept “can outperform a single stream.” Further study is underway with the universities of Illinois and Virginia as well as Purdue University, plus Rolls and NASA. One discovery by Illinois was that a vortex trapped between the inner and outer lips entrained flow and improved inlet performance at low speed. Ground tests were conducted on a G450 “to see if the trapped vortex was real—and it was.”
While Gulfstream is working on what NASA calls the “N+1”—or first-generation quiet supersonic transport—Boeing and Lockheed Martin are studying N+2 and N+3 concepts for notional 2025 and 2035 timeframes, respectively. Each generation is larger, with more-stringent targets for sonic boom and airport noise. The concepts tunnel-tested in 2011 were aimed at NASA's N+2 goals for a 35-70-seat, Mach 1.6-1.8 jet with an 85-PLdb boom and noise 12 EPNdb below Stage 4.
While the focus was sonic boom and cruise performance, Lockheed Martin's Phase 1 study included acoustic tests of two different low-noise nozzle concepts—a mixer-ejector design from Rolls-Royce Liberty Works and an inverted velocity-profile nozzle from. Both are intended for use with commercial variable-cycle engines developed using technology from the U.S. Air Force Research Laboratory's Adaptive Versatile Engine Technology (Advent) program.
Flagship of the's Versatile Affordable Advanced Turbine Engines (Vaate) research effort, Advent is a technology demonstration under which Rolls and will ground-test engines that combine high supersonic thrust with subsonic fuel efficiency. Advent engines feature adaptive fans that can vary their mass flow and pressure ratio, matching airflow to the flight regime, from takeoff to supersonic cruise. And where conventional turbofans have hot core and cooler bypass flowpaths, Advent introduces a cool “third stream” for power extraction and thermal management—and noise reduction.
Both the GE and Rolls variable-cycle engine designs use the third-stream capability to reduce jet noise by modifying the velocity profile of the nozzle exhaust. Where other sources of noise can be tackled within the engine itself, by controlling its generation and attenuation, jet noise is produced outside the engine, by the violent shearing between the high-velocity exhaust stream and the surrounding air. Modifying the jet plume before it leaves the exhaust is one of the only ways to reduce noise.
“The high-speed flow typical of jet exhaust produces noise from the velocity shear, so we have to modify the velocity profile in the exhaust to reduce shear, detune the frequency of the noise and focus the noise in different directions,” says John Kusnierek, business development director for the Liberty Works.
Rolls's solution is the mixer-ejector nozzle, which first mixes the core and bypass flows using a lobed mixer, then brings in slower freestream air through ejector doors to mix with the higher-velocity jet exhaust. “The concept of the ejector goes back a long way, but previously it was used for thrust augmentation and never the modulation and management of noise,” he says.
Under its Phase 1 task, Liberty Works defined an integrated inlet/engine/nozzle propulsion system meeting Lockheed's aerodynamic and acoustic performance goals. The advantages of variable cycle include more flexible airflow scheduling across the flight regime and the potential to better match aircraft thrust requirement at the takeoff, transonic and supersonic cruise points.
The nozzle has two ejector doors on the outside and two diverter flaps inside. The external doors pivot 5-15 deg., creating openings through which freestream air is drawn into the nozzle. The internal flaps pivot 8-12 deg., creating crescent openings during takeoff and allowing third-stream air to enter the main flow. The doors and flaps rotate proportionately using the same actuator, but with different settings from takeoff to cruise. Both doors and flaps are closed during cruise.
A 15% scale model of the variable mixer-ejector nozzle was tested in the anechoic-dome wind tunnel of NASA Glenn's Aero-Acoustic Propulsion Laboratory. Tests were run at Mach 0.3 rolling-takeoff conditions, with five ejector-door angles, three divergent-flap angles, three third-stream throat areas and three nozzle clocking angles.
The nozzle is designed for a 36,400-lb.-gross-thrust engine in which mass flow is split between the core (almost 15%), bypass (more than 58%), outer third stream (18%) and ejector flow (9%). Of the third-stream flow, 65% is used to pump the ejector to entrain freestream air, and 15% flows through the inner flap slots to enhance mixing of the main, third-stream and freestream flows.
Preliminary noise estimates for a four-engine aircraft were a cumulative 18.3 EPNdb below Stage 3 noise limits and 8.3 db below Stage 4. Lockheed Martin's predictions for the final three-engine concept developed for Phase 1 are a cumulative 22 EPNdb below Stage 3, short of its goal of 25-30 db, but still 12 db below today's Stage 4 limits.
“It's fairly complicated, but the concept performed as advertised,” says Coen. The tests uncovered tones that were dependent on door and flap settings. “There was an issue with the small passages in the model,” he says. Covering gaps, bumps and cavities created by movement of the doors reduced or eliminated the tones. “If high-frequency tones due to third-stream flow can be removed, noise levels may be reduced 1-1.5 decibels,” Rolls says.
GE discovered a similar issue with its model. Under its task for Lockheed, the company designed a variable-cycle engine using Vaate Phase 2 technologies that would be at a readiness level of 6 by 2018. Exhaust system features include variable geometry, to maximize performance throughout the mission; heat addition, to better balance engine size and fan pressure ratio; and noise reduction through an inverted velocity profile and fluid shield.
In an inverted velocity-profile nozzle, the main high temperature and velocity core flow surrounds the cooler and slower bypass flow—the opposite of a normal turbofan. “The core flow is forced to the outside and fan flow to the inside,” says Coen. This sounds counter-intuitive, because the shearing that causes jet noise is increased when the high-velocity flow is on the outside, but it also promotes faster mixing of the high-speed exhaust with the ambient airflow, reducing the peak velocity more quickly. This more-intense initial mixing increases noise at high frequency, but reduces it at low frequency.
The nozzle diverts the fan flow to the center of the exhaust system using struts in the bypass duct, “and has demonstrated significant acoustic benefits as well as very high aerodynamic performance,” says Steve Martens, principal engineer of inlet and exhaust systems. The fluid shield, meanwhile, is a thin layer of flow that surrounds the lower 135 deg. of the main jet and both attenuates and reflects high-frequency noise from jet mixing, offsetting the increase from inverting the flows.
The fluid shield is an active device and can be turned on in noise-sensitive areas and off at other times to reduce the thrust and efficiency penalties normally associated with passive noise-reduction hardware. The shield's circumferential position also can be adjusted depending on the emission direction requiring the most attenuation, reducing sideline noise on the takeoff roll, then reconfiguring to reduce flyover noise on climbout.
A substantial amount of engine flow, up to 20%, is required for the fluid shield to achieve significant noise reduction, and it is best suited to a variable bypass-ratio engine that can generate the additional mass-flow required during noise-sensitive operations. In the nozzle designed for Lockheed, the fluid shield is generated by the engine's third stream.
“It's quite complicated, and required extreme modifications to the hot-jet test rig at Glenn,” says Coen. But testing showed the nozzle works, says Martens. It also revealed the need to clean up sources of excess noise, such as nozzle flow separation and tones from the shield. Under Phase 2, GE is now rebalancing the engine size and fan pressure ratio and updating the nozzle flowpath to incorporate lessons from testing.
Under Phase 2, both Boeing and Lockheed will examine the propulsion effects on boom shaping in detail. “The primary objective of Phase 2 is to increase the fidelity of the propulsion system in our boom models,” says Coen. For Phase 1, the wind-tunnel models had simple flow-through nacelles; for Phase 2, new models will have “real” nacelles with inlet centerbodies or ramps, flow blockers to simulate levels of bypass and spillage, and nozzle closing angles to generate shock systems.
“The configuration optimizer makes some pretty fine adjustments to the aft end,” he says. “Shock position is pretty important, and small shocks from the nozzle flow could have an effect.” A focus for Phase 2 will be refining the designs to ensure a low boom across the 60-mi.-wide ground carpet. Here propulsion integration will be key. “There is room for improvement in the aft signature off-track,” he says. [Article edited to correct the spelling of John Kusnierek, Rolls-Royce Liberty Works.]