FR0812-GE Boeing 747 engine testbedGE Aviation´s high-flying laboratory
It is like time-hopping back to the 1970s: the spiral staircase of the jumbo jet, which was revolutionary back then, has lost nothing of its elegance to this day. In the upper entrance area, a chandelier-like lamp still gives off a cosy aura.
By Patrick Hoevler
One deck down, the giant, blue First Class seats are almost as comfortable as they would have been on their first day. But concealed behind these relics from the golden Pan Am era is state-of-the-art computer technology. This is hardly surprising, since this Boeing 747-100 is not a museum piece but GE Aviation's flying testbed, used to test out its latest engines.
But why in today's computer-dominated era should such tests be necessary, and why on a 747? Al Krejmas, head of GE's flight test organisation in Victorville, California, explained during a visit from FLUG REVUE. "We have been doing it since 1942. GE prides itself on flight-test. This is the way we have always done it. We have always though the best way to develop our product is to put it in the environment it is made to be in. As good as our computational fluid dynamic programms are, surprises are always possible. So we always validate them with real test data." Before this Jumbo, they used to use a Boeing 707. But then came the GE90 turbofan. "Our only options were the 747 and the Lockheed C-5. Only a few aircraft of this size can fly as fast, but also as slow and as high as the Jumbo Jet, which is also more readily available and easier to maintain."
GE had acquired the former "Clipper Ocean Spray", one of the first Jumbos to be built for Pan Am after 1970, the year after the illustrious airline went bankrupt in 1991. After a thorough overhaul and extensive modifications the first series of flight tests with the GE90 engine followed. They took place in Mojave, where GE's Flight Test Organisation (FTO) had been based since the 1970s. In August 2003 the unit then moved to Victorville, formerly George Air Force Base. Here, compared with their previous base, the 49-person team enjoys its own hangar and two usable runways instead of only one.
But the main advantages are the location and climate, as Krejmas explains. "When we take off here, we are over the test area in five minutes. For this purpose we rely almost entirely on the 200 square mile military operating area (MOA) of Edwards, which is off limits to commercial air traffic and where we can fly at any altitude. On top of that, we get 324 clear days a year here. Due to the sensitive measuring instruments in the engine, we want to avoid any moisture getting into the sensors, which means not flying in clouds or in rain." The 747 only ventures away from its home base for special test series, for example, cold weather tests in Alaska, icing tests around Seattle or high-altitude sorties from Colorado Springs.
At Victoville the FTO has at its disposal a machine shop, a workshop for treating composite materials, a paintshop and a welding shop. The broad spectrum of tasks requires quite a bit of versatility on the part of the small but perfectly coordinated team, as every new engine programme calls for new modifications. The latest example is the GEnx for the Boeing 787, which has two large starter-generator units that are used to start the engine and generate power for the aircraft. But the challenges already began during assembly on the wing. In order to attach the first strut from 787 production to the wing of the 747, the FTO team had to design a special adapter.
In the former cargo bay of the jumbo, along with a system for measuring air pressure there is also a modified version of the electrical control system for the 787 from Hamilton Sundstrand. This provides the GEnx generators with the right output and hence the torque required to start the engine. In the case of the test 747, the necessary power comes from the three other engines, paradoxically JT9D's from competitor Pratt & Whitney. As soon as idle running is achieved, the system switches, to put it simply, and the generator accepts power in order to generate electricity for the onboard system. And a huge amount as well: together the two generators produce an electrical output of 500 kVA. "We can actually overload them to 618 kWA", says Krejmas. This is done on some special trials, such as abrupt acceleration. Here the flow of fuel has to be adjusted to enable the engine to achieve the necessary speed. "The worst case for a 787 would be to fail one of the two engines right on rotation. Here, the airplane will take the minimum power required which is still fairly substantial to maintain pressurisation, hydraulics." The engine that is still working has to supply the additional energy between 180 and 260 kVA and still function perfectly. "That is the biggest challenge, but we are pretty much over it."
Naturally such emergency procedures are simulated during the flight tests, as Krejmas explains. "We do things like high-power fuel cuts: at take-off power we shut the fuel off, flip the fuel back on five to ten or fifteen seconds later. The engine has to self-recover. For a five second cut the generator stays online and generates power because it does not dip much in speed. But there will be a bang. The whole prodecure is very abusive and aggressive. We did ten of them yesterday and it went pretty well."
Meanwhile the enormous electricity output of the generators cannot be used in the 747 testbed and is converted to heat in load banks under the fuselage, rather like a gigantic toaster, and released into the air. "They allow us through switching to set specific load amonts for all those failure cases. The old set-up could only handle 180 kVA, so we had to fabricate the present one," Krejmas explains. "The build-up for the GEnx covered a lot of electrical integration because the 787 application is an all-electric airplane. This was a big change, a big difference on what we tested before." After many years' experience, Krejmas is convinced that the secret to success therefore lies in early preparation. "The key is to get the requiremnets early. About 24 months before we fly we start talking about how to attach it to the testbed and what, electronical interfaces are necessary. The first questions I always ask is: what is new about this engine?"
When a new powerplant arrives in Victorville, it will normally have completed about 30 hours on the ground test stands and will therefore be fully instrumented. Altogether about 1,000 sensors measure approximately 2,500 parameters. By the time the engine has been installed on the 747, there will be several kilometres of cables on board. The cables are not jumbled up: with the aid of small tags at the ends and methodical plans, they are laid out in a clean arrangement. The 15 cm thick cable harnesses wind around the aircraft like anacondas.
The data from the digital FADEC system is also fed into the measurement results. "By the time we get the engine, we can put the nacelle on, hook up the instrumentation and go flying in 18 days. The lion share of the work we do is to develop the control schedules that go into the FADEC for the entire flight envelope. We do air starts or operability tests like throttle bursts and chops. One series of tests for a new engine will entail about 150 flying hours on average. A typical test flight, involving a team of about 20 on board, takes between five and seven hours. Naturally this generates a vast amount of data which may already be in the region of terabytes and is saved to a removable hard disk. For reasons of flexibility, there is no real-time datalink to a ground station.
Depending on the mission, up to seven systems, which completely fill up the cabin with their approximately 20 workstations, are used to record the values. The first three rows contain computers that record the measured pressures. The next row has three screens which duplicate the Windows 2000-based display system for test data in the cockpit. Should any computer problems occur, they are resolved here. Next comes the data station, which records parameters such as external temperature, exhaust gas temperature and oil level. In the event of a power failure, batteries in the rear cargo bay can maintain an uninterrupted power supply to the data system for up to 30 minutes.
Thanks to several video cameras, the crew are able to observe what is going on outside the cabin. Cameras are trained on the test engine intake, the wing, the rear and the landing gear. Then below the fuselage sits the "R2D2", as the team likes to call the 360 degree camera equipped with zoom and controlled by joystick, on account of its resemblance to the head of the Star Wars robot. Every angle is recorded over the entire flight.
The FADEC monitoring system (FMS) is upgraded the most frequently, as the bus architecture of each new product is subject to frequent changes. It serves as the interface through which the control engineers change and record the control laws. Next to this console a new workstation has been set up for the GEnx, to control and monitor the generators. One row beyond is a high-speed tape recording system which can record a frequency range of up to 20 kHz. Behind it is a distribution station which directs each of the signals coming from the engine to the designated console or recording unit. Another four monitoring stations allow the engineers to gain an overview of the data.
Each console can be equipped any way, like a big computer case, and is fitted with circuit breakers on the upper side. If any smoke should be seen, the operators can rush over to the appropriate equipment item and disconnect it from the power. In a really big emergency, another workstation used to monitor the power supply system has a kill switch, which can close down the entire data system immediately. Another unit is used to record the tip measurements in the compressor and turbines. The final row of workstations is reserved for special applications such as noise measurements.
But the test repertoire also includes aggressive flight manoeuvres such as high angles of attack. At 45 degrees, the flow at the engine air inlet separates. "The fan must take that distortion which will induce higher stress on the fan blades. We have to make sure we do not exceed fan stress limits or stall the engine", explains Krejmas. The GE engineers decide exactly what to test on each flight in a telephone conference. Then all the test points to be worked through are listed and after the flight they are recorded in a report.
The next candidate for flight testing is already waiting in the wings in the form of the GEnx-2B. Once again the 747-100 will be used, as no substitute is planned in the near future. But if at some point a successor should appear, "N747GE" will have definitely earned a place in the museum.
Flight tests from the pilot's point of viewHow does a jet fly with two different engine models? GE test pilot and engineer Gary Possert describes the biggest challenges as follows. "On our 747, the thrust of the engine in the No. 2 position does not match with that of the other engines at all. So we have a lot of assymmetrical thrust particularily on take-off. Unlike some other engine testbeds we use the test engine for the performance of the airplane. So when we take off we set it to take-off thust. With the GE90-115 which produces 115000 pounds of thrust, much more than the engines on the other wing can produce, we have a higly modfied take-off and power setting technique." The initial phase of take-off is particularly tricky. "As only seven percent of the airplane weight is on the nose gear the steering is very ineffective. If the aircraft starts to go off course and you try to correct it with the nosewheel tiller, it will just skid the tires. The rudder is not fully effective until you get to about 80 knots." This means that all three crew members have to assist with the take-off. "The pilot has his hand on throttle number one using that thrust to steer the airplane. He sets it up to some nominal level, if the airplane is goeing right he will reduce the thrust. The flight engineer is controlling engines number 3 and 4, while the copilot is handling the test engine. When the rudder comes effective, number one is set to full thrust. We usually do our take-offs at full thrust.
Another problem is the weight and balance. "Because of the larger or smaller engine being installed on the number two position we are no longer laterally balanced. Very large rolling moments can come in because the center of gravity is now across the wing span as opposed to longitudinally along the airplane. So we have two weight and balances we have to maintain: a normal one, and a lateral one that we have to counter-balance with special fuel loading and fuel manament."
But the big engines can also be fun. "With the three normal engines and something like the GE90 or even the Genx at full power, the 747 climbs like a rocket. It is good performing airplane. With the additional thrust it even surprises the controllers. I have experienced climb rates in excess of 6000 feet per minute."
On the other hand, during taxiing one has to watch out since sometimes there is a lot less ground clearance. For example, landing lights or signs can lead to incidents. Sloped taxiways require special attention too. Thus there was one instance of a test engine on the 747 which was only 15 cm clear of the ground on pushback.
From FLUG REVUE 12/2008