A distant corner of the vast Florida everglades is reverberating to the sound of not just one, but four Joint Strike Fighter (JSF) engines on simultaneous test at Pratt & Whitney's extensive Large Military Engines site in West Palm Beach.

While it may be business as usual to the local alligator community, which shares the nature reserve site with the engine maker, it represents an unprecedented build-up for P&W. In a record time of 23 months, the company has designed, built - and is now testing - two JSF119 engines for the conventional take-off and landing (CTOL) versions of the JSF and two different JSF119 versions for the short take-off and vertical landing (STOVL) contenders.

All four engines are based on the F119 developed by P&W for the Lockheed Martin/Boeing F-22, but each is unique. JSF119 programme director Bob Cea says: "In June we grew twins and, by December, I had quadruplets."


The effort is aimed at clearing the engines for the first flight of the CTOL concept demonstrator aircraft (CDA) in the second quarter of 2000 and the STOVL aircraft around mid-2000. At the same time, P&W is engaged in testing engine configurations, technology and systems for the preferred weapon system concept, which is expected to culminate at the end of 2000 with the completion of plans for the engineering, manufacturing and development (EMD) phase.

The CDA phase began successfully, with the start of testing on the two JSF119s for the CTOL and US Navy carrier version (CV) aircraft. Testing of the JSF119-614 (FX651) for the Boeing aircraft, and the JSF119-611 (FX661) for Lockheed Martin, testing began in June. "There were 70% fewer assembly issues and it took us around 80% less time than previously needed to set the tests up," says Cea, adding that much of the improvement reflects the use of the solid computer-based modelling techniques that have become a hallmark of the CDA phase.

P&W has also been at the forefront of dealing with the issue of translating between the various computer-based systems used by all the manufacturers. "Part of the issue is to get all of us talking in the same language," says Cea, who adds that there is "fertile ground for improvement". The main differences lie between the Dassault CATIA-based system, used by most of Lockheed Martin and Boeing, and the Unigraphics-based system used by the former McDonnell Douglas elements of the Boeing team, as well as by P&W.

After almost 200h of initial testing, both CDA engines have recorded "green lights" in all four key areas: weight, performance, operability and thrust response. "During the first series of tests we hit maximum flow, and maximum rpm," says Cea, while turbine temperatures were "-lower than we anticipated". Although actual numbers are classified, the margin was "50-100° lower than expected". Component efficiencies were also "a couple of points better" than expected, says Cea.


The CTOL control software also "-worked well" and stall margins were fully demonstrated with the original versions of the two different sized fans adopted by the two engines. The higher flow requirements of the Boeing-powered lift STOVL version means that the fan is slightly larger - at around 1.12m (44in) diameter - than the engine for the Lockheed Martin JSF. The design commonality goal of the programme means that the fan is common to both the CV/CTOL and STOVL versions.

Because of the complex geometry of the compact inlets in both designs, the fan stability and surge margin of both main JSF119 versions are crucial areas of investigation during this early test phase. To help stabilise the intake flow, both engines have been fitted with a relatively large set of fixed inlet guide vanes. The trailing edges of the vanes are moveable, to give some variable-inlet control, although P&W hopes that the early signs of robust engine operations will allow it eventually to delete this feature from both powerplants.

Vibration levels recorded on the CTOL engines were "very low", says Cea, and weight was "-below goal-which is good", he adds. One of the engines, although he declines to identify which one, "-is within 8lb [3.5kg] of the estimate". The STOVL engines, for which the weight target was even more crucial, have met their targets after input from employees and a "weight control Czar". The company offered incentives to employees, in the form of bonuses and prizes, if they could come up with weight saving suggestions. JSF119 business development manager Lindy Shambaugh says: "It made everybody aware of the subject."

The two CTOL test units, termed generically "engine one" by P&W, were joined on test in November by the two STOVL test engines, "engine two". The Boeing engine, JSF119-614S (FX652), was mounted on an outdoor test site normally used for large PW4000 commercial engine runs. The higher mounting provided good ground clearance and room to measure thrust in six axes. The Lockheed Martin powerplant, JSF119-611S (FX662), is initially being tested indoors, but will eventually move to an outdoor site.

"We have run both to maximum flow and speed, and we have run the lift fan [on the -611S] to maximum flow and speed," says Cea. Some of the greatest concerns with the Boeing STOVL engine soon proved groundless after the initial tests. The -614S works in a similar way to the British Aerospace Harrier's Rolls-Royce Pegasus engine (Rolls-Royce is providing the lift modules for both JSF STOVL engines). Fan duct flow is discharged through a laterally and vertically aligned jetscreen to provide lift and prevent hot gas ingestion from the hot nozzles which discharge the gas jet. For conventional flight, the hot gas flow is redirected by large butterfly valves through the main vectoring exhaust.

P&W was anxious to prove that the transition from one mode to another would not cause a back pressure at the low pressure turbine, caused by asymmetry as the valve was opened. "The valve worked fine, however. We did not encounter any flutter and we've cleared the low pressure turbine," says Cea.

The vastly different -611S for the Lockheed Martin JSF uses a shaft-driven lift fan and a three-bearing swivel nozzle for the bulk of its lift. The shaft connects to the low pressure turbine and drives the lift fan through a clutch and gear mechanism. The clutch was originally expected to be a fluid-based design, but was changed to a simpler mechanical plate connection, supplied by BFGoodrich. "We encountered vibration levels in the shaft and clutch that were higher than we liked," says Cea, adding that "-Rolls-Royce is rebalancing the shaft and clutch and we will begin re-runs later in January".

Despite the initial vibration issue, P&W appears confident of successfully completing the 2,500h test programme by the end of the year. Walter Bylciw, F119 programmes senior vice president, says: "We knew it was going to be tough, and it's fair to say it is an aggressive programme, but everything appears to be on schedule. We have achieved 100% power in the STOVL configurations, and we have margins now beyond our expectations. Now we have to make sure of shipping the first flight test engines before the end of the year."


Assembly of the next pair of JSF engines ("engine three") will begin "within weeks", says Cea, while fabrication of the following pair is due to begin by March. The pair making up "engine four" will be used for accelerated mission testing, which will clear the way for engine certification and flight safety. The "engine three" set will be refurbished, following tests, and delivered at the end of the year for installation in the two CTOL CDAs under construction at the Boeing and Lockheed Martin respective sites in Palmdale, California.

In the meantime, P&W is hooking up flight simulators (one each for the Boeing X-32 and one for the Lockheed Martin X-35 CDA demonstrators) to its engine control software as a vital step towards controls development and integration. The simulators were delivered last December and will be used to run simulated missions "24 hours a day", says Cea. The concept is based on a similar simulation developed for the Advanced Tactical Fighter competition, and integrates the weapons systems, flight control systems and engine control software. "We will be testing in open loop or closed loop mode in the simulator," says Cea, who adds the simulation will "-remove any glitches".

The results of P&W's control system model simulations will be fed to more sophisticated simulators at Boeing and Lockheed Martin, where they will be used to perfect more accurate operating scenarios and help pick up any problems. In the case of the Lockheed Martin model, the software will say: "We are apparently landing. Open the doors, start transition to the hover, move the three-bearing swivel duct and start shipping horsepower to the lift fan. It has to happen in precise order, and one reason we are running the programme with pilots is to make sure that we carry out the changes they want to see," says Cea.

The ultimate goal of the effort is the integration of the simulators and the actual engines. "By the end of the year, we will combine the two. We will get the Boeing and Lockheed Martin simulators here and will use them to run the hardware. It is something we've never done before," he adds. In parallel with this effort, P&W is also working on plans for the configuration of the engine in the EMD phase. The configuration will be tailored to the specifications laid out in JIRD 3, the next Joint International Requirements Document.

One of the potential elements of the production configuration engine is an advanced diagnostics/prognostics system for improved safety and maintainability. The concept goes much further than the health and usage monitoring systems developed for helicopter engines and transmissions, or the self-diagnostic software in the most recent commercial engine full authority digital engine control (FADEC) systems. It aims to push the technological envelope with untried methods of warning and detection.

These include acoustic foreign object damage (FOD) detection, eddy current monitoring and even an electrostatic engine monitoring system. Together with other advanced systems, such as dual FADECs, oil conditioning and coking monitoring and fault isolation, the drive is towards an "autonomic" logistics support system. Unlike current support systems, in which the engine is supported at depot level by a set number of spares, the autonomic system would be self-regulating and rely on its own data to keep precise pace with the spares and maintenance demands.

The new devices have been tested on old F100 engines "-which we could use and abuse" says Shambaugh. These engines were deliberately "seeded" with faults, to see if the sophisticated new systems would detect the trouble and accurately predict the ensuing failure. Cea says: "We actually tried to break things. We introduced foreign-object damagee, we made the oil dirty and we made bearings run hot and dry. We tried to get the sensors to measure when things would break. It was a kind of ecumenical test."

The electrostatic system monitored changes in the "electrostatic quality" of the inlet. "We dumped stuff in there, and the sensors would measure it to see if a bad thing was about to happen," he says. The system was able to detect the difference in electrostatic reading between leaves, small stones and even surround-wrap plastic. The initial trials ceased in December, but further evaluations are planned for the second quarter of this year.

Tests will include further investigations of the eddy current and acoustic FOD systems, the latter aimed at detecting minute differences in the sound made by a bearing or a gear to give warnings of wear or failure. "We will be trying to determine if they are viable to bring into EMD. We will do the analysis of the data and see what we are going to do with it. We will also give it to the pilots and the maintenance crews and see what they make of it-we have to think through all that," adds Cea.

Another advanced system being developed for the engine is known as the self-tuning on-board real-time model (STORM). The JSF119 engines, like their F-22 sister powerplant, are controlled by two fan-speed sensors. The STORM runs full-time simulations of the engine parameters in real time and compares them with actual readings. If one, or even both of the fan sensors becomes inoperative, the STORM system is expected to be able to calculate what the engine "should" be doing, and feeds the data to the FADEC.


One of the conditions of the JIRD is the provision of assurances that the JSF119 has plenty of room for growth. "It's a question of power. Right now, we are meeting or exceeding the requirement," says Cea. Nonetheless, the company is putting together a growth plan and a technological roadmap of how to get there.

Currently sized in the 40,000lb thrust (178kN) area, the near term growth is expected to be achieved relatively easily and simply with a modified "overflow" fan and a throttle push. This could raise overall thrust levels by as much as 5%. A more significant thrust increase of up to 10% is achievable, from 2000, with another revision to the fan to boost airflow further. The rotor inlet temperature (RIT) would also be increased by introducing more advanced cooling into the high and low pressure turbines. This would include the supervane and superblade cooling techniques developed as part of P&W's work in the US Government's Integrated High Performance Turbine Engine Technology programme.

Beyond 2010 is potential for further growth towards 50,000lb thrust. This would involve a further RIT increase, more cooling and a larger fan diameter and associated fan airflow growth, as well as a bigger fan pressure ratio.

For the moment, however, most of P&W's efforts are focused entirely on the successful test and delivery of the initial JSF engines to Palmdale. With much to do before the all-important 2000 first flight dates, it seems unlikely that the current frantic pace of life at the West Palm Beach site is likely to slow down any time soon.

Source: Flight International