Lockheed Martin's mile-long fighter plant in Fort Worth, Texas strains today to deliver an average of one F-35 Joint Strike Fighter prototype every month. Since the second flight-test aircraft - the short take-off and vertical landing (STOVL) model BF-1 - debuted in June 2008, only one other flight-test aircraft has achieved "first flight" status.
Although Lockheed maintains that the first 13 flight-test aircraft will be complete by early 2010, which itself marks a slight delay from the latest revised schedule, the production system is clearly struggling to keep pace.
The second STOVL prototype - BF-2 - achieved first flight in February, but has since been grounded while it undergoes another build cycle. Lockheed chief test pilot Jon Beesley says the programme's management is considering a plan to keep the prototypes in final assembly longer to avoid such a lengthy, follow-on build cycle after the first flight. Such a plan, if approved, is likely to push first flight schedules for subsequent prototypes even further behind their current due date.
Somehow, Lockheed must quickly overcome these early growing pains. Over the next seven years, current acquisition plans call for dramatically raising output until a new fighter is delivered every working day, excluding holidays and weekends, or about 240 jets in a year.
In the world of modern military aircraft manufacturing, the F-35 is on an entirely new level. Even the F-16 production system peaked at about 15 F-16s a month. But the F-35's final assembly plant at its projected peak will produce 20 aircraft a month split among three major variants and perhaps a dozen international configurations.
Bob Fiorintini, Lockheed's head of F-35 global production, likened the expectations for JSF production to commercial industry. "This isn't a typical military programme," he says. "It's more like a commercial programme."
As such, Fiorintini says Lockheed has sought out lessons from the Toyota Production System, which is known as the forerunner of lean production methods aimed at eliminating waste and integrating just-in-time logistics.
By the end of the next decade, Fiorintini says, Lockheed's aim is to have branded its own successfully manufacturing style the "Fighter Production System", to be emulated by other producers.
The precise production concept remains in development, but elements of the commercial manufacturing model are already in evidence in Fort Worth. For example, Michael Williams is a production supervisor of 28 employees on position J475, which assembles the aft-inner section of the F-35's wing.
Previously, Williams co-ordinated any changes with functional engineers who were his peers. Now, all of the functional supervisors report directly to Williams on the production line, and are charged with ensuring that his team has what they need to assemble the aircraft.
Williams says Lockheed's production system is quickly improving. While the first prototype aircraft through the system fell behind schedule, the timelines at his station has shortened progressively for each of the last five aircraft in assembly, he says.
Williams' experience is backed up by the reported observations of the Government Accountability Office, the investigative arm of the US Congress. In May, the GAO reported Lockheed's projected labour hours for completing the final assembly process on the first 12 flight-test aircraft.
In the report, the GAO quoted Lockheed statistics showing that the fourth conventional take-off and landing flight-test aircraft - AF-4 - is expected to consume about 20% fewer labour hours in final assembly than the previous CTOL type. Meanwhile, the projected decline for the STOVL variant is even more dramatic. BF-5 is expected to consume 32% fewer labour hours than BF-4, although the latter is highly complex as the first aircraft dedicated to mission systems and radar cross-section testing.
However, the GAO's numbers also reveal that these projected declines reflect a major growth from estimates made only two years ago. In fact, AF-4 was originally expected to take 91,000 labour hours to complete, but is now projected to need 162,000. BF-5, meanwhile, is projected to require 174,000 labour hours to assemble, or about 70% more than forecast in 2007.
"Problems and delays are largely the residual effects from the late release of engineering drawings, design changes, delays in establishing a supplier base, and parts shortages, which continue to cause delays and force inefficient production line work-arounds where unfinished work is completed out of station," the GAO reports. .
"The 2007 schedule assumed a steeper drop in labour hours as more units are produced and manufacturing and worker knowledge increases. The new schedule, based upon actual performance, projects a less steep decline in labour hours, indicating slower learning and lesser gains in worker efficiency," the GAO adds.
Lockheed's manufacturing system is still struggling to overcome the design changes imposed by the 2004 STOVL weight attack team (SWAT), which removed nearly 2,270kg (5,000lb) of excess weight from the F-35B design. The redesigned airframe and propulsion system spilled over to the CTOL and carrier-based variants, as well as the manufacturing system. One major change from the original programme was the elimination of "quick-mate joints", which added complexity at the final-assembly stage.
But a perhaps more profound change involved a major redesign of the wing's production method. Instead of a single-piece wing, Lockheed broke the structure into four sections that must be mated with each other individually and with the centre fuselage in combination. At the same time, Lockheed also saved weight on the wing by reducing tolerances for the machined components.
At the time, Fiorintini says, Lockheed was aware that the new tolerances would create disturbances in the supply chain. Lockheed's suppliers for machined components lacked the tools required to meet the new specifications. As the supply chain introduced and experimented with new tooling, key components of the wing would be delivered to final assembly later than planned, he says.
© Lockheed Martin
Steven O'Bryan, Lockheed's vice-president of business development, also acknowledges the source of the problem on the wing.
"It's really about parts arriving on time by the suppliers and [more to the point] not arriving on time. It's not the producibility," O'Bryan says. "The parts did not arrive on time so you have out of station work. So parts arrived at the wrong time. So you're not doing it in the process you want and it takes longer."
Lockheed had laid out its production system to mate fully assembled wings with the fuselage sections in two-storey assembly structures on the final assembly line called the electronic mate and assembly system (EMAS).
The parts for the wings would have been integrated ahead of the EMAS station, where workers could install parts on wings suspended vertically in a tool, Fiorintini says. Many of those parts still do not arrive until after the wing structure is loaded horizontally into the EMAS station for mating with the fuselage. This requires Lockheed's workers to lie on mats atop the aircraft to install parts that should have already been assembled into the structure.
The GAO recognised the same issue in its assessment of the F-35's current manufacturing progress.
"The overlap in the work schedule between manufacturing the wing and mating [connecting] it to the aircraft fuselage has been a major concern for several years because it causes inefficient out-of-station work," says the GAO report. "The contractor continues to address this concern, but the new schedule indicates that this problem will continue at least through 2009."
In reality, Fiorintini says he expects that the out-of-station issue on the wing assembly will persist through at least 2011, at which point the supply chain will finally be equipped to meet Lockheed's post-SWAT design tolerances for machined components.
That is also the timeframe when several changes will be made to speed up the production system. With the F-22 unfunded in President Barack Obama's proposed fiscal year 2010 budget, Lockheed is already investigating how the F-35 could occupy the space now used for F-22 production in Fort Worth, Fiorintini says.
Around the same time, Lockheed also wants to create a single flowpath inside the final assembly centre for all sections of the F-35. Currently, some aspects of machining and assembling wing subsections are conducted on a secondary line adjacent to the F-35's EMAS and moving assembly stations. Lockheed's goal is to move wing subsection assembly to the beginning of the F-35 assembly flow, minimising and simplifying the movement of parts from one station to the next, Fiorintini says.
It will be a major change in operating style and philosophy for the world's largest defence contractor. Lockheed has abandoned commercial aircraft production since the demise of the L-1011 TriStar in 1985. Surveying the fate of current commercial programmes, such as the Airbus A380 or the Boeing 787, also does not inspire confidence.
Defence and commercial manufacturers face a different set of challenges. Defence manufacturers usually benefit from a locked-in source for development financing, but the customer has the right to define the aircraft's requirements and capabilities. Commercial manufacturers often must seek an independent source of development financing, but have the advantage of selecting the aircraft's performance levels internally.
O'Bryan points out that the F-35's production strategy is unique in the military aircraft business for reusing the tooling that was designed for the development phase in production.
"Why do we think we can produce the aircraft? Looking at legacy aircraft - the F-16 and everyone else - we built the SDD jets on a tooling and line [optimised] to get it out there as quickly as possible," O'Bryan says. "I used to say that we threw away that tooling, but the government never throws anything anyway. They just put them in a warehouse, and you never use it again.
"The F-35 is the same actual tooling that we used in SDD - that moving assembly line, that EMAS station, those auto-drive tooling systems that you have," he adds. "You have proven that out in SDD. It makes more work to do that, but it makes it so much easier because all of your processes, tooling and timeline are the same ones that you used in SDD."
As a military programme adopting a commercial-like production model, Lockheed's evolving production strategy must keep up a robust output while maintaining the flexibility to adapt to changing customer requirements that emerge from flight-test and real-world experiences.
Fiorintini says he sees a direct parallel with the nature of building narrowbody airliners, such as the Airbus A320 or Boeing 737. They are not originally designed for military applications, although both have been adapted for certain military roles - most famously the 737-derived Boeing P-8A Poseidon.
But there are obvious similarities between Lockheed's production plans for the F-35 and narrowbody airliners. The level of output is one example. The F-35 is expected to peak at one delivery every working day, Airbus and Boeing each deliver at least one new narrowbody every day, including holidays and weekends.
The F-35 has three major versions, while Airbus and Boeing each produce four major variations of the A320 and next-generation 737. Further, as the commercial airframers must tailor elements of each model's configuration to any one of dozens of international buyers, so must Lockheed adapt certain elements of the F-35 to different foreign governments.
Another added complexity of Lockheed's planned production system includes an alternative final assembly and check-out plant based in Italy. The Italian parliament has recently approved the $2 billion investment to build the plant, which will assemble jets purchased by Italy and perhaps the Netherlands. Similarly, Airbus is starting up final assembly of the A320 at a new plant in China.
Lockheed intends to study the A320 production system, in particular, as it also sees parallels in Airbus's distributed supply chain among several major industrial partners in different countries.
Fiorintini says Lockheed's top interest is studying how Airbus manages the A320 supply chain. As Lockheed is learning, he says, keeping up annual output despite all potential disruptions, including natural disasters, is no simple task.
"In the supply chain, things happen," he says. "How do you recover from that?"