Boeing has overcome unusual challenges while flight testing the 737-900 stretch

Guy Norris/SEATTLE

The terms "ground breaking" and "earth shaking" are not normally associated with the commercial flight test - particularly not the certification of a simple stretch of the world's most prolific jet airliner. Yet they have taken on new meaning to the Boeing 737-900 team that completed tests on the longest, and probably the last major 737 derivative in April, clearing the way for first deliveries to launch customer Alaska Airlines on 16 May.


Fighting to regain lost time after the discovery of early elevator tab vibration problems, Boeing's primary test -900, YD501, was completing autopilot work in eastern Washington on 28 February when a major earthquake measuring 6.8 on the Richter scale rattled through the Puget Sound area. Although more than 80km from the epicentre, both Boeing Field and the company's nearby Renton production site were badly hit.

YD501 was diverted to Everett's Paine Field, well north of Seattle, while feverish work began to repair the badly damaged runway at Boeing Field - the centre of the company's flight test and standard-body completion and delivery operations. "Within two days we were able to dispatch aircraft and in six we could land aircraft with some weight restrictions," says 737 chief test engineer Tony Aresu. The field was re-opened for business by mid-March.

The quake also damaged several older buildings at Renton, some of which had to be condemned. Also damaged was Boeing's Test Data Retrieval System (TDRS) which collects downloaded test data from the aircraft, adds Aresu. "The TDRS toppled over, but fortunately none of the equipment itself actually fell on its face," says Aresu.

Jon Robinson, 737 programme manager, adds: "In fact the biggest obstacle was getting access for our test engineers to some of our buildings." Despite the serious disruption, Boeing recovered nearly full functionality within "a couple of days, mostly because we were dealing with a well-motivated and highly flexible workforce," Robinson adds.

The earthquake hit just as Boeing was catching up on lost time caused by the elevator tab problems. These originally surfaced in the form of a 44Hz vibration on the maiden flight of YD501 on last 3 August. The resulting fix and additional testing forced Boeing to increase flight test hours by around 60% over the original schedule and added around 25% to the amount of ground test hours originally budgeted for the programme. It also inevitably led to a month's delay in certification and delivery.

US Federal Aviation Administration and European Joint Aviation Authorities certification was originally set for late February, but slipped to mid-March as the elevator problem was tackled. The earthquake further affected the schedule and FAA certification was finally obtained on 17 April, followed two days later by JAA approval. Boeing originally expected the -900 testing to last just six months, which would have made it the shortest programme of its kind.

Flight and ground tests

The plan called for a mere 380 flight test hours, plus 119 ground test hours. All was to be achieved using only one instrumented test aircraft, again a first for Boeing. The 737-800 test programme, for example, involved two instrumented aircraft and had been concluded after 692 flight hours and 443 ground test hours.


Boeing had good reason to be optimistic. The -900 is a basic stretch of its recently certificated predecessor, with a 1.06m (3.5ft) aft fuselage extension and a 1.57m stretch forward. The resulting 42.1m long aircraft seats 189 in a single class layout, is 13.5m longer than the 737-100, and even exceeds the overall length of the first production 707-120 by 0.4m. Developed as a fast-track competitor to the Airbus A321, it was formally launched in November 1997 with an order from Alaska Airlines' for 10.

Coming so close behind the -800, the fast-paced plan did not require engine inlet waterspray tests or Vmcg/Vmca (minimum centre of gravity and control speed) evaluation, as all could be achieved by analysis alone. It also did not require ground vibration or flutter testing, and needed only reduced testing for stability and control, as well as autopilot and basic aerodynamic qualities. In addition, the limited systems and avionics changes also meant that extensive testing in these areas was not required.

As a result, Boeing opted for only one instrumented test aircraft (YD501), plus a second aircraft with a standard interior for limited air and ground tests. The second aircraft, YD502, was allocated a block of 56 flight test hours and 22 ground test hours. However, the test issues encountered in the programme forced YD502 to be eventually used for 123 flight test hours and 41 ground hours, while a third non-instrumented aircraft was also drafted in for a limited number of additional tests. YD501 finally amassed 528 flight hours and 120 ground test hours.

Other test differences from earlier programmes included combined NAMS (nautical air miles fuel consumption tests) and initial airworthiness tests using the same aircraft. In addition, YD501 was deployed to Edwards AFB, California, for combined Boeing aerodynamic performance tests and FAA ground effect and stability and control certification tests. The revised elevator tab design, which was strengthened with a new hinge point and additional layers of composite material, passed flutter tests in a new series of tests that began in late January. Flight tests also evaluated the effect of the stiffened tab on the elevator hinge moment and air data calibration.

In the interim, YD501 was also used to test the aeroelasticity characteristics of the Next Generation wing to determine why early flight tests of a winglet-equipped -800 had come up with a higher fuel burn than expected. The tests showed the baseline Next Generation wing was itself not responsible for the phenomenon, which appeared to effect the wingleted -800 at particular weights. The Aviation Partners Boeing blended winglet was subsequently certificated in April for retrofit and forward fit, and entered service with German charter airlines Hapag Lloyd and Air Berlin on 8 May. Tests revealed "some aeroelasticity with winglets, but they still offer tremendous fuel savings," says Robinson who adds that the impact is "not as significant as we first envisioned". The first winglet-equipped -900 is expected to be offered to operators in 2002.

Aircraft on order

To date 45 aircraft have been ordered, including 10 for Alaska, 15 for Continental Airlines, four for KLM and 16 for Korean Airlines, the first of which is due for delivery later this year. Alaska is expected to take delivery of the first three -900s by the end of May, while the first aircraft for Continental is also due to be handed over by month's end. Further deliveries to the two US carriers are set for June, as is the first -900 delivery to Netherlands flag carrier KLM. Boeing meanwhile plans to complete 26 737-900s by the end of the year, of which 21 will be delivered compared to the original target of 22.

With long-awaited deliveries getting underway, Boeing is refocusing on three principal areas: future -900 sales, supporting entry-into-service and preparing for the possible launch of the -900X, a 200-plus seater charter variant of the 737. Initial market studies of the larger capacity, longer range derivative are being concluded with much of the emphasis being placed on trade studies between raising maximum take-off weight (MTOW) and limiting structural changes.

The company wants to keep MTOW increases to within 4,550kg (10,000lb) or so, yet hopes to achieve a range with full payload in the 3,700-4,440km (2,000-2,400nm) bracket. This is expected to be challenging since the only way to overcome the -900s exit limit would be by adding a pair of Type 1 emergency exits aft of the wing, or by enlarging forward doors. The extra weight of these changes, added to the structural strengthening needed to support the modifications (which could include a stronger wing and stronger undercarriage) is likely to make this a tough trade.

On the other hand sluggish sales of the baseline -900, combined with aggressive marketing by Airbus with its plans for higher weight A321-200 variants, are helping to push the prospects for a -900X go-ahead sooner rather than later. Following the pioneering work of the initial 737-900 test team, some at Boeing may well consider the risk well worthwhile.

Source: Flight International