Report in Business week

Bubbles in fuselage a hitch in FAA certification for 787
Glitch won't sidetrack program, Boeing says

A final composite fuselage section that was to be used in certifying The Boeing Co.'s 787 Dreamliner failed during manufacturing, sending engineers scrambling to keep the company's all-important new jetliner program on schedule.

A cross section of the Dreamliner's composite fuselage, which will be made by Boeing's partners in South Carolina, Wichita, Kan., and in Italy. The wings will be made in Japan.
Mike Bair, vice president and general manager of the 787 program, said in an interview Thursday that the failure will not sidetrack the program, and the first planes should still be delivered on time to All Nippon Airways of Japan in mid-2008.

But the problem shows the kinds of risks that Boeing faces at a time when the 787 program is entering a critical stage and the manufacturing and production system must work as advertised. The first flight of the new jet is expected late next year.

The 787 will have a composite airframe, including the fuselage -- a first for commercial jets. Boeing cannot afford any major snags in proving the technology to the Federal Aviation Administration and other regulatory agencies if the plane is to enter airline service on schedule.

Boeing has not been specific about the actual delivery date, but sources say All Nippon Airways is supposed to take delivery of the first planes in May 2008.

"We know what happened," Bair said. "We learned some things as a result, which is part of what this whole process is about. We know what we have to do to support our certification plan."

To develop, refine and prove new manufacturing processes needed for the Dreamliner, Boeing and its partners have been producing composite fuselage test barrels since late 2004 at the company's developmental center, the same facility that in the late 1960s was used for work on Boeing's supersonic transport before it was killed by Congress. The center is near the south end of Boeing Field.

Eight barrels were produced there without significant problems.

The failure came on the ninth, and what was supposed to be the final, fuselage barrel. Testing of that barrel was to have provided the data Boeing supplies the FAA as part of the 787 certification process. A commercial jetliner must be certified before it is allowed to carry passengers.

Ultrasonic testing found it was too porous -- too many bubbles were in the composite material after the fuselage section came out of an autoclave in April, Bair said. It took engineers more than a month to fully understand the issue.

Commercial jetliners have traditionally been made out of aluminum. For the fuselage, this requires pieces of aluminum held together by thousands of rivets. The composite fuselage barrels of the 787, however, are made in one large piece. They are the largest composite pressure vessels ever made. The barrel that had the bubble problem was the largest that Boeing had made yet -- about 33 feet long.

The manufacturing process involves laying down carbon fiber material on a huge mold. That mold, or mandrel, is mounted on a tool that rotates the barrel as the plastic carbon fiber tape is applied. The structure is then wrapped and placed in a huge autoclave oven for curing. While in the autoclave, the barrel is under enormous pressure, which essentially squeezes the layers of composite material together.

Every composite part that is cooked this way has a certain level of "porosity" that is acceptable, Bair said.

"You can't get it perfect."

But testing of the ninth barrel showed its porosity level made the barrel unacceptable, he said. The bubbles might have made the structure too weak, he said. The problem was a "subtle condition" and could not have been spotted by someone simply looking at the barrel.

The problem was the mandrel, Bair said. It was built some time ago and was too big. The mandrel changes size as it is heated in the autoclave and this must be taken into account when the fuselage barrel is made. Some of the material on this mandrel had been machined off to get it back to the proper size, Bair said, but as a result there were places that leaked. That's why too many bubbles formed during the curing.

Bair said engineers knew there was some risk of this happening.

"We made a judgment that we could make this work," Bair said. "It turned out we didn't."