By Dominic Gates / The Seattle Times
At a U.S. Senate hearing last month, a Boeing whistleblower claimed that small gaps in Boeing’s 787 Dreamliner could cause a catastrophic accident. The following week at the 787 assembly plant in South Carolina, Boeing insisted the gaps pose no risk and the jet is safe.
The dueling narratives, emerging as Boeing’s credibility is near an all-time low, left industry observers and the public at a loss as to the risk.
There may be no more independent a judge of that risk than John Hart-Smith, who made a name for himself standing up to Boeing management and who is also a world-renowned expert on airplane structural integrity.
At the request of The Seattle Times, Hart-Smith, a retired Boeing engineer and senior technical fellow, examined the data behind the allegations.
Hart-Smith is famous within Boeing for something more than his eminent scientific pedigree: For being a stout internal critic, a champion of the engineers, a technical expert who more than two decades ago warned management its shortsighted financial focus would be ruinous.
A few years later, when the 787 program began, he made an internal appeal to Boeing leaders that, when they failed to listen, also proved prophetic: He told them that building the 787 fuselage sections out of giant single-piece carbon-composite barrels would be a very expensive mistake.
For this unwelcome message, managers sidelined Hart-Smith, telling him not to interfere with the 787 program, which, while a sales success, is unlikely to ever turn a profit for Boeing.
So he feels some kinship with current Boeing engineer Sam Salehpour, the whistleblower who told the Federal Aviation Administration and then Congress that the gaps are a structural flaw in the 787’s major fuselage joins.
Salehpour, 70, has worked in aerospace for 40 years. A design and quality engineer at Boeing for the past 17 years, Salehpour previously worked at various companies mostly on missile systems.
On Thursday, he filed a complaint with the Occupational Safety and Health Administration, alleging his warnings inside Boeing went nowhere and brought “blatantly hostile and retaliatory treatment.”
Salehpour’s main allegations center on how the large 787 fuselage sections are fastened together with a “splice plate” around the circular join between them. (He made separate allegations about poor manufacturing quality on Boeing’s 777 jet.)
His concern on the 787 is how Boeing deals with tiny gaps at those major fuselage joins.
These gaps are not along the circumference of the join but inside the aircraft, between the splice plate and the skin of the fuselage.
Salehpour claims the forces Boeing mechanics apply to close these gaps during final assembly can damage the carbon composite skin around the fasteners at the join — risking a major structural failure.
“They are putting out defective airplanes,” he said during the U.S. Senate hearing.
Hart-Smith believes Salehpour is mistaken. His analysis supports Boeing’s insistence that the 787 fuselage gaps are not a safety risk.
Salehpour is “right in theory that if you have to apply too much force to close the gaps during assembly, that could damage the structure,” Hart-Smith said in a phone interview from Australia. “But it appears the gaps are not big enough to make that happen.
“It’s not a safety issue.”
‘They wouldn’t listen to me either’
A fellow of the prestigious Royal Aeronautical Society, Hart-Smith — now 83 and retired from Boeing since 2008, though he consulted for the jet-maker until 2015 — still writes scientific papers for aerospace engineering journals on how to economically build sound airplane structures.
After obtaining his engineering doctorate in Australia, in 1968 he joined Douglas Aircraft, the commercial airplanes division of McDonnell Douglas, which later became part of Boeing.
Over four decades, his research and analysis improved manufacturing techniques and reduced defects. Hart-Smith’s redesign of a problematic tail cone on the C-17 military transport, for example, is credited with elimination of defects in that part while saving $500,000 in manufacturing cost per airplane.
At Boeing, he was designated a senior technical fellow, the highest level of engineer at the company.
Hart-Smith doesn’t have access to all the internal Boeing 787 data. But he knows the strain on composites that will cause damage. And from the dimensions of the 787 fuselage he can work out with basic mathematics the strain on the skin caused by closing any given gap.
He applied this to the data from measurements of the 787 fuselage gaps revealed in internal Boeing documents obtained by The Times. This was the same data Salehpour spotlighted to support his claims.
Hart-Smith calculated the threshold gap size that would begin to cause damage at the fasteners at about just over 0.1 inch, significantly larger than any of the gaps Boeing measured and, he said, “way outside anything the FAA would tolerate or that Boeing aerodynamicists would tolerate.”
Hart-Smith hardly lets Boeing off the hook. He believes and sympathizes with Salehpour’s claims of retaliation for pushing his concerns internally.
“They wouldn’t listen to me either,” Hart-Smith said.
Moreover, meticulously measuring and filling the gaps during assembly of each 787 is as expensive as Hart-Smith warned it would be.
Boeing has projected the cost of inspecting and reworking the gaps on previously built 787s before it delivers the jets to airlines at $6.3 billion.
“It’s not a catastrophic safety issue, as Salehpour suggested, but only a major disaster for Boeing in regard to costs and assembly times,” Hart-Smith said. “As everyone now knows, it’s a financial disaster.”
Damage measured in microstrains
The strain on an airplane skin is measured in microstrains.
Consider a piece of metal or composite the length and shape of a footlong ruler. If you apply a force and stretch the material by one millionth of a foot, that’s one microstrain.
Hart-Smith says a force of 1,000 microstrains would be needed to produce the first “barely detectable” damage in the composite material around the fasteners. And it would take a force of 3,000 microstrains to actually break it.
He calculated that a circumferential strain of about 1,000 microstrains would be produced by forcing the sections together over a gap of 0.11 inch. The largest gap Boeing found on testing a sample of 28 Dreamliners was a third of that, 0.035 inch.
In Boeing’s South Carolina press briefing, Steve Chisholm, chief engineer for Boeing Mechanical and Structural Engineering, declared himself “absolutely not concerned” about a gap of that size.
One internal Boeing document reviewed by The Times has a single measurement of a significantly larger gap.
This was from a 2013 test on a 787 under assembly to determine whether increased forces would cause damage. The document notes a “relatively large gap” of 0.09 inch. This is a key data point Salehpour cites to justify his concerns.
Boeing said that test “was performed to intentionally induce an artificially large gap at a circumferential joint” and “is not representative of actual airplane builds.”
Informed of Hart-Smith’s analysis, Salehpour stood by his position. His lawyers, Debra Katz and Lisa Banks, said they “strongly disagree” with Hart-Smith’s conclusions.
Although no current Boeing employees have yet publicly supported Salehpour, the lawyers said his position is “confirmed by engineers who, unlike Dr. Hart-Smith, actually work at Boeing now and have observed these issues firsthand.”
Salehpour wrote that in the documents about the test on that large 0.09-inch gap, he’s seen “no indication that this was an artificial gap, despite Boeing saying so now.”
Katz and Banks said Boeing is asking for trust but after so many revelations about quality lapses “does not deserve the benefit of the doubt.”
Salehpour has provided documentation of his allegations to the FAA, which is investigating.
Small gaps, big consequences
It has always been a routine part of aircraft final assembly to deal with gaps at airframe joins.
Nothing ever fits perfectly, especially when the parts coming together are fabricated at different sites by different suppliers.
Mechanics pull the fuselage sections together by tightening temporary fasteners in a splice plate at the circular join. This leaves the outer skin surfaces of the fuselage sections flush, completely smooth across the join.
However, small disparities between the fuselage skins of the two sections may then leave a gap between splice plate and skin on one side.
Most of these gaps are closed by force as the temporary fasteners are tightened.
Any remaining tiny gaps — as thin as two sheets of paper or a human hair — are then plugged with machined pieces of material called shims. In carbon composite planes like the 787, the shims are typically made of fiberglass.
Finally, the temporary fasteners are replaced with permanent ones.
Correct shimming makes the structure solid and sound. Incorrect or missing shims reduce the life of the airframe and over time can cause structural damage.
Boeing officials routinely speak of how they took the 2020 discovery of 787 fuselage gaps and missing shims very seriously. They cite how Boeing largely halted deliveries for 22 months at huge expense to closely inspect its undelivered airplanes.
This wasn’t voluntary. Boeing was forced to do so by the FAA.
In a June 2020 letter obtained by The Times, the FAA gave notice to Boeing it was preparing to mandate action for the 787 over “multiple airframe unsafe conditions … due to systemic manufacturing deficiencies when joining principal structural elements.”
The FAA requested a “global evaluation of the shimming process” and asked Boeing: “What was the justification for removing Quality Assurance from the various stages of assembly?”
That prompted Boeing’s extensive inspection and reworking of about 120 airplanes. That work continues today on about 40 remaining undelivered 787s.
During this massive project, Boeing held training presentations for employees on “why shimming is important.”
Late 2021 internal Boeing emails obtained by The Times note the explicit intention of the training: to convey “how a seemingly small/inconsequential item can have catastrophic consequences” and “how can something the thickness of a sheet of paper be so critical?”
That’s illustrated by an FAA airworthiness directive issued last July in response to reports of cracks discovered on 747-8 jumbo jets.
The FAA said a Boeing investigation found the root cause was “un-shimmed or incorrectly shimmed gaps that were larger than engineering requirements.”
The FAA mandated detailed inspections of all 747-8s because the cracks could grow and “lead to a failure of the fuselage skin … which could adversely the structural integrity of the airplane.”
The first 747-8 where the cracks were discovered was just over a decade old and had flown just 5,517 flights, the FAA said.
Boeing’s Chisholm said his engineers design and test airplanes to be structurally sound at least for a 20-year service lifetime. For the 787, the design target is a conservative 40,000 flights.
The first 747-8 entered service the same year the first 787 was delivered.
Carbon-composite structures
Though the 747-8 comparison sounds alarming, an important difference between these two jets is that the 747-8 has an aluminum skin while the 787 is the first airliner made largely from carbon composites.
Composites, fabricated from plies of carbon fiber embedded in hardened epoxy resin, don’t crack like metal. Instead, when damaged the plies can separate, referred to as delamination, and microcracks develop in the resin.
Salehpour’s contention is that the increased force Boeing has been using to close the gaps between the splice plates and the fuselage skins when joining the barrel sections could cause delamination around the fasteners — and that this might spread around the circumferential splice.
A key question is whether such local delamination will spread with the repeated stress loads of subsequent flights, the way cracks in aluminum grow.
Hart-Smith cites research that shows when the load around a fastener is increased enough to initiate damage to composite material, the resin softens and as the softened area widens the stress concentration decreases. The force on the fastener diminishes and the damage is self-limiting.
“The loads on the airplane skin in service are negligible compared to the forces used to join the barrels,” he added. “So if composite subassemblies do not break during assembly, they are unlikely to break in service.”
Responding to this point, Salehpour said via his lawyers that this is true only if an aircraft is properly shimmed. He said the issue is that “Boeing is not properly shimming the vast majority of the time.”
In counterpoint, Boeing points to evidence that the 787 structure is robust and remains so after many flight cycles.
In South Carolina, Chisholm described tests where the composite fuselage was deliberately hammered to cause delamination at the damage site, which didn’t spread after being put through 40,000 simulated flight cycles.
Boeing put one of the first 787s built through fatigue testing on the ground that simulated the loads and pressurization from 165,000 cycles. And it conducted detailed inspections of in-service 787s, including some coming in for their 12-year maintenance checks.
In all of this, Boeing found “zero fatigue issues in the composite structure,” Chisholm said.
Standing up to Boeing management
Speaking with a refined Australian accent — influenced by his Oxford-educated father — Hart-Smith has an endearingly professorial disregard for the practicalities of everyday life.
He struggles to keep his cellphone and laptop up to date and working. Managing his home in Long Beach, California is a perpetual challenge.
His gentle manner cloaks a steel spine, as shown by a paper Hart-Smith presented in 2001 to senior engineers and executives assembled at Boeing’s annual Technical Fellowship meeting in St. Louis.
He had just had surgery to insert a stent just below his heart and so had to deliver his presentation by phone, with a colleague flipping the slides at the venue.
He presented essentially an economics paper, though expressed in clear, commonsense terms and with a dry humor.
It scathingly critiqued the Boeing strategy since the 1997 McDonnell Douglas merger of outsourcing work and divesting core design and manufacturing assets. He explained how this would ultimately increase costs, lower profits and jeopardize Boeing’s ability to develop future airplanes.
Management denounced the naiveté of a structural engineer intruding on such matters. But the paper went viral among employees in the Puget Sound region and throughout Boeing.
Hart-Smith got into trouble with management again as the 787 development program took shape in the early 2000s, this time with a technical opinion. He warned top-level engineers and executives not to go ahead with the plan to fabricate the jet from single-piece composite barrels, each made by a different major subcontractor.
He told them that perfectly joining these large fuselage sections would prove extremely costly.
Hart-Smith wrote a report, which Boeing ignored, outlining a better approach that would avoid any circular joins: making the sections from curved, fuselage-length panels fastened together.
Some years later, Airbus decided to use large, curved panels instead of barrel sections to manufacture its all-composite A350 fuselages.
While the Airbus design still has three circular joins in final assembly, Hart-Smith said it’s easier to build because the panels allow more flexibility at the join.
“I tried tooth-and-nail to prevent them using this barrel manufacturing method,” Hart-Smith said.
By 2014, when the outsourcing of the major 787 parts was a clear failure and the program was billions of dollars over budget, then-CEO of Boeing Commercial Airplanes Jim Albaugh admitted that Hart-Smith’s 2001 paper had been “pretty prescient.”
Since 2020, Boeing has been paying the enormous cost of ignoring Hart-Smith’s advice on how to build the 787.
But the company can take comfort from his assessment that for the 787s it’s delivering, however expensively built, there’s no safety risk with the gaps at the fuselage joins.
Talk to us
> Give us your news tips.
> Send us a letter to the editor.
> More Herald contact information.
