Aircraft Accidents and Lessons Unlearned IX: United 232
On July 19, 1989, at 37,000 feet over northwest Iowa, United Airlines flight 232, a McDonnel-Douglas DC-10-10, suddenly lost hydraulic fluid to all three main hydraulic systems. Hydraulics, being the lifeblood of an airliner, is used to actuate all primary and secondary flight controls, e.g. Ailerons, Flaps, Elevators; Hydraulics are required to lower and raise the landing gear, steer and brake the aircraft on the ground. Despite the heroic efforts of the three pilots and an off-duty check airman, limping the airliner to Sioux City, Iowa airport, control of the aircraft was lost on Final approach.
United 232 suffered a catastrophic uncontained failure of the #2 engine’s Stage One Fan Disk, the center hub of the engine’s most forward fan section. At rotational speeds of 1800 rotations per minute – any random fan blade of the eight-foot diameter fan section spins past the same point 1800 times per minute – the center hub came apart. The separated sections’ centrifugal forces included enough energy to slice through the engine’s protective fan shroud, pierce the aircraft’s horizontal stabilizers and sever the #1 and #3 Hydraulic System lines routed inside the horizontal stabilizers; the #2 Hydraulic System was compromised at the #2 engine itself; thus, all three Hydraulic Systems failed.
In the case of United 232, the complete loss of Hydraulic System fluids was the cause of the accident. It was, however, not what caused the accident.
Why did the #2 engine’s Fan Stage One Disk (FSOD) fail? Per National Transportation Safety Board (NTSB) report AAR 90/06, the suspect FSOD was, at the time of the accident, seventeen years old; it had accumulated 41,009 flight hours and a total of 15, 503 flight cycles (one take-off through landing equals one flight cycle). The year previously, it underwent its sixth Florescent Penetrant Inspection (FPI), a Non-Destructive Inspection (NDI) method.
Non-destructive Testing (NDT) and Non-Destructive Inspection (NDI) are a means for finding cracks, corrosion and other structural anomalies; NDI and NDT are used to test these components without having to destroy the component, part, structure, by destructive stress testing; the part can be reused if found to be structurally sound. It’s the equivalent of X-rays versus exploratory surgery
The quality of the NDT process and materials is vital to accurately locating metal deformities and defects. Without properly trained NDT personnel, NDI is completely ineffective. Florescent Penetrant, Ultrasonic, Eddy Current, Magnetic Particle and X-ray inspections aren’t taught through common means; specialized training and qualifying to a high standard are critical to finding different types of defects; proper use of meters and equipment are imperative.
During the post-crash analysis of the recovered FSOD pieces, the metallurgical findings included radial cracks stemming from a ‘nitrogen-stabilized hard alpha inclusion’. Per sciencedirect.com, ‘hard alpha inclusions (HAI) are low-density, hard, brittle regions of spuriously high nitrogen and oxygen that occur in titanium alloys;’ the HAI was introduced during the melting or forging process. Titanium alloys become more brittle from the addition of nitrogen and/or oxygen, breaking down the integrity of the metal. HAI is easily detectable by NDI as surface cracks and voids. However, they are extremely difficult to detect when subsurface and require NDI that deals with magnetic fields to detect.
Still, though the Stage One Fan Disk separating so violently from the engine is the reason for the Hydraulic failures, the Fan Disk is an inanimate object; its integrity is dependent on outside influences; it can’t detect its own flaws nor repair them. What initially caused the accident? We must go deeper by examining other similar accidents.
On July 6, 1996, a Delta Airlines MD-88, flight 1288, suffered an uncontained engine failure on take-off roll in Pensacola, Florida. Engine debris from the #1 pod (left) engine’s Fan Hub, exited the engine’s shroud before penetrating the fuselage. Delta 1288’s Fan Hub’s failure began during manufacturing; the drilling and boring of holes caused radial crack damage. The crucial defect: hard alpha inclusions (HAI), this time involving Oxygen contamination, not Nitrogen. The titanium alloy was overheated and exposed to oxygen during machining. Drill bit chips, not removed during the drilling process, may have made micro-cuts in the holes.
On October 28, 2016, American Airlines flight 383, a Boeing B767, experienced an uncontained engine failure during takeoff roll in O’Hare airport in Chicago, IL. The engine’s High-Pressure Turbine Stage 2 Disk broke into three large fragments, plus several smaller fragments; the second largest fragment, weighing fifty-eight pounds, punched through the right wing and soared a half-mile onto a warehouse’s roof. Though the final report hasn’t been approved, the investigation is leaning towards defects in the High-Pressure Turbine Disk’s manufacturing.
The Lessons Unlearned from these three investigations should focus, not only on improving non-destructive inspection methods, but in how well they are being applied. In United 232, it is easy to find fault with limited NDI technologies. But, the eight years between United 232 and Delta 1288 saw improvements in Florescent Penetrant Inspection (FPI) technologies. Why would the defects not be found?
Per NTSB accident report AAR-98/01, during post-accident inspections at Delta’s Atlanta facility, Federal Aviation Administration (FAA) inspectors found that Delta’s FPI testing area for components and parts was ineffective, e.g. the integrity of FPI materials were corrupted by contamination and improper housekeeping methods; FPI personnel training was deficient; inadequate quality controls; all led to seven violations that could have each contributed to insufficient inspections.
This problem is cultural; it points to multiple abuses in an airline’s culture, from cost-cutting attempts to manipulating the mechanic seniority system. The NDI technology is there, but the air operator ignores the benefits. The NTSB cannot dictate culture changes; an air operator’s philosophy can’t be changed by a Recommendation or put to rights by an Airworthiness Directive. A culture must be fixed from within.
The twenty years between Delta 1288 and American 383, saw a rise in Eddy Current Inspections (ECI), a different type of NDT. ECI improved the detection of sub-surface defects, e.g. cracks or cavities. In time, Amplitude Analysis, an older version of ECI, was being overshadowed by a more sensitive ECI, Phase Analysis.
Culture was not a strong contributor to American 383; the technology was there; American made use of it. But what industry changes took place in the years between 1996 and 2016?
In November 2001, while investigating the American 587 (NTSB Report AAR-04/04) accident, I flew to Tulsa, Oklahoma, to look at American Airline’s maintenance facility. The Tulsa facility was representative of a vanishing maintenance norm; an era where airlines conducted their own maintenance, in-house, e.g. heavy inspections, component rebuilding; practically everything done to keep an airliner airworthy was accomplished by the airline itself.
Through the late 1900s, as the passenger and cargo airline industries continued to expanded overseas – slowly at first, but rapidly increasing their flights – maintenance and inspections were contracted out to other airlines and Repair Stations in Europe, East Asia, the Middle East and South America. This saved money, e.g. labor and facility costs, and prevented unnecessary movement of the airliner just to accomplish maintenance. As a result, airlines e.g. American, began downsizing their maintenance facilities and contracting out important inspections, e.g. Non-destructive testing. Although the quality of inspections did not suffer during the transition, the airlines began to turn over control of their inspection programs and processes to a third party; oversight went from first-hand (single level) oversight to multi-level oversight.
This represents major changes in the control of: training, personnel, inspector pay scale, types of NDI offered, timeframes for work accomplished, etc.; a major transformation of the aviation industry with regard to the quality control of inspections, e.g. who performs inspections and who oversees them.
These changes are beyond the NTSB’s abilities to recommend safety. Few in the industry could have foreseen airline industry milestones, e.g. Deregulation, Hub-and-Spoke, Regional Airlines, International Contractors, that are commonplace today. We should go back to the early accident reports to assure safety measures applied then when the industry was different, still apply today; perhaps to learn lessons unlearned. If not, American 383 may not be the end of the line of such accidents, a line that should have stopped almost thirty years ago with United 232.