This is my third addition to my Lessons Unlearned series. I have been trying to aim at lessons unlearned from particular accidents; however, I feel it is important to look at the root of all accident investigation and ask the question: Are they being done correctly?
A young colleague recently asked me: Can the National Transportation Safety Board (NTSB) be audited? I responded that auditing the NTSB would be hard. For instance, the Federal Aviation Administration (FAA) audits their individual departments; the findings are constructive when acted upon. One can argue that self-auditing is counter-productive by design, but then who can better audit the FAA than itself through its other detached divisions?
The NTSB is different; they are accident investigators for the five transportation modes: Marine, Rail, Aviation, Highway and Pipeline. The NTSB’s job is not to conduct surveillance, understand the daily occurrences or even track the transportation mode they are overseeing (because they don’t oversee anyone); they don’t track the goings on of industry or perform trend analysis based on root cause data. If the Fire Department doesn’t track homeowners’ battery spending to verify home smoke detectors are operational, can anyone audit them for the preventable fires they put out? How about Homicide Detectives; if they can’t track the drug traffic accurately, can they be audited for the unsolved drug related murders?
The NTSB serves a unique function in the transportation community; it is problematic to audit an entity that serves that kind of disconnected purpose. Hard? Yes, it would be difficult. Unprecedented? Probably so; and yet it would (and should) not be impossible. Considering the NTSB’s odd influence over the entire transportation community, it is my opinion that it should be absolutely vital. It’s confounding that if not previously proposed, the time is long overdue. Consider this: all transportation modes have evolved with the times; computers have taken over all aspects of Rail, Aviation, Marine, etc. yet the NTSB doesn’t stay abreast of the latest technologies training; they’re not required to. So, how does one bring the NTSB into the 21st century?
One way would be to compare its accident report history against its successes. Has the NTSB changed the different transportation industries for the better? Do the Probable Cause(s) and Recommendations address the contributing cause(s) of the accident as well as what caused the accident? Are the NTSB’s Recommendations staying ahead of the transportation industry’s changes with technology? Do they understand composites? How about 3-D printing?
Here’s the problem: the NTSB has changed very little with the times since becoming an independent agency in 1967. Over the last fifty years, all five transportation industries have strived to change with the times while the NTSB has remained unaltered, fixed in an LBJ-era mentality of how it does its job and who it hires: investigators that should understand the transportation industries.
To define the NTSB’s influence on the transportation community, let’s use billiards as an analogy. After an accident, the cue stick represents the NTSB, while the cue ball serves as the transportation oversight agency, e.g. the Federal Railroad Administration (FRA), the Federal Maritime Commission (FMC), the Federal Aviation Administration (FAA), etc. The fifteen billiard balls would symbolize the respective operators, e.g. airlines, shipping firms, rail lines, etc.
The cue stick drives the cue ball in any direction it wants with as much (political) force it chooses. Motivated, the cue ball directs the other billiard balls to orderly ricochet off the walls, bee-line into a specific pocket or put all of them into complete chaos. This is the power and influence of the NTSB. The irony is that the NTSB ultimately advises those agencies with far more experience, e g. FRA, FMC, FAA. In other words, those who know how their respective industry works better than the NTSB could possibly grasp.
For this article’s purpose, the emphasis will be on aviation. Any aircraft accident investigation relies on the integrity and experience of its investigator(s); at least it’s supposed to. The NTSB employs mostly engineers as major investigators for the aircraft side of the investigation; they review everything from airline system safety to structural integrity to maintenance records. They comb the wreckage looking at the effects the accident aircraft had upon the flight, the crew, passengers and cargo. Engineers can pinpoint how an aircraft is supposed to perform, its limits and what improvements can do for the aircraft.
Engineers cannot, however, anticipate the human or environmental element. Moreover, they cannot anticipate an operator’s culture.
Most major accidents aren’t a result of the aircraft’s design or its failure to meet that design. The quality of the design is what the aircraft is repaired to: equal to or greater than. The design is what the pilots are trained and fly the aircraft to. Very rarely does a major aircraft accident occur due to the airplane NOT performing to design; it does happen, but not often. ValuJet 592 crashed despite the fact the aircraft was mechanically sound, the crew properly trained. Air Midwest 5481, was flown outside of its design parameters. China Air 611 crashed, but not because of a faulty design, but because the operator failed in a minor inspection timetable. US Air 1549 ended up in the Hudson, not from a poor design, but from environmental influences acting outside the design, namely a flock of birds that exceeded tested allowable bird ingestion limits. American 1420, Pan Am 759, Delta 1141, Southwest 1248, Eastern 66 and many more were due to outside factors.
Engineers, like most NTSB investigators, aren’t part of the everyday operation of the airlines. In some cases, they aren’t even employed by the airline, but are contractors. Here are some things an engineer does not do for an airline: calculate weight and balance on a live flight; rivet a new structural member; swap out an engine; check baggage; fly the aircraft; change a main brake; swage a hydraulic line; perform a pre-flight inspection; or schedule an aircraft for maintenance. So, the question is: If engineers have nothing to do with the day-to-day operations of an airline, how can an engineer investigate why the operational failure within an airline caused an airliner to crash?
In addition, the NTSB engineers are investigators in charge of each NTSB investigatory group. Having run one of these investigatory groups, it has been my experience that participants, e.g. the airframe manufacturer, the engine manufacturer, the unions, the contract maintenance provider; while assisting with the accident findings, are trying to assure their respective employers don’t have the ‘finger of blame’ pointed at them; it’s human nature – and good business – not to bear the fault of innocent lives lost. While leading said investigatory groups, I had the advantage of knowing how each party ‘operated’; where their skeletons were hidden and how to find them. In the end, the tricks played to sidetrack the investigation were not successful. But the groups led by inexperienced engineers were not so successful; at least the NTSB investigator did not prevent any monkey business. However, the FAA was there to prevent the games.
An airline’s culture is paramount when determining the cause of any accident; it would be irresponsible to ignore it. The NTSB usually finds what brought down the airliner, but they [rarely] find the true cause(s) of the accident. When they do, they defer to the obvious, e.g. pilot error, instead of the truly important cause. It’s enough that industry buys into whatever they are selling, no questions are asked. And if they do, the five Board Members – novices all – put their feet down, relying solely on the NTSB’s name.
Why do we investigate accidents? To assure they don’t happen again; that the causes are addressed – properly. Perhaps it is time to audit the NTSB. It would be proactive of the transportation industries to affirm that the NTSB’s half century-old procedures work; that their B707-era, analog-instrument, aluminum-alloy structure Age practices of the 1960s work as well in a B787-era, digital-instrumentation, composite-structure Age of today. I’d be surprised … very surprised … if they do.