Hacker News Reader

Comment by andrewcooke | Hacker News Reader

more info - http://www.flightglobal.com/news/articles/ntsb-urges-groundi...

(this is serious because - as far as i can tell - this engine has a 330 minute ETOPS rating http://www.geaviation.com/press/genx/genx_20120308.html http://en.wikipedia.org/wiki/ETOPS which means that a twin engine plane (like the dreamliner) can fly 5.5 hours (the number in the report above) from "safety" (eg over the atlantic), so if one engine fails the other must continue working for that time. if the problem is common then that is not so reasonable an assumption...)

(not all dreamliners are affected; rolls royce also provide engines)

View on HN

A little more info on ETOPS and the challenges created by multiple similar failures within a small population...

There are two parts to ETOPS (tongue in cheek "Engines Turn or Or People Swim")..The first is the certification to design standards, including flight testing. The second, and perhaps more crucial element, is the operations/procedures certifications. In the first stage, the entire system (engines, airframe, fuel system and everything else) is evaluated as a comprehensive whole. Essentially, you're looking to eliminate single points of failure.

The second part is arguably more important as it deals with the statistical failure rates. The assumption is that a modern jet (turbofan) engine is highly reliable. As the basis of such procedures, multiple, similar failures are worrisome. An airline must have approved ETOPS operations to fly routes requiring ETOPS (a certified airplane is necessary but not sufficient). Common operational procedures include not replacing both engines at the same time, servicing similar fuel system components together, etc. Basically, avoid the possibility of introducing a single point of failure into otherwise redundant systems.

Single points of failure can emerge but hopefully they're rare. The classic case is something like British Airways Flight 38 at Heathrow in 2008:

http://en.wikipedia.org/wiki/British_Airways_Flight_38

(By sheer good luck and good piloting skills the crew were able to make it over the runway threshold after both engines lost power during final approach: injuries and a hull loss, but no deaths. Problem was a design flaw in the fuel oil heat exchangers, only triggered by unusual meteorological conditions on a long flight. Both engines had the same design issue and were drinking the same fuel ...)

And if it had been a 4 engine plane the same issue would have been present (or even a 24 engine plane).

Sadly I can't find the article any more but did read one about how each new generation of commercial aircraft is increasingly aerodynamically efficient. The side effect of that is how much further they can glide from the same starting point (less drag). Its hard to find numbers that aren't speculation but it seems like 1980s era designs (eg 767) are around 15:1 and current generation are around 20:1.

Wikipedia even has a page about flights that required gliding: http://en.wikipedia.org/wiki/List_of_airline_flights_that_re...

> The assumption is that a modern jet (turbofan) engine is highly reliable.

I was really surprised to find how true this was. One of my first jobs was working in a manufacturing plant for a company remanufacturing various parts of various jet engines (rebuilding blades, putting new airfoils into damaged vanes etc). The joke on the floor for new employees was that once you saw how badly damaged the engines were by the time they got to us you'd never want to fly again. It was also very interesting to see how much FAA regulation was involved in the manufacturing/remanufacturing process and how often people tried to skirt those rules to save parts due to how expensive they are.

I'm glad they're paying attention to single points of failure in maintenance: echoing cdtross, I can remember reading about a flight taking off from Miami, I think it was, going over water, where in pretty short order all engines failed hard due to their all having been incorrectly serviced in the same way, something to do with oil seals. Fortunately it wasn't far out when this happened and it had enough energy to make it back safely to the airport.

That would be Eastern Airlines Flight 855: http://en.wikipedia.org/wiki/Eastern_Air_Lines_Flight_855

All three engines on the L-1011 failed due to an incorrect maintenance procedure.

> All three engines on the L-1011 failed due to an incorrect maintenance procedure.

There was absolutely nothing wrong with the maintenance procedure. It was simply not executed properly.

Apparently there was a maintenance procedure being performed quite consistently on these engines. But it was a bad one.

You're making the exact same mistake. Check the wikipedia article on this particular flight, under the heading 'Cause':

""The National Transportation Safety Board determined that the probable cause of the accident was "the omission of all the O-ring seals on the master chip detector assemblies leading to the loss of lubrication and damage to the airplane's three engines as a result of the failure of mechanics to follow the established and proper procedures for the installation of master chip detectors in the engine lubrication system, the repeated failure of supervisory personnel to require mechanics to comply strictly with the prescribed installation procedures, and the failure of Eastern Air Lines management to assess adequately the significance of similar previous occurrences and to act effectively to institute corrective action. Contributing to the cause of the accident was the failure of Federal Aviation Administration maintenance inspectors to assess the significance of the incidents involving master chip detectors and to take effective surveillance and enforcement measures to prevent the recurrence of the incidents.""

A faulty procedure means that every plane of this make and model would have this defect. A faulty execution of a correct procedure would only affect those aircraft that were handled by the people executing the procedure incorrectly.

It is the difference between a bug in the software (the procedure) or a computer with a faulty ram chip. Both will cause your program to crash, but the first will cause every instance of that program to crash, the second only affects the specific instance run on that hardware.

There is a huge difference when it comes to root causes. If perfect procedures are executed sloppily then it doesn't matter how good the procedures are. And that is why there was a huge flap over this particular incident because if there is one thing that the whole airline industry revolves around then it is adherence to procedures.

You started picking nits saying:

There was absolutely nothing wrong with the maintenance procedure. It was simply not executed properly.

But the text you cite says:

"as a result of the failure of mechanics to follow the established and proper procedures" and "repeated failure of supervisory personnel to require mechanics to comply strictly with the prescribed installation procedures"

I'm pointing out that there were, in fact, multiple procedures involved here. Some were well-defined and supported the reliable operation of the engine, some procedures were probably poorly defined and led to rapid failures.

But the mechanics do appear to have been following a repeatable process as evidenced by the fact that all the engines failed simultaneously. This is a procedure in the general usage of the term, even if a bad one and different from the one prescribed by the regulatory authorities.

Reply on news.ycombinator.com