Composite structure issues at Airbus?
One nice thing about metals in critical functions is that many of them fail slowly. Composites tend to fail catastrophically, like this (via AirDisaster):
.jpg)
Here’s what happened, according to the online Mail and Guardian:
At 35 000 feet above the Caribbean, Air Transat flight 961 was heading home to Quebec with 270 passengers and crew. At 3.45pm last Sunday, the pilot noticed something very unusual. His Airbus A310′s rudder — a structure over 8m high — had fallen off and tumbled into the sea. In the world of aviation, the shock waves have yet to subside.
This isn’t the first time that an Airbus A300/310 has had a rudder failure. The whole vertical stabilizer fell off on American 587, which crashed just after take off two months after 9-11. According to AirDisaster: “Separation of the vertical stabilizer, which was found over two miles away from the main impact site in Jamaica Bay, is suspected as the precipitating event, leading to an in-flight breakup of the aircraft.”
Flawed inspection procedures?
American 587 and Transat 961 are not the only flights with possible structural issues regarding the rudder and vertical stabilizer of the A300 family. The M&G reports a FedEx A300 that experienced a serious rudder malfunction in 200. Further, a scanning of the NTSB database of A300 incidents shows a number of incidents, of which this May 1997 American Airlines Flight 903 incident seems to bear some similarities.
According to the M&G, rudders and vertical stabilizers are visually inspected for problems, which, according to a leading expert, is inadequate:
In an article published after the flight 587 crash, Professor James Williams of the Massachusetts Institute of Technology, one of the world’s leading authorities in this field, said that to rely on visual inspection was “a lamentably naive policy. It is analogous to assessing whether a woman has breast cancer by simply looking at her family portrait.”
Airbus’s response isn’t very reassuring:
Barbara Crufts, an Airbus spokesperson, said visual inspections were “the normal procedure” and insisted Williams’s case was unproven. “You quote him as an expert. But there are more experts within the manufacturers and the certification authorities who agree with these procedures.” She disclosed that the aircraft used in flight 961 — which entered service in 1991 — had been inspected five days before the incident. She said did not know if the rudder had been examined.
Issues at Airbus?
Continuing with the Mail and Guardian report:
The separation of the rudder may have further implications for the cause of the 587 crash. In its report, the NTSB said the tail and rudder failed because they were subjected to stresses “beyond ultimate load”, imposed because the co-pilot, Sten Molin, overreacted to minor turbulence and made five violent side-to-side “rudder reversals”. The report said the design of the A300 controls was flawed because it allowed this to happen.
However, the NTSB investigation has been criticised by many insiders. Ellen Connors, the NTSB chair, told reporters last January that the report was delayed because of “inappropriate” and “intense” lobbying by Airbus over its contents, adding: “The potential for contaminating the investigation exists.” In America, the NTSB staff is small and manufacturers provide many of the staff employed on air-crash investigations into their own products.
The report of inappropriate and intense lobbying effort by Airbus is extremely disturbing, particularly given a possible pattern of failures of composite rudders and vertical stabilizers. The authorities should be treating this safety of flight issue with the seriousness it deserves. Further, if there is a structural problem with any A300 component, Airbus should be out in front of it — not doing lobbying behind the scenes to tone down critical reports.
Rand Simberg, Short Final, and Rovian Conspiracy have more.
UPDATE
An Emergency Airworthiness Directive was issued by Airbus a week after this piece.
May 28th, 2005 at 2:11 pm
Flight 961 Rudder Failure
05-22-2005
First, composites are good choice for application on many aircraft structures.
Rudder structures are good candidates for applying a well designed honeycomb sandwich composite structure.
The rudder on the A310 is honeycomb sandwich. It is not foam.
A honeycomb sandwich structure or, for that matter an aluminum structure will fail in service if, never inspected!
Honeycomb sandwich structures are prone to absorb water over a period of time and when it freezes, it will cause progressive facing to core disbonds!
These sandwich structures meet design and safety requirements providing an adequate inspection program is in place to ensure that core to facing disbonds are discovered at onset. Then, repair dispositions are made using approved repair procedures or rudder replacement. What are the Airbus inspection requirements, inspection schedule and the test methods required by Airbus to discover facing to core disbonds on the inner and outer facings of the rudder specified prior to this failure??
To my knowledge, Airbus does not require any inspection of the tail and rudder to be performed prior to five years in service. After five years only visual inspection is required.
Ref. Air Safety week, June 17, 2002, page 3 and July 29, 2002, page 3. I don’t recall the source of the “visual inspection after five years of service”.
We all know that the pilot on a walk around inspection can only tell if the rudder is attached.
Airbus requires only visual inspection of the tail and rudder. Facing to core disbonds, the primary modes of failure on the rudder will not be discovered by close up visual inspection. A one inch square core to facing disbond will not be detected by visual inspection. When water freezes in the core cells it expands enough to break the facing to core bond in that area. The facing remains in contact with the core after the ice melts. There may be a .001 of an inch or less space between the facing and core and the facing is not bulged in that area. This .001 in. gap is not detectable visually. This one inch square disbond will increase in size over a period of time and could grow to a 2 x10in.area and not be detected by visual inspection.
Once a disbond occurs an adjacent disbond will occur in lesser time than the first. Now the stress in the facing must go around the disbonded area, and causes a stress riser at the adjacent core to facing bonds. Also a peel force is introduced as the facing will try to buckle over the disbonded area, thus disbonding occurs at a much faster rate as the disbond area increases.
These disbonds should be detected at onset before large areas of the rudder are affected. When skin waviness is detected visually, this is a good indication that a large area of disbonding has occurred and the structure is not flight worthy. Small disbonded areas should be detected at onset by some means of testing other than visual: because you can’t see them. Fix these small disbonds before the large areas of skin disbonds can be seen visually indicated by waviness.
The manufacturer will specify the maximum area of core to facing disbonds allowed for the rudder to stay in service. This area will be relatively small. The largest single disbonded area allowed, my guess, Five inches square on each side of a sandwich panel or 20 square inches per rudder.
Large disbonded areas put forth in some discussion forums as acceptable that include a total disbonding of one face form the core or 30% of the total bonded surface will cause the rudder to separate form the tail.
Reference AD 97-04-07 allows up to 5000 square cm or 1,968 square inches of disbonds per rudder. The rudder is constructed of two sandwich panels consisting of an outer and inner facing each or four bonded facings per rudder. Allowable disbonds of approx 500 square inches per bonded facing amounts to a 10in. X 50in. disbond per facing. A 10 in., X 50 inch disbond on one facing would be enough to unravel the rudder.
These high numbers allowed for disbonds are not realistic and show a lack of understanding of how sandwich structures work. Are we talking mm?
In short, core to facing disbonds are the primary cause of failure in sandwich structures.
It follows then, that engineering provides an inspection method applied to the rudder at scheduled inspection intervals to determine if facing too core disbonds exceed the maximum allowable disbonded area allowed by engineering.
What if any inspection methods other than visual methods were applied to this rudder over the last 13 years of service?
A tap test applied to the exterior facings will easily detect facing to core disbonds on the exterior surface.
The interior facings are closed to tap tests. What test do we use to determine disbonds on the inner surface?
There should be an ultra sound test to do this or some test, specified by engineering that will accurately and in a timely manner discover disbonds at the inner facings.
There is one Airbus test called ELCH “Repetitive Elasticity Laminate Checker” It is designed to discover core to inner facing disbonds. This seems to be sensitive to test procedures as a vacuum is applied to the outer skin and pulls it in into the cavity above. The amount of deflection of the skin then indicates if there is a disbond.
Overlaps of the skin prepreg could show less deflection indicating a good bond. Another problem is porosity in the skin which could change the readings.
If there is no record of testing for disbonding history for this rudder for the last 13 years: It’s highly possible disbonding exceeded the maximum area allowed by engineering.
Please click on the link below for honeycomb sandwich info from Hexcel.
Please click on Benefits on the right and click Attributes and Properties.
http://www.hexcelcomposites.com/Markets/Products/Honeycomb/Sand_design_tech/p04.htm
Below, is shown the upper part of Flight 587 rudder. The outer facing (skin), honeycomb core and inner facings are shown.
See photo No. 5
http://www.usread.com/flight587/Coast_Guard_Pix.html
The #7 hinge is not connected but is attached to tail with pieces of the rudder attached. The tail photo is not shown.
If #7 hinge was still attached to the A310 tail with bits of the rudder spar attached as the other hinges are, then the cause of separation from the tail would point to unraveling of the rudder. But!!!!! The cause of the #7 hinge failure at the rear spar of the tail must be determined before cause of the separation can be resolved.
The current ADs requiring tap test on these rudders should be withdrawn or revised to include a test method that will discover outer and inner facing to core disbonds.
My advice to the pilots is to insist, through your Flight Safety Organizations, on a full report detailing the inspections conducted to date on the rudder of your aircraft. If the report documents visual inspections only: and show no history of testing to determine if disbonding is within engineering allowables:
Don’t fly the ship.
Any comments are appreciated:
Thank You:
Eugene Schulte
261 PR 3333
Bridgeport, TX 76426
940-683-2855
griptite@wccs.net
Enclosures:
U.S. Read-coast Guard Pix of Flight 587 Rudder.
(2) Hex Web Honeycomb Sandwich Design Technology
Air Safety Week March 21, 2005 Publication, “A310 Losses Rudder, Prompts Fleetwide Inspections, Inquiry”.
Airbus “Repetitive Elasticity Laminate Checker”
November 1st, 2005 at 11:51 am
Has anyone responded to my offering that I believe is a creditable reason for the Flight 961 rudder failure?
December 17th, 2008 at 10:51 pm
ur missing jal 123 :S august 12 1985…….520 death. 4 survivors. mistake by a bad rudder lost its hydraulics and crashed
December 17th, 2008 at 10:53 pm
do not talk bad of airbus because 123 jal lost it too just without any thing… this a 300 lost it because of the !”$&(()??$% pilots
June 10th, 2009 at 12:56 am
Marcel, the JAL accident identified a flaw in the design, as well as the repair carried out on the pressure bulkhead of the JAL aircraft. Boeing acknowledged this, and modified the hydraulic system to try and prevent future losses. The Airbus rudder is similarly flawed, in my opinion. I have not noticed Airbus rushing to modify the basic structure, despite several worrying incidents. My experience of AB is that they are convinced their product is perfect, despite evidence to the contrary. Dangerous!