Main Gear Boxes and The Grand Lottery

[SASless]

We have had two Main Gear Box failures in the past few weeks that destroyed two aircraft and killed thirty three people causing much sorrow and loss to the families and friends.

Knowing Gann, I have to accept things like this will befall us as we pursue our career in aviation but yet we can question the reasons as to why they happen.

Is it "Fate"....a simple roll of the dice or are we assisted by the system letting us down?

Some questions that came to mind while thinking about this situation....


What does "Extremely remote" mean, and how was it missed by EC? By the DGAC? By Sikorsky? By the FAA?

That gear inside the 332's transmission is a "Primary Structural Element" and as such was certified in the 332L2 and the EC 225 as having a failure probability that was "extremely remote".

That means the probability of failure is less than one in a billion - a fleet of a thousand aircraft flying 1000 hours per year flying for 1000 years should not experience such a failure.

Yet, in spite of Eurocopter's promise and the aircraft certification, the gear broke, suddenly, and with little if any warning.

Clearly they missed it - both Eurocopter and the JAA/Easa, precisely as the oil leakage path on the S92 was missed by Sikorsky and the FAA.

How does one miss such things? Is there a blame? or is this something we have to live with?

[This aint Jim Beam]

If youre looking for 100% mechanical reliablity then youre in the wrong job.
It will just be another AD like the other mechanical issues causing fatalities.
Remember when the blades were flying off the A109s and the mast were coming off Bell mediums, when Arriel 5 modules were blowing up on Astars, and Robbie blades had all the inertia of a balsa model. Tail rotor tip caps came off 500s and all pretty much of the LTS101s early history?
Its that sort of thing.

[C.Korsky-Driva]

Jim, maybe the questions SAS is posing is valid for all those failures too. If we just sit back and accept it as part of the wallpaper of life then where, when and how do we progress.

There are stories about things going wrong at the factory that only come out years later. The saga of how the S76 tail rotor problems were suddenly solved overnight with a piece of kit that must have taken months to develop, the spindle saga, Bells combining gearboxes. Are they being straight with us or what?

When the Bell 412 gearbox (epicyclics) seized a couple of years back they managed an autorotation - why not the S92 and the 332L2?

CCD

[spinwing]

Mmm ...

To an extent JB is correct ... we are at the mercy of the mechanical engineering .... BUT is that not the reason to examine the result of incidents and accident in our industry and then IMPROVE the design and reliability of of our machinery .... to allow the improvements of computer design tools to make our machinery better ....

SASless as far as transmissions are concerned ... was not the BA Chinook disaster also related to a transmission failure??.

Perhaps ... now the certifying authorities will revisit the criteria for Public Transport Helicopter transmission design and certification in view of recent events!


:eek:

[Oldlae]

We do not know the precise failure of the 332L2 epicyclic gearing but I have concerns about Sikorsky using titanium studs on a vital part of the MGB oil system, why change from the usual steel studs used for years? Titanium is not very tolerant, the MRH on the Lynx has to be treated with extreme caution, any slip of a spanner causing a ding on the material is a big problem. I suspect that titanium might be similar to stainless steel (SS) in day to day use, we had to be very careful when using SS bolts not to strip the threads and sometimes it couldn't be helped.
With a filter cover that had to opened several times during the tbo of a S92 mgb perhaps the studs were inadvertantly overstressed causing an ultimate failure. We really need to know why the studs failed on the NL S92.

[This aint Jim Beam]

i think to some extent its a "learn from your mistakes" sort of industry.
A bit like, "wow Kapton wiring" and "..all aircraft will be made of composites in the future.."
You dont know about alot of this stuff until its been in service for a while. But then the 332L has been around for a long time, maybe my argument is only valid for the S92 accident...

[bondu]

SAS,

Quote:- "That means the probability of failure is less than one in a billion - a fleet of a thousand aircraft flying 1000 hours per year flying for 1000 years should not experience such a failure."

It does state where in that one billion an incident will occur: it could be during the first 100 hours or equally it could be in the last 100 hours. Who knows? Don't get me wrong; this event must be thoroughly investigated and steps taken to try to make the odds better than one in a billion. As a current EC225 driver, I have a vested interest!

bondu

[SASless]

The BAH Chinook crash was a transmission failure.

JimB,

You failed to catch the point I was making.

The FAA has a definition of "Extremely Remote" as does the DGAC and CAA and JAA and EASA....just to mention a few of the certifying organizations.

My question is how they apply that definition to critical parts and how they determine when to apply the definition to those parts as compared to requiring documented testing of the items.

Look at the Sikorsky situation with the 92 gearbox.....it runs for hours with almost any amount of oil left in the sump....but only minutes with no fluid. It might have run longer if the aircraft had been flown at a reduced power setting than was used. No one anticipated the oil filter failing and causing such a failure as has occurred twice now.

The EC gearboxes have shown themselves to be subject to an unexpected cracking of gears based upon one known event and probably this latest tragedy. Despite HUMS and top notch maintenance engineering we see this failure occur and we now have to question the risk assessment that was applied to the potential failure of the gear box and its critical components.

I am not suggesting we can achieve perfection although that should be the goal otherwise we should not be involved in aviation to begin with.

What we need to do is assess the thinking behind some of the decisons and promises made by the builder and certifying agencies.

In light of this MGB failure at Bond.....the "Extremely Remote" concept just flew out the window!

Bondu,

You posted as I was editing my post.

I feel you prove my point about the definition....if it can happen in the first hour or the last hour....then what does that say about statistical probability.

If we had two failures on the same type aircraft, on the same operation, on the same day.....would that definition still hold? All we would have to do is to have the fleet fly the rest of its service life without a another failure and everything would be okay I guess.

[Rigga]

The "Problem" with aviation is that it is constantly changing.

Equipment, Materials, Personnel, Processes, Repairs Schemes, Regulations, Research, Development, Standards, Companies, Manufacturers. All changing, all the time.

And nearly all the time it is to save money.

I cant remember which astronaut it was (John Bell?) who I paraphrase as saying:

"All the time I was up there, I couldn't help thinking all the capsule's parts were made by the cheapest bidders!"

The same goes for all parts made for any aircraft. If you think a broken Main Rotor is a critical item - try flying a plane with a broken Spar!

Yes, I am being simplistic! But I am trying to say that it can be a matter of Luck. As it can be lucky just to get to work every day, and then get home safely too.

Every care is taken with getting these critical products right - Manufacturers do worry about getting it wrong!

Perhaps it's time to change the "business equations" of safety levels - but I think that would be sooo costly as to prevent future development of any aircraft.

The main point is that we all want to make money as easily as possible, with the most risk permitted. We all accept inflated amounts of money to take part in those risk mitigations too.

We don't accept a high level of risk in anything we do - because we partake in daily risk assessment and mitigation.

Sometimes, someone gets something wrong - but not often. Often, it has been found, the person who 'does it wrong' is not aware of what has happened or why it happened until years after it happened.

Aviation, like medicine, is not a science. It is a "best guess" with some maths thrown in as a check.

We must learn from it - and then move on. As we always do.

We are all as good as the last accident/incident proved.


My respects to those that died, and to their families.

Rigga

[This aint Jim Beam]

I take your point SASless.
All these "one in a billion" statements are dervived mathmatically though.
Two things spring to mind, bigger and heavier.
Aviation parts, and particularly helicopter components, are built as light and as weak as possible.
Thats probably the reason Sikorsky used titanium studs, lighter than steel.
Build in more overload, more safety features, larger saftey margin.

I re read your first post about the death toll, 33 in two weeks, id imagine the oil companies (who pretty much drove up the aviation standards single handedly) will have quite a bit to say on this matter and things may change from their side rather than the regulators.
It will be interesting to see how this might affect the EC175s production testing.

[The Sultan]

Step One:

Put chip burning chip detectors to adequately monitor the gearbox and link it to pilot caution warning system. You do not need a $500K HUMS to do this. Let the crew either command a burn (Bell 222 technology from the 70's) or automate it with a CAS message alerting the crew. Multiple successful burns should be indication of bad things and prompt immediate crew actions. Unsuccessful burns should have the same result.

Step Two:

Implement step one as part of the regs (i.e. do step 1)

From the L2 design chips may have been accumulating on a HUMS read mag plug (not annunciated) which if annunciated could have given this crew actionable information.

The Sultan

[Flyt3est]

Sultan,

Your post assumes the defective item is making metal.. It is possible, and reasonable that a crack in a gear will make an alteration to vibration signals, before it makes significant amounts of metal.. Thus $ 1/2 million HUMS systems.

Your dismissal of the system is in itself symptomatic of the kind of attitudes that prevent system improvement. If half a million bucks worth of HUMS isn't quite good enough, then why not spend another 1/4 of a mil making it better?? At least then when we start making the oily bits out of non-magnetic materials, we will have system maturity on the fault detection.

I'm not Rubbishing your statement, just highlighting limitations. I guess though, nothing that involves a human interpretation or judgement will ever be fool proof.


PS to whoever mentioned "spacecraft made of parts supplied by the lowest bidder"

One rather well known American manufacturer openly tells its supplier base that it will be insisting on a 2% year on year price reduction for the next 5 years.. How do you think that is going to be achieved?? A bit of lean manufacturing? Six sigma? Improved investment in new technologies? That must be it.. For sure.. Phew! No need to worry!

[The Sultan]

Flyt3est

I am making no assumptions. I am very knowlegdable in HUMS and I am not dismissing it as not beneficial, as history has shown it is one piece of the safety puzzle. It is an extremely useful tool, and if it is the primary tool for chip detection it is critical.

The issue is that those who claim vibration monitoring is superior to, or can eliminate, debris monitoring border on delusional. It is only a great addition to the tool kit. For vibration to be generated you have to have an actionable level of damage. Problem with cracks on high stress gears are that they may propogate to a critical level before any loss of stiffness is noticed in a daily check of HUMS. It is basically good, good and bang. Relative to root cause of many failures debris generates before enough surface damage occurs to provide an actionable vibration signature. If the chips are magnetic and the chip detectors are well placed, chip lights will be the first indication. Obviously, if the result of the Bond accident is causing a more frequent check of the mag plugs, the root cause is made from magnetic material.

As stated above all detection methods have a place. The industry and regs need to pick the low hanging fruit which has been missed.

The Sultan

[Geoffersincornwall]

One of the major oil companies has gone on record asking a major US manufacturer (and the message was, I believe, passed to their French, German and Italian brethren) ......"Here is a flight bag with 10 kgs of metal - please put that back in your helicopter where it will do the most good, we will gladly accept 10kgs less payload as a price worth paying".

That after having to live with fragile tail-rotor systems and components that are scrapped if they are even scratched.

G

[Tmbstory]

A sad accident from the past, on Google type " Puma Helicopter accident 16th December 1980 near Kuala Belait Brunei ".

About 30 years ago.

Tmb

[Special 25]

Am I right in thinking some aircraft including some Eurocopter models have a weak link in the Main Rotor Shaft, that will break and allow autorotation in the event of a sudden gearbox failure ??

If this is the case, why is it not a feature of all helicopters. As a pilot, I've never missed it before, but as of February, I am beginning to consider it a necessity (if viable and available)

[Flyt3est]

Sultan,

Ta for clarifying, slightly misunderstood your first post.

Concur with your thoughts on the 2nd one.

FT

[mtoroshanga]

I think we are missing the point here. The original cause of these failures is often an overtorque that has caused damage to a load bearing surface. This eventually results in stress fractures or similair failures. Correct me if I am wrong but I seem to remember that after an event the IHUMS is redatumed and virtually continues from this new baseline.
Regarding the 330J Puma failure I was there and involved both in the recovery and investigation. There is more to the story than revealed.

[peter manktelow]

I know this thread has concentrated on the two most recent accidents but can any 76 drivers please comment on what the 76 RFM requires us to do ie we can conceivably have all three MGB caution lights illuminated + gauge presssure zero + gauge temperature high and it is a LAND AS SOON AS POSSIBLE ! When we get Grinding , Abnormal Vibration or Torque increase then we go to a Land Immediately or Ditch. I've always thought that by then , even if we are at zot feet and reduced speed , we are stretching our luck.....comments ?

Answer to a previous query. The only machine I am familiar with that had , inside the gearbox) , a waisted section on the main mast was the AS350. Supposedly the gear box could fail/seize and the mast would shear below the main upper gear box bearing thus allowing an autorotation. I wonder how they tested that ???

Peter

[mtoroshanga]

Ref wasted portion in gearbox, the Alouette series had that feature in 1963 when I did the course so they have had plenty of time to test but have never heard of it been used in earnest.

[Agaricus bisporus]

Spinwing, the BAH Chinook crash was caused by a transmission failure.

Briefly a big ring-gear (about 2 feet across) is fitted to the vertical mainshaft with a lot of bolts to take power from the horizontal high-speed shaft. Some, that is some, of these bolts were found to be losing torque slightly over the x000hr inspection period. A hugely complex and costly fitting of oversized bolts with higher torque capability was instigated.

Primarily due to a change in the washers used under these new bigger bolts (a lubrication channel was deleted) salt particles got into the thread area and started a corrosion pit which propagated a circumfrential stress fracture in the ring gear which finally failed. The 'box was fitted with a comprehensive chip detector system. Chip detectors cannot detect a stress fracture until metal starts spalling off, by which time it's probably far, far too late.



Post modified in deference to the remarks in the post below, though very little of the original was apparently incorrect.

Addendum...

At the time it was widely & popularly believed within the BAH community that the modifications were unnecessary, that the BA and Boeing engineers had developed a perfectly acceptable solution to the torque loss that did not involve oversizing bolts (and thus the incorporation of the non-scalloped shim. This, iirc, did involve Loctite) and that this mod had been forced against their better judgement by the CAA despite strident objections. Further, it was stated categorically within my hearing by people who should have known that one feature of the scalloped (grooved) shim was that it allowed lube oil to flush salt contaminants out of the joint, thus removing a potential cause of corrosion. The new-mod solid shims could not do this, and the view was that this lack of flushing almost certainly promoted, if not caused, tha corrosion pit that started the process.

Cetainly not a forseeable error chain, but a good example of the complexity of such events.

[squib66]

Agaricus bisporus
AAIB tell a different story: Air Accidents Investigation: 2/1988 G-BWFC (http://www.aaib.gov.uk/publications/formal_reports/2_1988_g_bwfc.cfm)

[froggy_pilot]

With the new helicopters like S92 EC225 AW139 we fly the latest technology, full glass cockpit, fully computerised, fully monitered (CVR,FDR, HUMS), and basic magnetic plugs ... :ok:

In flight we only have access to basic informations like temp and pressure and chip warning. :confused:

What would be the cost to bring all those datas centralised and available in flight with a proper monitoring.

As an example in the 139 I can know my hydraulic level at any time in flight, the level is monitered, but any significant change in a short time (leak) won't be noticed until the level is low. We always now too late :ouch:

If something goes wrong in the MGB IGB or TGB any significant change in vibration level and increasing in a short time should be displayed on the screen and not only "recorded".:ugh:

Don't tell me it's difficult to do, what computers can't do that ? :eek:

When you find yourself in one of those extremely remote probabilities, the few minutes or even seconds advance warning may change everything.


Manufacturers have to think differently when they design new aircrafts.
3rd millenium technology with last centenary thinking and design :yuk:

[SASless]

You engineers out there correct me if I am wrong.....you certainly have in the past!

Do not chip detectors tend to find "bearing" bits rather than "gear" bits?

Gear wear is more likely to produce fewer but larger chunks as bits of gear are broken off thus fewer bits are produced and are of larger sizes than bearing bits.

Chip detectors don't do squat for finding cracked gears.

[mtoroshanga]

Hi SASless you are quite correct. although the chip detectors/mag plugs collect any ferrous material floating about in the oil. You may recall that a couple of years ago I found an entire gear tooth from a combining gear box in Port Harcourt that showed up only as an intermittent flash on the chip detector warning light.

[JimL]

I was asked to post this on another thread but it really belongs to this one - it comes from the HARP report of 1984:The most complex element of a helicopter, after perhaps the engine, is the transmission gearbox, taking the drive at high speed from one, two or more engines generally in a horizontal plane, and gearing it down by 80 or 100 to 1, or so, to a single vertical shaft carrying the rotor head, with another drive going rearward to a tail rotor (or another rotor head system), the engines having free Wheels, and accessory drives being taken off for generators and oil pumps.

These gearboxes are remarkable examples of the art of the mechanical engineer. Although simplicity is a good starting point in the conceptual stage, complexity is inevitable, and it is not surprising that each designer has found his own solutions. The panel has seen spur gear trains, epicyclics, bevels; it has seen normal straight teeth, 'conformal' tooth profiles; it has seen ball bearings, rollers, tapered rollers, shafts with separate ball races and shafts with rollers running direct on them; we have seen aluminium and magnesium alloy cases. The only common element noted was in the gear material, a traditional low carbon nickel-chrome case hardening steel, either carburised or nitrided, and invariably made of high purity vacuum melted stock.

Notwithstanding the exceptional facilities for gear production we were shown, and the great attention to quality control being paid we are bound to record that in our view it will never be possible to eliminate completely the possibility of failure of some part of so complex a mechanism in service, between specified inspection or overhaul periods.

Although minor surface damage to a gear or bearing is unlikely to be instantly catastrophic, and should be detectable, and although gearboxes are tested to run for a period such as 30 minutes with the oil supply failed, the possibility of a major failure within the box is ever present.

Since with present knowledge a main rotor cannot be duplicated although a tail rotor gearbox might (conceivably a helicopter may survive the failure of the latter) we now come up against the fundamental difference between normal aircraft and rotorcraft -the inability to guard against a possible defect by duplication.

The Panel believes that, while gearbox development and testing should be as stringent as practicable, other means must be introduced to monitor the condition or "health" of the gearbox, preferably in flight. Elementary condition monitoring is already practiced (e.g. oil chip detectors), but much more attention should be paid to this. We make our detail recommendations on this subject later in para. 8.5.

There have been failures of transmission systems taking the drive to the tail rotor, or coupling twin rotor heads. We believe it may be possible to achieve a degree of redundancy here, or at any rate 'damage tolerance'. On the other hand it may be simpler to have some other method of cancelling main rotor torque to use in an emergency (bleed gas jet, rudder surface •••••• )There is no need to repeat anything in this quote but my underlines emphasize a number of salient points in the text. In original post I also indicated "The text in paragraph 8.5 (referred to above) contains recommendations for 'condition monitoring'; methods include: chip detectors; vibration monitoring; thermal detection or imaging in flight; oil sampling; telemetry (i.e. instantaneous download of HUMS data); and usage monitoring."

The exchange of information between 'The Sultan' and 'flyt3est' was interesting because they both have extensive (design) knowledge of HUMS - and both admit to weaknesses in the system of monitoring of the epicyclic modules. I tend to agree with Sultan, and the message contained in HARP report, that there any number of tools - each of which has its part to play. There is also the issue briefly touched upon in the HARP report that the role of monitoring tools is to deal with unexpected conditions within the inspection and overhaul periods. The setting of such periods, although important, becomes less critical when monitoring is effective – but it must be shown to be effective.

With respect to the S92 accident, no amount of monitoring (using the tools mentioned above) would have prevented that accident - but we have already extensively dealt with that issue in the S92 thread.

Just as we have inspection and overhaul periods in continuing airworthiness, we also need to have confidence that monitoring and early detection of incipient faults will provide sufficient warning to avoid unecessary stress/actions from the pilot in flight. We have also recently reset our tolerence levels to faults extremely low following these two accidents (and hence other threads reporting diversions). We must not put pilots in a position where cockpit indications result in additional hazards; having said that, Sultan's point about fuzz burning and further monitoring is valid.

Unlike the S92, the L2 accident appears to have come out of the blue. It is likely that the precursors to this accident, although subtle, were there; if they were and are pinpointed by the AAIB and EC (as I am sure they will be), the interim policy outlined in the AD provided by EC and EASA will be augmented/replaced by other procedures.

Helicomparator has a valid point that there has been little research into improvement of the HUMS monitoring, and specifically HUMS interpretation, for some time. The fact is that the teams which initially promoted HUMS have exhausted themselves with efforts in ensuring HUMS is implemented and, more importantly, that it is endorsed by the manufacturers (as enthusiastically as it has by Sikorsky). Another factor is that a number of the personalities who were the driving force behind the HUMS initiative have left the scene – some in retirement. HHMAG has also been disbanded since legal competence for airworthiness passed to EASA (and has not been replaced).

In spite of the success of the work of key personalities of the HHMAG and the resultant acceptance of HUMS, it should be noted that although VHM (for larger helicopters) is an existing requirement in the North Sea States and a Recommended practice in ICAO (and work was in progress to include requirement in the JARs by 2012 – which had been accepted by all manufacturers), there is no sign of a requirement in the new EASA OPS Proposal. When these regulations trump the national ones of the UK, Holland, Norway and Ireland, the requirement will be no more.

Why, is that? Is a question we might ask!

Jim

[Lt.Fubar]

With most of the chip detectors I come by, they worked only on ferrous materials, and of size approximately 100µm and more. As even the electric ones work on principle of collecting chips with magnetic field, and then shorting an electric circuit inside with that said chip. If the chip is not ferrous (coatings, light alloys, polymers etc.) or too small - it will not indicate.

Better chip monitoring systems include spectrography and lubricant flow monitoring... known in aviation industry by common name: "expensive".

Unfortunately in a scenario of protective coating failure it can lead to relatively fast destruction of the gear teeth, although producing small debris - around 20-40µm - for such case only constant gear monitoring system that includes magnetic field monitoring could pick such failure in time... although this require Hall sensors sets on each and every gear, making it very expensive.

It is possible to design an "intelligent" gearbox that will sense any type of failure, and warn crew about it, although I'm not sure anyone would like to pay 3 or more times more for a helicopter than one with "dumb" design.

[FH1100 Pilot]

The thread kind of drifted a bit. SASless originally asked:
Is it "Fate"....a simple roll of the dice or are we assisted by the system letting us down?

It's probably fate.

In reading the accident report of the BV-234 G-BWFC in 1986, it became clear that the manufacturer had been having issues with that forward transmission since the introduction of the civilian model 234 in 1980. (From that we can surmise that Boeing-Vertol was having the same issues with the CH-47, but were less evident to the general public.) They were working on it; they thought they had a solution.

Nevertheless, the failure of G-BWFC's forward transmission happened so quickly and unexpectedly that the crew had no time to analyze and do something about it. A similar fate seems to have befallen the crew of the Puma that just crashed in the North Sea.

No amount of testing and computer-generated data can predict every failure. A manufacturer can run a transmission in a test rig for hours and hours, but can it accurately simulate all of the inflight and harmonic vibrations and loads that the airframe either generates or is subject to? Hardly.

Helicopters keep finding new and/or different ways of killing us. It occurs to me that no matter how "safe" the manufacturers tell us these products are, those of us who fly these wacky machines on a routine basis are still very much test-pilots, much more so than our fixed-wing counterparts.

That is a sobering thought.

[JimL]

FH1100 Pilot,

The purpose of collecting data on flights is not to simulate but to measure; and with that measurement comes the ability to spot emerging anomalies (clusters and trends) and predict when failure is about to occur.

The great thing about neural nets is that they support machine learning and do not need the setting of thresholds; eventually, after establishing what normality is, and without human intervention, they can spot abnormality and therefore the potential anomaly.

Perhaps what some of us are bemoaning, is that we saw a demonstration of at least one of these programs (the GE software running over the Bristow data-store) a couple of years ago. In fact we also saw the same software running over the HOMP data. What this can spot is something like a pilot who constantly approaches faster than the rest of the pilots - not fast enough so that it triggers an alert (because thresholds have to be set high enough to avoid nuisance alerts) but a group of data points that sit in a cluster and so become an abnormal pattern.

There will always be faulty parts (bad material, poor machining, incorrect assembly etc) and they will conspire to break outside the normal inspection pattern or before overhaul. The breaking of such parts will always take the occurrence outside the 'extremely remote'. What collection of data and monitoring does is to allow us to eliminate premature failure by spotting the precursors.

In the same way, HOMP (FDM) permits us to identify behaviour patterns that, when isolated are not themselves dangerous but, when put in the mix with other elements might be the final link in the causal chain of the accident.

What most of us really want, is to take the human out of the heavy process that is post flight data analysis; much better that software systems do that for us and then alert when the abnormal pattern is observed - they are much better at this than we are. Leave the humans to the standard intervention (nuts, bolts, mag-plugs, inspection) - they are really good at that.

The problem I have with the arithmetic of a probability of 'extremely remote' is that it has to encompass the knowledge of continuing airworthiness (when do I target my inspection, when do I do my overhaul); hence the target figure is preserved because the faulty element, when found, is removed from the calculation. For a practical example look to the introduction of the EC155 to Nigeria; there never was going to be an engine failure because no engine sat in the aircraft for more than 200 hrs. It is setting these intervention intervals that is the real skill and the one which permits a very small figure like 'extremely remote' to exist.

In an extremely complex system, we can mitigate the errors made in the establishment of such intervals only by monitoring. It was this very point that was the basis of the quote from the HARP report.

Jim

[loav8r]

Someone had mentioned that titanium is a fragile metal. I’m not a metallurgist by any means but it seems that every Blackhawk crash that I have seen, the rotorhead is wholly intact. By the way, the entire rotorhead is made from titanium. I’m only making an observation, not a conclusion.

[Lt.Fubar]

Titanium alloys may have low strength - lower then Aluminium ones, although only when the % of titanium is higher than 98% - those almost pure Titanium alloys are used only in medicine, as are extremely resistant to corrosion. Not useful for aviation industry though - here are usually alloys with less than 90% Titanium, with the rest being Aluminium, Vanadium, Chromium etc. Those alloys are very tough, light, and both corrosion and heat resistant. The only reason why aircrafts are not made entirely of Titanium alloys (Black Bird and X-15 excluded) is because of high cost of the TiO ore, and manufacturing - as it is not very efficient - milling is very expensive because of high material strength, welding was not possible on the wider scale until 1990s, and casting is problematic because of other materials usage, and time.

I played with few things made of titanium alloys, and those were almost indestructible. Also take a look at the wreckage pictures of F-14s - main central wing spar that is also a main fuel tank and the whole aircraft hang from it - is always intact. I can't imagine that Titanium parts designed with aviation industry safety factors could be broken under normal operation loads, although as we all know it happened few times already, but I don't believe the material selection had anything to do with it.

[AdamFrisch]

Sometimes I think manufacturers are stuck in very old ways of solving problems. Gearboxes being one of them.

It would obviously be fully feasible to design a rotor system without a metal-against-metal-soaked-in-oil grinding gearbox - either with belts or by having a bigger diameter turbine stage and direct drive the rotor.

Maybe manufacturers of big helicopters should take a look at how far the RC heli's rotor systems have improved over the years.

[nodrama]

For a practical example look to the introduction of the EC155 to Nigeria; there never was going to be an engine failure because no engine sat in the aircraft for more than 200 hrs.

Get your point and, except for the poor example,very well put.....

Nigeria is not the only place the EC155 was introduced and it has been more successful (with much longer engine 'sitting' times) in those places. So there was still the probability of an engine failure with that aircraft/ engine type, just not in Nigeria.

[SASless]

What is the technical reason that prevents putting a fluid level sensor in a gearbox sump and connecting it to at least a warning light or contents gauge and warning light?

I have never heard of one on a helicopter transmission ever thus there must be a good reason why would not work.

[This aint Jim Beam]

I think the reason theres no xmsn fluid indicator is the oil is really getting thrown around in there and it may give false readings. If youve ever watched a sight glass in a gearbox with the rotors turning youll know what i mean.
Thats the problem with using a sump as a reservoir
Theres also the problem of blocked jets and oil starved bearings which would not be picked up by a fluid level.

[EN48]

those of us who fly these wacky machines on a routine basis are still very much test-pilots, much more so than our fixed-wing counterparts.

That is a sobering thought.

Sobering, yes ... but also somewhat to be expecetd given that the acft our fixed wing counterparts are flying have several more decades of practical cumulative development experience - about 100 years for FW vs about 60 years for RW (give or take a few). And, a similar but even more exaggerated experience gap in terms of cumulative units produced. The next 40 years will likely see considerable advances in most aspects of helicopter safety and reliability.

[Flyt3est]

I personally feel that given the types of people who frequent our industry, with their engineering know- how, HUMS knowledge, pilots and those with experience and influence in regulatory spheres, the biggest crime here would be to fail to seize the current situation and use it as a catalyst to kick new life into flight safety initiatives.


My biggest fear is that this will not happen, and in 12 months, this whole tragic episode will be forgotten, and we will not have learned and improved to the best of our collective abilities. That would be a failure.

[Brian Abraham]

JimL, your post #27 has that piece of information which has caused some measure of debate. I've got the scars to show for it!!!! :) I know context is everything, but its little wonder people may have erroneous beliefs.
Although minor surface damage to a gear or bearing is unlikely to be instantly catastrophic, and should be detectable, and although gearboxes are tested to run for a period such as 30 minutes with the oil supply failed, the possibility of a major failure within the box is ever present.

[JimL]

Brian,

Perhaps some clarification is necessary here; at the time that this report was produced - i.e November 1984, the members of the HARP (a committee set up by the Airworthiness Requirements Board (ARB)) would have been aware of the contents of the proposed amendment to FAR 29 but would have had no direct interest because Certification in the UK was to an alternative set of requirements - i.e. BCARS. (Without checking, I would make the assumption that this requirement was contained in BCARS. It was also not clear at that time that the UK would eventually be forced to abandon BCARS in favour of FARs (in the form of harmonised JARs).)

Perhaps a more interesting fact is that in the proposed NPRM, the text was slightly different.3-46. By amending Sec. 29.927 by revising paragraphs (c), (d) (introductory text), (d)(2), and by adding new paragraph (f) to read as follows:

Sec. 29.927 Additional tests.
* * * * *

(c) Lubrication system failure. For lubrication systems, the function of which is required for operation of the rotor drive system, the following apply:

(1) Category A. It must be shown by tests that each rotor drive system, where the probable failure of any element could result in the loss of lubricant, is capable of continued operation, although not necessarily without damage, for a period of at least 30 minutes at a torque and rotational speed prescribed by the applicant for continued flight, after indication to the flightcrew of the loss of lubricant.By the time the requirement was finalised as amendment 29-26 in 1988, the rule contained the conditioning text we now see. Normally one can examine the explanatory text of the final rule and establish why the text has been subsequently amended. In the case of the text of 29.927(c), the explanatory text is as follows:Proposal 3-46. The notice proposed several amendments to Sec. 29.927. Paragraph (c) is changed by revising and extending the rotor drive system lubrication failure test requirements for Category A rotorcraft and by clarifying the corresponding test requirements for Category B rotorcraft. Category A aircraft must have significant continued flight capability after a failure in order to optimize eventual landing opportunities. However, indefinite flight following the lubrication system failure is not expected. The changes to the Category B rotorcraft drive system lubrication failure test requirements are largely for clarification and are not substantive.

A commenter notes that paragraph (c), as proposed, could be interpreted to preclude credit for auxiliary lubrication systems or to require consideration of lubricant failures to self-lubricated bearings. This was not intended, and the wording of paragraphs (c)(1) and (c)(2) has been revised to eliminate this possible ambiguity.
The 'law of unintended consequences' in action?

By putting in this clarification, there is no intent to reopen the discussion on the 30 minute run-dry capability (and the reason why it was not underlined in my previous post) only to clarify the extract from the HARP report.

Jim

[K48]

Sultan wrote:

The issue is that those who claim vibration monitoring is superior to, or can eliminate, debris monitoring border on delusional. It is only a great addition to the tool kit. For vibration to be generated you have to have an actionable level of damage. Problem with cracks on high stress gears are that they may propogate to a critical level before any loss of stiffness is noticed in a daily check of HUMS. It is basically good, good and bang.

To detect a crack the only concept I can see as practicable would be a delta resistance detection system of some kind. (thermal or magnetic resonance imaging being cited as too costly above). Strain and stress guages have been around for a long time. In simple terms I envisage a system that runs a current through the gearbox from entry to exit and gives a resistance level readout which would have a known behavioural pattern for load/torque/temp settings. A crack or significant fatigue (chip loss)would hopefully give a warning fractionally in advance of chip /temp/pressure warning systems.

Does anyone know of such a system?

The added benefit of such a system is that it may be applied in non magnetic material systems.. as long as they conduct a current. (Ceramics can be made conductive or not as required.. I believe.

[ericferret]

You can't design out the MRGB but maybe you can reduce the effects of a failure.

Look at how the gearbox works in the 500 and I believe the Apache.

Most gearboxes support the main rotor and have flying controls bolted to them. A major failure can lead to loss of rotor and/or control.

In the 500 design the gearbox is underslung and in theory could fall out completely with no effect on the flying controls and the abilty of the aircraft to autorotate.

For large aircraft this would be a radical redesign, resulting in either less internal space or a taller aircraft. Not impossible though just different.

[outhouse]

As I recall the engineer has a number of tools at his disposal, from the basic chip detector and the Mk one eye ball, periodic micro analyse of oil (every 50 hours as I remember) analysis of any partials picked up by the mag plug, and filter. Trending of the results of the 50 hr regular check. On the 61 this was extremely important as the white metal bearings started to shed indicating a potential problem.
This simple system of periodic sampling and trend analyses gave a good indication of gearbox condition. Now add the hums and vibration analyses system performance and add these trending results to the ones above and hopefully a good monitoring system both by engineering chaps and computer and what do we end up with??
Seemingly we go back to those old days of sudden and unexplained catastrophic failure of a major link in the drive train. The investigation will provide some answers and will recommend a series of quick responses. The authorities will huff and puff and seemingly repeat past history, by actually doing very little, and in another 20 years we will be adding to a similar thread in pp with nice to do stuff and how it could all have been avoided ‘’IF’.
Waiting to be pilloried, but someone had to state the obvious.

Outhouse.
:ugh:

[Oilandgasman]

Thanks for those two posts Ericferret and Outhouse as they sum up the deficiencies in this matter. Neither of you will be pilloried by me. The mechanical outcome of the transmission failure in this case is almost identical to the outcome on Chinook FC. The AAIB have already reported on that and made recommendations to stop a recurrence. But a loss of rotor head has occurred again with the loss of all on board. There are at least too problems here, the failure of the gearbox and the resultant failure of the rotor head and control of the A/C. The latter should not follow the failure of the former and must be engineered out. I appreciate that in this case it was inevitable. (It must have been considered at the design stage and as nothing was done must have been ignored) Yes pilots must know what is going on in their gearboxes, yes engineers must have the tools to predict failure before it occurs and yes pilots should be able to autorotate following a transmission failure. A lot of work will need to follow and a lot of money spent to remove this risk from helicopter travel, we now know what happened, we now need to get after it. What caused it to happen I leave to the AAIB, but their findings and recommendations must eliminate a recurrence. ( and please do not use the word "ensure" in your recommendations as it is not sufficiently specific, please tell us exactly what needs to be done.) Radical re-design? Yes if that is what it takes. Would this known fatal failure mechanism be allowed to persist in any any other industry? No!

[spinwing]

Mmmm ...

Eric .....

You mention the 500 ... it too had some gearbox issues in the past ... vaguely remember an AD which required removing the Xmsn oil level sight glass and checking all the main gear attach bolts for loss of torque .... AND I found one (bolt) which had failed and was hanging by its lockwire!

Stopped taking transmissions for granted after that ( I think about 1977/8 IIRC).

:}

[Two's in]

Sobering, yes ... but also somewhat to be expecetd given that the acft our fixed wing counterparts are flying have several more decades of practical cumulative development experience - about 100 years for FW vs about 60 years for RW (give or take a few). And, a similar but even more exaggerated experience gap in terms of cumulative units produced. The next 40 years will likely see considerable advances in most aspects of helicopter safety and reliability.

I would argue that rotary versus fixed wing safety is not primarily a function of experience, but simply a function of the fact that a fixed wing aircraft has primary lift generation surfaces and controls that are attached firmly to the fuselage by an impressive array of bolts, pins and rivets that don't move. The helicopter on the other hand has primary lift generation surfaces that revolve at 400rpm or so, generate massive g forces, and are constantly trying to depart the rotor mast, despite opposing force of the blade pins.

It takes little imagination to see that any design, engineering, or mechanical failure in that environment is followed by virtually instant and catastrophic loss of the aircraft. In contrast high speed rotable failures in fixed wing aircraft occur at the powerplants, so although that brings its own range of problems, containment technology and FOD resistance usually allow for a more favorable outcome when comparing the loss of a engine fan blade with a main rotor blade.

[AAKEE]

Just trying to think a bit.

If a bearing fails, or a gear starts to fail, making bigger parts come loose thats to big to hit the magnetic plug/chip detection system (like parts of cogs or so)...if these parts jam between the gears...could that cause the epicyclic gear to brake ? ..and/or possibly make the MGB suffer a sudden decrease in speed and/or sudden stop ?

My thought is that it is not sure that the origin of the fault was in the epicyclic ? If so, there shouldnt have helped if HUMS could detect cracks in the epicyclic gear ?

[212man]

A pilot should be checking his deferred defect log and if any MGB HUMS related defects are noted he should be demanding to see the trend graphs for his own piece of mind (and question them). No matter how good an engineers pre flight or pilots walk around, We are not able to detect the state of play inside the most crucial part of the machine through normal visual inspection. Only HUMS can do so and if there is a rising trend above a threshold that raises concern then a pilot should be quite within his rights to refuse that aircraft.

1st, why should a HUMS trend be in the Deferred Defects section of the ATL? The only items that should be in there are MEL deferrable items.

2nd, if the HUMS events (if there are any) are within the thresholds, and are being monitored in accordance with the AMM/AMP, and therefore manufacturer's?/Maintenance Organisation's approved protocols, why would you expect the pilot to overide this? Would you apply the same logic to Track and Balance figures, engine HSS vibrations, SOAP samples?

[ericferret]

SPINWING

The 500 main box has indeed had it's share of internal woes over the years.
So even better is a design that (in theory!!!) allows the gearbox to be a redundant item still giving you autorotation under control.

[dieseldo]

AGARICUS BISPORUS I never worked for BAH or BIH at Aberdeen. However I was lead to believe that your version of events is correct.It could be just hearsay but the engineer who told me worked on the Chinnooks for several years.
I reread the AIB report recently and was surprised that it conflicted with what I believed was the story behind the gearbox mod.

[dieseldo]

AAKEE A Hiller 12E suffered the type of failure you discuss in the 70's. Operated by Bond Helicopters in a previous incarnation I believe. It can be found on the AAIB website in the archive section. Sorry can't remember the year.

[DOUBLE BOGEY]

A pilot should be checking his deferred defect log and if any MGB HUMS related defects are noted he should be demanding to see the trend graphs for his own piece of mind (and question them). No matter how good an engineers pre flight or pilots walk around, We are not able to detect the state of play inside the most crucial part of the machine through normal visual inspection. Only HUMS can do so and if there is a rising trend above a threshold that raises concern then a pilot should be quite within his rights to refuse that aircraft.


This is an interesting issue. At present HUMS data is used only to support the Continued Airworthiness of the helicopter and not employed operationally in the sense that the MEL is at the moment.

The implications of this is that the Flight Crew do not at present have the ability to manage the totalily of risk when applied to the actual flight requirement, as you would do say for an deferred defect that might affect the type of flight you need to carry out.

Whether this is viable/sensible for the future is a far wider debate.

What may be more appropriate (and I have no expertise at all in this area) is whether sufficient HUMs data was being picked up immediatley prior to the failure that, if it had been made avialable to the crew, they could have taken immediate intervention to prevent the full catastrophic conclusion that we have just witnessed.

Should/can certain types/levels of HUMS data be provided to the crew in the form of a warning/alert system that, allied to a specified in-flight abnormal/emergency procedure, would prevent such an incident reaching its ultimate conclusion.

Are there any experts in the HUMS field out there who think such a system is viable and in particular, is the databank of alerts sufficiently developed to provide appropriate thresholds for warning and intervention?

DB

[DOUBLE BOGEY]

ERIC FERRET * OILANDGASMAN

Having just read your posts again I think you both have made a very very, valid points. The design concept of makiing the MGB "multitasking" certainly leaves the helicopter utterly susceptable to catastrophic failure and one could argue, that in doing all these jobs (load bearing in the air and on the ground) it is probably subjected to a whole bunch of unecessary stresses.

It is starnge that this concept has remained a feature for many helicopters yet the safety advantages, as you guys have mentioned, allied to the reduction in downtime (as the MGB could be removed without having to remove head and blades) would make you think it should have been thought about before.

Any engineering wizzards out there who could explain this anomaly???

[JimL]

It is probably in bad form to quote oneself however:As Flyt3est has indicated, the HHMAG (a group which contained the world-wide great and good of the HUMS community - what a pity the advent of EASA caused its demise) discussed cockpit indication at length.

Because the issue was so complex and in view of the lack (at that time) of the required level of reliability, and the simplicity of the systems (events only - and even those provided on a heuristic basis) - a working group was tasked to consider it in detail. This group made a finding that, for the time being, cockpit indications would only complicate the life of the pilot - not simplify it. It also reiterated that the aim of HUMS was to spot developing faults, clusters and trends in time for them to be addressed on the ground and before they could impact upon a flight.

In common with others, I still have my doubts that cockpit indications provide the answer. Perhaps the problem is believing a primary indication that appears irrational and cannot be confirmed by some other means. It could be that, in these days of digital systems and software control, more effort is required in the provision of algorithms that provide (and indicate) secondary confirmation based upon alternative diagnosis paths.
Jim

[unstable load]

Double Bogey, Eric Ferret, et al

Those are interesting points that are raised, however we need to bear in mind that no matter how thorough and detailed a manufacturing, inspection and maintainance process is there is still a REMOTE possibility that a catastrophic failure of this nature could have happened suddenly without a noticeable timeous trend on the HUMS or without making chips/metal.

These are machines that at the end of the day are subject to sometimes tragic random failures. I am not trying to start a fight here, merely making a point of view that IMO needs to be remembered.

As for the alternate load bearing structure of the Apache rotor system, bear in mind that it may not be feasable from a design point of view if you want to carry passengers to do it like that on the modern civil helicopter although it would make my job a lot easier for swapping out an MGB if it were.

[DOUBLE BOGEY]

JimL, I take the point that there are/have been some real "brains" working the issue of HUMS indications. Having said that, a starting point for pumas would be:

COCKPIT THRESHOLD ALERT for all MGB, TGB, IGB and ECU HOT-END bearing pickups. These being the weak points in the design concept.

I would consider being oblidged to land at the nearest safe site a considerable advantage rather than a complication!!!

Like I said before (and on the other thread) the jury is effectively still out on this one and until we know the EXACT circumstance of the Operator's response to the data they may or may not have had available prior releasing the aircraft to service we cannot know if the the requirement for an instant indication is valid let alone viable.

However, if it is the case that the HUMs only began squeaking during the fatal flight then there surley must be an argument for letting the crew know what is going on in the MGB during flight as the implications in this case would be that rapid and catastrophic deterioration can occur within the space of one flight. I have always hoped that the HUMS would have speaked long before the failure propogates to its full conclusion.

DB

[outhouse]

As an old fart and one who has like :(the posters in this thread enjoyed the discussions re the new and developing technology available to engineers in the quest of safety in the critical areas mainly the rotor drive system. I really don’t want to continue to be pessimistic but do understand the limitations to any fault monitoring system. As has been shown HUMS is a basic trending system, it will generate alerts; identify trending and progressive deterioration of bearings, tooth mesh, gear condition and changing general harmonics within selective parts of the drive train. The effectiveness of the system is down to the system design, computer programme and the ground operator monitoring the system.
Question, can this system in its present mode of operation pick out and alert the pilot/engineer to a possible instant failure, guess the answer is NO. (Based on the limited info available)
Question 2, is the pilot in the loop regarding the data and results of the HUMS system while in flight, answer NO.
Question 3, should he be? How could this be achieved and how could this be included in his ECL action response.
O

[Bertie Thruster]

As for the alternate load bearing structure of the Apache rotor system, bear in mind that it may not be feasable from a design point of view if you want to carry passengers to do it like that on the modern civil helicopter although it would make my job a lot easier for swapping out an MGB if it were.

If it can be done on a small civil machine (MD 902) why can't it be feasible, from a design point of view, for a larger helicopter?

[EN48]

It takes little imagination to see that any design, engineering, or mechanical failure in that environment is followed by virtually instant and catastrophic loss of the aircraft.

My point was not that helicopters will become as simple (and reliable) over time as airplanes, but rather that cumulative experience is a significant factor in identifying how to improve designs, and that when there is as much cumulative experience with helicopters as with airplanes (40 more years plus or minus) , we should expect to see great improvements in safety and reliabilitry vs where we are today.

[Geoffersincornwall]

I wouldn't be surprised to hear that the designers at EC and Sikorsky and maybe elsewhere are rising to the challenge of creating a gearbox design that will deliver the kind of failure modes that do not result in catastrophic failure. Multiple load paths, controls protected (probably electrically actuated), and a guarantee of autorotation even if the whole lot goes pop. I'm sure it can be done it requires some motivation and 25 years after HARP I think it's time to do something that will deliver that paradigm shift needed to move helicopter airworthiness another rung up the ladder.

G

[Mark Nine]

Geoff,
Most of the manufacturers could give you that kind of gearbox next week. Only trouble is, it would weigh about 5 tonne and be the size of a Puma.

[Geoffersincornwall]

Hyperbole aside if such a gearbox weighed 100kg more (1 passenger) wouldn't the trade-off be worth it.

Many years ago I attended a course at Bristol University on helicopter design. One of the lecturers was the Westland head of transmission design who had just produced the Lynx 'conformal' gear system. I remember him saying that If we designed a helicopter gearbox using the design principles of a car gearbox then the helicopter would be too heavy to get airborne. That wasn't hyperbole because in those days car gearboxes were made of cast iron and hugely overengineered to achieve the rigidity needed to tolerate the poor accuracy of the machining used in mass production. These days a typical car gearbox is computer designed and the casting process uses minimal material and is probably made using a weight-saving alloy. Modern mass production also produces components made to very fine tolerances.

My point is that if the car industry can step up to the plate I'm sure the helicopter industry can too. My feeling is the reluctance to go down a new path that will take the current designers down a new path given that 30 years with the same basic design gives them a thorough understanding of the strengths and weaknesses of it. Why take the risk (and spend huge amounts of time and money) going down another pathway when there is no regulatory or customer pressure to change.

The progress made since HARP is unlikely to have been realised if manufacturers had been left to do their thing with no pressure from those that use and regulate their products.

Can a transmission designer add to the debate please.

G.

[unstable load]

If it can be done on a small civil machine (MD 902) why can't it be feasible, from a design point of view, for a larger helicopter?


In short, I don't know.
I've not wrked on any of the MD products so cannot say if or why not. Off hand though, it would most likely do with the weight of the structure versus the AUW of the final machine versus expected payload/range/cost with cost probably being the decider as always.

[littlewing]

Bless the souls that were lost, may the engineers figure out how to keep this from happening again and may the oil companies give them the cash to do it.

[JimL]

WAH,

We saw a presentation by Brian Larder of the results from the (then Smith's - but now GE) system at the CAA Helicopter Research Management Committee in May 2007 - following the six-month trial by Smiths and Bristow. Following that I cannot recollect seeing anything except a mention in the CAA Research Report:Publish the final report on the development and demonstration of enhanced Health & Usage Monitoring System (HUMS) Vibration Health Monitoring (VHM) data analysis

The initial six-month in-service trial of the enhanced HUMS VHM data analysis with a major UK offshore operator concluded in November 2006. The trial was successful and highlighted a number of defects which the current analysis system did not identify, including instrumentation defects. Further enhancements to the system are being progressed and sponsored jointly by the FAA, CAA and UK industry. A further six month in-service trial, scheduled to complete by June 2008, will be undertaken once the key sections of this work have been completed.

Publication of the report has been deferred in order to incorporate the results of additional work identified and the six month trial extension.
Jim

[500e]

Interesting read and time line WAH.
The HUMS graph appears to show the failure time line especially from 193 to 40, was this ignored \ filed due to lack of understanding I presume the times were inserted after failure of item, and with hind sight.
In 2001? was so little known that it did not raise questions in engendering\ HUMS department?.
As a pilot and engineer it would have made me ask the questions regarding the leap in readings at the 193+ point even if I did not understand the graph enough to pick out the previous information

[Agaricus bisporus]

Dieso, thanks for that snippet. My understanding of the event also came from the engineers - and at a pretty umimpeachable level too; I have little doubt that the AAIB report was, well, shall we say rather skewed?

Even so, infinitely more factual than the shamefully dishonest coverage by the repulsive local rag, the Aberdeen Press and Journal.

[HeliComparator]

WAH

Ref your post #64 I think you are on a different HUMS planet from us lot on the N Sea. There is no way that the tech log would have a deferred defect entry saying "dodgy HUMS M6* reading on gearwheel x, keeping an eye on it...". It might say that an element of the HUMS system itself was U/S, with a total of 25 flight hrs being allowed until the problem is fixed.

So your comments about pilots needing to be concerned about looking at HUMS trends is totally wide of the mark.

Regarding the Norne accident, I believe that HUMS was not mandatory in Norway at that time, and perhaps it was not taken as seriously as it should have been?

Its worth bearing in mind that engineering actions in the event of an adverse HUMS trend depends on the equipment manufacturer. For example, since the IHUMS system was invented by Bristow, when there is a HUMS issue on our 332L, S76A etc, we (our HUMS type engineer) will decide what to do about it. However if its EuroARMS or M'ARMS on an L2 or 225, as soon as an amber or red traffic light is produced by the groundstation, its straight on the phone to Eurocopter and they decide what to do about it.

But at no point in that loop is the pilot involved, nor should he be since he generally doesn't understand the issues and is in no position to decide whether the aircraft is serviceable or not.

Can I also point out that whilst you said
If HUMS was within the thresholds, then there's obviously nothing wrong
that presupposes that the thresholds were correctly set. A very brave assumption since the thresholds are generally set by guesswork, modified only in the light of excessive false alarms or failure to detect a failure. Its for this reason that the GE work is so important.

HC

[mtoroshanga]

I was not directly involved but I seem to remember that a system was being developed in the days before the Airlog intrusion that involved taking a downlog from the DAPU and transmitting it back to base on the flight positioning system to be monitored by the IHUMS engineer. This would occur about every ten minutes. There was also an inflight SOAP facility involving a laser snining through a transparent tube. I feel that this would be a more practical system than the pilot being confronted with a diluge of information.

[500e]

"But at no point in that loop is the pilot involved, nor should he be since he generally doesn't understand the issues and is in no position to decide whether the aircraft is serviceable or not."

Why not!! he is the person who is responsible for the craft, who says he does not or should not be trained to understand at least the basic issues?.
Looking at the graph it would appear to me that there was a significant change in levels at 193, if I was confronted regularly with the typical daily graph & saw this steep increase I would want an explanation & want it before I flew.
It should be possible to produce a daily trend graph that shows the variation from the norm then changes such as that would stand out as an anomaly to be examined

[HeliComparator]

mtoroshanga

I was involved, and there was never any plan nor actuality of sending DAPU/ HUMS data by data link back to base. There was some trialling done of a full-flow spectrographic oil analysis system, but it was unsucessful. There was also a trial of an exhaust gas debris monitor (detecting metal swarf coming out of the engine exhaust) but again, unsucessful and in general the emphasis was on the health of the one gearbox, the two engines being less important.

500e - sorry, wrong planet. He is not the person that is responsible for the airworthiness of the aircraft - he is just responsible for driving it safely.

HC

[HeliComparator]

OK I had better change my point to WAH to being "I think you are on a different HUMS planet from Bristow".

I suspect that the fundamental difference might be that I believe CHC (?) has a paper techlog, whereas Bristow has a computer-based techlog that records almost everything except for what goes into the paper-based part of it which represents the exchange between the pilots and engineering. What we pilots call the techlog contains only such things as daily/turnround inspections, pilot signing for the aircraft, any pilot reported defects, any deferred defects, configuration changes, groundruns/airtests required etc. It has no information about maintenance work carried out on the aircraft except a number of hours to run to the next inspection.

But I am puzzled how you can allow an aircraft to fly with a HUMS parameter exceeding a threshold? As far as I know we never do that, unless its decided that the problem is a sensor rather than the transmission itself. I certainly don't see how you can have a deferred defect that calls into question the airworthiness of the transmission. That is definitely not in the MMEL!

Are there any pilots out there from WAH and Mitchaa's company that would care to comment on whether they see and are in a position to interpret HUMS trend data and over-ride an engineering decision?

(WAH - posts crossed but never mind!)

HC

[widgeon]

http://www.aibn.no/items/247/144/6203097590/LN_OPG_eng.pdf

here is link to complete report on ln-opg.

[maxwelg2]

HUMS, GE neural model, existing flight instrumentation, new gearbox design. We can always improve from where we are today...

As an engineer I am used to embracing new technology. Too much information or too little knowledge is dangerous, but with proper system engineering and end-user training I can see how SKY and EC should be able to improve on their existing condition monitoring systems.

What what I've gleaned on the various threads the technology is there, it just needs to be firmed up to better provide the engineer and ultimately the end-user (flight commander) pre-flight check data to decide if the equipment is serviceable and as safe as reasonably practicable.

Is it not better to err on the side of caution e.g. if HUMS thresholds are exceeded because they are not "tuned" then surely non-acceptance of helos by engineers and pilots will up the ante to get this issue resolved. If this methodology is impractical then the GE prototype system should be further developed to enhance the current available HUMS threshold setting methodology.

Who pays for this way forward? Might I suggest a joint venture between the operators, helo manufacturers and the oil companies, who generally end up owning the helos and leasing them out to the operators.

It also sounds like various operators are applying HUMS exceedences differently. Why is there no synergy here, surely the system design is universal, therefore the executive actions should also be? Collectively there must be enough collated historical HUMS data globally to allow the manufacturers to issue better guidelines and thresholds, even to the level of each helo's custom thresholds, or is the onus squarely on the operator's HUMS support team?

It seems crazy that SKY is only now developing an updated MGB to satisfy their Canadian military contract obligations versus having this capability for the S-92a units. Prevention is better than cure, especially when their 1 x10^9 extremely remote classification is gone....

[212man]

We are obviously following different procedure paths, but yes you are correct we fly our aircraft with exceedances above thresholds every day otherwise they would never get off the ground. I would be mightily surprised if you did not either as exceedances are regular occurences.

Are you talking about 'real exceedences' or spurious spikes that indicate a faulty accelerometer? If the latter, then that is simply a normal use of the MEL to allow flight with a defective HUMS system (10 hrs or 25 hrs depending on whether close monitoring is in place.) If you are talking about HUMS exceedences that are considered real (following a trend for example), then I'm somewhat astonished - to put it mildly! Or, are you talking about low level thresholds, in which case you are simply close monitoring the HUMS system, as it is designed to be.

Either way, I see no reason why the pilot should get involved in the process, and don't believe the logic bears much scrutiny. The same logic would apply to wear tolerances on mechanical componants, leak rates, crack propogation etc etc.

[JimL]

WAH et al,

Here is the latest position on the GE system:Current work comprises an in-service demonstration of an Artificial Intelligence (AI) based anomaly detection and diagnostic system to enhance the performance of current HUMS. A demonstration system covering all 35 shafts in the Super Puma main rotor, accessory, intermediate and tail rotor gearboxes has been developed and tested using all available IHUMS data up to March 2006. The first six-month in-service trial of the system at Bristow Helicopters was completed in November 2006 and was very successful. Further enhancements to the system were identified, developed and implemented in the trials system, largely under FAA funding, and a further six-month in-service trial undertaken which was completed in June 2008. The results of this trial demonstrated further improvement. The final two research tasks covering data mining and data reasoning have been completed and are to be reported by end March. A final summary report is to be produced for publication in the public domain in the summer. The last progress meeting with GE Aviation took place on 15 January 2009; the next is scheduled for 29 April 2009.

Presentations on this work were given by GE Aviation at the 10 June 2008 Oil & Gas UK ASTG Symposium and at the 11/12 June 2008 RAeS Maritime Operations of Rotorcraft conference. A paper proposing full implementation of the research was presented to the Oil & Gas UK Board on 15 October 2008 and approved. GE Aviation will be progressing implementation via OEMs for aircraft with OEM supplied HUMS, and this would be assisted if a clear indication of ‘customer’ demand were available. GE Aviation will deal with helicopter operators direct for ‘legacy’ HUMS. GE Aviation is proposing to hold a seminar in Aberdeen to help launch implementation.
The GE Aviation meeting referred to in the briefing is to be held at Aberdeen on the 29th April - the AAIB have been invited.

212man, I totally agree with your statements on this - these are not matter for the casual consideration of pilots; HUMS are complex systems - too complex in fact, the setting of the threshholds is more of an art than science and awaits the development of appropriate systems for trend and cluster modelling.

If the aircraft is declared servicable in the tech-log (or substitute) and satisfies the contraints of the MEL it is fit to go - only issues discovered on the preflight walk-around and during start-up (and, of course, operational considerations) are within the remit of the pilot.

Jim

[500e]

The graph shown early in the thread shows a large upward trend, are people saying that the manufacturers said this was within tolerances?? or the HUMMS was faulty ??.
If we say the indication was correct (as it turned out) why was it not acted upon or if a fault rectified as a matter of urgency.
Why is it not OK for the person strapping themselves in to ask questions about ANY method that helps make flying safer, I do not think or expect all pilots to understand in depth the system but there should be a custom graph showing expected readings and if the days \ flights data shows a marked deviation engineering should be there to explain.
HeliComparator why bother with walk round check then, you says can I fly it safely, if I am current this should be taken as read, as don't want to die so my walk round will as thorough as I can make it & any further information that is available will be factored into the go\ no go equation.
maxwelg2 The comments posted by you seem to be along the same lines as I think, everyone should be in the loop if you remove 1 link the whole chain is broken.

[SASless]

The British Walkaround Inspection does not an American Preflight Inspection make!

I was looked at with great disdain when I began to open cowlings and the like and was told "that just isn't the done thing, mate!"

At some operations there was no need to even do the walkaround ( of either type) but at some it was amazing what could be found lurking within the cowlings.

On one occasion I pulled a full bed sheet out of a hidey hole underneath the tail rotor drive shaft on a Bell 212. The reaction to my demanding a proper inspection of the aircraft by the engineering staff was interesting to say the least....and not because it was based upon a lot of embarrassment or professional concern.

A second interesting event was looking over the aircraft after a tail rotor change and airtest and a signed release for flight (as indicated by the tech log) to see one of the tail rotor pitch change links dangling loose from one end.

[HeliComparator]

500e

I bother to walkround because its a required part of the maintenance schedule/pilots preflight check as detailed in the Flight Manual.

I can't help thinking that you have never actually looked at a HUMS groundstation. On the EC225 M'ARMS system for example, it monitors over 30 items in the transmission. For each of these items, there are on average about 10 different parameters to be graphed. Are you saying that before you go flying you want to check each of those 300 graphs? You had better come in early for your 07:00 departure then!

On the other hand, are you saying that you will allow the experienced licenced engineers trained in HUMS to trawl through the data, find a dodgy looking graph that in fact they asess as OK, show it to you so that you can say

"I don't care that I am not a licenced engineer, I haven't done the HUMS course, I have no experience of looking at HUMS graphs, but nevertheless I disagree with your assessment and refuse to fly the aircraft"?

If so you would have a very unhappy existance if you worked for a mainstream operator. This is my last response on this matter.

HC

[maxwelg2]

"I don't care that I am not a licenced engineer, I haven't done the HUMS course, I have no experience of looking at HUMS graphs, but nevertheless I disagree with your assessment and refuse to fly the aircraft"?

HC, I believe that the key point is that HUMS interpretations should be made "user-friendly" either via either the new GE software or yet another new diagnostic tool that's not even invented yet. Time is of the essence, I hope your upcoming symposium moves this issue forward.

Using the available tools to maximum advantage is the only way to move this technology forward.


Everybody from the HUMS engineer to the pilot to the PAX should have confidence that the condition monitoring and preventative maintenance schedules are as good as they can be within our current technological (and budgetary) limits. If this means being able to present the data in laymans terms/with minimal additional training then so be it.

Post-diagnosis is key to preventing future occurrences. My current perception is that HUMS is not user-friendly enough to be confidently used and does not have a universal application structure, hence differing interpretations of the data sets. Until that is resolved I'd rather trust the invasive inspections via boroscope and increased inspection frequency on critical items as/where required.

[HeliComparator]

maxwelg2

I quite agree on that one - as I said in an earlier post, we invented IHUMS in 1990 and it was a great step forward, but since then there has been no real progress (and in some ways, some of the newer OEM HUMS systems are not as good as IHUMS). It still behaves a bit like a research project with "black art" and "interpretation" being key concepts. It needs a complete rethink in terms of the usability to bring it into mainstream, but of course that takes money, which is never available (despite oil at $140) until after some people have died.

HC

[JimL]

HC,

A little bit harsh - particularly as Bristow has been part of the ongoing improvement project with GE. It proceeds slowly (more slowly than we would prefer) but that is a function, not of money, but of the scarcity of personnel to 'manage' and 'work' on these research and development systems. How much return do you think that GE (or others) could make on the sale of any such system?

Remember the dream with this particular development system, that it could sit as an anomaly detection tool above any HUMS (or any other data rich) system and provide a uniform interface.

Before this accident, there was no imperative to speed up development in any area. Let us see what the AAIB/EC/GE analysis of the HUMS data provides in this case - it might result in a clear focus that will provide an acceleration and concentration of effort. The recommendations might well be in the HF area but, until we see the report, we cannot possibly know.

maxwelg2,

It's OK to make these statements but you need to understand that there are dangers associated with physical intervention - you need to remember what brought you into this forum in the first place. There is need for interpretation in all but the simplest systems; we need all the tools in the toolkit and all appropriately utilised (as the Sultan previously indicated in his remarks about mag-plugs) but we also need qualified personnel.

Some years ago (at the last HHMAG meeting attended), the FAA indicated that they intended to spend several millions on HUMS implementation; I for one have not yet seen the fruit of that investment. Has anyone else?

Jim

[ShyTorque]

I would ask why there aren't enough people with the knowledge to utilise HUMS to its full advantage.

Perhaps the government could do more to encourage young people to take up engineering. They are talking about apprenticeships - the aviation industry could gain a lot from an increase in the numbers of properly trained and sufficiently paid specialists.

Perhaps manadatory action is required.

[500e]

Thank you all for your replys to my post , I feel that there are 2 ways forward the single track, " we know what is best" or the one "we are all in this together", some take one route, others try and look at the overall picture.
HC you are correct I have never seen a HUMS ground station or a M'ARMS station, not being a professional helicopter pilot this thread is only of academic interest to me, other than hoping that discussion will push forward safety, but I have looked at a fair amount of telemetry information from in car systems, still don't understand all of it, would totally agree it is a specialised field, but big trends are noticeable, as in 192. why was it not acted upon, I would have walked away if someone could not have given me a very sound reason for jump in reading.
If there is a dodgy graph, I would like to have it explained to my satisfaction.
maxwelg2
We are thinking along the same lines, I feel that physical intervention produces its own dangers, so the best of both worlds is the way forward.
JimL
Thanks for your offer will PM you, don't know how much will sink in but will try to get my head around as much as possible.
I appreciate the 5 tonnes and over, presumably due to cost , complexity & commercial pressure, perhaps with further work and a larger user base, the cost should come down allowing lighter helicopters to utilise maybe a cut down version of the technology,

[SASless]

JimL,

Some years ago (at the last HHMAG meeting attended), the FAA indicated that they intended to spend several millions on HUMS implementation; I for one have not yet seen the fruit of that investment. Has anyone else?


As you well know.....promises from government are not exactly something would care to bet the house on especially coming from the FAA whose concept of urgency comes closer to being a Slug Race than a Hare Chase.

Unless they put a time certain in the promise and then followed it up by hundreds of PR fete's.....then expecting the usual delays and over runs.....don't ever look forward to seeing something from that bunch.

Their bureaucratic chain is advanced of the Italian style and they had an 800 year head start on the FAA.

[HeliComparator]

Jim

Your point
Before this accident, there was no imperative to speed up development in any area
sums up my point really well - but do we really have to rely on people dying to trigger a move to improve? We all knew that HUMS had not moved on from its infancy - Smiths/GE at least tried to do something but not many were listening.

You are right that its perhaps not primarily a function of money, its probably primarily a function of interest from the industry. Someone (a body or an individual) has to "own" a project like that and push it along, otherwise it stagnates with or without money.

HC

[JimL]

HC,

As with all of these systems; they work better when they are someone's hobby-horse - as I have said to you many times.You are right that its perhaps not primarily a function of money, its probably primarily a function of interest from the industry. Someone (a body or an individual) has to "own" a project like that and push it along, otherwise it stagnates with or without money.As I said in an earlier post, the body that used to provide the focus was the Helicopter HUMS Management Group (HHMAG); when EASA became the legally competent body for airworthiness, the CAA could not progress the issue any more, and no longer had a budget for arranging, facilitating and hosting the meetings. Virtually the last act of the HHMAG was an attempt to ensure that the UK/Norway Requirement was adopted by EASA.

EASA neither adopted the group (the excuse was that their methodology did not extend to standing committees; hence the disbandment, as well, of all JAA committees - before taking advantage of their knowledge for the production of OPS Regs), nor did they take over the role of continuing to require HUMS. As I also indicated earlier (without any comment on PPRune), the EASA Ops Proposal does not contain a requirement for HUMS - which puts it out of compliance with an ICAO Recommended Practice. Without a State invoking Article 14 of the Basic Requirements, HUMS will cease to be a requirement at the time that EASA OPS comes into force.

Progress indeed!

Jim

[shortfinals]

Super Puma crash: Air Accidents Investigation Branch issues another compulsory inspection notice (http://www.flightglobal.com/articles/2009/04/17/325312/super-puma-crash-air-accidents-investigation-branch-issues-another-compulsory-inspection.html)

Whether it's improvement to HUMS, iHUMS, or to the system for interpreting the data or whatever, something has to happen.

[Oilandgasman]

EASA Airworthiness Directives Publishing Tool (http://ad.easa.europa.eu/ad/2009-0087-E)
EAD issues by EASA

[shortfinals]

Oilandgasman, that's the original.

The new release, so far published only by the AAIB (soon EASA I expect) is not just about mag plugs, it's is about taking the epicyclic of every Super Puma apart and looking for something we don't know about.

[Dibs 22]

Hello. A few of my thoughts on HUMS for my first post.

Health and Usage Monitoring Systems are advisory, not executive. I can’t really see and alternative whilst HUM is an inexact science; and it is still far from being anything more than that. It certainly helps though, and I think most of those involved in RW (and indeed FW) engineering would agree that it currently makes flying safer, and in time will make it much safer. And cheaper – it’s on the path to CBM.

Meanwhile, it is immature technology and there are problems. Rotorcraft are hostile environments for delicate diagnostic systems, and HUMS are rarely 100% serviceable. It has to be working and someone has to look at, and understand, what it is saying for it to give protection. That person has to recognise the difference between ‘no exceedences’ and an unserviceable system. They should know that alert levels are statistical probabilities, not engineering/operating limits (unless actual fault data exists – the same fault shouldn’t be missed twice). They have to understand that alert levels will be different on similar aircraft with different HUMS and that the confidence levels in those alerts will vary, as will the criticality attached to the many cautions that their HUMS will provide.

Developments (like those at Chandlers Ford) in data mining, and possibly neural networks, to provide more accurate fault indications are particularly exciting and, from what I have seen, are very promising (google ProDAPS). However, I still think we are a few years away from allowing a vibration based diagnostic computer system to ground our aircraft without recourse to a well trained and HUMS educated engineering human brain.

[Oilandgasman]

Short finals, yes you are of course correct. Have rechecked EASA and there is nothing new there at the moment. But the inspection work you have mentioned is being reported in the media with a 7 day time limit. I wish you all well with this task.

[coalface]

shortfinals said The new release, so far published only by the AAIB (soon EASA I expect) is not just about mag plugs, it's is about taking the epicyclic off every Super Puma apart and looking for something we don't know about.

Is this necessarily the case or could it be a boroscope inspection?

[JimL]

Athough we have already seen a drawing of the MGB, this URL brings up a presentation which has (on slide 20) the drive train of the AS 332 with the superimposed EUROHUMS sensors indicated.

http://www.luftfartstilsynet.no/multimedia/archive/00002/16_03_06_Smiths_Aeros_2084a.ppt

Jim

[nodrama]

Air Accidents Investigation: Initial Report 2 - Super Puma accident (http://www.aaib.gov.uk/latest_news/initial_report_2___super_puma_accident.cfm)

[Special 25]

Posted on other thread relating to the Bond accident. Full article at BBC NEWS | Scotland | North East/N Isles | Helicopters grounded for 48 hours (http://news.bbc.co.uk/1/hi/scotland/north_east/8004382.stm)

[maxwelg2]

This may be a stupid question, but why is the AS332L1 not included in the MGB increased inspection requirements? Different design or based on lower power/load requirements?

The GE PowerPoint on ProDAPS was an interesting read. Many areas to be improved e.g. re-tuning sensor baseline data post-maintenance to account for revised quiescent harmonics etc.

JimL, I appreciate your point on the less invasive inspections the better, but in this specific case there does not appear to be any other choice until the root cause failure mode is determined. The only good thing that comes out of these tragedies is increased awareness and focus on preventative actions.

Has anyone considered a possible manufacturing issue with the gears on this specific MGB batch? I would expect QA/QC to be very tight, but will boroscope inspections be enough or will the gears require x-ray etc. to determine serviceability?

[212man]

Has anyone considered a possible manufacturing issue with the gears on this specific MGB batch

I'm sure a lot of people have - I certainly have.

A few years ago, in my previous company, a Bell-212 had a blade grip failure on a 1200 hour rotor head. Fortunately for the crew they were about 75 ft during the landing phase onshore, and 'got away with it,' though I think the underwear were written off! The point being that a vey old, established and proven design was the victim of poor manufacturing QA, and the head contained latent stresses which eventually resulted in the lower tang completely shearing accross its width.

[Hilife]

This may be a stupid question, but why is the AS332L1 not included in the MGB increased inspection requirements? Different design or based on lower power/load requirements?


I’m told the gears in both the A332L2 and EC225LP are subject to increased hardening during the manufacturing process, I suspect as a result of increased torque levels.

I'm also told some of the Dauphins require repetitive boroscope inspections of the internal gearbox areas, but due to the seriousness of this inspection, I would be surprised if boroscoping alone would satisfy the level of inspection required by the Airworthiness Authorities.