Cost/Benefit Study of CVR and FDR

[Graybeard]

Has it been done? Airliners carry these boxes, associated structure and wiring, and maintenance costs for their entire lives. How many airplanes and lives have been saved by data from the CVR or FDR? Knowing why an airliner crashed may be comforting, but is it worth the cost?

Taking AF447 as an example, will it in the end have been worth the cost if the recorders are retrieved? How much more would it be worth to assure the next crash into the water could be pinpointed in near real time?

GB

[TopBunk]

GB

One thing that recording of data brings to the table is fact. Without facts everything is basically conjecture.

There are numerous events where analysis of CVR and FDR data have been of benefit to everyone through changes of procedures on the ground and in the air and in design that would suggest your question to be simplistic at best.

[ab33t]

In the AF case even if the FDR did not give evidence Im pretty sure the crew would have commented on the conditions

[kijangnim]

Greetings, Happy New Year
When it comes to safety, and investigations, it is more Cost and Price.
To ask if things could be improved? yes, there is a need and AF447 has given the wake up call, but costs and benefits I dont agree

[mm43]

GB

If AF447 had been sending its position via SATCOM every minute instead of the 10 min interval initiated by AF Operations/Maintenance, the search area would have been reduced to a maximum of 8NM radius. That's assuming it crashed in the 59th second, which gives an area of 201 square NM (690km^2). We do know that it was still airborne 4 minutes later, and the BEA are using 0215z as their crash time. This represents the 8 x 5 = 40NM radius they are ostensibly confining their search to, i.e 5026 sq. nautical miles (17,240km^2).

This boils down to one simple fact, i.e. the search area increases by the square of the time between position reports.

Just phoning home at smaller intervals increases the SATCOM costs proportionately, but reduces a search area by the square root of the changed interval. The costs of one versus the other are too large to even contemplate.

In meantime, implementing the above is the cheapest thing that AF or any other carrier operating Oceanic routes outside of SSR coverage could do right now.

Any other changes, e.g. GELB (GEPIRB) and ULB pinger frequency, are going to take time and cost a lot more money.

mm43

[Mansfield]

It is really not a matter of comforting everyone with the knowledge of the cause. It is more a matter of how to make precise changes in order to mitigate future accidents. There is a cost benefit study done on every proposed rule change in the US. These efforts can be interminable due to industry/authorities argument over both cost and benefit. The benefit argument is dramatically strengthened by precise knowledge of previous accidents. When that knowledge is less precise, as in no recorders, the argument can be fruitless.

Following the Roselawn accident in 1994 we had over one hundred DFDR parameters downloaded within a couple of days. We had 80% of the knowledge we needed to reach some preliminary conclusions within a week. Subsequent rulemaking efforts have benefitted greatly from the comprehensive and specific data yielded in this investigation as well as that of Monroe, Michigan in 1997.

Contrast that with TWA 800...in this case, all we had from the recorders was a very specific point in time when they ceased recording. That in itself was useful in validating the breakup sequence, but the rest of the work had to be done by careful reconstruction and analysis. Fortunately, thanks to the impeccable work of a group of young Navy divers who remain largely unrecognized, we were able to do that. But the process was enormously expensive, painstaking, and remains controversial to the public. Independent power sources for those old recorders would have made a very large difference, not only in public perception, but in the dynamics of subsequent rulemaking in areas like fuel tank inerting and fuel tank safety analysis.

If you want to get a serious change made following an accident, you'd better be able to build an argument that the beancounters cannot squirm out of (and squirming tends to come naturally to them). Good recorder data can do that.

[Graybeard]

mm43
Just phoning home at smaller intervals increases the SATCOM costs proportionately, but reduces a search area by the square root of the changed interval. The costs of one versus the other are too large to even contemplate.


Does anyone know the per message cost for Satcom?

[bucket_and_spade]

I have a problem with the premise of the question - doing cost/benefit analysis of carrying the FDR/CVR doesn't sit well with me. By all means do the analysis on practically every other facet of the operation. The CVR/FDR infrastructure is a very minor expense.

I suspect you might be playing devil's advocate...

There are a plenty of accidents which would have remained a complete or at least partial mystery if a CVR/FDR wasn't carried. Actually, I'd suggest the majority.

How do you make progress without knowing what to improve? In the few examples where there are witnesses, they're notoriously naff at telling investigators the aircraft's speed, angle of attack, thrust setting, aileron position, exhaust gas temperature, etc.

On a more morbid note, families have a deep need to know how and why their loved ones died.

B&S

[mm43]

gb

There are Inmarsat 4 providers, e.g. Satcom Direct Aircraft Satellite Communication for Private Aviation, Iridium, Inmarsat, Aircraft Phone (http://www.satcomdirect.com/main/aviation) that offer packages per aircraft on a fixed fee basis. I can't help with the actual costs at this stage, but will have a dig around.

Update:: 4.8KB data rates would cost less than US$5.00 per MB - depends on the provider and the actual service plan. SATCOM installation is about US$100K per aircraft, but that's a given on most fleets these days.

mm43

[Piltdown Man]

It depends what value you put on human life. Fortunately, we don't have to do that calculation because governments have mandated their installation so they are not optional. But they are expensive to own and many airlines would probably prefer not to have them - but the data they provide is priceless for investigators.

PM

[mm43]

Graybeard

Update2:: A typical ACARS/ATSU/ACOM message at 2.4kbits/sec using the Inmarsat Broadband Global Area Network would use less than 128bytes of bandwidth (the average email is 3Kb), but to that you need to add the message headers etc., so allow for 512bytes per session. Note: traditional ACARS message size was 3,300 characters (maximum). With increasing use of the system for transmission of WX data/images etc., compression and data encoding protocols have now been introduced.

Bearing in mind the small bandwidth being utilized for an ACOM transmission coupled with Fleet Rate packages, we are now talking 20c per message. In the AF447 case, this would equate to say 4 hours at 1 per minute = 240 x 20c or US$48.

I'm sure there is someone out there who could validate this - one way or the other.

mm43

[mm43]

Graybeard

The BEA instigated "Flight Data Recovery Working Group Report" has indicated that the cost of an ACARS message via SATCOM is $US0.25, and that to include a AOC position report with each ACARS message would add about 5 percent to the cost.

It is something that they recommend implementing immediately for those already using ACARS via SATCOM and flying Oceanic routes.

http://www.bea.aero/en/enquetes/flight.af.447/flight.data.recovery.working.group.final.report.pdf

mm43

[Graybeard]

Thanks a lot, mm43. A report a minute works out to $15/hr, of course. Say there are 500 airliners beyond VHF on average, that works out to $7500/hr, right? That totals a staggering US$65,700,000 a year.

When was the last time a plane was lost like AF447? Such a recommendation isn't to be taken lightly.

Back to my original question: It's easy to come up with a SWAG, Scientific Wild Arsed Guess of the cost of CVR and FDR in every plane, most of which never crash. Figure $200K for hardware and installation, and $10K per year to haul the boxes and wiring, and another $10K maintenance. Multiply that by what, 10,000 planes?

Now, how often has a CVR or FDR revealed previously unknown information that led to prevention of a future accident?

GB

[mm43]

GB

You are absolutely right about the costs of upgrading DFDR/CVR etc., but the revised standards would only be mandatory to those a/c engaged in flying Oceanic FIRs. I'd expect that administrations would give exemptions for ferry flights.

The implementation of the AOC report with every ACARS message for SATCOM equipped a/c is a very cheap means of implementing a vastly reduced search area in the case of an accident. We know that in the case of a recent incident that the ULBs parted company from the prime recorders.

The AOC report with every ACARS message would not have been of any help in the TWA800 incident - if it had been outside of SSR coverage. There could be others like that in the future, and we then have the AF447 type search scenario all over again.

In the real world there will not be 10,000 a/c affected and the costs involved will drop quite dramatically.

How do you know that every accident has happened before? You don't, and wont until the recorders from the latest one are recovered.

In the AF447 case, there will be a series of precursors to the problem that 'bit them in the bum', but the recorders are needed to determine that that was the case.:sad:

mm43