Jazz CRJ "Oops" in Toronto...

[Moon The Loon]

And CRJ's haven't been grounded by EAD's because...?

[nozel]

Just looking at the ADs and the areas for inspections it does seem like a lot of them are concerned with that area but

The AD/SB's ask us to inspect the gear leg not wing structure...

The gear main fittings are intact, wing hard points are not...by what we see in these pictures.

No side loading evident here.

Its a miracle the fuel tanks stayed sealed with the damage to wing structure...

[MURRAY]

MTL:

Awesome avatar..

Mitch: I'm building a butterscotch Tele...Kneel and revere...

Not topical? Sorry...

[boestar]

Just hearsay but I under stand the aircraft bounced after initial touchdown 45 ft. then hit ground at 4.7G. This is beyond the design limit I would think. I will actually check on that but seems high.

This would lead one to SPECULATE that perhaps the aircraft touched down and for whatever reason bounced and in the recovery the aircraft stalled ( we all know the CRJ does not like high AOA) and impacted HARD.

Also the trailing link design of the landing gear does not transmit stresses to the airframe in a straight line there is actually a torsion applied to the gear fittings and surrounding structure. This explains the gear failin rearward (again SPECULATION).

I must say that I like the way the aircraft structure failed in such a way that there was minimal damae to the passenger cabin. Indicative of a good damage tolerant design.

B

[Mitch Cronin]

Just hearsay but I under stand the aircraft bounced after initial touchdown 45 ft. then hit ground at 4.7G.

"Hearsay"? Why, because you didn't extract the data from the DFDR yourself? ...never mind... rhetoirical questions....

[J.O.]

The part about a bounce on landing appears to be factual. From what I've been told, the G-load value hasn't been determined. Most FDRs only sample G-load once per second, so it's quite possible that an accurate G-load value won't be determined from the FDR. There's a way to calculate a fairly accurate value, but that takes an advanced knowledge of math and physics that I simply don't have.

[Don Hudson]

JO;

Re, "Most FDRs only sample G-load once per second, so it's quite possible that an accurate G-load value won't be determined from the FDR."

Yes, that's the case. However, the dataframe for many flight data analysis programs provides sampling rates for vertical acceleration ('g' ) of eight times per second.

This brings up a related point regarding DFDRs and FDA Recorders.

While the law in Canada requires that DFDR's be tested and certified once per year in Canada, and while these DFDRs only sample relatively few parameters and at slower sample rates, FDA recorders with suitable parameter maps and dataframes sample thousands of parameters and at sample rates of up to 16x per second depending upon the installed dataframe and software. What's more, what is being recorded on the FDA equipment is the same information as is being recorded on the DFDR and such data is examined daily so data verification is daily not yearly.

That begs a big question for airline managements then...why can't airlines use FDA data to dispatch an aircraft which may have had a hard landing, a flight control problem or some system MEL issue which requires maintenance action for dispatch? What can sometimes take a day or more can be resolved in hours with data certification and wireless celluar technology.

Certification of FDA data should be a no-brainer as it is the same 0's and 1's as exist on the DFDR. Yet no FDA data is legally permitted, by itself, to be used to dispatch an aircraft. We have been informally told that it is "okay", but unless it's stated specifically, what airline would do so?

The cost savings alone are enormous. An aircraft tied up at an outlying station, awaiting maintenance inspection or the pulling of the DFDR for reading (likely back in Ottawa) costs huge bucks, when in fact the FDA data can be quickly pulled and read to verify if there was/is a problem with the aircraft.

In fact, wireless cell-phone FDA technology now makes it possible to interrogate the aircraft either in real time or once it gets to the gate so that a problem could be solved in the time it takes to normally turn the aircraft around.

Why this ruling isn't being pursued and the technology installed is a mystery as the potential cost savings are huge especially when compared to equipment costs and cellular transmission costs.

At a parameter sampling rate of eight times a second, during a hard landing one can see the 'g' load actually building to a peak, dropping off and building again in the case of a bounced landing. Peak g loads are easily seen, at least for loads-analysis as well as investigative purposes.

Lateral 'g' forces (remember AA587?) are sampled at 4x/sec.

Cheers,

Don

[boestar]

Actually Don on the CRJ we do use the FDR data to determine if the landing was in fact "Hard" the calculation is made looking at the time prior to touchdown until WOW is detected which gives a decent rate at touchdown. the calculation gives the ft/s which woould relate to a G Load at touchdown. The Maintenance manual procedure says that if the rate is below then dispatch is acceptable with no further action. If it is higher than the inspection must be completed prior to further flight.

B

[Don Hudson]

Hi Boestar;

Thanks...I removed the comment earlier because of this...the data IS indeed available from the DFDR.

However, the frequency of the DFDR recording does not compare with what can be achieved using higher sample rates (bigger dataframes). Jazz isn't doing FDA I know, but that said, the information, such as it is, is indeed available from the DFDR.

I know of the calculation you're speaking of as we have engaged that dialogue with our maintenance guys who have used it from the same sources. The numbers in the manual are, I believe, 640fpm IVSI or about 10 feet per second at touchdown or 2.5g's, (whichever comes first... )

But when in this kind of territory (hard landing), an event is close to "instantaneous" so the results of the calculation are in the realm of an approximation, and the peak of the event can be missed. The time of WoW is only as accurate as the number of times that parameter is recorded. If it's only recorded once-per-second the actual IVSI at that moment may not be known and would therefore not provide the fps rate of descent. The recording of the parameters (at once per second) is either too early or too late and the calculation will reflect that inaccuracy.

The A320 dataframe we use provides WoW (oleo compressed/not compressed parameter) eight times per second (nosewheel is four times per second) so along with the VACC parameter (vertical acceleration) at eight times a second, the touchdown point and g-loading is highly accurate. That said, sixteen and even thirty-two times per second for some parameters is now being discussed. It's a matter of storage capacity and new recording equipment. Further, there is a g-datum calculation which must be made (I don't do this and dont' know how...I leave that to the mathematicians and our Service Provider! ). Due to flex, the 'g' experienced by the airframe at the wingtips,, tail/nose or the pylons is different than right at the center of the aircraft.

An early case in developing the Program illustrates this: We have an event for the flight control check. The rudder pedal position parameter and the rudder position is recorded once per second but slightly after one another (it is a serial recording, after all! ). A number of flight control check events triggered because the rudder pedals weren't taken to the required degrees of pedal deflection left or right. So they recorded an "event".

I discovered in further examining the data however, that the rudder had indeed travelled all the way to the stops. It's just that the pedal-parameter was "in between" recordings. It's exactly like a 32-frames per second video being viewed at (say) 26fps...stuff gets missed, or using a yardstick to measure a foot...same thing - the eyes aren't open.

For the purposes of FDA, (and for investigations), determining the exact point of takeoff and of touchdown is indeed mathematical in nature and reliant upon a number of parameters, just the way you say. The inaccuracies merely come from the sample rates.

Most of the time this kind of an issue isn't a problem and the FDR data is very useful in an investigation but here, where there may possibly be indications of very high 'g' loadings over an extremely short period of time, there are these things to consider.

[conehead]

That begs a big question for airline managements then...why can't airlines use FDA data to dispatch an aircraft which may have had a hard landing, a flight control problem or some system MEL issue which requires maintenance action for dispatch? What can sometimes take a day or more can be resolved in hours with data certification and wireless celluar technology.

.

.

.

The cost savings alone are enormous. An aircraft tied up at an outlying station, awaiting maintenance inspection or the pulling of the DFDR for reading (likely back in Ottawa) costs huge bucks, when in fact the FDA data can be quickly pulled and read to verify if there was/is a problem with the aircraft.

In fact, wireless cell-phone FDA technology now makes it possible to interrogate the aircraft either in real time or once it gets to the gate so that a problem could be solved in the time it takes to normally turn the aircraft around.

Why this ruling isn't being pursued and the technology installed is a mystery as the potential cost savings are huge especially when compared to equipment costs and cellular transmission costs.

Don; some very good points. I feel frustration at pulling an Embraer off-line for an inspection that may or may not be needed after an "over-speed" event is reported. I can download the DFDR onto a flash-card, and e-mail the data to the playback center in YUL for analysis. The problem is, they only work bankers' hours. My event usually occurs outside of "normal" office hours. (Murphy's law?) Hours are wasted calling in a technician to analyse the data, when we could easily do it in YYZ, or where ever. In every case, the ship winds up in the hangar for an inspection that is sometimes not neccessary, resulting in lost revenue.

I find many instances of our employer not taking advantage of available technology that could help the operation run better. As I said, it's frustrating.

[wizard]

So when can we expect to get some solid evidence on what really happened?

[Don Hudson]

Conehead;

Fully understand your frustration and have seen it time and time again when an A320 is grounded at an outlying station because of the aforementioned snags.

Re, "The problem is, they only work bankers' hours."

Again, fully understand. The FDA Team at AC is "on duty" 24/7 in the sense that it's reachable by cell and can start the process of investigation very quickly. With cellular technology (WGL: Wireless Ground Link), we would be able to interrogate the aircraft, download the data, examine it and then forward the necessary parameters for maintenance to make a dispatch decision, all very quickly. All that is needed is support for WGL and the decision that FDA data is as good as DFDR data. With the data cards as you may know, the process is a PITA for everyone concerned not the least of whom are the ACM people pulling and replaying the cards daily in Toronto. It makes complete commercial sense, but...

The Embraer won't be online for a long time yet and I know this information could assist. The flight safety value notwithstanding, the commercial value is tremendous.

[jump seat]

Was this a check ride as well for the FO?

[Guest 06L06R]

So do we have anything new or factual on this yet?

[dagger]

So do we have anything new or factual on this yet?

The plane, a CRJ-100, was written off, and Air Canada - not Jazz - received a $14 million insurance payout.

[Canus Chinookus]

so how does that work? I thought Jazz owned their fleet.

[dagger]

so how does that work? I thought Jazz owned their fleet.

Many planes, but the CRJ-100s are leased by AC to Jazz.

[conehead]

The plane, a CRJ-100, was written off, and Air Canada - not Jazz - received a $14 million insurance payout.

Wow, that's pretty good.

[dagger]

Wow, that's pretty good.

I bet they couldn't have sold that plane undamaged for $14 million!

[conehead]

That's exactly what I was thinking, although I'm sure the engines are worth a few bucks.

[Fido]

Many planes, but the CRJ-100s are leased by AC to Jazz.

I bet the airplane was owned by Alberta-------- limited which may be a subsiduary of AC or ACE. I know not which.

[wizard]

So I ask the question. What type of aircraft will Jazz get to replace the lost CRJ-100 ? The question was raised at the Jazz second quarter conference meeting and Joe ruled out replacing it with another CRJ-100 as all he said was it will be a jet.

[Super 80]

hah, if they want to fly it away we will give them one.

[dagger]

So I ask the question. What type of aircraft will Jazz get to replace the lost CRJ-100 ? The question was raised at the Jazz second quarter conference meeting and Joe ruled out replacing it with another CRJ-100 as all he said was it will be a jet.

A CRJ-200 or a CRJ-705?

[conehead]

There's lots and lots of -200's sitting around doing nada.