Looking for an incident

[Gabriel]

Last year (I think) there was an incident where an Airbus (I think and A320) departed from some US airport (I think near water, maybe New Jersey) and shortly after had an important failure where they lost all of their primary instruments, returned to land, asked for the longest runway, but there was some equipment in it, the tower called the guys to get off the runway ASAP to let the Airbus land but they couldn't do it on time so the Airbus landed on another shorter runway and overrun it (or maybe veered off).

I was thinking that enough time elapsed since the incident and perhaps there are interesting findings already.

This incident had its thread in these fora.

As you can see, I don't remember date, airline, flight number, or even location, in short, all the necessary info to search in the Internet. So any help will be appreciated.

[Fear_of_Flying]

I think I found an article on the incident, though haven't looked for the thread. Was it United Airlines 497?

http://www.nola.com/news/index.ssf/2...t_airport.html

[3WE]

Originally Posted by Gabriel

...an Airbus had an important failure...

In other news, half of the earth was plunged into darkness yesterday shortly after sunset, depending on the time zone.

[Gabriel]

Originally Posted by Fear_of_Flying

I think I found an article on the incident, though haven't looked for the thread. Was it United Airlines 497?

http://www.nola.com/news/index.ssf/2...t_airport.html

Yep. That's the one. Thanks.
The relevant AvHerald report has not been updated since a few days after the incident. So I guess that no more information has been made public.

[Evan]

Originally Posted by Gabriel

As you can see, I don't remember date, airline, flight number, or even location, in short, all the necessary info to search in the Internet. So any help will be appreciated.

United UA-497, A320 Reg N409UA, 4/4/2011 out of New Orleans. Only a preliminary report exists, which is just a synopsis.

The crew received first an auto-thrust related message, then an avionics smoke indication on the ECAM. They took checklist actions to shut down electrical busses and requested an immediate return to the airport. No smoke was ever detected by the crew, only the indication on the ECAM.
They landed overweight without burning fuel on the shorter runway (the longer was closed with vehicles on it) and overran the runway, stopping in the grass with the nose gear collapsed.

A news report printed this:

Despite the danger, the tenor of the recording between air traffic control and the pilot remains remarkably nonchalant.
"The ability to stay so calm," Esser said. "Hat's off: That was great."
This is what they train for, Sevier said.

I wonder about that last statement. The crew definitely seemed cool under pressure from the ATC transcripts. But a few things cause me to think they overreacted and failed to follow the correct procedure.

THE FOLLOWING IS AFAIK BASED SOLELY ON THE FCOM AND QRH DATA I HAVE:

An avionics smoke indication appeared on the ECAM.

If an avionics smoke indication appears on the ECAM, the correct procedure is then determined by whether there is actual evidence of smoke in the cockpit. If there isn't, as in this case, the procedure requires them to first shut down the AC 1 BUS, then the AC 2 BUS, to determine which is involved. The QRH procedure reads as follows:

Consider shedding the AC BUS bar on one side. Then, if unsuccessful, on the other. When it is clear that the shedded side is not involved, reconnect it.

I take this to mean one AC bus should be powered during the process. In this case apparently both AC busses were shut down.

The RAT was deployed. The RAT drives the emergency generator which provides power to the AC ESS bus. The AC ESS BUS powers the CPT PFD and the E/WD (ECAM). So I don't understand why those would have been lost. The fact is, I think reports that they lost ALL instrumentation was inaccurate. The crew did have at least basic back-up instruments: the report on Aviation Herald indicates that the Capt had airspeed, altimeter and attitude information.

In order to lose the LEFT main displays, the AC ESS BUS would have to be shed. The pilot can only select it to ALTN, where it feeds from the opposite AC BUS. It is automatically fed from the static inverter or the EMER GEN (RAT) if both AC 1 and 2 are selected off.

The RAT deploys automatically if the normal AC supply is lost. Thus it would have deployed if both AC 1 and 2 were manually shed.

The RAT would deploy manually if they had selected EMER ELEC PWR > MAN ON. (I'm not sure if this will also automatically shed both main AC busses, since the engine generators were still functional. If so, that would also explain why they were both shed.) But AFIAK, they should not have selected EMER ELEC PWR > MAN ON unless they actually detected smoke.

So, as far as I can reason, the crew either shed both AC 1 and 2, or they selected EMER ELEC PWR > MAN ON (or both). Either would seem to have been a procedural error since there was never any actual smoke detected.

Certainly, they handled this with the utmost professionalism and reacted quickly to the situation. I just wonder if they had been trained properly (and adequately) for the situation, because it appears that they erred on the side of caution where to err on the side of caution resulted in a much more dangerous situation.

I wonder when we will see a final report on this one.

[Fear_of_Flying]

Originally Posted by Evan

United UA-497, A320 Reg N409UA, 4/4/2011 out of New Orleans. Only a preliminary report exists, which is just a synopsis.

The crew received first an auto-thrust related message, then an avionics smoke indication on the ECAM. They took checklist actions to shut down electrical busses and requested an immediate return to the airport. No smoke was ever detected by the crew, only the indication on the ECAM.
They landed overweight without burning fuel on the shorter runway (the longer was closed with vehicles on it) and overran the runway, stopping in the grass with the nose gear collapsed.

A news report printed this:

Despite the danger, the tenor of the recording between air traffic control and the pilot remains remarkably nonchalant.
"The ability to stay so calm," Esser said. "Hat's off: That was great."
This is what they train for, Sevier said.

I wonder about that last statement. The crew definitely seemed cool under pressure from the ATC transcripts. But a few things cause me to think they overreacted and failed to follow the correct procedure.

THE FOLLOWING IS AFAIK BASED SOLELY ON THE FCOM AND QRH DATA I HAVE:

An avionics smoke indication appeared on the ECAM.

If an avionics smoke indication appears on the ECAM, the correct procedure is then determined by whether there is actual evidence of smoke in the cockpit. If there isn't, as in this case, the procedure requires them to first shut down the AC 1 BUS, then the AC 2 BUS, to determine which is involved. The QRH procedure reads as follows:



I take this to mean one AC bus should be powered during the process. In this case apparently both AC busses were shut down.

The RAT was deployed. The RAT drives the emergency generator which provides power to the AC ESS bus. The AC ESS BUS powers the CPT PFD and the E/WD (ECAM). So I don't understand why those would have been lost. The fact is, I think reports that they lost ALL instrumentation was inaccurate. The crew did have at least basic back-up instruments: the report on Aviation Herald indicates that the Capt had airspeed, altimeter and attitude information.

In order to lose the LEFT main displays, the AC ESS BUS would have to be shed. The pilot can only select it to ALTN, where it feeds from the opposite AC BUS. It is automatically fed from the static inverter or the EMER GEN (RAT) if both AC 1 and 2 are selected off.

The RAT deploys automatically if the normal AC supply is lost. Thus it would have deployed if both AC 1 and 2 were manually shed.

The RAT would deploy manually if they had selected EMER ELEC PWR > MAN ON. (I'm not sure if this will also automatically shed both main AC busses, since the engine generators were still functional. If so, that would also explain why they were both shed.) But AFIAK, they should not have selected EMER ELEC PWR > MAN ON unless they actually detected smoke.

So, as far as I can reason, the crew either shed both AC 1 and 2, or they selected EMER ELEC PWR > MAN ON (or both). Either would seem to have been a procedural error since there was never any actual smoke detected.

Certainly, they handled this with the utmost professionalism and reacted quickly to the situation. I just wonder if they had been trained properly (and adequately) for the situation, because it appears that they erred on the side of caution where to err on the side of caution resulted in a much more dangerous situation.

I wonder when we will see a final report on this one.

To your point, here's an article with some of the same questions:

http://online.wsj.com/article/SB1000...googlenews_wsj

[snydersnapshots]

Originally Posted by Evan

Certainly, they handled this with the utmost professionalism and reacted quickly to the situation. I just wonder if they had been trained properly (and adequately) for the situation, because it appears that they erred on the side of caution where to err on the side of caution resulted in a much more dangerous situation.

Full Disclosure here first: I've done my 10 hours of duty time and then had a couple of pints at the Polar. Don't worry--I'm very conservative and have another 20 hours before I work again. Having said that, my reply might be a tad opinionated and cranky...

That being said, here's my reaction to your quote above. "...they handled this with ...professionalism and reacted quickly." I.e. they did a good job. Then "...they erred on the side of caution where to err on the side of caution resulted in a much more dangerous situation." I.e. they screwed the pooch.

We can't have it both ways here. You wonder if they were trained properly and adequately for the situation. OK. You're analyzing the situation--and I will say you know Airbus systems far better than I--from the perspective of a room and a desk and a chair that is not moving and certainly not giving you a message that says it is possibly burning.

Those guys, on the other hand, were analyzing the situation from the perspective of a long aluminum tube that is giving them an indication that there is smoke in places there shouldn't be. Yes, the checklist says to shed one bus, and if that doesn't work, reconnect it and shed the other. That's great if you're in the middle of a 180 minute ETOPS segment and have three hours before you can get to land. On the other hand, you don't spend a bunch of time troubleshooting an electrical issue when you have a smoke indication and are right next to an airport. You get the SOB on the ground. Think Air Canada, Swissair, and Valujet. I'm reasonably sure that's what they were thinking. Did they panic? No--I don't think they did. If they had had RWY 10/28 available, we wouldn't be having this conversation, but they didn't. They felt they needed to get the airplane on the ground before 10/28 could be cleared of equipment, so they used runway 19, which is quite a bit shorter. Yes, they ran off the end. Yes, they collapsed the nosegear. Not the optimum situation. In retrospect, the airplane was NOT on fire, however they didn't know that at the time. Had they really been on fire and lost the airplane because they tried to troubleshoot the situation, you'd be asking why they didn't land sooner...

Alright...enough of my cranky, opinionated analysis. I'm always up for a spirited debate and would love to discuss this over a pint of beer or a cup of coffee (depending on when I have to work again). If we ever meet up, I'll be happy to buy the first round...

[Evan]

Originally Posted by snydersnapshots

You wonder if they were trained properly and adequately for the situation.

Yes. That's the entire point. I'm not coming down on them for being very intent on getting the source of the fire offline and getting it back on the ground ASAP. I'm taking the constructive point, as always, of trying to understand why this happened and what can be done to prevent it in the future. I don't think they panicked, but I think we are seeing some human factors-related errors.

Originally Posted by Fear_of_Flying

To your point, here's an article with some of the same questions:

http://online.wsj.com/article/SB1000...googlenews_wsj

Very interesting...

The crew apparently failed to turn a generator back on prior to starting the final approach to the New Orleans field, according to people familiar with the investigation's preliminary findings.

This could describe a failure to select LAND RECOVERY before touching down. The LAND RECOVERY switch re-enables certain vital circuits needed to land and roll out. It's in the checklists, but it should be known to them through training.

The co-pilot's primary instrument screen also went blank apparently due to loss of electricity, and the intercom between the cockpit and the cabin stopped functioning

If true, this is as I suspected. The Capt would still have his PFD and E/WD functioning as they are fed from the RAT via the ESS BUS. However, the crew apparently reported losing "all our instruments" and asked for radar guidance back to New Orleans. Strange. By shedding the AC 1 and 2 BUSSES they might have lost both ND's (Capt's ND is on the AC ESS SHED BUS however), which means the nav info might not be unavailable. That might explain the request for radar guidance.

United has since revised a portion of the emergency checklist to help clarify how pilots should deal with electrical malfunctions, according to people familiar with the matter.

That's a step. But has United revised its commitment to recurrent training for emergency situations like this? It seems (from my armchair) that these pilots had no preparation in their heads for a false smoke alert. I'm not saying a pilot should memorize every procedure, but something more general about first taking a moment to determine if there is actual smoke present before proceeding with the checklist. The checklist requires this.

It seems that they skipped over this part:

A team of safety experts led by the National Transportation Safety Board has tentatively concluded that after skipping over a portion of a checklist, the pilots of the Airbus A320 also failed to restore power to some equipment, making it more difficult to carry out the emergency landing, the people familiar with the matter said. No one was hurt in the April 4 incident.

FOF, when was this article published?

[Gabriel]

Snyder, a few comments:
- The plane didn't overrun the runway. It blew both right-hand tires (apparently antiskid was inop due to loss of systems) and veered off the side of the runway in the last phases of the roll (apparently NWS was also inop, so when they lost speed they lost directional control). It is also not clear to me that they collapsed the nose gear. Maybe it got damaged, but certainly the airplane was not left resting on the forward fuselage.
- The crew said that they were doing an overweight landing and that had partial stopping means, so requested the longest runway (which was not available), but in the end landing in one runway or the other did not make the difference.
- I am not sure that in a plane that is so much electron-dependent (including a joystick with no physical connection with the flight controls) switching off all the mains is a good idea.
- You say it as it was "either they did the right procedure OR they returned quickly". Shedding one at a time would not have prevented them from returning quickly to the airport. In fact, it could have made it easier and quicker since it seems they were flying in solid IMC with no available nav instrument (but for heading indication and ATC vectors) for a bunch of time (all the time since they switched off both buses to the point where they descended below the overcast of just 600ft.
- And, apparently, the procedure was not very clear: "United has since revised a portion of the emergency checklist to help clarify how pilots should deal with electrical malfunctions, according to people familiar with the matter."

[TeeVee]

yes, those pesky humans did it again, and yet another perfectly designed and built aircraft has been damaged.

makes me wonder why airbus was so negligent in failing to design an aircraft with smoke detectors and a designated system of computers running on a completely separate bus that serves only to detect smoke and fire and then take over control of shutting off the correct buses in the correct sequence, deploying the rat and restarting the correct buses in the correct order. oh, and it should probably take over during the landing sequence since it will know better than anyone what systems have been shut off that need to be re-engaged. after all, detectors are far more sensitive than a human nose, and would detect smoke at a very early stage allowing the tertiary genius smoke detected takeover the aircraft system to do its job.

[Evan]

Originally Posted by Gabriel

Snyder, a few comments:
- The plane didn't overrun the runway. It blew both right-hand tires (apparently antiskid was inop due to loss of systems) and veered off the side of the runway in the last phases of the roll (apparently NWS was also inop, so when they lost speed they lost directional control). It is also not clear to me that they collapsed the nose gear. Maybe it got damaged, but certainly the airplane was not left resting on the forward fuselage.

Yeah, my bad. It seems that they veered rather than overran, and the nose gear either dug in or partially collapsed, but the fuselage was not resting on the ground.

[- And, apparently, the procedure was not very clear: "United has since revised a portion of the emergency checklist to help clarify how pilots should deal with electrical malfunctions, according to people familiar with the matter."

One other theory - they might not have selected AC ESS FEED to ALTN before shedding the AC 1 BUS. Perhaps that could cause a transient dropout of all displays (though I doubt it).

I'll post the A330 FCOM procedure below (probably the same one they had).

[Evan]

This is from a copy of an outdated KLM A330 FCCM:

[Evan]

Close ups of the previous FCCM page:

Green box = immediate actions
Yellow box = initiate diversion
Purple box = consider smoke/fumes removal, consider EMER ELEC CONFIG, consider immediate landing
Blue box = identify source of smoke (or smoke detection)

[Gabriel]

Evan,

Airplanes have become very complex machines, with many very complex systems, and thus, lots and lots of failure modes.

It is IMPOSSIBLE to keep the crew trained to the point where they are very familiar with, and feel natural handling, all the abnormal and emergency procedures.

I agree with you that memory items should be learned and trained until it's second nature. After all, they are memory items for one thing: you don't have time to grab and read a checklist.

Other things that are not memory items but that require a lot of training is handling the plane in abnormal situations like engine failure on take-off. Well after you are stabilized in the take-off climb, or you have stopped on the runway, you will run the relevant procedures, but you must get to that stabilized situation first.

Upset recoveries, stall recoveries, windshear, TCAS avoidance maneuvers, CFIT escape maneuvers, approach down to minimums, low level go-around's, all require that the pilots be very crisp handling the plane.

Thus, these things consumes a lot of time of the training and simulation exercises. And training time is a limited resource.

For the rest of the things that are not as critical in terms of urgency (they might be in terms of consequences), what I'd expect, and I think it's the tendency in the industry, is that the pilots are very familiar with the airplane systems and the cockpit indications and controls, so they feel comfortable with, and can reliability, read and follow a step-by-step procedure that they have studied and trained perhaps a few times.

So the question is, why didn't this worked as expected in this case?
I don't think that lack of training is necessarily the answer. The above requires very clear and unambiguous procedure, since the crew will be relying in what is written, and not in their own (simulated) experience, to sort it out.

[Gabriel]

Originally Posted by Evan

FOF, when was this article published?

The date is there: April 25, 2011.

[Evan]

Originally Posted by Gabriel

It is IMPOSSIBLE to keep the crew trained to the point where they are very familiar with, and feel natural handling, all the abnormal and emergency procedures.

As I said:

I'm not saying a pilot should memorize every procedure, but something more general about first taking a moment to determine if there is actual smoke present before proceeding with the checklist.

Also, when flying a FBW aircraft, the pilots better have a good sense of how to shut down electrical busses and what the consequences are.

The above requires very clear and unambiguous procedure, since the crew will be relying in what is written, and not in their own (simulated) experience, to sort it out.

Agree completely. I don't know what exact QRH procedures they had, but they need to be very clear about shutting down busses.

[Gabriel]

Brian, this is the original thread. Please do your magic.

[snydersnapshots]

Originally Posted by Gabriel

Evan,

Airplanes have become very complex machines, with many very complex systems, and thus, lots and lots of failure modes.

It is IMPOSSIBLE to keep the crew trained to the point where they are very familiar with, and feel natural handling, all the abnormal and emergency procedures.

I agree with you that memory items should be learned and trained until it's second nature. After all, they are memory items for one thing: you don't have time to grab and read a checklist.

Other things that are not memory items but that require a lot of training is handling the plane in abnormal situations like engine failure on take-off. Well after you are stabilized in the take-off climb, or you have stopped on the runway, you will run the relevant procedures, but you must get to that stabilized situation first.

Upset recoveries, stall recoveries, windshear, TCAS avoidance maneuvers, CFIT escape maneuvers, approach down to minimums, low level go-around's, all require that the pilots be very crisp handling the plane.

Thus, these things consumes a lot of time of the training and simulation exercises. And training time is a limited resource.

For the rest of the things that are not as critical in terms of urgency (they might be in terms of consequences), what I'd expect, and I think it's the tendency in the industry, is that the pilots are very familiar with the airplane systems and the cockpit indications and controls, so they feel comfortable with, and can reliability, read and follow a step-by-step procedure that they have studied and trained perhaps a few times.

So the question is, why didn't this worked as expected in this case?
I don't think that lack of training is necessarily the answer. The above requires very clear and unambiguous procedure, since the crew will be relying in what is written, and not in their own (simulated) experience, to sort it out.

Gabriel--I wholeheartedly agree with what you're saying here. We can't train for everything. We need the emergencies that can kill you immediately trained to the point where the reaction becomes automatic: Engine failure on takeoff, stall recovery etc. There are so many iterations of the other 99% that it's impossible to train pilots to react automatically to each event. The next best thing is to hire pilots who know how to think and give them adequate tools and information to do the job.

One aspect of training I would like to see included is some human factors training about how the body reacts to stress and an adrenaline dump (Lt. Col. David Grossman does an excellent analysis on this subject in his book "On Combat."). This training could give pilots an insight into and tools to combat normal physiological reactions that may decrease their effectiveness in an emergency.

[MCM]

Evan,

Just read the checklists that you linked to.

Where does it say that both busses should not be turned off at the same time?

The procedure states to turn off a bus, and if that is unsuccesful, turn off the other. It does not say it is essential that a bus is reconnected prior to trying the other one.

It THEN says, in the small notes (not in the individual items) that you reconnect a bus after you know it is not involved.

What about if you do not know for sure that a bus is involved or not?

[Evan]

Originally Posted by MCM

Evan,

Just read the checklists that you linked to.

Where does it say that both busses should not be turned off at the same time?

The procedure states to turn off a bus, and if that is unsuccesful, turn off the other. It does not say it is essential that a bus is reconnected prior to trying the other one.

It THEN says, in the small notes (not in the individual items) that you reconnect a bus after you know it is not involved.

What about if you do not know for sure that a bus is involved or not?

As I said before, I'm just extrapolating from what I have (for instance, I don't know what the ECAM actions were prior to losing the display), but the procedure reads as follows:

Consider shedding the AC BUS bar on one side. Then, if unsuccessful, on the other. When it is clear that the shedded side is not involved, reconnect it.

So, you notice a few things. First if all, it says to shed the AC BUS on one side—not shed both AC busses—for a reason. Secondly, it says 'if unsuccessful, [shed] the other', which implies that the opposite bus should not be shedded until it is clear that shedding the first bus did not alleviate the problem. Thirdly, it says that when it is clear that the 'the shedded side' is not involved, reconnect it, not 'a shedded side', which, taken literally, means that only one side should be shed at any given time.

Even if those assumptions are correct, I agree with Gabriel that the checklist is far from clear to a crew running a checklist under stress. It should be revised and made much more clear and effective. However I also realize that the checklists should be reminders to procedures that have been trained, not an instruction manual to learn at the time of the incident.

For example, as you see in the very small type, shedding the AC 2 BUS will knock out the F/O's displays, so you need to know that this will occur when considering that action. Also it is important that the proper display switching is done before shutting down either AC bus because the System Display will be lost and you still want to ELEC page in front of you. I think these things should be prominent.

Aside from the number of AC busses shed, the other issue is sequence. For instance, it seems pretty vital that the AC ESS FEED switch is selected properly depending upon which bus you shed BEFORE you shed it.

And lastly, the checklist boxed item is only for actual smoke in the cabin. It is the only part of the checklist that calls for EMER ELEC CONFIG, i.e the RAT deployment via the manual switch. If there is no detectable smoke (as in this case), it is not called for (although it will also deploy automatically if you shut down both primary AC busses).

So, by skipping parts of the checklist, it seems that they made a manageable situation into an accident. I don't necessarily blame the pilots here, but something is clearly not as it should be. That could be the confusion of the checklists/QRH procedures and/or lack of trained awareness. If so, this was a learning experience for Airbus and its operators.

I'm glad to see United is dealing with this checklist/QRH issue, but what about training and what about all the other operators out there. Why is this not coming from the manufacturer and the CAA's? Why is the reaction to so many incidents like this always so piecemeal?