Pelican's Perch #82: The Dreaded Three-Engine 747
I have read more nonsense in "The Media" about the British Airways 747 that made the Los Angeles-to-England crossing on three engines than I can recall about most aviation "events." I long ago gave up on "media" when it comes to aviation reporting, but this one is leading a lot of pilots astray, and we're seeing something of a lynching.
John Goglia, a highly respected former member of the NTSB, roasted the crew and the airline in Aviation International News (June 2005).
My good friend and colleague Barry Schiff just did a column on the incident in AOPA Pilot (Nov. 2005). His central point about yielding command authority is valid, but I don't think it is correct in this case.
From various media accounts, the FAA appears to have tried very hard to come up with some way to file a violation. As far as I know, they have not found a basis for such an action. The FAA's extreme haste to file a violation -- any violation -- was unseemly, in my opinion. I suspect the good FAA folks in that agency (and there are many) are more than a little embarrassed.
With all due respect for all those fine folks, permit me to disagree.
Right up front, loud and clear, and based on what I've seen:
I would have done the same thing.
I applaud the BA pilots for the decision to "go."
I applaud British Airways for their part in approving the flight before it left the Los Angeles area. I hope they continued to support the decision, and their crew after the return, and especially after the fallout in the media -- which is the only real problem.
Well done, chaps!
You can Google on "British Airways 747" and come up with representative material.
To me, the facts appear to be these:
- The 747 departed LAX for London with about 350 people on board.
- Somewhere after liftoff, a flash and/or sparks were observed from the number two engine, most likely a "compressor stall." This was preceded by sparks from that engine before liftoff.
- The crew reported shutting down the engine.
This left the crew with several options. My favorites, in order:
- Continue heading for London, while exploring the issue;
- Continue to some on-line BA station; or
- Return to LAX.
All are viable, safe options, given a few assumptions. But as usual, there are "best" and "worst" choices.
After some consultation with management back home, the crew elected to continue the flight to London. Returning to LAX would have been my absolute last choice, by a considerable margin.
Let me tell you why.
The media accounts most closely describe a "compressor stall" caused by some disruption in the airflow or fuel flow through the engine. While these can be quite spectacular, and severe ones can cause damage, most engines will handle the momentary fault with aplomb, and in some cases will continue to operate normally for tens of thousands of hours. These events were very common in the early models of the 747 with the JT-9D-3A engine, mostly during startup and taxi, but occasionally in-flight. In fact, abnormal starts with that model engine were the rule, not the exception, and we became quite blasť about them. Many older jet engines will snort and bang if the throttles are handled carelessly during throttle movement. The 727 will protest loudly in a crosswind as takeoff power is applied. Some will "backfire" and blow a huge gout of flame out the front of the engine, then mind their manners and run just fine. Post-flight inspections will usually show nothing wrong, and no maintenance is needed beyond the inspection. Some engines are likely to be damaged by a compressor stall of any kind. It seems to me that the newer jet engines are less likely to do this, and are less tolerant, as well.
Jet engines have up to three "spools," or three separate shafts, with turbine blades and compressor blades on each. The relative RPM between these "spools" can be quite critical, and if one of the "spools" gets a bit off-RPM, the airflow can be disrupted, and the usual reaction is the good-old compressor stall again. Some engines will bang once, then return to normal, some may not. In some, the compressor stall itself will cause subsequent compressor stalls, halted only by reducing thrust, then bringing it back gently.
Jet engines are wonderful vacuum cleaners and tree chippers, and if there is FOD (foreign object damage/debris) on the runway, it can cause problems ranging from simple airflow disruption to chipping or breaking a blade, or worse.
When starting down from high altitude, some early jet engines will "rumble" in protest if the throttles are brought back too quickly (or too far on the old 747s with -3 engines), and might flash and bang in a compressor stall. Completely harmless, or so we were told at the time.
If it's more than momentary, the compressor stall will cause an abnormally fast EGT rise, and if it continues as some will, the EGT will quickly soar past the limits and do major damage.
If any of the blades on those spools fail, or a piece breaks off, it will rattle around the engine and cause sparks out the tailpipe, with the severity proportional to the size of the chunk and where it came from. Some engines will tolerate a missing part (a blade) for thousands of hours, some won't.
Fortunately, most of these problems have been solved with better technology, and compressor stalls have become quite rare. So rare, in fact, that when they do occur, they are much scarier than they used to be. The knee-jerk reaction now seems to be, "Let's get this sucker back on the ground!"
Back to BA
I'm guessing now, because the details have not been released. But I'm betting a blade let loose inside the number 2 engine and rattled around, causing the sparks observed during the latter stages of the takeoff. That may have done some additional damage, and finally the engine coughed with a big compressor stall. That may have killed the combustion taking place, failing the engine, or the crew may have just shut the fuel off. EGT sits up very near redline on a full-thrust takeoff, so it wouldn't take much rise to trigger the warnings. It's even possible that pulling the thrust lever partly back and reducing the thrust would let the engine run normally, even with a blade or two missing. For this and other reasons, almost all modern jet engines have very sensitive vibration detectors, and these and other abnormal indications will usually tell the tale.
I do not know if the BA crew simply shut the engine down, or if they operated it at partial thrust for a time, or what the indications were, or what the vibration was, but it doesn't matter. They knew what was going on, and a very senior, highly experienced crew -- together with management back home -- made a decision to bring it on home.
A Little Perspective
Take the single-engine, piston aircraft: lots of heat, moving parts, reciprocating motion, vibration. It is a device that is doing its best to tear itself apart. It is amazing to me that we have managed to get some of them to run up to 3,600 hours between overhauls when properly put together and properly operated (as in the big, late-model R-3350s in the Connies and DC-7s). Many flat engines today are not properly put together, and have problems during the TBO run that can be catastrophic if not detected and repaired.
The wise pilot (I have not always been wise in this regard) will attempt to avoid situations where a failure of any of those flailing parts is likely to kill or injure. In the event the engine does fail, it's obvious the single-engine airplane is going down. While this can raise the pulse rate, "forced landings" usually turn out pretty well. Single-engine (piston) flight across the North Atlantic is not within my personal limitations these days, though many do it. I have taken my own family from Key West over Cuba to Grand Cayman and back in my Bonanza, figuring the seas are fairly conducive to ditching, the water is warm, and a quick pickup is likely. (Besides, the kids were being miserable that day). The time over land is significant, and the odds of a failure during the brief over water portions are very slim indeed. Still, some may choose not to do it. I'd be a lot more concerned going across Lake Michigan in the winter. Few would suggest commercial operations like this. We owe a higher duty, and a higher level of safety, to the paying public.
Let's move on to piston twins. Or, let's not! Piston twins do not provide the huge jump in safety that seems so obvious to the non-pilots in the family, or even to the piston-single pilot yearning for a bigger airplane and more levers in the right hand. Having two more than doubles the odds of an engine failure, and operating on one remaining engine is a lot less safe than flying a piston single, because the odds of another failure generally go up, and there are other major complications as well. On one engine alone, the twin will need a much higher power setting. The airspeed will be a lot lower, with less cooling. The aircraft may be unable to maintain altitude, forcing a descent. There are also the issues of talent, training and proficiency, but that's for another column.
For these reasons, the generally-accepted advice is, "When one engine fails on a piston twin, land at the nearest safe airport." Good advice, I agree.
There were once three-engine piston transport aircraft, but let's not complicate this.
With a four-engine piston aircraft, life gets a lot better. Since most of these will operate just fine on three, we can go anywhere on four, or even three, under some circumstances. In fact, all those I'm familiar with are authorized for ferry flights with one engine inoperative right from the ramp, albeit with restrictions (e.g., no passengers). If one fails in flight, we can usually complete the flight on three. However, it's not quite that simple. When one engine is shut down, the piston transport will have to descend to altitudes where icing and weather can be a major problem, and fuel burn goes through the roof. This was not fun on the North Atlantic! An engine failure right after takeoff in any multiengine piston aircraft was not trivial, because the remaining engine(s) have to be operated at much higher-than-normal power, again making the loss of another engine much more likely. The loss of a second engine at very heavy weights will not permit extended flight unless fuel or payload is dumped. As I recall, the manuals for four-engine piston aircraft said something like, "Consider landing at a suitable airport with one out," and, "Land immediately with two out." Different companies had different philosophies on this.
I love flying the big, old radials, but for real work out there in the cold, nighttime clag, I'll take the turbines even if jet fuel does stink. "Props for play, jets for work." There's no reciprocating movement, no flailing parts; they just roll on, spinning endlessly.
The single-engine jet pilot is in a far better position than the single- or twin-engine piston pilot. The aircraft will generally fly higher, the glide distance is much better, and the engine is orders of magnitude more reliable. I personally won't ferry a piston single or twin across the North Atlantic, but I would cheerfully ferry a modern single-engine jet that I knew well, with precautions. (Our guys in the military do this all the time). I wouldn't put my family aboard, and I wouldn't carry paying passengers, of course.
Which brings us to twin jets like my current mount, the magnificent Gulfstream IV. Give me a little room on the numbers, but I've been told there have been five total shutdowns in something like three million hours of operation on these Rolls Royce Tay engines, or once every 600,000 hours. I like those odds! I assume some of those have been "precautionary," where the engine might have continued to run just fine, but never mind that. If one engine does fizzle, the odds of failure for the remaining engine do not change much, if at all. The thrust on the remaining engine doesn't change, we just drift down from, say, FL450 to FL330, where the normal cruise thrust from the single engine will maintain speed and altitude nicely. Still way above any icing levels. The speed will be a bit slower, so we'll arrive later, but with the same fuel, or pretty close. That's a massive increase in safety over pistons, but our advice in a twin jet with one out is still "Land at the nearest suitable airport." If I lose one right after takeoff, I'll know in advance exactly where I'm going, and how to get there (usually right back to the same airport). Over "The Pond," there is Newfoundland, Greenland, Iceland, Bermuda and the Azores, and the "magic" shows us how far, how long, and how much fuel to each. The odds of one failing are very remote, and the odds of having one fail, then the other, are astronomical. To me, those are acceptable for my family, and for the paying public. Obviously, the FAA agrees, for they permit 757s, 767s, and 777s to do it, too.
Let's move on to the Tri-Motors like the L-1011 or DC-10 (and variants). (And no, it is not a Boeing, it's a Douglas DC-10, just as it's a Douglas DC-3! Don't get me started.)
Both those airplanes will fly just fine with any engine out, and as I recall, they'll both fly on one engine at the lighter weights. If I will fly a twin jet across "The Pond," I'll sure fly a trimotor there, and I might be a little more casual about diversion airports. There are far more "diversions" to my liking in Berumuda while getting the airplane fixed than in, say, Greenland. With the twin, I pay attention to where the nearest (in minutes) non-water airport is located, and how to get there; with a Trimotor I'd have the information handy, but wouldn't worry about it too much. With one out right after takeoff, I probably wouldn't head across "The Pond," for life would get too difficult with the loss of another. But there's no big hurry to land, either.
With four jets, all those worries pretty well go away. Flight with one engine shut down is not an emergency in the 747; it simply becomes a Tri-Motor. It's an "abnormal" just like an inoperative generator. Cruising flight is completely normal, although it can't climb quite as high, or go quite as fast. The 747 on three will still blow away some of the twinjets on the North Atlantic. To my knowledge, there is no guidance in the 747 manual to do anything non-routine with one shut down or failed. There is no degradation of any of the systems. Some operators may have added something, but that's often a personal matter by a chief pilot who is more worried about "What people might say" than real safety.
Pilots will (rightly) ask, "What about collateral damage?" That is, if one has failed, suppose whatever caused that failure causes another?
Enter the beauty, the sheer elegance of the pod-mounted jet engine, and the design genius of Boeing, particularly on the 747. Unless the engine explodes in the classic "uncontained engine failure," there is nothing that can cause a related failure. As long as the pod remains visually intact, and not trailing 400 feet of flame or smoke, the problem is solved when the engine is secured. All fluids, gasses, and electrical to/from that pod are shut off. It is totally isolated, or "contained." The very rare "uncontained failure" will make itself known very well, believe me. I'm one of the few privileged enough to have had one, and it is spectacular. It ended up being a non-event, and we returned to JFK in about 10 unhurried minutes, no problem. It wasn't so easy for Al Haynes and the crew of UAL 232, for that was not a pod-mounted engine, and the flying parts gutted his airplane, including flight controls. But let's keep those separate and distinct from "simple engine failures," which are harmless outside the pod-mounted engine itself.
By the way, suppose we had an eight-engine transport? Or a hundred engines? Would the simple failure of one call for an immediate return on those?
What People Might Say
What to tell the Nervous Nellies? How about, "Don't fly at all. Take a Conestoga Wagon like the settlers did. If going to the West Coast, your odds of dying are 40%, and if going to Europe you'll drown in the first 10 feet, but you'll avoid those dangerous airplanes!"
Let's put it another way:
L-1011s and (Douglas) DC-10s routinely do the Hawaii run on three engines. As I recall, there's one point on that flight where you are further from land than at any other point on the globe. 767s and various Airbii do it on only two! Anyone who says the BA 747 shouldn't have continued to London on three engines cannot possibly justify those flights at all! A lot of pilots do make an issue; they huff and puff, and claim they wouldn't fly the overwater twins, pontificating, or joking ("ETOPS" means "Engines Turn Or Passengers Swim"). Easy to say, but wait until the airline downsizes and "juniority" forces them to bid the 767 routes to Europe, where the highest pay is. Not a single one will refuse. Huff and puff all you want, but the FAA has (rightly) ruled that twin-jet operations with paying passengers is safe over the water, and I fully agree.
There is the matter of BA landing short due to low fuel. That's a separate issue entirely. I don't know the details, but I can speculate. First, they apparently stooged around the the LAX area for some period of time making up their minds, communicating, etc., burning fuel. Second, there are rumors that they had plenty of fuel approaching London, but didn't know how to get it to feed all engines. That's a training issue. The full report may say something about standpipes, jettison pumps and misunderstanding of the configuration, but that is sheer speculation. In this column, I am assuming the best available data showed enough fuel for the flight before they set off for London with three engines. This is the only question.
The fact is that very long-range airline flights very often use legal tricks to make these those flights. Some may require 10% extra fuel to be loaded to allow for abnormal winds, temperatures, and weather diversions. If the normal burn fuel is 330,000 pounds, a requirement for an extra 33,000 pounds is a financial killer, and may even prevent the flight. Most airlines will dispatch that flight to some intermediate point (New York, Gander, Rejkjavik, Shannon) with no intention of landing there. Then, approaching the "Reclearance Point," the flight will be "Recleared" or "Redispatched" to the real destination, and will have the 10% margin at that point. Just as a flight from that point would.
Once in a while, that trickery runs afoul of winds or other factors, and the flight actually has to land somewhere for fuel. No problem, no safety issue, we land, refuel, and away we go. In fact, there are flights within the U.S. making unplanned fuel stops every day! It is not a matter of safety, but of economics and irritation. If you've heard the captain announce, "Well folks, the winds are stronger than forecast, so we're going to stop in Phoenix and take on a bit more fuel," then you've been on one of these flights.
I don't know if the BA flight even filed a "Reclearance Flight Plan," or what the numbers were. But they might have, and all things considered, it may have left them uncomfortably low on fuel at the end of the flight on three engines (or four, for that matter). No big deal, just another unplanned fuel stop.
Making some of the assumptions I've mentioned, I hope I would have headed right out on the planned course, knowing that even if I wanted to land back at LAX, I'd have to spend close to an hour dumping fuel down to maximum landing weight. If you're going to do that, why not head off in the right direction? Then if you must return, dump fuel, time it right, and turn around to land at maximum landing weight. Not a really good choice.
Rather than turn around, it might be "better" to pick the next BA on-line station (Las Vegas? Chicago? JFK?), and fly there. Maintenance would be available, perhaps other flights to take the passengers home, and (very important), the foreign carrier has diplomatic approval to land there. I've landed a JAL 747 in Boston, where JAL did not have landing rights, and it was ugly. No one can get off or on for any reason short of a heart attack. Not a good experience.
And by the way, those who say "Land immediately at LAX?" The tires (sorry, "tyres," they're British, you know) would have been hot from the takeoff, and they would only dump to max landing weight, leaving them with a very heavy, very fast landing. That's very likely to blow the tire safety plugs ("fuses"), ruining all 16 main tires. Sure, it's legal, and doable, but I'd much rather continue to London (or Chicago, or New York), and land with the normal landing fuel, perhaps 200,000 pounds lighter than max landing weight.
Most airlines do very precise fuel planning, and modern jets have excellent data. It's very easy to program for some minimum value at landing, with many using 25,000 pounds for the 747 (3,000 for the Gulfstream). I think some carriers will go less than that. Bear in mind that on a long flight, adding 10,000 pounds of fuel will increase the burn by 5,000 pounds, which means you've only got about 5,000 extra at the end of the flight.
It doesn't take much change in the winds to steal 5,000 pounds, or even 10,000. At some point, the intelligent captain will say, "Wait, this just isn't working, we're going to land at Shannon, or Manchester, or whatever, and take on some fuel." No big deal, it's a natural consequence of tight planning. If that happens a couple of times a year, it's well worth it, because of all those other flights that burned much less fuel. Airlines have this process down to a science, because most airline executives would murder their mothers for a 1% fuel savings.
Finally, there are rumors of fuel mismanagement on the BA flight, or not understanding how much of the fuel remaining was truly available. I cannot speak to that. I can guess at a couple of scenarios, but that's sticking my neck out a bit too far. I'd love to see the official report, when it's all done.
Barry Schiff's major point is the modern degradation of PIC responsibility, and I agree with him entirely. However, he may have it wrong in this case.
I think that crew was very wisely doing some CYA, not abdicating authority. Modern CRM and "company communications" is no different from the "shared responsibility" between the Dispatcher and the PIC. Notwithstanding the comment one of the pilots made to ATC, I would guess that the PIC got on the horn ("blower" in Brit), and said something along the lines of:
"I say, chaps, here's what's happened, and I don't see any reason we can't press on with stiff upper lips to London on three engines. Lacking that, perhaps land in New York and move the load to another machine, or maybe dump of bit of petrol and alight in Chicago. I recommend not returning to LAX, that dreadful place; we're too heavy, we'd have to dump too much fuel, and we'd still end up landing very heavy with hot tyres. The crew wants to go to Hawaii, but it seems a bit out of the way, don'cha think?"
This is in no way abrogating authority as PIC. By bringing in the company from the start, he was preventing any trouble from the company later. Oh, and by the way, with modern jets and monitoring, management in Jolly Auld London probably had more and better data from the "failed" engine than the crew did, another good reason for a consultation.
Now, if Dispatch had said, "Continue to London" and the PIC didn't like that choice, then the PIC would rule, and properly so.
The only reason I know of to not go to London in this case (assumptions as above) is simply the PR factor, which is all that we're seeing now. To the layman, continuing flight after an engine failure is unthinkable. Obviously it's incomprehensible to a lot of pilots, maybe even a few 747 pilots. I still think I'd have done the same thing the BA crew did. It's not even close.
There are points worth making beyond just this one incident.
I've now been a working pilot for just 50 years, and in that time I've seen many swings of the FAA pendulum. I've seen a harsh and difficult FAA under Quesada, and I've seen a more benevolent FAA that tried to "promote" aviation. I've seen an FAA that is quick to hammer pilots for the slightest violation, and an FAA that tried counseling with a bit of forgiveness for the genuinely repentant. I've seen an FAA with a fair amount of trust for those appointed to serve as DPEs, DARs, etc. There is no doubt in my mind as to which is the more effective, which creates better safety.
I don't think today's FAA is the best it could be. We have more regulations, greater complexity and confusion, more bureaucracy, and more hostility on both sides than ever before. Pilots deeply mistrust the FAA because the FAA (as an agency) is perceived by many pilots as an agency which deeply distrusts everyone. It is truly a vicious circle.
This BA flight is a case in point. The FAA appeared frantic to find any rule they could to prosecute and persecute the crew and the airline, long before they (or anyone) knew what had transpired. I think it would have been better to grant a little credit to an extremely experienced cockpit full of professionals who had no wish to die, and an airline management who fully supported their actions. By all means, find out what happened, and why, and then use that to educate the rest of us. In such a culture, data would be much more forthcoming (even if detrimental), and we'd all learn.
Another effect of this is that when pilots are faced with "difficulties," the first thought will be, "What is the best choice to avoid certificate action by the FAA?" That's not a good way to run anything, much less an airline, or an airplane.
By taking Draconian action (immediate violations), the FAA pushes all crews everywhere to hide or erase data, tell "alternative stories," and stonewall everything. It is now an "Us against Them" situation, and safety suffers greatly.
It doesn't take much imagination to see our own FAA getting more and more like the Brazilian authorities in the recent mid-air in that country. The Brazilians have buried the data, stonewalled the whole thing, seized the airplane and assets of the owners, and worst of all, arrested the crew, seized their personal assets, and are sitting on it. For shame, Brazil! I don't think I want to fly in a country like that.
Despite the common perception, there are an astonishing number of very, very good people in the FAA who deplore what we are seeing even more than I do. Unfortunately, these good people are being driven out by the few who seem to favor tactics that are perceived by many on the outside as being too much like those of a police-state. We must reverse this in the FAA, and in the country as well.
Be careful up there!
More from AVweb's Pelican is available here.