Wednesday, September 15, 2010

I felt like I didn't give yesterday's post enough meat so to follow up the Biggin Hill accident, I want to clarify a couple of things and maybe ramble a little bit as well. Read yesterday's post first, then this one :)

First of all, in the Citation 500 series of jet we are taught that if an engine is still providing thrust, we are in no hurry to shut it down. For example, an engine fire after takeoff - the first item on our checklist is to ignore the fire indication until we climb to a safe altitude, then calmly deal with it, without rushing.

Why don't we freak out? We have no real reason to.

On our baby jet, the engines are in the back, attached to the fuselage. If worst comes to worst, they are gonna burn off and depart the airplane, which doesn't hurt the structural integrity of the airplane at all. In my previous ride, the Mitsubishi MU-2, the engines were built into the wings, so if you left an engine fire too long it could easily burn through the wing spar (which is bad), but on our jet that's not the case. Just like on lots of Airbuses and Boeings, if one of the engines falls off they will leave the airplane with reduced thrust, but they won't cause the airframe to break up. If you are going to lose an engine soon, you might as well milk it for any excess thrust you can get before you shut it down, especially if you have just departed and are close to the ground.

That's the part I really don't understand about the accident - the engine indications would have shown that the engines were still producing thrust, but I guess the vibration from the back of the airplane caused them to think that the engine indications were faulty, and that at least one of their fans was busy digesting itself. Maybe they were concerned that an uncontained engine failure might spray the cabin with fragments of fan blades - a pretty rare thing, but if it does happen, that usually means anyone sitting in the back of the plane is going to have a really bad day. The vibration must have been bad, but there's no way it would have caused the airplane to shake itself apart in the air or anything.

Why did they pick the right engine as the culprit? I don't know - they either saw something in the engine gauges that made them think it was the right engine, or they flipped a coin. Either way, we are also taught to pull the power lever back to idle to see what happens first BEFORE shutting the engine down, and that's a standard procedure for every twin-engined aircraft I have ever flown. So suppose they did that.

My take on it is that once they were convinced that it was the right engine that was sick (the left engine would have still been producing enough thrust to keep them climbing, so maybe that's how they decided it was the right engine), and when pulling the thrust lever back to idle didn't stop the vibration (again, the vibration was the air cycle machine, which had nothing to do with the engines, and it would have continued as long as either engine was operating), they decided to shut it down completely.

So far, I can see how that would happen.

The vibration continues after the right engine is shut down, the crew goes "unfortunately it appears we shut down the wrong engine" or words to that effect, and they decide to go to plan B.

Now here's the part I don't understand:

If I have the good engine shut down, I'm not even going to bother pulling the sick engine back to idle thrust as long as it's putting out a single pound of forward thrust. I wouldn't be touching the sick engine, I'd be flying as best I could on what thrust I had, and working my ass off to get the good engine relit.

In their case, the 'sick' left engine was actually working just fine, so I don't understand why they would even try to pull the thrust back at all. However, they did, and unfortunately the missing rivet head on the left thrust lever allowed them to pull the lever all the way back to fuel cut-off, killing both engines. What an unpleasant surprise that must have been.

So they are a glider then, and have maybe a minute or so before they hit the planet. The part where they had really back luck was that the dual engine failure checklist seems to say that you can try to light up both engines as long as you wait ten seconds in between, when in reality it takes 35 seconds for an engine to go from 'dead' to producing useful thrust, and even worse, if you try to light up both engines at the same time using battery power, it will kill both engine start sequences.

They try to light up the right engine, but they aren't going fast enough (200 knots) for an airstart, so they need to use the starter, which runs on the battery if neither engine is operating and they are in the air. It takes more than ten seconds, so in a panic (the ground is rushing up to meet them), they hit the left engine start button to try to get that back and running. There's no way the small aircraft battery can handle a simultaneous double-engine start sequence, so it shuts down the flow of electrons to both starters, and all that's left is picking a soft spot to land. Unfortunately they were over a bunch of buildings, so that was that.

Anyway, I find accidents pretty fascinating, and I read about as many as I can, my logic being "If I can remember to not do all the things they did that resulted in their demise, maybe I will break the accident chain links".

What I came away from this accident with are a few things:

1. Don't assume that vibration in the back of the plane is an engine. There's other stuff back there too.
2. Confirm which engine is the bad one. Is it obvious? How? Am I absolutely sure?
3. Confirm it again before touching anything.
4. Take my time in an emergency. There are only 2 things in my airplane that require split-second action - cabin depressurization at altitude, and thrust reverser deployment in flight. For anything else, I will take a deep breath before touching anything.
5. If I ever suffer a dual engine flameout, I will only concentrate on starting one engine at a time.
6. Check the thrust levers from time to time (on the ground) to make sure I can't pull the levers to the fuel cut-off position accidentally.

Any of those steps would have helped mitigate this accident, and maybe the people involved would still be with us. Unfortunately they can't speak any more, but we can still learn from what they left behind.

I don't see aviation as dangerous but I do see it as unforgiving. I'm not that smart, but I do work hard to minimize the occasions I need to beg forgiveness, and I hope the same for you.


Chris said...

Perhaps the vibrations were so severe they thought that structural damage might occur if they allowed it to continue?

Either way, hell of a spot to be caught in :(

dick fisher said...

Great post. You gave us a real insight into the thought processes of pilots when a problem occurs.

But is the real problem that men (in particular) are more process driven and their instinctive reaction is to "do something?" Even if the something turns out to be completely the wrong thing to do!

Anonymous said...

I think pilots read accident reports and say, "I would never do that." But under the right circumstances you might, so it's best to be humble.

I believe the training did not serve them well. Simulator training consists too much of procedures and checklists. Had they been thinking wider, they might have realized that at 1000 AGL with no engines it was time t think about an off-airport landing.

Scote1992 said...

The thing that got me was why they would bring the left engine back to idle before starting the right one, which was obviously not causing the problem (first b/c it was the air thing, and second nothing happened when they shut it down). Especially considering the low altitude they were at.

Anonymous said...

Both your posts are definitely harm reduction in action. Thak you for caring enough about the safety of others to take the time and parse the meaning of the sad event.

Splendor said...

Brilliant posts, as usual.

I'm a little baffled at the base design of the thrust levers - it seems to me that anything which can get pulled to off when you want idle, even if there supposedly is a catch, is a disaster waiting to happen.

The plan I fly has dedicated engine master switches which are very well guarded, there is no possible way of accidentally switching them off. We also have reverse and there again it's a separate handle on the thrust level, not a resistance detent behind idle, so selecting reverse accidentally is also physically impossible. Fail-safe, I hope...


Anonymous said...

" Just like on lots of Airbuses and Boeings, if one of the engines falls off they will leave the airplane with reduced thrust, but they won't cause the airframe to break up."

Tell that to the crew of the ElAl B747 that crashed off Amsterdam 20 years or so ago (was it that long?).

Engine "departed" complete with pylon (some bolts gave way, causing the entire assembly to break away from the wing), severely compromising the structural integrity of the wing and eventually causing the aircraft to crash into a highrise apartment building.

Of course that wasn't fire, but structural failure, but then again fire can cause structural failure (as you mentioned).

It's a tradeoff. Go on with the burning engine running in the hope to use what thrust it still provides, or try reducing the damage to the ship by pulling the fire levers now and hoping the remaining engines can get you to a safe spot for an emergency landing.

For the ElAl crew, it ended bad, they didn't live to tell their tale.
But a lesson was learned. If a flight crew states that "we've lost an engine", ATC shouldn't assume the aircraft has just lost power to it (which is where Amsterdam ATC went wrong, they severely misjudged the nature of the emergency based on that one line, or they'd have given the aircraft a different emergency approach, not aided by the apparent lack of panic on the flight deck as witnessed from their communications with ATC).

Aviatrix said...

Anonymous, I'll use "Engine failure, number two" versus "number two engine has separated from the airframe." ATC is still going to have trouble grasping that second one, though.