747-8 787 with GE GEnx engines - icing

Yeah, the new GEnX engines have had quite a few rollbacks (temporary power losses). I believe Cathay Pacific lost either 3 or 4 engines at once a few months back. This month Atlas lost 3 or 4 engines also. I believe both were in cruise flight. In both cases engine damage was found after inspection. While at Boeing a few days ago, the reps were talking about these problems. There's a new version of the engine, that has a better fuel consumption. Approx 1.8% I think. But the engines haven't been put on the wings yet as they haven't been FAA ice certified.

I was told that the FAA was proposing a 50nm rule instead of the current 20nm rule concerning thunderstorms. For those that don't fly in thunderstorm areas often, the 20nm rule is really never followed. It can be impossible to do. 50nm would really limit flying. That said, the 20nm rule is.....avoid thunderstorms by 20nm. Sort of a guideline. But the new rule for the GEnX may state: PROHIBIT by 50nm. And it gets worse. I believe it may state PROHIBIT flying within 50NM of RED or AMBER weather returns. There are areas of weather that aren't thunderstorms that show AMBER.

Also, the engine anti ice (EAI) doesn't help you in this case as only the lip on the nacelle is heated. Apparently the ice builds up somewhere behind the fan blades. I'm not sure if it's really known yet where the icing occurs. At NCA, we have been operating the air conditioning PACKS in Hi FLOW, which draws more air from the engine. This helps inprove the stall margin of the engine, but at a .63% fuel cost.

We do have a checklist on the -8 for suspected engine ice. I'll look it up and post it later.

Something does have to be done though. Losing 3 or 4 engines could make for a really bad day.

 

Yeah, the memory items for multiple engine failure is: Fuel Control Switch Cutoff, then Run. That's it as far as memory items go. You then run the checklist, which on the -8 involves deploying the RAT to keep flight controls.

But obviously airspeed is a concern. At heavy weights, a single engine loss would result in a descent to the high 20's. 2 engines out and you'd be flying pretty low. 3 engines out and you're landing somewhere.

The FMS does the calculations. You basically set max cruise power and when you slow to your "driftdown" speed, you descend at that speed at max cruise power.

In both the 400 and -8, there is no "Loss of all engines" checklist. There used to be in the 200. There is a ditching checklist though.

Both the 400 and -8 have auto-relight, so in theory they should start back up automatically. In the 400 we can select "continuous ignition" to keep the igniters firing. We use it occasionally in turbulence or heavy precip. The -8 doesn't have that switch anymore.

I found out the Atlas Air encounter in November involved 2 engines, #2 and #4. They were seperated by 10 mins. They lasted 20 secs each. So I don't believe they had to leave their cruise level. 6 days later the engines were borescoped and 3 engines received damage and had to be replaced.

I'm hoping a fix can be found. It's unrealistic to fly around most the clouds in the sky...

 

It would seem to me that they could run water pipes near the hot section and then inject a mist of warm water in the critical section, just warm enough to keep anything from icing.

 

That sounds logical and workable to me, I'm glad that I thought of that. Good idea, Jim.

 

Not Ice, but a true story of a 747 loosing engines and making a safe return

[ame=http://www.youtube.com/watch?v=OpuxqBp-CXw]British Airways Flight 9 (Jakarta incident) - Falling From The Sky - YouTube[/ame]


Lou is there something different about the GE engines on the -8 that make them "create" there own ice build up ?? Is the air expanded more rapidly in that section than other engines ? Is the 747-8 flying with the RR-Trent engine ? and is that experiencing the same phenomenon ?

 

Getting a lot of info from Boeing, NCA, and GE. The ICI (Ice Crystal Icing) is behind the main fan. They say the reason the 747-8 (GEnx-2B) has a higher risk of icing is because of the high flowpath angle change. ????

1st pic to show flowpath angle change......

Also a pic of where they think the icing is occuring.


For those not used to icing conditions, typical icing conditions are as follows:

Supercooled water droplets tend to freeze upon impact with surfaces cooler than freezing. The wings, tail, engine inlets, fan and spinner.

Ice Crystal Icing in the engine is like this:

Frozen ice crystals impinging on a warm surface in the engine.
Some crystals melt, wetting the surface.
Crystals impinging on the now wet surface stick, cool the surface to 0C.
Ice begins to form.

I don't believe the 747-8 is available with any other powerplant other than the GE right now. I could be wrong about that.
Image Unavailable, Please Login
Image Unavailable, Please Login

 

I still remember that BA incident. Almost took the whole plane down when they lost all 4 engines.

 

7 years later, KLM had a similar incident over Alaska with a B-744 of only 6 months old.

KLM Flight 867 - Wikipedia, the free encyclopedia

 

Good thing it did not crash. Very lucky.
I thought the regulators acted quickly after the first incident.

 

Great info Lou. Looks like an expensive fix for GE if the have to heat that section. Maybe they can do some aerodynamic magic just prior to that angle increase.

 

They might already have a fix, one of the nerd wed sites I follow had a picture of a new undelivered 747-8 in Nippon Cargo colours having its engines changed out at Boeing field, for some "unknown" reason. It could be the upgrade that GE promised for the engines to deliver the fuel burn that Boeing promised to the original customers, or the fix for this........or of course any one of a million other reasons

 

Ice only forms within a narrow temperature range. Usually, by the time the air gets through the first couple of stages of the compressor, the temperature increases enough that ice can't form anymore. Once you get to cruising altitude it's pretty cold out there (-50F + some mach effect but since it isn't above mach 1 it isn't a huge amount).

You don't want to introduce heat into the compression process since that would really whack the engine performance if you had to use bleed to get the heat. Hotter air takes more work to compress. The bigger the fan the lower the pressure ratio and the lower the pressure ratio the lower the temperature rise due to compression. Bigger fans turn at lower speed. This means the tip speed in the LP compressor is lower and you can't do as much work (pressure ratio) in those stages, so again less work equals less temperature.

Bigger fans have higher aspect ratios and lower hub speeds. In smaller diameter fans the hub speed is higher and this tends to centrifuge all of the water out of of the inner flow path and out the fan duct. In the JT-8D, almost no water went into the core even if you were blasting a fire hose into the inlet because of the centrifuging effect. With these higher aspect ratio fans that isn't happening and water droplets are getting into the core. Also there is a pressure ratio distribution between the hub and tip of the fan, the pressure ratio at the hub is lower, so again lower temperature leaving the hub area of these big fans.

The areas where they are saying ice is forming are the LP compressor second and third stage stators, the leading edges of the mid frame struts and the core inlet guide vanes. That doesn't have anything to do with the shape of the transition duct so that stuff about it being related to the shape of the transition duct doesn't make sense. It wouldn't take much ice buildup on those surfaces to create a power loss since ice would effect the stator shape, the exit flow velocity and mess up the incidence on the next stage.

It's not going to be an easy fix.

It is hard to anti ice those parts since they are thin metal. If the heat buildup was on the transition duct, due to the transition duct shape, they could heat a bit of the flowpath and take the performance hit when it was on, since it wouldn't get used much and the water in the inlet actually improves performance, but you have to be careful since if ice breaks off it could go into the core and do damage to the small blades in there. Once the air gets into the core it gets hot pretty fast since those early stages have a higher pressure ratio.

Rolls probably doesn't have the problem since they have a three spool engine and their LP compressor is spinning faster and the temperature rise in the first stages of that machine are higher.

Some high bypass turbofans have the core inlet down behind the fan, at a lower diameter, and that tends to keep stuff like FOD and water out of the core, but this engine seems to have a pretty straight inlet behind the fan so the water is staying in the core stream.

Like I said, no easy fix here.

 

As well as being very informative that was a really interesting read. Thanks I enjoyed reading that.

 

That is really some good information and well presented! Things have gotten beyond the " Suck, Squeeze, Bang, Blow" level of explanation. Thanks for that post.

 

Agree, good information. All events happened in cruise flight above 31 ,000 ft in conditions too cold for normal airframe icing. I believe they think the shape of the duct has an impact because most events were on the 747-8 engine, with only 1 or 2 on the 787. I believe the duct shape is a difference between the 2 engines.

I've read that the fix may be a "software" fix, with sensors to detect the condition and software to operate VBVs (Variable Bleed Valves). Not sure how true this is.

I've seen pictures of significant blade damage.

Happy Thanksgiving to all.

 

Blade damage from something like this can come from any number of causes, not just the release of ice in the flowpath. For instance there could be increased aerodynamic excitation of the rotating blades because of the larger vane wakes that the ice buildup on the vanes caused. Or the reshaped vanes (from the buildup of ice) could cause flutter of the downstream rotor if the incidence is pushed too far, or similarly there could be surge caused by the change in flow angles. Unless you have an instrumented engine it is very very difficult to figure out what causes it. You can usually figure out what stage broke first, but often it is very difficult to figure out what started the avalanche.

Opening the VBV's would make the engine run harder (increase rotational speeds and therefore pressures and temperatures) and that could indeed address the problem but it does seem that you might move it further forward and that getting rid of it might be harder than that.

In a perfect world you would try to have the temperature gain from below the icing range to above it occur within one rotating stage. Even if the water droplets don't totally vaporize, the temperatures downstream are too hot for any ice to form on the structure. I just don't know how you do that with the limited pressure ratio capability of a low speed stage. Stage pressure ratio for an axial stage is a function of tip speed. Since these LP compressor stages are running at fan speed the fan tip speed limits the shaft speed and you can't get much pressure ratio out of the LP compressor stages. With a high speed axial compressor you can get pressure ratios of between 1.6 and 1.8 (generally) and see a temperature rise of over 200 degrees F across a stage. With a low speed stage you are looking at a pressure ratio of only 1.1 or 1.2 if you are lucky, and with that low a stage pressure ratio the temperature rise across any one stage isn't going to be very big. In this case they are apparently seeing two stages where the there is a chance that icing can occur and that's hard to get rid of. If the speed is increased, the problem might just move forward in the LP, but not go away.

Tough problem

 

Very interesting read. Whatever the fix if they increase the rotational speed the fuel efficiency numbers will go down at cruise altitude. I bet it will be a mix between software and repostioning the VBV's so that the efficiency won't go down as much when in use. Those -8's really rotate at a low speed and climb like crazy. I love to watch them take off. I am no expert like some of you.

 

In thinking a bit more about it, it is going to depend on where they are dumping the bleed. There may well be an provision for anti-icing the splitter. The splitter is the piece just behind the fan that splits the fan and core airflows. If they turn on anti-ice protection for the splitter they can maybe reduce the icing along the flowpath.

In thinking about the fan speed, that isn't going to climb much no matter how much you bleed. Fan speed pretty much = thrust because the bypass ratio is so high. I don't think they can increase fan speed very much because the thrust would go up and you don't want that. So perhaps there is something that they are doing with the bleed that will provide the anti-icing that they need. All that happens is that the core speed increases and the engine gets less efficient, but the fan flow and speeds aren't going to change much.. There's got to be something else going on.

 

It's funny, from the cockpit the 747-400 sounds like a "jet engine" during the takeoff roll. The -8 sounds like a vacuum cleaner. Sort of a low rumble.

While I'm pretty sure it's impossible to tell whether you're being pushed or pulled, the 400 feels like you're being pushed through the air while the -8 feels like you're being pulled. Strange. Probably due to the different sound and possibly acceleration rate.

 

Solofast,

Love hearing your inputs. As a pilot, I have never thought of engines icing at altitude. Usually at SAT -40, you're good to go. Not enough moisture at that temperature and what is there is already frozen. So this is a new one to me. I have a lot of info and pictures I don't have access to right now. I'll post what I can later. Some of it is sensitive and I don't think I can post it here.

 

I know of 2 occurrences on the 787. I believe JAL has pulled them off certain routes. Im not sure about other airlines. The flow path angle I showed before was a thought as to why it happens more on the -8 engines according to GE.