APU air inlet door

[Hardy Heinlin]

I too think in the last videos it's hard to see when the door exactly starts and stops. The 767 video is clearer; I guess when the 767 video begins, the door is about 3/5th of full open and reaches 5/5th after 8 seconds.

Another question ... can we assume that the APU's airspeed dependent N1 and N2 windmilling RPMs are similar to those of the engines, and that they vary slightly with the door angle? The EICAS indications should be plausible when the ground test is enabled in flight while the door is open and the APU is shutdown.


|-|ardy

[Avi]

Hi,

I played a little bit with the new APU malfunctions and have 2 things to say.

With the APU off, I activated the door malfunction and tried to start the APU. As expected it didn't start but when I repaired the malfunction, it did. I don't think the APU should start before setting the selector to START again.

I understand that after activating and repairing the N1, N2 and low oil pressure malfunctions and re-starting the APU there is a test cycle (start, cooldown and shutdown which removes the inhibit status if everything is ok – John reply 10) but I noticed that after the APU is started during this cycle, we get the APU RUNNING memo message and AVAIL lights for a fraction of a second and then they are removed (it doesn't look good).
I don't think we should get them if it is only a test cycle. A short delay may be needed here.

Cheers,

[Hardy Heinlin]

Hi Avi,

I had noticed this too -- at 95% N1 the generators and the APU bleed air valve get active for a second. In the next update I'll inhibit this during the test.

Regarding the door: When the start sequence is waiting for the door to open, and it doesn't open, it keeps waiting ... and when it opens, the start sequence continues. I don't know if there's any timeout.


Cheers,

|-|ardy

[John H Watson]

Another question ... can we assume that the APU's airspeed dependent N1 and N2 windmilling RPMs are similar to those of the engines, and that they vary slightly with the door angle? The EICAS indications should be plausible when the ground test is enabled in flight while the door is open and the APU is shutdown.

Plausible, but you would have to ask the experts if a normal jet engine (designed for thrust) windmills differently from an engine designed primarily for bleed and electrics?

The APU has quite a few quirks:

The N1 and N2 rotors are counter-rotating.

The APU air inlet feeds air into the middle of the engine. Some air goes forward to the "load compressor", some air goes backwards to, perhaps, a more traditional type engine with compressor and turbine stages. I don't know how much the air going into and out of the load compressor would affect a windmilling engine: The load compressor supplies air to the bleed system. The airflow going into the compressor is controlled by variable inlet guide vanes. These vanes allow more/less air into the compressor depending on bleed air load and ambient conditions. The air coming out of the compressor is controlled by a "diverter valve" which dumps excess air directly into the tailpipe (especially noticeable when you switch all the packs off at once).

[Hardy Heinlin]

I think for the windmilling I'll just use some random values ...


Re no start due to closed door:

With the APU off, I activated the door malfunction and tried to start the APU. As expected it didn't start but when I repaired the malfunction, it did. I don't think the APU should start before setting the selector to START again.

Do you think there should be a timeout of, say, 60 seconds after which the start inhibit mode should activate?

In version 10.30 the "closed door" detection (<13°) will only activate the start inhibit mode when a normal start has been in progress for at least 8 seconds (N2 rising etc.). Otherwise there is just no start in progress. If there is nothing in progress, there is nothing to stop.


Cheers,

|-|ardy

[simonijs]

Just came across this document on the internet (http://203.72.21.28/arm/images/files/Download/001-B747-400-05/49___045.PDF), which is a 747 Maintenance Manual Chapter 49 APU (China Airlines). CH. 49-15-00: APU Air Inlet - description and operation.

I believe, KLM's FCOM now also contains the 2 % fuel penalty. In the obsolete Aircraft Operations Manual Volume III, however, this fuel penalty was not mentioned. Instead, it read: "Decrease maximum engine out altitude by 1000 ft."  (Condition: In Flight, APU DOOR remains displayed)

Regards,
Simon

[Hardy Heinlin]

By the way, I've seen this same drawing of the open door in other documents. It looks like the door opens towards the inside. But as we can see on the videos it moves towards the outside. Is anything wrong with my eyes? :-)


|-|ardy

[Avi]

Do you think there should be a timeout of, say, 60 seconds after which the start inhibit mode should activate?

I don't know. You may be right that the APUC simply waits for an OPEN signal.

It looks like the door opens towards the inside. But as we can see on the videos it moves towards the outside. Is anything wrong with my eyes? :-)

Probably yes :-)
It may move outside a little bit during transit but when it is open, it opens inside (unlike B767 and B777).

Transit:


Open:


B777:


Cheers,

[Hardy Heinlin]

Aha, it's different on the 767 and 777. That was confusing me :-)


Another question, maybe related to engine start too:

When we have a strong headwind at the gate and N2 is windmilling way above zero, what effect is to be expected when the starter motor mechanically connects to the gear shaft while the starter's own RPM is initially zero? Will it brake the N2 rotor momentarily? Or will it only connect when the RPMs are equal? Or is the system so flexible and quick that there's no special effect?


Cheers,

|-|ardy

[torrence]

747 APU Door opens in vs out question:  I assume this affects the fuel burn penalty for a stuck door at some level?  I'd naively figure that the 747 door opening in would be somewhat less of a drag problem that the 777 out position, but don't know if it's a major effect.

Cheers,
Torrence

[Hardy Heinlin]

In version 10.31 I added a drag that causes a fuel penalty of ca. 2% as mentioned in some 744 manuals. In THR REF climb and idle descent there are no fuel effects, but the climb will be shallower and the descent steeper in order to maintain the command airspeed. Cruise requires more thrust to maintain the command airspeed. Consequently, max altitude is a little bit lower. The FMC doesn't know this.

The effect varies with the door angle. 2% refers to 45° open (inward).


Cheers,

|-|ardy

[John H Watson]

In version 10.30 the "closed door" detection (<13°) will only activate the start inhibit mode when a normal start has been in progress for at least 8 seconds (N2 rising etc.). Otherwise there is just no start in progress. If there is nothing in progress, there is nothing to stop.

I don't understand your logic here. The APUC will not even begin the start process if the door is not open at least 13 degrees. There will be no turning at all. The APUC controls the starting relays. This is stated in my engineering training notes and in the (BA) AMM.

If the <13 degree switch is activated at any time with the with APU running, this will cause an autoshutdown. i.e. the fuel valves will be commanded closed.

[Hardy Heinlin]

Yes, it will not even begin. That's why nothing happens. Once started with a normal door angle, the door can only be moved towards closed by an add-on through the PSX network -- theoretically (not recommended). PSX alone cannot move the door to a non-normal position, it can only keep it at a position. Those 8 or so seconds before are just sort of an APUC power-up phase.

[John H Watson]

Roger... Now understood.

When we have a strong headwind at the gate and N2 is windmilling way above zero, what effect is to be expected when the starter motor mechanically connects to the gear shaft while the starter's own RPM is initially zero? Will it brake the N2 rotor momentarily? Or will it only connect when the RPMs are equal? Or is the system so flexible and quick that there's no special effect?

I haven't got much detail on the APU starter motor except that it has a "sprag" type clutch which engages and disengages at specific rpms (disengages at 69%, reengages at 47%). Surely an electric starter can't be engaged all the time at these speeds? Wouldn't it act as a generator?
Google mentions that sprag clutches engage in one particular direction, but not in the other. Perhaps the clutch only engages when the starter motor catches up with the speed of the windmilling engine? But why then would starter re-engagement speed be so critical? My notes say that the APUC will not activate the starter until the N1 and N2 has fallen below 20% for 20 seconds ( I assume to allow the N1 and N2 to reach near zero)

I'm not sure how much windmilling there will be on the N2 stage. The N2 stage drives the accessory gearbox. This powers an oil pump and a fuel pump.

[Hardy Heinlin]

I've noticed in some EICAS videos (APU spool-down) that N1 and N2 almost stop decelerating when N2 drops below 20% and the door closes. 45 seconds after spool-down start from 100% N1, the EICAS APU data blanks. At that time N2 is still at ca. 6 %. It must take some minutes to drop to zero. I guess 6% N2 is a normal RPM at which the APU can be restarted. I set the windmilling force so that it generates about 6% when the headwind is 70 kt (in PSX max surface windspeed). When the starter spools up, N2 remains at 6% until the starter goes above 6%.

[John H Watson]

It must take some minutes to drop to zero.

I wonder what causes this slowdown in acceleration. Looking at the same video, the slowdown is rapid down to about 12% (more than 1 rpm/sec) , then the slowdown reduces dramatically.  Perhaps the stages stop compressing air at these lower speeds (i.e. stop doing work)

What do you use as a normal door close signal after a normal shutdown? Supposedly, the door is commanded closed at the end of the cooldown period (60 seconds after the APU selector goes to OFF). The Analysis page shows the door closing when the rpms are quite low.

[JRBarrett]

I'm not sure how much windmilling there will be on the N2 stage. The N2 stage drives the accessory gearbox. This powers an oil pump and a fuel pump.

It's my understanding that the core of the PW901A APU was derived from the PWJT15-D turbofan engine, extensively used on small biz jets like the Cessna Citation, Beechjet etc. The JT15-D is somewhat unique in that it uses a single-stage centrifugal compressor - no axial stages. When a JT15-D is used as an aircraft engine, it does of course have a fan stage ahead of the compressor inlet.

It's very common to see aircraft engine fans spinning from wind blowing through the bypass duct, since the N1 fan is free to turn, and it doesn't take much breeze to get it moving. If the wind is coming from the rear of the engine, the fan will happily spin backwards.

I'm not sure if the APU on a 747-400 would windmill at all from anything less than a hurricane-force wind impinging directly into the open inlet. The centrifugal compressor has a lot of mass, and there would not be the same kind of direct path for a wind to travel "through" the APU, like there is when wind travels through the bypass duct of a turbofan engine.


Edit HH: Quote tags corrected.

[Hardy Heinlin]

JR, thanks for your input; it agrees with my model: My APU windmilling RPMs are much lower than those of the engines.

JHW, my door model closes below 20% N2. I think this agrees with the books? I think it should allow some air flow through the rotors during spool-down to get rid of the heat and to allow a smooth deceleration. If the door closed at 95% N1 already, the system would probably get a surge effect because there is no air inlet anymore at high RPMs.


|-|

[John H Watson]

JHW, my door model closes below 20% N2. I think this agrees with the books?

After further reading, I did find reference to 20%. It didn't sound right closing at 100%, but I found two references saying it closed after the cooldown period.

For future reference, the training notes (Elec Book 9/ page 150) do say:

Normal Shutdown Sequence

Following a 20-second time delay [? ? ?], both [FMU] solenoids are de-energized and a check of N2 speed is made to determine if the Shutdown Test passed...  [lots of technical babble follows]

.... After N2 or N1 is less than 20%, the APUC performs the following:

*APU Inlet Door to close
*APU Fuel Valve to close (There are two other fuel valves in the fuel metering unit which are shut off earlier)
*APU DC Fuel Boost Pump to off
*Low Oil Pressure test (after time delay of 20 secs)
*APUC powered down (after time delay of 156 secs)

Oddly, the abnormal shutdown sequence mentions the same 5 events, but there is no reference to N1 or N2

*Primary [FMU] Solenoid to OFF
*Stepper Motor set to min flow
*Secondary Solenoid to OFF (after 3 second delay)

The shutdown output causes the following:

*APU Inlet Door to close
*APU Fuel Valve to close
*APU DC Fuel Boost Pump to off
*Low Oil Pressure test (after time delay of 20 secs)
*APUC powered down (after time delay of 156 secs)

[Hardy Heinlin]

... After N2 or N1 is less than 20%, the APUC performs the following:

*APU Inlet Door to close
...

Yes, N2 or N1. I check N2 alone because in my spool-down model N2 is always the first parameter to drop below 20%.