Why does putting Eng Anti-Icing on (increase air bleed use) help control engine rpm
 

Unloads the compressor stage, depending on engine type it stabilizes bleed demand and avoids rapid swaps between HP and LP bleed sources on the engine at the high altitudes.

It would help if we knew the engine/aircraft you're talking about.
BTW, is it safe to assume by "high density altitudes" you mean high altitudes/low density? Because "high density" usually means low altitudes.

Regional technical language barrier.. at least in North America, there's the term "density altitude" meaning what the altitude would be, on a standard day, that corresponds to the current density of wherever we're talking about.

So his situation (a "high density altitude") is a density altitude that is high, i.e., low density.

Sure On and not Off.
 

I am puzzle by the instruction you read from pilot's notes for prevention of over speeds, my understanding:-
In a multi spool compressor using air off take would unload the forward spools causing them to speed up!
Same config at T/O would decrease thrust if an increase in fuel/EGT were not made!
Same config. a low air density (high altitude) would also cause front spools to run faster and perhaps over speed!
The only good reason for using air off takes,as far as the engine is concerned, is to improve the compressor stall margin for engine!
Welcome corrections.

The C-17 engine is derivative of the PW 2000 used on the B757. In the C-17 they are flat rated to 40,400 lbsf to 30 C. I am sure there are other differences, but the biggest obvious difference is the C-17 has a core thrust reserver in addition to the fan reserver that the B757 version has.

The thrust reversers are used on the ground to back the aircraft, a useful feature in an unimproved landing zone. The C-17 can back up a 2% grade at up to maximum gross weight (585,000 lbs). This capability was demonstrated during development on McChord AFB taxiway H. C-17 flight crews routinely back the aircraft into parking using reverse thrust.

The flight manual (T.O. 1C-17A-1) states that while backing the aircraft, further protection from engine surge may be achieved by selecting engine cowl anti-ice on for the period of time reverse thrust must be used above idle. I have seen similar procedures (turning on engine anti-ice) to clear compressor stalls in other aircraft.

I've always thought that applying a bleed load like the engine anti-ice reduced the compressor load (meaning that the compressor wasn't producing the same compression ratio). However, it would seem that the engine control would compensate to keep the engine operating at the same RPM.

Would be nice to hear from a propulsion engineer. Thank you

Sorry, I'd missed when Hueymeister posted the engine in question (I've been traveling a lot lately).
Never worked the C-17, but I was responsible for the PW2000 on the 757 for several years - the differences the F117 and the PW2000 are fairly minor.
As noted, bleeding air from the engine unloads the compressor and generally improves the stall margin. While the FADEC will work to keep the rotor speed constant, there are response times involved - the lag between when the FADEC senses a speed change, the fuel flow responds, and the resultant change in rotor torque can not only allow the rotor speed to change significantly, it can actually result in an oscillatory speed response.
Educated guess here, but I'm thinking that, under the circumstances in question, the HP compressor is on the edge of stall - and the variations in flow separation in the compressor can cause the rotor speed to fluctuate more rapidly than the FADEC can compensate for - possibly even resulting in an oscillatory response as noted above. Turning on anti-ice bleed unloads the compressor sufficiently, moving the compressor away from stall and helping the FADEC keep the rotor speed constant.

I doubt if bleed extraction changes the performance of the speed control loop in any significant way. If the engine is rated on LP spool speed (I'm assuming here...), maybe there is a debit applied when bleed is requested (to reduce turbine temp) which decreases rated LP speed - thus giving greater margin to NL:max limit..

Also worth noting that at an LP speed, extraction of bleed (from the HPC) would increase HP speed, so unless the debit idea above is true, it would seem that putting AI on reduces margin to NHmax....

I don't understand the use of the term "rated on" How does this relate to a term describing a controlling function burried in the FADEC? like pressure output or temperature generated?

For a given pilot demand (throttle lever position) engines are controlled to ('rated to', or 'rated on') a looked-up value of a particular measured engine parameter. That looked-up value is the result of much modelling and discussion between Engine and Airframe people, and relates to delivered power or thrust at a particular flight case.

Some engines are controlled to LP speed as their prime rating parameter; some are rated to EPR; some are rated on turbine exit temperature.

OK

but it seems to me the issue from the OP is not related to a pilot controlling action directed at LP speeds, so I assumed the question was about what the engine control did or did not do. and the how and whys.

The PW2000 and F117 engines are controlled to EPR at power, not N1 (assuming normal operation - there are failure modes where control reverts to N1).
N2 speed control is used at/near idle, as is minimum burner pressure (PS3) (i.e. at idle, the control will control to N2 or PS3, doing a 'select high').
Turning on bleed doesn't change the control loop response, but if the N2 rotor is flirting with separation or stall (not an uncommon problem with this engine type, mainly the earlier builds), the rotor speed changes can simply be more rapid than the control can respond to - there are significant lags in the time between the control senses N2, the fuel flow changes, and the resultant torque change results in an N2 speed change. We're talking fractions of a second here, but that's all it takes to set up an oscillatory rotor speed response. When controlling to EPR, the time lags get worse, because of the inherent lags in sensing EPR compared to rotor speed, making an oscillatory response even more likely.

I was talking about what a controller might do at high power in response to AI being selected. No pilot action was involved - sorry if I gave that impression.

OK - so we are taking about an engine rated on EPR rather than N1, but the principle is similar. The OP mentions AI is claimed to be advantageous in mitigating overspeed. We are asked why. One way that this could be true is if AI selection dials in a rating debit (an EPR debit in this case) which would reduce rated power and hence increase speed margin (margin which may be partially consumed during the transient response onto the EPRrating from idle).

It was an idea - not necessarily the right answer .... Are rating debits applied for AI for this engine at takeoff?

Yes for the PW2000, I'm not sure about the F117, but since it's rated a little higher (43k vs. 40k for the PW2000) I would expect it does.