feat(bleed): Implement wing anti ice system

[omrygin]

Fixes #[issue_no]

Summary of Changes

This PR tracks the implementation of the wing anti ice consumer.

• Pneumatic & thermodynamic logic:

  • Controlled valve and container for the slat pipes
  • Valve connected to the precooler
  • Valve pressure regulation to 22 PSI (source needed for NEO)
  • Static exhaust to the ambient atmosphere
  • Valves operate pneumatically
  • WAI correct pressure
    • When valve is closed
    • When valve is open

• Misc:

  • Ground 30 seconds test logic
  • Simvars documented

Even though some tests are not checked, it doesn't necessarily mean no tests were run on that particular subject. It just means that I am still running tests.

• Tests:
  • Unit tests:
    • WAI valves normal operation
    • WAI consumer pressure
    • Ground 30 seconds test logic
    • Simvars
  • In-sim tests:
    • Ground 30 seconds test
    • WAI valve normal operation
    • WAI consumer pressure
    • Simvars

Screenshots (if necessary)

References

Additional context

Discord username (if different from GitHub):

Testing instructions

How to download the PR for QA

Every new commit to this PR will cause a new A32NX artifact to be created, built, and uploaded.

1. Make sure you are signed in to GitHub
2. Click on the Checks tab on the PR
3. On the left side, click on the bottom PR tab
4. Click on the A32NX download link at the bottom of the page

[BlueberryKing]

Great work so far! I really like this StaticExhaust concept and want to be able to use it in my code, so I suggest focussing on the physical aspects of the system for now and leaving out the ECAM logic. Some basic overhead pushbutton interaction logic is obviously needed, but let's leave out the FAULT light logic in this first step, so I can merge this soon.

Also, you don't have to invest too much time into tuning parameters. I think we should wait with this until most of the system is complete to avoid having to re-tune the entire system.

Let me know when you think the fundamentals are in with some unit tests, so I can review the system as a whole. I want to avoid leaving feedback on stuff you're still working on. Of course, you can always ask if you need an opinion on something. :)

[omrygin]

Thanks for the comments.
I agree with what you said, and removed the ECAM and FAULT off the checklist. I unticked the Static Exhaust implementation, as there is a conceptual issue with the temperature when it reaches a pressure equilibrium.
I'll think and come up with a solution, and then we can discuss this.

[omrygin]

There was a conceptual issue with the StaticExhaust that I have a way around, though it is not totally self consistent physically.
What I did so far was to move air from the consumer to the ambient atmosphere via method change_volume, which assumes an isentropic process.
When applying it to the exhaust when the valve is closed, the anti ice and the ambient atmosphere reached pressure equilibrium, however their temperatures were different. This of course cannot be true as the ambient atmosphere is a huge thermal bath.

This happens because the exhaustion is not isentropic, as the air must lose heat to reach temperature balance.

So I had to model the cooling of the air inside the WAI container, separately from how the pressure change. This is not totally self consistent.

I used Newton's cooling law as presented here, with a heat transfer rate (tau in wikipedia) which will need to be tweaked.

[BlueberryKing]

This sounds good to me. The whole isentropic idea is mostly just so we can use simple formulae for the transfer of air between containers. We might even want to apply Newton's cooling to all containers at some point.

[omrygin]

It's been a productive couple of days!
I think basic unit tests are done. The valves are now pneumatically operated, meaning that if the precooler is not pressurized, the valves won't open and the pushbutton will light FAULT.
This is noisy (i.e. the FAULT light flashes in-sim) due to the unstable PID. Perhaps when moving to Davy's mod it will be better. But this is easily fixed.

I have noticed that the precoolers are not symmetric - have you noticed that?

Also, I commented out the pressure change for temperature change. We need to discuss this as I think there is an error there that causes my pressure values to be NaNs

[BlueberryKing]

Awesome! Really great to see it. I think it's realistic anyways to use some sort of latching logic for the fault light to prevent the flickering.

I have noticed that the precoolers are not symmetric - have you noticed that?

Oh really? In what way do you mean not symmetric? Not the same for both engines somehow?

Also, I commented out the pressure change for temperature change. We need to discuss this as I think there is an error there that causes my pressure values to be NaNs

Sure! I don't actually think this issue was introduced with my "pressure change for temperature change" commit, but is an issue caused by the update_temperature_for_volume_change method. Namely this happens because we try to remove more volume from the pipe than is actually in it, meaning we are raising a negative number to a rational power, which causes the NaN. This happens while the system is still stabilising in the simulator. I have actually introduced a workaround/fix for this, but have unfortunately not pushed this yet. The fix is just to move the volume incrementally in a while loop, moving at most the volume of the pipe in each step. The underlying issue, in my eyes, is that we are treating volume as an "amount of fluid", which is - as you have previously pointed out - not ideal, as we can see here. This is why I've introduced this fix for now.

[omrygin]

A couple of things:

1. If we look at update_move_fluid, then the way volume is moved is

self.move_volume(
         from,
         to,
         self.open_amount()
             * equalization_volume
             * (1. - (-Self::TRANSFER_SPEED * context.delta_as_secs_f64()).exp()),
);

If context.delta_as_secs_f64() is small, then the amount of fluid moves in a single tick is approximately self.open_amount() * equalization_volume * (Self::TRANSFER_SPEED * context.delta_as_secs_f64()))
Which is already small, so I wonder how different that would be with the while loop. Or perhaps I misunderstood you.

2. In update_pressure_for_temperature_change, you basically wrote

P_new = P_old * (T_old/T_new) ^ (gamma / (gamma-1))

However, from here, it seems that the exponent should be gamma / (1-gamma), i.e. negative what you wrote. High temperature should correspond to a higher pressure due to the frequent collisions of the particles with the wall. Since gamma is always greater than unity, in the formula you have used as T_new gets bigger, P_new gets smaller.
Do you agree?