Necessary components to build a generalized engine simulation?

[Voyager]

Once again, not sure if this is the correct forum for speculative theorizing, but here we go again... 

 

So, from recent discussions on engine behavior in game, it sounds like the current engine engine is focus on modeling to the pilots' manuals, but does not yet predictively model engine operation and failure modes. I've been thinking for a while that an engine detonation detonation engine may be one of the next areas of WWII flight sim modelling, with persistent wear models. 

 

With that, I'm wondering, what core algorithmic engines, and what core primary source material would be needed to develop such a model and populate it for the aircraft in the sim? 

 

First thought would be likely a solid catalogue of the ways piston engines likely fail. Postulating the core ones would be failure to deliver torque to the prop, catching fire, and rupture spawning damaging debris.

 

I'll keep putting together thoughts and continue this later, but am interested in what other people think. 

[AndyJWest]

I'm not sure that cataloguing failure modes will tell you much. What really matters is the probability of such failures under specific conditions. Possibly the modern automotive industry can simulate such things, but the degree to which this can usefully be done in real time in IL-2 GB, with limited CPU resources, is another question. Likewise, the developers only have limited resources, and it might be a mistake to put too much effort into creating a model which would of necessity have to be based on a very incomplete historical record. So while I'm sure that engine modelling is still open to improvement, and while I suspect that the developers may wish to look intro this further at some point, I'd caution against trying to do the developers job for them. If you have specific suggestions, there is a subforum for it, but a long list of things we think 'should' be modelled might not go down well. Particularly when we don't really know how the existing engine modelling works anyway.

[Voyager]

First stage, ID failure modes. Later stage, put together causes. Overall objective: build an algorithm for converting engine inputs, from fuel to shells, into engine outputs, with primary source documentation to support. 

 

Was also thinking of this as more a forum project that could be useful in the future rather than a "devs do this now". Flight model documentation took years to get to the caliber things are at now, before this game was even a glimmer in the devs' eyes; figure engine model long of this degree could easily take just as long. Or we might turn up handy tools that make it simple? Never know without looking, and figure more fun than debating semantics of pilots' handbooks. 

[=475FG=_DAWGER]

It is actually pretty simple. Heat or pressure causes all failures related to how the engine is operated. Engine limitations (What is written in  manual) are developed to enable the engine to reach its service life, not prevent an instantaneous failure.

 

If engines were properly modeled, we could abuse the heck out of them because we get brand new ones every sortie and it is pretty hard to cause catastrophic failure on a new engine.

 

In my mind that is the way they should model the engines and maybe build some layers of complexity later to simulate wear and tear.

 

I personally have experienced several engine failures in Pratt and Whitney radial engines. The first was an exhaust valve that broke because the engine had sat in a rice field for a few years before they hung it on the airplane I was flying. The valve bounced around inside the cylinder until it punched a hole in the head and the resulting heat buildup in that single cylinder slowly deteriorated engine performance over a 45 minute period until it was producing about 15% power when I landed. It took 130 hours for the corroded valve stem to fail. The corroded area produced a hot spot that weakened the stem over time.

 

The second was a cylinder head that separated from the barrel as the result of an over pressure event. The thing is the over pressure occurred about 200 hours earlier. I got caught in a microburst on a missed approach and had applied "radar power" to stay out of the trees. Months later the cylinder gave way, breaking the interconnecting oil lines. It took about 20 minutes to pump 8 gallons of oil overboard. I shut the engine down when it got below 20 psi.

 

WEP limits are designed to allow repeated applications without causing undue wear and tear. There are not hard stop failure limits. In truth, hardly any limits in an airplane are hard stop limits in that sense. Most limits are there to keep the system serviceable in the long term.

 

Games have, for at least two decades, interpreted those operating manuals as being filled with hard stop, binary limits. Engines blow up, wings break in half, gear legs rip off, flaps tear away never to be seen again all in a very unrealistic fashion. The reality is abuse now causes problems later. Hard to simulate when you are issued a factory fresh bird every sortie.

 

I would hope anything daring to call itself a simulator would attempt to understand this and model engine performance limits much more realistically but they don't. It may be a game play decision to keep folks from getting silly or it may be a lack of understanding of what pilot handbooks actually are. Possibly both.

[Herne]

3 hours ago, Voyager said:

Once again, not sure if this is the correct forum for speculative theorizing, but here we go again... 

 

So, from recent discussions on engine behavior in game, it sounds like the current engine engine is focus on modeling to the pilots' manuals, but does not yet predictively model engine operation and failure modes. I've been thinking for a while that an engine detonation detonation engine may be one of the next areas of WWII flight sim modelling, with persistent wear models. 

 

With that, I'm wondering, what core algorithmic engines, and what core primary source material would be needed to develop such a model and populate it for the aircraft in the sim? 

 

First thought would be likely a solid catalogue of the ways piston engines likely fail. Postulating the core ones would be failure to deliver torque to the prop, catching fire, and rupture spawning damaging debris.

 

I'll keep putting together thoughts and continue this later, but am interested in what other people think. 

 

Unless I'm mistaken, which is entirely possible, I thought I read either in a dev post or a DD recently, something regarding detonation, the code has apparently been in the engine since Rise of Flight, but for some reason was not enabled.

When I'm at work tomorrow I will try and find the post.

[Rattlesnake]

6 hours ago, Voyager said:

Once again, not sure if this is the correct forum for speculative theorizing, but here we go again... 

 

So, from recent discussions on engine behavior in game, it sounds like the current engine engine is focus on modeling to the pilots' manuals, but does not yet predictively model engine operation and failure modes. I've been thinking for a while that an engine detonation detonation engine may be one of the next areas of WWII flight sim modelling, with persistent wear models. 

 

With that, I'm wondering, what core algorithmic engines, and what core primary source material would be needed to develop such a model and populate it for the aircraft in the sim? 

 

First thought would be likely a solid catalogue of the ways piston engines likely fail. Postulating the core ones would be failure to deliver torque to the prop, catching fire, and rupture spawning damaging debris.

 

I'll keep putting together thoughts and continue this later, but am interested in what other people think. 

Detonation should be complex but calculable. The difficulty I'd think would be getting an accurate idea of exactly how long it takes cylinder head at X atmospheric conditions to reach the temp at Y power setting that will cause detonation at Z MP, make sense? I have no idea where one would even start getting a ballpark figure on that for all these airplanes.
Persistent engine wear? Nah.
Frankly if every plane in the game were able to run at it's WEP till it ran out of gas that would be just fine by me. The relative speed and climb rate differences would remain fairly constant, and we could concentrate on maneuvering the airplanes. And since engines were frequently run way longer at WEP than called for without breaking it would arguably be just as "realistic" as current system.

[Rebel_Scum]

3 hours ago, CMBailey said:

Frankly if every plane in the game were able to run at it's WEP till it ran out of gas that would be just fine by me. The relative speed and climb rate differences would remain fairly constant, and we could concentrate on maneuvering the airplanes. And since engines were frequently run way longer at WEP than called for without breaking it would arguably be just as "realistic" as current system.

An excellent idea until something better comes along.  

8 hours ago, =475FG=DAWGER said:

WEP limits are designed to allow repeated applications without causing undue wear and tear. There are not hard stop failure limits. In truth, hardly any limits in an airplane are hard stop limits in that sense. Most limits are there to keep the system serviceable in the long term.

 

Games have, for at least two decades, interpreted those operating manuals as being filled with hard stop, binary limits. Engines blow up, wings break in half, gear legs rip off, flaps tear away never to be seen again all in a very unrealistic fashion. The reality is abuse now causes problems later. Hard to simulate when you are issued a factory fresh bird every sortie.

 

I would hope anything daring to call itself a simulator would attempt to understand this and model engine performance limits much more realistically but they don't. It may be a game play decision to keep folks from getting silly or it may be a lack of understanding of what pilot handbooks actually are. Possibly both.

This.

[Rattlesnake]

At least one WWII game used to automatically switch WEP off when limits were reached and auto-retract flaps when over-sped. When you think about it that seems like a reasonably compromise.

[Guest deleted@83466]

7 hours ago, =11=Herne said:

 

Unless I'm mistaken, which is entirely possible, I thought I read either in a dev post or a DD recently, something regarding detonation, the code has apparently been in the engine since Rise of Flight, but for some reason was not enabled.

When I'm at work tomorrow I will try and find the post.

 

If true, that would be great.  Flying around in the P-47 this afternoon, thinking about octane, and mixture, and carb temps and stuff like that got me thinking about engine knocking too.  I know these planes aren't 'Accusim' aircraft like some of us have flown in FSX, but I think some additonal things like detonation would be a cool addition to the simulation.

[Matt]

I think they should just get rid of the time limits. Then eventually add a more complex damage modelling later.

 

I still don't understand what the problem is. Fuel load and limited supply of MW-50/water (on planes that use that) already limits the time you can fly at full power. Sure, some planes can theoretically fly around for longer than 1 hour at maximum power, but then people would need to carry a ton of fuel. So in the end, most of the time you would be better of using less fuel and not running WEP all the time.

[Legioneod]

5 hours ago, Matt said:

I think they should just get rid of the time limits. Then eventually add a more complex damage modelling later.

 

I still don't understand what the problem is. Fuel load and limited supply of MW-50/water (on planes that use that) already limits the time you can fly at full power. Sure, some planes can theoretically fly around for longer than 1 hour at maximum power, but then people would need to carry a ton of fuel. So in the end, most of the time you would be better of using less fuel and not running WEP all the time.

 

Agreed. Here's something I posted in another thread, while it's still a simplification I think it'd be a better model than what we currently have.

 

 

 

"Here's an idea for the engine model. It's still a very very simplification but it'd be much better than the current time limit/ cooldown we currently have.

 

Why not have temperature be the overall limiting factor for time at power? Use the current temperature model for each aircraft and use it to determine engine life. The hotter you run the engine the more likely it is to cause detonation and the more likely it is to fail. 

This would practically eliminate the timers and cause heat to be the concern, players would be forced to reduce power once engine overheats or risk failure.

 

Water/WEP could work in a similar way, if you run out of water there is no more WEP and the same heating mechanics would apply though the overall temps may be different due to the use of water."

Edited November 27, 2018 by Legioneod