Jump to content

[1.0.5] Advanced Jet Engine v2.6.1 - Feb 1


Recommended Posts

Essentially, my subsonic airliners would fall apart due to Mach effects if I don't reduce throttle as I speed up. I remember that in real life, airliner's maximum level-flight speed is limited by thrust + drag, not overheating or airframe integrity.

What's your dynamic pressure? If it's anything above 30 kPa, you're not at realistic cruise conditions. Also check your static TWR - airliners typically don't exceed 0.4 and are often considerably less.

Link to post
Share on other sites
Then how do you propose the engines would work? What evidence do you have to defend that? IRL, planes don't get hypersonic because of a number of reasons, not necessarily lack of thrust. I get many reports about how planes fly unrealistically, but how a plane flies is not something I can work on, only thrust and Isp.

Turbojets, ramjets and Turbofans must decelerate the incoming air to subsonic speeds (relative to the aircraft) before it enters the combustion chamber. As a result their net thrust drops as your airspeed approaches the exhaust velocity of the engine. Thermal issues aside, an engine which decelerates 1kg of air coming in at 1000 m/s to 300 m/s and then accelerates it to 600 m/s isn't going to produce net thrust.

Going mach 1 is particularly dubious using a high bypass turbofan from an airliner. Because they use a convergent nozzle, the highest velocity the exhaust can reach is mach 1 (obviously this is a bit more than 341 m/s because the exhaust has been heated by compression, but it won't be too much higher).

Link to post
Share on other sites

There might be something to that last point. AJE assumes that all of the available thermal energy of the exhaust is converted into kinetic energy, which isn't really the case for fixed geometry nozzles. Fixing it would require completely reworking the way that nozzles are handled though.

Link to post
Share on other sites

It might be an idea to verify the issue by having camlost recreate it in his save. To that end, miyuruasuka, would you mind sharing the .craft file that would be easiest to see this phenomena, and directions to replicate it? (I'm a noob to flying in ksp and not very good...>,> so I would appreciate it)

Camlost, I lovevthe mod... and I like that I can mess around with various aircraft, from props all thr way through jets! Thank you for your work!

Edited by Fizwalker
Link to post
Share on other sites
Turbojets, ramjets and Turbofans must decelerate the incoming air to subsonic speeds (relative to the aircraft) before it enters the combustion chamber. As a result their net thrust drops as your airspeed approaches the exhaust velocity of the engine. Thermal issues aside, an engine which decelerates 1kg of air coming in at 1000 m/s to 300 m/s and then accelerates it to 600 m/s isn't going to produce net thrust.

That is called ram drag and calculated of course.

There might be something to that last point. AJE assumes that all of the available thermal energy of the exhaust is converted into kinetic energy, which isn't really the case for fixed geometry nozzles. Fixing it would require completely reworking the way that nozzles are handled though.

There's a nozzle efficiency coefficient.

Link to post
Share on other sites

camlost, Seems that it still needs to be updated to support 1.0.2. Right clicking on an engine shows low fuel flow, constant Isp and no data on intake. So it seems that stock engine physics are used.

Link to post
Share on other sites
camlost, Seems that it still needs to be updated to support 1.0.2. Right clicking on an engine shows low fuel flow, constant Isp and no data on intake. So it seems that stock engine physics are used.

Some major under-the-hood changes are going on right now based on the stuff NathanKell mentioned 2 pages ago. If you're desperate for an update, the dev version on Github is mostly stable (based on my tests), but only stock parts are working right now - all the other configs need major changes that haven't happened yet.

Link to post
Share on other sites
It might be an idea to verify the issue by having camlost recreate it in his save. To that end, miyuruasuka, would you mind sharing the .craft file that would be easiest to see this phenomena, and directions to replicate it? (I'm a noob to flying in ksp and not very good...>,> so I would appreciate it)

Camlost, I lovevthe mod... and I like that I can mess around with various aircraft, from props all thr way through jets! Thank you for your work!

That is something I will definitely do once I get back home. It would be on 0.90 with the last stable version of AJE, though.

- - - Updated - - -

There's a nozzle efficiency coefficient.

It's a constant, though. In real life, for a converging nozzle, it should start near 1, and approach 0 as the speed increases.

Currently, there is no speed (according to AJESolverTester) at which a turbofan would start to produce "negative" thrust. The NetThrust decreases as ram drag increases, but after a certain point, the increasing thrust overpowers it, and goes into speedy ramjet mode. I'll demonstrate this once I get on my KSP computer.

Link to post
Share on other sites

This is a nice one, hoping to see it back so I can test it. I have couple of questions:

- With the new curves for stock jet engines, based on speed and atmosphere, how un-realistic it still is? (considering it has to be a simulation/game)

- Do you intend to implement some proper engine parts? eg: It always felt wrong to me how the game interprets that very stub part of concentrated mass an engine, when it should be just an exhaust.

- Turbo-fan and Props are in the plans, at least a code base so others and add models and parts to use it?

Link to post
Share on other sites
This is a nice one, hoping to see it back so I can test it. I have couple of questions:

- With the new curves for stock jet engines, based on speed and atmosphere, how un-realistic it still is? (considering it has to be a simulation/game)

In stock, thrust now varies with speed, which is an improvement, but Isp is now constant over all flight conditions (whereas it should vary a lot with speed and a little with altitude). The old but of intakeAir being counted toward Isp has been fixed, but the Isp of stock engines is still a lot higher than what you'd get in real life, particularly since the max thrust is more like what you'd get on afterburners for real engines.

- Do you intend to implement some proper engine parts? eg: It always felt wrong to me how the game interprets that very stub part of concentrated mass an engine, when it should be just an exhaust.

For most engines, AJE moves the CoM of the part closer to what it'd be for the real engine (outside of the model). Of course, there's still nothing preventing you from filling that space with fuel rather than leaving it for the actual engine, but that's a harder problem to solve.

- Turbo-fan and Props are in the plans, at least a code base so others and add models and parts to use it?

Turbofans work just fine and there are many already included in AJE. There's currently no solver in place for turboprops, but perhaps that will change in the future.

Link to post
Share on other sites
That is something I will definitely do once I get back home. It would be on 0.90 with the last stable version of AJE, though.

- - - Updated - - -

It's a constant, though. In real life, for a converging nozzle, it should start near 1, and approach 0 as the speed increases.

Currently, there is no speed (according to AJESolverTester) at which a turbofan would start to produce "negative" thrust. The NetThrust decreases as ram drag increases, but after a certain point, the increasing thrust overpowers it, and goes into speedy ramjet mode. I'll demonstrate this once I get on my KSP computer.

That's cool! I still havr my RSS/RO install and a stock install

Link to post
Share on other sites

I've thought about the question of subsonic engines and nozzles a bit and I'm proposing the following solution:

First, how it currently works: The exit velocity is calculated assuming an isentropic expansion to ambient pressure. The exit pressure is ambient unless the flow chokes.

My solution would have two nozzle types: convergent and adjustable. A convergent nozzle would have its exit velocity capped at the speed of sound. Exit pressure would be ambient unless the flow chokes (same as current). Adjustable nozzles would operate the same as convergent nozzles at subsonic speeds. At supersonic speeds, the exit velocity would no longer be capped and the exit pressure would always be ambient (since the divergent part of the nozzle expands the flow beyond sonic conditions).

If this solution is agreeable to everyone, I'll implement it. It doesn't account for all of the effects involved but it seems like it would satisfy the complaints that have been levied against the current system, and it's a pretty easy change.

Link to post
Share on other sites
I've thought about the question of subsonic engines and nozzles a bit and I'm proposing the following solution:

First, how it currently works: The exit velocity is calculated assuming an isentropic expansion to ambient pressure. The exit pressure is ambient unless the flow chokes.

My solution would have two nozzle types: convergent and adjustable. A convergent nozzle would have its exit velocity capped at the speed of sound. Exit pressure would be ambient unless the flow chokes (same as current). Adjustable nozzles would operate the same as convergent nozzles at subsonic speeds. At supersonic speeds, the exit velocity would no longer be capped and the exit pressure would always be ambient (since the divergent part of the nozzle expands the flow beyond sonic conditions).

If this solution is agreeable to everyone, I'll implement it. It doesn't account for all of the effects involved but it seems like it would satisfy the complaints that have been levied against the current system, and it's a pretty easy change.

If I'm reading this right, you're saying that the adjustable nozzle would act as a convergent nozzle at subsonic speeds and a convergent-divergent nozzle at supersonic speeds.

I think real engines with variable nozzles change their nozzle configuration based on what the engine is currently doing. When you watch an afterburning jet throttle up the nozzle first closes to form a convergent shape (Its open at idle because having a wide open exhaust at low throttle minimizes thrust and stops the plane from rolling forward) and then when the afterburner ignites it opens to form the convergent-divergent shape. The defining factor is presumably flow rate- if the flow rate is high enough to choke the nozzle in the convergent-divergent configuration then the nozzle opens. In a normal turbojet or turbofan that only happens when the afterburner is lit and the mass of all that extra fuel is added to the exhaust. Seeing as non-afterburning engines always seem to use a convergent nozzle, maybe you could link the change in nozzle mode to the afterburner being ignited?

Also, the speed of sound in the hot exhaust is quite a bit higher than in the ambient air. You might get some weird complications when modelling turbofans, particularly high bypass ones, because the bypass proportion of their exhaust is relatively cold and has a lower speed of sound than the gas coming out of the core. For extra pain, airliner turbofans are using two nozzles, one for the bypass and one for the core...

Edited by Nerd1000
Link to post
Share on other sites
If I'm reading this right, you're saying that the adjustable nozzle would act as a convergent nozzle at subsonic speeds and a convergent-divergent nozzle at supersonic speeds.

I think real engines with variable nozzles change their nozzle configuration based on what the engine is currently doing. When you watch an afterburning jet throttle up the nozzle first closes to form a convergent shape (Its open at idle because having a wide open exhaust at low throttle minimizes thrust and stops the plane from rolling forward) and then when the afterburner ignites it opens to form the convergent-divergent shape. The defining factor is presumably mass flow rate- if the mass flow rate is high enough to choke the nozzle in the convergent-divergent configuration then the nozzle opens. In a normal turbojet or turbofan that only happens when the afterburner is lit and the mass of all that extra fuel is added to the exhaust. Seeing as non-afterburning engines always seem to use a convergent nozzle, maybe you could link the change in nozzle mode to the afterburner being ignited?

Also, the speed of sound in the hot exhaust is quite a bit higher than in the ambient air. You might get some weird complications when modelling turbofans, particularly high bypass ones, because the bypass proportion of their exhaust is relatively cold and has a lower speed of sound than the gas coming out of the core. For extra pain, airliner turbofans are using two nozzles, one for the bypass and one for the core...

The real situation is certainly more complicated than I proposed. The biggest limitation for the current system is what we only have an approximation of the mass flow rather rather than actual compressor maps. The documentation I've seen suggests that actuation of the adjustable nozzle is mach dependent (mostly based on documentation of the J58 and Olympus 593), but there may be other conditions where actuation happens. This is complicated by the fact that (AFAIK) some engines actuate the converging and diverging parts separately, whereas others don't. The nozzle flaps may be passive in some cases too, so that their position is set by balancing ambient pressure with exit pressure (if possible).

So do you have a proposal for how this should work? The two things that need to be determined are the exit velocity and exit pressure.

Link to post
Share on other sites
The real situation is certainly more complicated than I proposed. The biggest limitation for the current system is what we only have an approximation of the mass flow rather rather than actual compressor maps. The documentation I've seen suggests that actuation of the adjustable nozzle is mach dependent (mostly based on documentation of the J58 and Olympus 593), but there may be other conditions where actuation happens. This is complicated by the fact that (AFAIK) some engines actuate the converging and diverging parts separately, whereas others don't. The nozzle flaps may be passive in some cases too, so that their position is set by balancing ambient pressure with exit pressure (if possible).

So do you have a proposal for how this should work? The two things that need to be determined are the exit velocity and exit pressure.

I'm not qualified to do the maths (Dammit blowfish, I'm a biologist, not a fluid dynamics expert!)- I was basing my comments on some background info (such as this: http://exploration.grc.nasa.gov/education/rocket/nozzle.html) mostly, so I'll defer to your knowledge in this area.

Perhaps a good compromise on nozzle operation would be to have an option to set how the nozzle behaves in the part file. Would a "Velocity vs nozzle position" curve and a "Throttle vs nozzle position" curve would be applicable?

Link to post
Share on other sites

To add to my previous post, exploration.grc.nasa.gov/education/rocket/rktthsum.html contains all the equations you need to model a nozzle, including exit pressure and exit velocity. If we know the throat area, the exhaust area (For a convergent nozzle exhaust area = throat area), the temperature at the turbine exit, the mass flow rate and the constants (gamma and R- the ratio of specific heats and the ideal gas constant) we should be able to calculate everything else.

Link to post
Share on other sites
So many mods are only 0.9 compatible. I'm a little frustrated because I'm almost unable to enjoy the game without modding it now. And as such have stopped playing it. :(

Patience. All of those modders doing what they can in their free time. They own us nothing to make things anymore faster that they already do.

But all users own them great appreciation for doing it at all in first place.

Link to post
Share on other sites

Before an official update comes out we need to completely restructure the propeller and rotor code. It's not that these changes are strictly necessary for compatibility, but much of the basic functionality of engine solvers was split into a separate plugin so that other mods (mainly RealFuels) can use it, and this means rewriting a large portion of AJE. If you're desperate to play with AJE though, the dev version on github _is_ working (save a few minor bugs), though not all the mod configs have been updated yet.

Link to post
Share on other sites
If you're desperate to play with AJE though, the dev version on github _is_ working (save a few minor bugs), though not all the mod configs have been updated yet.

Well heck, I think I'm going to give that a shot! Just got RealFuels working last night. :)

Edit: So... do I want to download the solver_redo branch?

Link to post
Share on other sites
Guest
This topic is now closed to further replies.
×
×
  • Create New...