[1.12.x] SystemHeat - a replacement for the CoreHeat system (November 20)

[Nertea]

5 hours ago, yjchun said:

I feel sorry a lot of heat is just wasted away. If I were a space engineer, it would be my highest priority to develop something useful with those wasted energy. Btw, do 'order' of components matter? I found no difference of radiators' location.

10 hours ago, Nertea said:

Nope there is no ordering.

Oh, people try to do useful things. You could use some heat for thermal processes, heating your ship/base, etc, but those things are not modeled in KSP, and they're not going to be a large fraction of the energy generated. Plus, all that energy is going to need to go somewhere eventually - if you use it to heat your ship, it'll eventually need to radiate out or you'll cook the crew as you keep adding heat. 

I started writing a huge blob on thermodynamics here but really best to go look up heat engines on wikipedia. 

5 hours ago, yjchun said:

I think the heat loop and atomic reactor itself should radiate some heat naturally, same way radiators work and also melt down process itself would use a lot of energy. Natural radiation, heating the ship itself, and melting down reactor, it would be difficult to design all these correctly, we could just say heat is 'convect'ed away in some way. Umm but it could be very small percentage and neglegible. It is difficult to argue technically but feel like things are fragile and lightweight when a few tones of huge space system melt down in just a few seconds.

10 hours ago, Nertea said:

What are you trying to achieve? Even modeling the whole reactor core and coolant system as a open, perfect radiator is not going to add more than a dozen kW of radiation - not enough surface area.

6 hours ago, yjchun said:

My beautiful ship is just covered with radiators. System Heat forces use of lots of radiators but radiator choice is farely limited. I wish to have more aerodynamic radiators and thermal generators.

 

That is pretty much as intended. 

Always fun to look at 515 kilometres of radiators

 

[lookolookthefox]

Is there any way to patch MKS into systemheat? When i use the genericHarvesters for the drills, it only allows me to mine dirt. I'm not very good at writing code for stuff like this, but i'd guess that the harvesterswapoption module throws off the genericharvesters config? I wouldn't know how to fix it though.

[AtomicRocketBooster]

Hello - I think I have encountered a bug.

I have a nuclear engine and 2 radiators which used to work just fine - plenty of cooling capacity.

However, after:

• Saving the game
• Warping a day or so
• Installing the update
• Warping some more
• Switching to that vessel

For some reason, everything is going absolutely haywire - my reactor overheats basically instantly.

Here's a screenshot of right before it melts down:

I've looked at the logs (KSP.log, Player.log, and debug.log) and nothing seemed obvious, but here's the player.log

 

[Nertea]

20 hours ago, AtomicRocketBooster said:

Hello - I think I have encountered a bug.

I have a nuclear engine and 2 radiators which used to work just fine - plenty of cooling capacity.

However, after:

• Saving the game
• Warping a day or so
• Installing the update
• Warping some more
• Switching to that vessel

For some reason, everything is going absolutely haywire - my reactor overheats basically instantly.

Here's a screenshot of right before it melts down:

I've looked at the logs (KSP.log, Player.log, and debug.log) and nothing seemed obvious, but here's the player.log

 

Hmm there was another issue reported by another user with that but I can't reproduce it. You're sure you updated all the extras patches as well as the main mod?

[AtomicRocketBooster]

20 hours ago, Nertea said:

Hmm there was another issue reported by another user with that but I can't reproduce it. You're sure you updated all the extras patches as well as the main mod?

Just used CKAN to install all available updates.  So I think that should have gotten everything.

Is there save file surgery you can do to get it back up and rolling?

[Nertea]

1 hour ago, AtomicRocketBooster said:

Just used CKAN to install all available updates.  So I think that should have gotten everything.

Is there save file surgery you can do to get it back up and rolling?

No, because I can't reproduce anything on my end so I don't know what is wrong.

Can you post

• MM cache
• Save file
• Full mod list
• An indication of what the problem craft is

And maybe I can try to look at some other things.

[DichromaticLotus]

Having a bug where changing heat loops is inconsistent and scary. Also the nuclear reactor heats up when I turn it off. All radiators are on loop 0, resource converter on loop 1, and nuclear reactor on loop 2. Associated vessel is Mun Station. It's the 2nd fresh station cheated into Mun orbit because the original was made in 1.8.1 and I was unable to integrate a functioning heat loop because of the same bug. The first replacement station was "infected" because I docked a legacy 1.8.1 vessel that has a cursed nuclear engine and it broke the station's heat loop. In this 2nd one the legacy vessels that are docked doesn't have any parts that produce heat but it still broke the station.
 

Logs:
https://drive.google.com/file/d/1k1lUNviTjmfWdfWm9IReZ05eDJkV9DC-/view?usp=sharing

Edited April 4, 2021 by DichromaticLotus
Accidentally repeated a sentence

[Egg Electron]

Hello,

I really enjoy all of Nertea's mods! Very impressive work!

Since the latest update, I've had an issue in the VAB with the System Heat simulation when using nuclear rocket engines like the NERV and some others from Near Future mods. What happens is the heat Total Generation will fluctuate between two values, which correspond to the heat output of the engine and the heat output minus the exhaust cooling of the engine, causing the total system temperature to fluctuate. When the simulation rate is around 40x or 100x, the fluctuations are very rapid. This doesn't happen with the electrical nuclear reactors. Only the engines. But if I slow the simulation rate to 1x after placing the engine, the fluctuation (sometimes) stops, and other variations in the order of placing the engine and changing the simulation speed can make the fluctuations stop or start again. Is this intended or is this a potential bug? I'm happy to post a video or logs if you'd like to check it out.

As a side note, I was finding it difficult to figure out how to use the nuclear engines in my ssto's and small crafts after this latest update because if I shut the engine down the usual way, going from full throttle directly to zero throttle at the end of a burn, then the system flux would immediately jump from 40kw to 2000kw (for the NERV for example) and by the time my radiators cooled the engines their cores were half destroyed. But I figured out that by down throttling the engine gradually I could avoid this, and also by making the engine flame out by switching mode without the necessary propellants.

[Nertea]

4 hours ago, DichromaticLotus said:

Having a bug where changing heat loops is inconsistent and scary. Also the nuclear reactor heats up when I turn it off. All radiators are on loop 0, resource converter on loop 1, and nuclear reactor on loop 2. Associated vessel is Mun Station. It's the 2nd fresh station cheated into Mun orbit because the original was made in 1.8.1 and I was unable to integrate a functioning heat loop because of the same bug. The first replacement station was "infected" because I docked a legacy 1.8.1 vessel that has a cursed nuclear engine and it broke the station's heat loop. In this 2nd one the legacy vessels that are docked doesn't have any parts that produce heat but it still broke the station.

Logs:
https://drive.google.com/file/d/1k1lUNviTjmfWdfWm9IReZ05eDJkV9DC-/view?usp=sharing

I spent about two hours tonight going over all these files and trying various ways of reproducing things by building weird and strange systems. Found nothing - no idea what could be wrong. Will keep thinking about it but on that vessel all the internals look just fine and there's nothing amiss. 

This only occurs on vessels from a previous version loaded in this version?

13 minutes ago, Egg Electron said:

Since the latest update, I've had an issue in the VAB with the System Heat simulation when using nuclear rocket engines like the NERV and some others from Near Future mods. What happens is the heat Total Generation will fluctuate between two values, which correspond to the heat output of the engine and the heat output minus the exhaust cooling of the engine, causing the total system temperature to fluctuate. When the simulation rate is around 40x or 100x, the fluctuations are very rapid. This doesn't happen with the electrical nuclear reactors. Only the engines. But if I slow the simulation rate to 1x after placing the engine, the fluctuation (sometimes) stops, and other variations in the order of placing the engine and changing the simulation speed can make the fluctuations stop or start again. Is this intended or is this a potential bug? I'm happy to post a video or logs if you'd like to check it out.

Thanks, I know about this but it can't really be fixed with the way the simulation currently works. It's harmless but annoying.

13 minutes ago, Egg Electron said:

As a side note, I was finding it difficult to figure out how to use the nuclear engines in my ssto's and small crafts after this latest update because if I shut the engine down the usual way, going from full throttle directly to zero throttle at the end of a burn, then the system flux would immediately jump from 40kw to 2000kw (for the NERV for example) and by the time my radiators cooled the engines their cores were half destroyed. But I figured out that by down throttling the engine gradually I could avoid this, and also by making the engine flame out by switching mode without the necessary propellants.

I should write documentation for this, but not until system is stable.

Looks like that flameout thing is a bug to be addressed. 

[pp3d]

The. log is inundated with messages as shown below for most the ships in the game... is this normal and does it cause a gameplay slowdown? 

[LOG 17:35:14.348] [SystemHeat][SystemHeatVessel]: Resetting Simulation for MK-88 Probe (unloaded)

[LOG 17:35:14.348] [SystemHeat][SystemHeatSimulator]: Building heat loops from 0 ModuleSystemHeat modules

[LOG 17:35:14.348] [SystemHeat][SystemHeatVessel]: Refreshing VesselData from Vessel event

[LOG 17:35:14.348] [SystemHeat][SystemHeatVessel]: Resetting Simulation for SW4 Ship (unloaded)

[LOG 17:35:14.348] [SystemHeat][SystemHeatSimulator]: Building heat loops from 0 ModuleSystemHeat modules

[Nertea]

No, not a performance issue, probably just some incorrectly gated debug messages. 

[Wyzard]

I'm trying this out for the first time today, and seeing a quirk that might be a bug, or might be that I'm misunderstanding something, but I'm not sure which:

I have a test spacecraft with a Garnet reactor and four of the XR-500 high-temp radiators from Heat Control.  When I extend the radiators, the SH summary shows a total rejection of -575 kW.  Then when I start the reactor, total rejection immediately changes to zero while the reactor heats up.  When the reactor reaches operating temperature (800 K), the total rejection jumps from zero to -2329 kW, and the reactor is stable at 800 K.

Then, when I turn the reactor off, rejection goes to zero again — while the reactor is still producing heat.  The reactor's flux steadily decreases, but its temperature goes up (above 800 K) since it's producing heat that's not being rejected.  This causes it to overheat and take some core damage while shutting down.  When reactor flux reaches zero, rejection jumps from zero to -3014 kW briefly, then quickly ramps down to -575 kW, and the overheated reactor cools down.  Notably, while the reactor is "fluxing down" but heating up after I disable it, the System Heat section of its PAW shows "/ NaN K" as the end of the "System Temperature" line.

My test craft is a stock RC-L01 probe core, a Garnet reactor on top of it, a B-12K battery (from NFE) below it, and four XR-500 radiators attached to the battery.  I also get the same behavior on an actual space station that has the same reactor and radiators.

(Also, I'm not sure where that -2329 kW number comes from.  The XR-500 radiator's temperatureCurve says it radiates 500 kW at 1000 K, which interpolates to 400 kW at 800 K, and the SystemHeat panel in the VAB says -1600 kW, which looks like the correct number for the four radiators.  But on the launchpad, it's -2329 kW instead.)

I can provide a log if needed, but I don't see anything that looks like an error.  No logs from SH at all while starting and stopping the reactor.  (There are more than a hundred lines that say "[SystemHeat][SystemHeatVessel]: OnVesselRollout", though.)

Edited April 7, 2021 by Wyzard

[Nertea]

11 hours ago, Wyzard said:

I'm trying this out for the first time today, and seeing a quirk that might be a bug, or might be that I'm misunderstanding something, but I'm not sure which:

 

I mean everytime you see an NaN anywhere something is probably bad. It looks like there is an issue related to parts generating no 'temperature' that I need to address. I wonder if that is the cause of some of the other issues in this thread. 

11 hours ago, Wyzard said:

(Also, I'm not sure where that -2329 kW number comes from.  The XR-500 radiator's temperatureCurve says it radiates 500 kW at 1000 K, which interpolates to 400 kW at 800 K, and the SystemHeat panel in the VAB says -1600 kW, which looks like the correct number for the four radiators.  But on the launchpad, it's -2329 kW instead.)

 

Convection! If you set the situation in the VAB to Kerbin SL you should see -2335K. Because convection is very situational, I don't have a good way to represent it in the part info window.

 

 

 

 

[pp3d]

On 4/6/2021 at 12:29 PM, Nertea said:

No, not a performance issue, probably just some incorrectly gated debug messages. 

Ok .. but now, when updated to the latest version in github, i see these when loading a ship:
WRN 13:55:48.597] [SystemHeat][Settings]  not found, using default coolant
[LOG 13:55:48.622] Packing Ninsun for orbit
[LOG 13:55:48.839] Unpacking Ninsun
... then repeats

I suspect that this constant packing and unpacking of a loaded ship is creating a gameplay issue---- ship is generally unresponsive to interaction (it's not that large  .. less than 150 parts) 

 

Edited April 7, 2021 by pp3d

[Nertea]

2 hours ago, pp3d said:

Ok .. but now, when updated to the latest version in github, i see these when loading a ship:
WRN 13:55:48.597] [SystemHeat][Settings]  not found, using default coolant
[LOG 13:55:48.622] Packing Ninsun for orbit
[LOG 13:55:48.839] Unpacking Ninsun
... then repeats

I suspect that this constant packing and unpacking of a loaded ship is creating a gameplay issue---- ship is generally unresponsive to interaction (it's not that large  .. less than 150 parts) 

 

Well... this mod doesn't pack or unpack ships, so should be safe there.  You can disable the log messages in the settings file, a debug flag was left on by mistake. 

Nobody else seems to have your issue, suggest you investigate other mods (or even post an actual log, snippets are not helpful). 

[Wyzard]

9 hours ago, Nertea said:

Convection! If you set the situation in the VAB to Kerbin SL you should see -2335K. Because convection is very situational, I don't have a good way to represent it in the part info window.

Aha!  And that explains why it didn't happen on my space station (I get the expected -1600 there).

I was looking at the "Total Rejection" line in the SH toolbar panel, rather than the PAW.  As a suggestion for making it more clear in the UI, maybe put a pair of indented lines below that one saying "Radiation -1600 kW" and "Convection -729 kW" — perhaps controlled by a checkbox in the lower part of that panel, hidden by default.

[Nertea]

14 hours ago, Wyzard said:

Aha!  And that explains why it didn't happen on my space station (I get the expected -1600 there).

I was looking at the "Total Rejection" line in the SH toolbar panel, rather than the PAW.  As a suggestion for making it more clear in the UI, maybe put a pair of indented lines below that one saying "Radiation -1600 kW" and "Convection -729 kW" — perhaps controlled by a checkbox in the lower part of that panel, hidden by default.

Unfortunately the simulator at that level has no understanding of flux sources and it would be very involved to get them out. I do regard convection as an always positive response as well as an advanced thing, so just making sure that the default settings always show the worst case is best IMO.

[AtomicRocketBooster]

Hello - I was able to create a craft that is reliably encountering overheating issues.

It works fine if the engine is running however it overheats basically immediately (flux of >3k) once the engine cuts off.  In the VAB, everything seems fine - it stabilizes at 850k.

While engine is running:

Seconds after throttling down:

Here's the modlist.

Here's the savefile (It has saves from both before and after the overheating)

And here are the logs:

• Module Manager Cache
• Module Manager Log
• KSP.log
• Player.log
• debug.log

 

Edited April 8, 2021 by AtomicRocketBooster

[aristurtle]

2 hours ago, AtomicRocketBooster said:

Hello - I was able to create a craft that is reliably encountering overheating issues.

It works fine if the engine is running however it overheats basically immediately (flux of >3k) once the engine cuts off.  In the VAB, everything seems fine - it stabilizes at 850k.

While engine is running:

Seconds after throttling down:

Here's the modlist.

Here's the savefile (It has saves from both before and after the overheating)

And here are the logs:

• Module Manager Cache
• Module Manager Log
• KSP.log
• Player.log
• debug.log

 

With the current thermal model, you will need to keep propellant flowing through the engine in order to keep it from overheating (unless you use a truly very large amount of radiators.) The engine gets ridiculously hot, and is cooled by a flow of ridiculously cold liquid hydrogen.

So, turn the reactor down to 5% (or lower) while the throttle is still on, then ease off the throttle slowly. You should see the reactor cool down, and specific impulse slowly decrease as the reactor cools down. (This is the way nuclear thermal engines work in real life, it's one of the challenges in building a mission around them). If you're doing your burns with Mechjeb or whatever, try turning the reactor off a second or two before the burn is supposed to finish, this will usually get you pretty close.
 

If you're very careful about this, you can even fly one without any radiators at all, by managing heat through reactor control and propellant flow. But it's easier if you have enough radiators to manage the heat generated by the reactor at its lowest setting, so you can "idle" it there to warm it up before the burn, and shed the last bit of waste heat after the burn without wasting a lot of propellant.

[AtomicRocketBooster]

1 hour ago, aristurtle said:

With the current thermal model, you will need to keep propellant flowing through the engine in order to keep it from overheating (unless you use a truly very large amount of radiators.) The engine gets ridiculously hot, and is cooled by a flow of ridiculously cold liquid hydrogen.

So, turn the reactor down to 5% (or lower) while the throttle is still on, then ease off the throttle slowly. You should see the reactor cool down, and specific impulse slowly decrease as the reactor cools down. (This is the way nuclear thermal engines work in real life, it's one of the challenges in building a mission around them). If you're doing your burns with Mechjeb or whatever, try turning the reactor off a second or two before the burn is supposed to finish, this will usually get you pretty close.
 

If you're very careful about this, you can even fly one without any radiators at all, by managing heat through reactor control and propellant flow. But it's easier if you have enough radiators to manage the heat generated by the reactor at its lowest setting, so you can "idle" it there to warm it up before the burn, and shed the last bit of waste heat after the burn without wasting a lot of propellant.

So if I'm reading you correctly, then this is working as intended?

If so, I have some questions:

• What is actually being modeled when you look at the loops in the VAB? The UX here isn't totally clear.
• What is the intended purpose of nuclear engines (like the Neptune) that have the built-in reactors if you can't actually use them? The Neptune specifically calls out its passive power generation as a feature, so something seems off.
•  This seems like a situation that should get handled automatically? It seems like a pretty rough UX to have throttling down to 0 (like you would with any other engine in the game) be an insta-meltdown when using nuclear engines. What does that "Manual Control" button actually do? Is there a way to actually use these things without enabling manual control?
• Relatedly, what does the "Hibernate" feature actually do? It doesn't seem like it does anything?
• Did something change in the last update to introduce this behavior? I don't recall this happening before the most recent version of system heat.

Love the mod and everything Nertea does, so just trying to figure out what's working as intended vs. bug vs. feature request. Also happy to move this discussion over to the Kerbal Atomics thread if that's a better place.

Edited April 8, 2021 by AtomicRocketBooster

[aristurtle]

26 minutes ago, AtomicRocketBooster said:

So if I'm reading you correctly, then this is working as intended?

If so, I have some questions:

• What is actually being modeled when you look at the loops in the VAB? The UX here isn't totally clear.
• What is the intended purpose of nuclear engines (like the Neptune) that have the built-in reactors if you can't actually use them? The Neptune specifically calls out its passive power generation as a feature, so something seems off.
•  This seems like a situation that should get handled automatically? It seems like a pretty rough UX to have throttling down to 0 (like you would with any other engine in the game) be an insta-meltdown when using nuclear engines. What does that "Manual Control" button actually do? Is there a way to actually use these things without enabling manual control?
• Relatedly, what does the "Hibernate" feature actually do? It doesn't seem like it does anything?
• Did something change in the last update to introduce this behavior? I don't recall this happening before the most recent version of system heat.

Love the mod and everything Nertea does, so just trying to figure out what's working as intended vs. bug vs. feature request. Also happy to move this discussion over to the Kerbal Atomics thread if that's a better place.

 I mean, I didn't write the mod; I'm just working from the description and behavior here, but I think this is as intended?

• heat flow through a heat loop, including heat sources (e.g. nuclear reactors) and sinks (e.g. radiators) and stuff like coolant tanks
• The built-in power reactor tops out on power production at a pretty low reactor percentage (around 5% if I recall correctly?), so if you want to use it as a power generating reactor, you'd leave it at that percentage and carry enough radiators to radiate 5% of the reactor heat, then crank it up to 98% or 100% when doing a burn
• I couldn't tell you, I've only used these with manual control. Maybe adding coolant tanks to increase loop volume would help it not overheat so much? The reactor still has a lot of residual heat that needs to go somewhere when you turn it off.
• as far as I can tell, it turns off the reactor when you're not focused on the ship, to save on reactor fuel, then turns it back on when you jump back to the ship. I think this would be good for probes and bad if you have a life support mod enabled and your reactor is the only power source keeping your crew alive?
• This is all recent as of the last update to SystemHeat, yes; prior to that a nuclear engine would require the same amount of radiators regardless of how much propellant was flowing.

Nertea, I'm sorry if my propellant flow heat request made this a usability nightmare for everyone else!

[Nertea]

1 hour ago, AtomicRocketBooster said:

So if I'm reading you correctly, then this is working as intended?

No not really, the bug Wyzard found is affecting things.

1 hour ago, AtomicRocketBooster said:

• What is actually being modeled when you look at the loops in the VAB? The UX here isn't totally clear.

Heat needs at full throttle. Use the thrust slider to change that.

1 hour ago, AtomicRocketBooster said:

• What is the intended purpose of nuclear engines (like the Neptune) that have the built-in reactors if you can't actually use them? The Neptune specifically calls out its passive power generation as a feature, so something seems off.

You still need radiators for the passive power generation. However this is affected by the bug (if you turn off the reactor radiators don't work properly). 

1 hour ago, AtomicRocketBooster said:

• Relatedly, what does the "Hibernate" feature actually do? It doesn't seem like it does anything?

Turns off the reactor when you go into time warp, turns it back on automatically after.

1 hour ago, AtomicRocketBooster said:

• Did something change in the last update to introduce this behavior? I don't recall this happening before the most recent version of system heat.

Lots of things changed in 0.4.0.

I mean fundamentally what changed from NFE was the following items:

• Some automation was introduced to reactors overall, there was none before.
• Reactor power actually adjusts slowly instead of instantly, this was intended but I never got it to work correctly without the stock thermal system murdering everyone and their children.

It's obviously taking some work to dial in the correct level of automation here.  Currently (assuming the bug gets fixed):

• Engine reactors run at power levels of (usually) 5-100%, with power uprates of 10%/s and down rates of 20%/s
• Turning reactors on (when no engine is on) will increase their throttle to minimum (5% ). This is 100 kW for the LV-N. They will melt down eventually at this level without radiators or propellant flow.
• If the reactor can generate power, the minimum provides you with maximum power generation. I think the minimum is offhand 2% for these reactors.
• When you increase engine throttle, the reactor will throttle up automatically to match the engine throttle, at 10%/s so will take ~10s to get to full power
• When you decrease engine throttle, the reactor will throttle down automatically to match the engine throttle, down to the minimum value at 20%/s so will take ~5s to get to minimum power. 
• You need to do something in that 5s or the reactor will melt down. Currently, the lower power reactors should survive a hard cooldown (e.g, the temperature increase in 5s of zero cooling will be below the meltdown threshold) without doing anything special. Higher power reactors will need a slower cooldown, using extra radiators or by slowly decreasing the engine throttle. 
• Highest power reactors (gas core stuff) will need radiators to run at all.

What is missing here?

[UltraJohn]

On 3/8/2021 at 9:13 PM, JadeOfMaar said:

This new issue is due to my expecting that users would have System Heat's generic extras patch that changes all supported ModuleResourceConverter into ModuleSystemHeatConverter. The errors point at my patch for a B9PS module that changes the options of an EC-consuming converter in the Dark Drive engine.

So I'm sorta running into this problem too. I've got SystemHeat and all the extra patches installed. From what I understand, the extra generic patches is conflicting with OPT? I've tried enabling your SystemHeat reconfig from OPT as well, but it didn't seem to help. I'm not sure what I should do. Would removing the SystemHeat generic patch, but keeping the other extra patches fix this (which means SystemHeat support for OPT is removed)?

[JadeOfMaar]

22 minutes ago, John007qwe said:

enabling your SystemHeat reconfig from OPT

Nope. That config would not solve this problem. It should, if anything, make things worse for you as it is disabled and has very clear notes or warnings in comments at the top of it. It's WIP and unfinished, and waits for someone who uses System Heat (which excludes me) to complete it. You'll have to open GameData/OPT_Reconfig/Resources/OPT_DarkDrive_FFT.cfg and delete lines 29 to 376 as highlighted here.

 

[Wyzard]

8 hours ago, Nertea said:

What is missing here?

FWIW, the mechanics you described sound pretty reasonable to me.  Seems like there are issues with unclear UI, but I don't see anything "missing" mechanics-wise.  (And the "unclear UI" thing is halfway just that the documentation is unfinished.)