[1.12.x] Heat Control - More radiators! (August 22, 2021)

[GabrielG.A.B.Fonseca]

Hope this works for all of ya .

I, personally, go with whatever you think is best. I trust your judgement.

[TheSaint]

Hope this works for all of ya .

So let it be written. So let it be done.

[KAO]

since the new plan is to roll with globalized heat, will convection pumps still be in the works, or is their purpose defeated now?

[Captain Sierra]

What's the status of the plugin code in 1.0.3+? WOrking fine or heavily borked? I can fix the radiator issues with a config change but my luck with recompiling plugins is about zero.

[Svm420]

Hope this works for all of ya .

Whelp squad's Half baked ideas ruined another mod for me. Hope someone somewhere pick up the ball up and rids us of magic radiators. I will miss 1.02 as my fav version. Stupid steam forced updates

[Captain Sierra]

Hope this works for all of ya .

I knew rhetoric was not going to win this battle, so I will answer as follows.

Not only does this not work for me (I don't respect you any less as a modder/person, I just disagree with the decision) it doesn't work at all.

I present:

EMPIRICAL EVIDENCE!

Proof positive that the deficiencies in the stock radiator module (which are not shared by Nertea's custom module) cause them to be incapable of handling NFE reactors.

The theory:

-The lack of refrigeration system prevents stock radiators from stabilizing reactors below their nominal operating temperature.

-The stock radiator cooling mechanism (being global-selective in nature) is incapable of handling a real mission scenario with multiple active heat sources.

The testing:

All testing was using Hyperedit to place the craft in a 120km kerbin orbit. Infinite fuel was enabled, as was thermal debug colors and right-click data.

Test ship 1:

This is what I would assume is sufficient based on past experience.

It failed to stabilize at all (resulting in exploding reactor) lead alone below nominal temperature.

Test ship 2:

I went massively overkill to ensure I could stabilize the reactor.

Interestingly enough, both ships stabilize the reactor at ~700k for a time until it hits about 400Ec/s power generation. At this point reactor temp begins to climb above nominal temperature.

This particular ship stabilized the reactor just over 1000k.

Once the engine was ignited and burned for several minutes, it too reached a high temperature threshold. Now the radiators had to cool both the reactor and engine. This task proved too much even for this mass of radiator area and eventually the reactor went bang, followed shortly by the engine which somehow remained on even after losing power (infinite fuel at work perhaps; this behavior did not occur pre-1.0 and is perhaps new).

It had been suggested previously that increasing reactor operating temperature would allow for the radiators to stabilize at their critical operating temperature before the reactors pass nominal. Below is my test with the same ship, but with the reactor modified to be over 80% hotter.

Nominal temp: 1400

Critical temp: 1900

Part max temp: 2200

60% additional heat generation from reactor

All of these are to help the radiators stabilize (the radiators reach equilibrium at over 1500k).

The results of this are below.

Test ship 3:

The reactor stabilizes at 850k as the radiators warm up.

The radiators reach the same temperature as the reactor, and it all goes downhill from here.

This is right before the thing goes bang. It failed to stabilize below max temp.

The results:

-The stock radiators cannot stabilize the reactor because the reactor's operating temperature is far below the radiators' thermal equilibrium temperature. Even with a max heat tolerance of 2200k on the modified reactor, it could not stabilize. The lack of refrigeration prevents stock radiators from being effective with NFE reactors without beyond-ridiculous quantities of radiator panels.

-The addition of multiple heat sources disrupts the stock radiators after reaching stability, as they did on test ship 2. Adding a hot engine to the mix causes them to pay less attention to the reactor, failing to keep it stabilized, resulting in cooling failure.

P.S. All of these images are clickable for full resolution. I downsized them due to the sheer number contained in this post.

In light of these results, I make my appeal again.

Nertea, please do not convert to the stock radiator module.

EDIT: It should also be noted that I did edit the reactors to include the necessary parameters so the radiators would kick in when the reactor reached ~200k below nominal temperature.

Edited June 25, 2015 by Captain Sierra

[Gribbleshnibit8]

IMO global heat removal makes more sense than localized. Granted the ISS doesn't deal with nuclear reactors, but you also don't see every module on the station stacked with radiators because they use circulating heat exchangers to move the heat into the radiators from all over the station. Global heat works well for people who don't want to manage sticking radiators on each part that will be hot, and does not limit people who DO want radiators on hot parts from just sticking them there anyway.

On a side note, while I've had this mod installed for a while, I have not yet gotten to the point in my career save where I've needed heat rejection (I only just unlocked nuke engines). I'm glad the decision was made to stick to the stock mechanic, as I was prepared to (somewhat sadly) reconsider my mods, as I will always prefer the pure stock method over a mod method any day.

Looking forward to the updates and changes to NF mods, since I'm getting very close to being able to use them I've had to put my game on hold while these get updated.

[Svm420]

IMO global heat removal makes more sense than localized. Granted the ISS doesn't deal with nuclear reactors, but you also don't see every module on the station stacked with radiators because they use circulating heat exchangers to move the heat into the radiators from all over the station. Global heat works well for people who don't want to manage sticking radiators on each part that will be hot, and does not limit people who DO want radiators on hot parts from just sticking them there anyway.

False now my design consideration are pointless. It limits my creativity and says innovation is the same as slapping crap everywhere. DO NOT WANT!

[mikegarrison]

Not only does this not work for me (I don't respect you any less as a modder/person, I just disagree with the decision) it doesn't work at all.

It's pretty clear he's going to have to change how the reactors work. Maybe just drop the whole idea that the reactors create heat. Anyway, the fact that the current NFE doesn't work with the new stock radiator system (even if you give it a few tweaks) isn't all that much of a surprise.

- - - Updated - - -

False now my design consideration are pointless. It limits my creativity and says innovation is the same as slapping crap everywhere. DO NOT WANT!

Hey, I preferred the other choice too, but it's Nertea's mod.

[Captain Sierra]

It's pretty clear he's going to have to change how the reactors work. Maybe just drop the whole idea that the reactors create heat. Anyway, the fact that the current NFE doesn't work with the new stock radiator system (even if you give it a few tweaks) isn't all that much of a surprise.

- - - Updated - - -

Hey, I preferred the other choice too, but it's Nertea's mod.

I know, you were on my side the entire way. My point was to prove that this isnt a balanceable problem. There's a serious code time investment. As Nertea said, he's a modeller by passion, a programmer by necessity. I think the amount of code time involved in updating the reactors would be more investment than allowing this to continue to exist. I think we're finally beginning to put the true cost of that decision on the table.

[Svm420]

It's pretty clear he's going to have to change how the reactors work. Maybe just drop the whole idea that the reactors create heat. Anyway, the fact that the current NFE doesn't work with the new stock radiator system (even if you give it a few tweaks) isn't all that much of a surprise.

- - - Updated - - -

Hey, I preferred the other choice too, but it's Nertea's mod.

If that becomes a possibility then I'll have to hope someone else puts out reactors

[Captain Sierra]

So after doing that testing I was thinking about where it gets us. Hopefully, it can open some degree of uncertainty back up and get discussion on the table again.

The problem:

Nertea wants to stay close to stock. Some of us disagree with the stock behavior of radiators. These two ideas are conflicting heavily.

A possible solution:

Right now, Heat Control is defining a custom part module for active radiators. Stock now has a part module for active radiators. Is it possible to extend the existing stock module to be capable of replicating the current Heat Control radiator function (or close to it) without defining a whole new module? By leveraging stock instead of ignoring it, Nertea gets to take the mod closer to stock than it is now, those of us desperately wanting the precision heat management still get that, and retaining refrigeration capabilities prevents the need to overhaul the NFE reactors.

Minimal work, minimal plugin code, same popular functionality.

[Fraz86]

Captain Sierra & Mikegarrison,

There are plenty of variables that can be tweaked to balance the reactors. Sierra's tests barely scratch the surface and don't actually prove anything about balanceability.

Conforming to the stock system is balanceable; it may be a bit challenging, and may involve some compromises, but it's doable. Nertea will not need to abandon reactor heating, nor is the new system likely to require much (if any) new plugin code.

It's clear that the two of you (and others) aren't happy with Nertea's decision, but ill-informed naysaying isn't constructive.

[mikegarrison]

Captain Sierra & Mikegarrison,

There are plenty of variables that can be tweaked to balance the reactors. Sierra's tests barely scratch the surface and don't actually prove anything about balanceability.

Conforming to the stock system is balanceable; it may be a bit challenging, and may involve some compromises, but it's doable. Nertea will not need to abandon reactor heating, nor is the new system likely to require much (if any) new plugin code.

It's clear that the two of you (and others) aren't happy with Nertea's decision, but ill-informed naysaying isn't constructive.

You know, I do wish people would stop putting words in my mouth in this conversation. I didn't say it couldn't be made to work. In fact, I said pretty much the exact opposite of that. I said he was going to have to change how the reactors worked.

OK, so I probably shouldn't have speculated about just dropping heating altogether, but based on what Nertea just wrote about being more interested in modelling than modifying the game mechanisms, I'm not sure why he doesn't just eliminate reactor heating. It would let him stop worrying about this and go back to what he likes to do.

Edited June 25, 2015 by mikegarrison

[Fraz86]

You know, I do wish people would stop putting words in my mouth in this conversation. I didn't say it couldn't be made to work. I said he was going to have to change how the reactors worked.

OK, so I probably shouldn't have speculated about just dropping heating altogether, but based on what Nertea just wrote about being more interested in modelling than modifying the game mechanisms, I'm not sure why he doesn't just eliminate reactor heating.

My post was written in response to both you and Captain Sierra, so only some aspects were directly applicable to your comments. I apologize for any confusion. It was Captain Sierra who claimed his tests prove that NFE reactors are fundamentally un-balanceable with stock radiators.

As for changing how reactors work, I doubt it will require much (if anything) more than altering some variables in part configs. Of course I could be wrong, but let's just wait for a basic 1.03 NFE test build, and let Nertea and the testers give it their best shot tweaking config variables before we start speculating regarding the need for serious reactor overhaul.

Edited June 26, 2015 by Fraz86

[mikegarrison]

My post was intended to respond to both you and Captain Sierra, so only some aspects were directly applicable to your comments. I apologize for any confusion. It was Captain Sierra who claimed his tests prove that NFE reactors are fundamentally un-balanceable with stock radiators.

As for changing how reactors work, I doubt it will require much (if anything) more than altering some variables in part configs. Of course I could be wrong, but let's just wait for a basic 1.03 NFE test build, and let Nertea and the testers give it their best shot tweaking config variables before we start speculating regarding the need for serious reactor overhaul.

OK. I'm probably willing to do some testing too.

[Captain Sierra]

OK. I'm probably willing to do some testing too.

Please do. More people gathering genuine data the better assessment we can make. THe decision thusfar was made with little actual hard data on the table, so lets see how this may actually pan out.

Both NFE and NFP seem to work fine in 1.0.3/4. Heat Control is heavily borked.

Fraz, I will do some testing tomorrow evening when time permits to show what it will take. If it takes what I think it will, it will shock you.

Evidently I am alone in this fight now, as all my common allies have resigned to accept that this is happening. I, however, will not. At this point I am not fighting for the function because its what I prefer, or because its what I believe is better. I fight for the idea that because the car ahead of you turns down a road towards a cliff doesn't mean you have to blindly follow it. Squad made a mistake. They made a gloriously detailed thermodynamic simulation (albeit terribly uninformative as always) and then added a bandaid solution of "add this part to ignore this system." Just because they chose to sacrifice all potential for good gameplay and engineering challenge does not mean this mod needs to follow suit. This mod brings real gameplay & engineering challenges. That's something to be admired. It should not be thrown to the wayside because the stock game chose not to adopt that possibility. It should not be sacrificed to appease the ideal of following the stock game. Until such time as the deed is done I will continue to fight for this idea.

[Crzyrndm]

This mod brings real gameplay & engineering challenges. That's something to be admired. It should not be thrown to the wayside because the stock game chose not to adopt that possibility. It should not be sacrificed to appease the ideal of following the stock game. Until such time as the deed is done I will continue to fight for this idea.

Well then maybe someone who is interested should enquire as to whether they can continue with the plugin. Nertea seemed fairly clear above that he wanted to focus on making models (which he does an amazing job of, and has plenty of work going). If Nertea goes with the stock system, it wouldn't be difficult to MM patch all radiators back to a continuation of this system as you so strongly seem to desire.

[Fraz86]

Fraz, I will do some testing tomorrow evening when time permits to show what it will take. If it takes what I think it will, it will shock you.

I don't think that will be necessary. I wrote the following ModuleManager config as a proof of concept:

@PART[*]:HAS[@MODULE[FissionReactor]]:Final
{
    @MODULE[FissionReactor]
	{
		HeatGeneration /= 2
	}
}
@PART[*]:HAS[@MODULE[ModuleGenericRadiator]]:Final
{
	%crashTolerance = 12
	%maxTemp = 2500
	%bulkheadProfiles = srf
	%emissiveConstant = 10
	%heatConductivity = 0.001 //They have built in insulators	
	%skinInternalConductionMult = 2000
	%radiatorHeadroom = 0.75
	%thermalMassModifier = 2.5
	MODULE
	{
		name = ModuleDeployableRadiator
		animationName = Radiator_Extend
		retractable = true
		pivotName = Rotator
		raycastTransformName = Rotator
		windResistance = 5
		trackingSpeed = 0.1	
	}
	MODULE
	{
		name = ModuleActiveRadiator
		maxEnergyTransfer = #$../mass$
		@maxEnergyTransfer *= 50000
	}
    !MODULE[ModuleGenericRadiator] {}
}
@PART[radiator-conformal-2]:Final
{
	@MODULE[ModuleDeployableRadiator]
	{
		@pivotName = Midlink
		@raycastTransformName = Midlink
	}
}
@PART[radiator-conformal-3]:Final
{
	@MODULE[ModuleDeployableRadiator]
	{
		@animationName = RadiatorExtend
		@pivotName = LinkNub
		@raycastTransformName = LinkNub
	}
}

I wouldn't consider it "polished" by any means (I slapped it together in <15 minutes), but it successfully converts conformal radiators to the stock module, and adjusts reactor properties such that everything works together roughly as intended (reactors, stock radiators, and Nertea's radiators using the stock behavior). I used the "advanced radiators" concept that I mentioned in previous posts for conformal radiators, giving them increased emissivity at the expense of retaining their higher mass and cost relative to the stock radiators. The result is seen here:

Basically, an MX-1 operating at 100% with four GR-1's stabilized at 612K. That's pretty good!

Here's what it looks like with an equivalent mass of stock radiators:

As you can see, the MX-1 with two Large Thermal Control Systems stabilized at 844K. That's a bit hotter than what I'd like it to be, but still under nominal temp, and with a little more tweaking I could get it right where I want it.

Edited June 26, 2015 by Fraz86

[Starbuckminsterfullerton]

but because of a pre-0.90 desire to allow players to place radiators anywhere on their ship.

Hope this works for all of ya .

I remember being very excited for this ability and then never had time to use it. When the radiator-less stock heat came out you were the hero we needed, but it did mean taking a step backward with the placement thing. I hate to see all the work that went into balancing heatcontrol go, but just like last time sometimes progress means leaving something behind.

I'm sure whatever you come up with will be just fine.

What will happen to the pipes, insulators, and conductors? Does the stock module allow for such functions? I know that they might not be necessary because of the radiator mobility, but there are still plenty of applications.

Edit: removed weird siamese twin post

[Streetwind]

@Captain Sierra, thanks for your empirical testing, but please keep in mind that your results aren't necessary conclusive. You cannot reliably test the expected behavior with the setups and systematic you chose. You'll notice that I haven't done any testing myself yet, and instead asked for an official test build with proper reactor stats, and that's for a reason.

- The reactors you tested with have no (or bogus) values for skin conductivity under 1.0.4. Therefore the reactors don't handle heat as they should; they have the wrong thermal mass and radiate the wrong amount of heat. The self-radiation in particular can make or break a stability point, and without it, you cannot reasonably conclude that a reactor stabilizes or fails to stabilize at a certain temperature. Conductivity is another factor; 1.0.2, reactors weren't very good conductors, but that may change because it's no longer necessary to keep heat confined to a single part. Or it may change the other way, making them conduct even less, because the reactors might aim for a higher nominal temperature in the future. Until these values are intentionally defined with 1.0.4 in mind, no test can give more than a rough indication.

- In your second test, you mounted the engine with the second-highest heat output in all of NF Propulsion. It outputs waste heat equal to some of the larger reactors, and even in 1.0.2, it would cause temperature issues for ships without additional radiators dedicated to it (just ask Darth Lazarus!). The fact that it destabilizes the equilibirium between reactor and radiators is not a problem, it is expected behavior if the radiators are sized for the reactor alone. (Also: yes, 1.0 made infinite fuel also provide infinite Ec. It was very useful during NFP testing.)

- In your third test, with the elevated maxTemp values, you increased the reactor's heat output by 60%. Why? There's no reason to expect this value to change in any way for this update; you should have tested at default heat output. The unnecessary increase is single-handedly responsible for the test to fail. From a scientific standpoint, this is a serious systematic error and could be interpretated in... certain ways. Just saying.

@Nertea: since we're most discussing NF Electrical right now, do you think it would make sense to move the discussion over to the Near Future thread? Or maybe your dev thread, but there it might derail the mk4 development discussion.

Edited June 26, 2015 by Streetwind

[Fraz86]

you increased the reactor's heat output by 60%. Why? There's no reason to expect this value to change in any way for this update

I wouldn't rule out a change in reactor heat output, though obviously the direction of change would be a decrease rather than increase. Compensating for the lack of refrigeration behavior may prove challenging if we don't decrease reactor heat output. Yes, nominal & critical temps can be increased as you suggested, but it may not be enough.

Keep in mind, we actually doubled reactor heat production late during testing for the recent NFE release. I had advocated for that increase, primarily due to concerns regarding structural parts being overly capable of dissipating reactor heat. I think it would be reasonable to revert to the prior heat output if needed, and the issue of excessive structural heat dissipation could be addressed by other means (e.g., by giving reactors very low conductivity).

Edited June 26, 2015 by Fraz86

[Streetwind]

That's a change that ripples through all of Nertea's heat-related parts (including NFP engines), though. It seems to me that there are other, less intrusive changes that can be made first, and gut feeling says that they're probably enough to make things playable.

The heat loads we have on reactors now are also more believable in terms of comparison to the LV-N. Even after all the adjustments it's seen. I think they're in a fairly good place, personally.

[Fraz86]

Streetwind,

I don't necessarily disagree with you. I share your opinion that reactor heat output is currently in a good place. If we can balance everything without decreasing reactor heat output, I will be happy. I'm only saying that I'm not confident we can take it off the table.

Anyway, I would love to hear if you have any thoughts regarding my proposal to give Nertea's radiators high emissivity (probably >4) as a means of allowing them to retain their ability to provide sufficient cooling for their corresponding reactors, as well as justifying their high mass and cost (relative to stock radiators).

[Streetwind]

Honestly? I'm not a fan of "fudging physics", as it were, even if it works ingame. That's not to say I will argue against it if it turns out to be useful, but you won't see me calling for it, sorry It just feels counterintuitive to me to take this incredibly complex thermodynamics simulation that we have and then go and make it behave in an unrealistic fashion. Yes, I know the stock game does it all over the place, but that doesn't make it better.

The thing to keep in mind here is that doing this wouldn't have to be done to make the Near Future radiators work with the stock modules. They would work perfectly fine with no changes other than perhaps price and mass, and they'd in fact be minimally better per area offered because they already have the highest possible (realistic) emissiveConstant, higher than what the stock radiators offer. But they would not truly offer anything new besides fancy emissives. (Which are, admittedly, one of my favorite things.)

It's precisely that which irks me. Build a part that does something novel and useful, and it requires fudging physics? Okay, may be acceptable. I eventually went with advocating fudging the physics on the DS01 Charon MPDT because it turned out that the size-0-to-size-0 bulkhead connection arbitrarily limits thermal conductivity to a huge degree. But build a part that is redundant, then fudge physics in order to justify its existence? That just feels contrived... Now of course, Nertea's parts came first and the new stock radiators merely sniped their reason to exist, so nobody can blame anyone for seeking new and novel ways to keep them relevant. I just think that fudging physics should be a last resort, not a weapon of choice.

Perhaps we might even get away without fudging physics. I have some vague ideas, but they're all pointless to discuss as long as I don't properly understand how NF reactors will run with stock radiators and proper skin stats. Once there is a test build available, this will certainly be the focus of my testing activities.