[1.8+] Real Fuels

[Starwaster]

2 hours ago, Oromis said:

Hi everyone!
I've designed a manned lunar mission in RSS/RO and have decided to use fuel cells instead of solar panels.
The smaller FC (750W) says in the description that it uses something like "41 liters of LH2 and 29 of LOX per week, including boiloff". I only needed two weeks but to be sure, I had decided to include enough fuel for three weeks... Well it didn't last even close to that. I think it lasted three or four days.
Can anyone explain me why? And yes, I've used cryogenic tanks.

Use a Service Module tank instead. Those assume vacuum bottles and/or vapor cooling.

[Wallygator]

A quick question: Is there a minimum TWR /Time required to assure sufficient ullage prior to engine start?

The background to this question relates to a lander using the lunar module descent engine - which, with a starting TWR of 1.31 seems to randomly end up with feed line vapour.

Looking at the engine config file I can only see "ullage = true", also found nothing further in the various RO config files that I can see, so my initial assumptive answer to the question above is NO.

(Note: I have not gone the next step to search through source code)

Thanks in advance for a response - and once again thank you for a fantastic mod!!!

[Oromis]

22 hours ago, Starwaster said:

Use a Service Module tank instead. Those assume vacuum bottles and/or vapor cooling.

So you're saying that SM tanks are better than cryogenic ones? Is this true in RL? And can this be applied even to big tanks like 1st/2nd stage tanks?

Btw, your suggestion worked! I've used SM and boiloff was muuuch better now, although with resources calculated for 3 weeks of working fuel cells, they only lasted 10 days (due to LH2 in particular, obviously).

Here are some screenshots... LH2 is 371L (41x3weeksx3FC) and LOX is 87Lx3Tanks(=261L, 29x3weeksx3FC). Only one LOX tanks in monitored here but the other are the same.
LH2 and LOX after 24h (fuel cells not active)

Spoiler

After 7d (6d of active FCs)

Spoiler

After 11d (10d of active FCs)

Spoiler

And these are my tanks and equipment (the gold ones are LOX, the central orange one is LH2)

Spoiler

As you can see they drain way too quickly. My only guess is that it would be better to use spherical tanks (lowest surface/volume ratio)?

Edited July 12, 2016 by Oromis

[Starwaster]

1 hour ago, Oromis said:

So you're saying that SM tanks are better than cryogenic ones? Is this true in RL? And can this be applied even to big tanks like 1st/2nd stage tanks?

You can apply it to big tanks but you shouldn't if your goal is realism. The SM tanks were modeled after Apollo's LH2 and LOX SM tanks and they're somewhat complex affairs: double walled vacuum bottles mounted on shelves inside the SM compartment. Additionally, thermal conduction is limited because of reduced contact area and some amount of radiation.

There are practical limits  as to size from an engineering and cost standpoint. None of those limitations are conceptually present in Real Fuels so there's no enforcement on size limits so there's nothing stopping you from making launch stages using service module tanks. IRL it would be impractical and costly at best and for very large stages maybe impossible. (I can't imagine doing something like that on a Saturn V scale)

Quote

As you can see they drain way too quickly. My only guess is that it would be better to use spherical tanks (lowest surface/volume ratio)?

I'm not sure what the effect of that would be. I don't think PP shape has an effect on surface-surface conduction which is the only heat transfer when those service module doors are closed. (I think radiation is either non-existent or reduced for cargo bay contents. I can't remember actually)

[Starwaster]

@Oromis

Well I've had a few bits of revelations here. I'm comparing all this to Apollo for a baseline and a few things occurred to me. One of which is that fuel cells don't use oxygen or hydrogen in liquid form. They use it in gaseous form. IRL, as the tank contents boiloff, they're kept (at 900 PSIA for O2 or ~240 for H2) either in the tanks themselves or in a surge tank. So that boiloff is likely to end up in the fuel cell instead of being wasted. (in fact, only SOME of the heat came from heat leakage. The rest of it came from an internal heater and the tank fan motor)

Conceptually, liquid->gaseous conversion doesn't exist in Real Fuels as propellant is boiled off. It's just gone - though there was some discussion awhile back of  maybe converting the boiloff into gaseous resource but there just wasn't a big enough demand to make it a priority.

[h0yer]

Hi!

The UniversalStorage cores come with integrated LH2/LOX->gaseous converters (RO configs?), but they as well don't take their input from the boiloff, but I guess some code could be written, that just converts the boiloff into gaseous resources, if an appropriate empty tank is available.

Basically a continuously running resource converter, that converts the boiloff amount into an empty matching tank. Boiloff 'losses' would then only occur, if the surge tanks are filled and the fuel cell not running. Maybe something like this could work somehow^^

 

A few thoughts/ideas regarding the RF core:

Wouldn't it be an interesting challange to have to use some electricity to keep the catalyzers inside hydrazine thrusters heated, so they can work/not freeze?

 

Thank for awesome stuff

[Oromis]

@Starwaster Awesome, your answers are very interesting!

Even though in RL boiloff is used (and encouraged with fans and heaters, as far as I understood), I think the simplest way to simulate it it's just as it is now, using lh2 and lox as the primary resources. That's because in-game boiloff itself probably is too low for the fuel cell, so some code should be written to make the fc increase the boiloff of that specific tank... unless you make it necessary to attach the fc onto the tank... maybe make it heat the tank... well, too complicated in my opinion!
A compromise could be that "in-game fc consumption" = ("fc fixed RL-based consumption" - "in-game boiloff"), although this would probably mean a dynamic consumption rate and more cpu work to calculate it.

Universal storage's approach is more realistic, there's an LH2->H2 and LOX->O2 converter and the fuel cell uses these gaseous end-products. But in their tanks there is little to no boiloff at all.
Therefore another solution would be to eliminate boiloff, and use a converter (conceptually it could represent those fans and heaters). But a new type of tank should be added, cause SM tanks are also used for pressure-fed engines and they should have boiloff.

LOX->O2 conversion could be useful to store breathing oxygen... It has actually happened to me to have the oxygen running out quickly, and a big fat tank of LqdOxygen underneath, completely useless ^^

That still doesn't explain why my FCs didn't last long enough though

[Starwaster]

3 hours ago, Oromis said:

@Starwaster Awesome, your answers are very interesting!

Even though in RL boiloff is used (and encouraged with fans and heaters, as far as I understood), I think the simplest way to simulate it it's just as it is now, using lh2 and lox as the primary resources. That's because in-game boiloff itself probably is too low for the fuel cell, so some code should be written to make the fc increase the boiloff of that specific tank... unless you make it necessary to attach the fc onto the tank... maybe make it heat the tank... well, too complicated in my opinion!
A compromise could be that "in-game fc consumption" = ("fc fixed RL-based consumption" - "in-game boiloff"), although this would probably mean a dynamic consumption rate and more cpu work to calculate it.

Universal storage's approach is more realistic, there's an LH2->H2 and LOX->O2 converter and the fuel cell uses these gaseous end-products. But in their tanks there is little to no boiloff at all.
Therefore another solution would be to eliminate boiloff, and use a converter (conceptually it could represent those fans and heaters). But a new type of tank should be added, cause SM tanks are also used for pressure-fed engines and they should have boiloff.

LOX->O2 conversion could be useful to store breathing oxygen... It has actually happened to me to have the oxygen running out quickly, and a big fat tank of LqdOxygen underneath, completely useless ^^

I think (to some extent) that you're overthinking things a little. A lot of what you're saying could be realized through config changes. The actual  code that would be needed wouldn't be terribly tricky and would amount to maybe 2-3 lines of code. (Plus a little more code to read the config changes that would be necessary to assign boiloff byproduct)

Quote

That still doesn't explain why my FCs didn't last long enough though

Aside from what's already been discussed I think maybe a lot more of it is through FC draw than boiloff. I made a test craft roughly equal to what you had in your example though with half the LH2/LOX and after 12 days I still had 90L of hydrogen. Which is still a bit high but not to the extent of what you're seeing. Maybe I'll know more after you've tried the debug plugin I sent you.

[OliverPA]

@NathanKell I've not seen this TweakScale-related issue mentioned anywhere other than @ferram4's FAR mod where the error description is that vessels start floating at launch and in some cases the game even crashes. But it seems ferram's mods are involved but not the root cause.

Now the floating at launch only occurs when parts are re-sized with TweakScale and it is actually introduced by KerbalJointReinforcement during physics easing.
But I've tracked the root cause down to these parts having a negative part mass and a rigid body mass of 0 which only occurs when RF is enabled! The same vessel without RF but with TS/FAR/KJR works fine.
I've reduced my ksp install to bare-bones (RF, RF-stockalike, TweakScale, KerbalEngineer) and every tank or engine that is sufficiently down-scaled has a dry mass <0 sooner or later.

It's taken a couple of hours to get to this point but now I'm quite stuck. I can't figure out how the RF mod influences TweakScale behaviour and/or dry mass calculations
Any ideas where to go from here?

[blowfish]

@OliverPA Tweakscale and RF are incompatible right now, I believe that's a known issue.

[NathanKell]

I was not aware of TS-RF incompatibility; I have implemented the interface @pellinor asked for.

[OliverPA]

9 hours ago, NathanKell said:

I was not aware of TS-RF incompatibility; I have implemented the interface @pellinor asked for.

Oooh, that's a bit embarrassing, I totally missed that there's a TS interface in the plugin

But that raises the question - is TweakScale_RealFuels.dll supposed to be distributed? Because it's not in the RF release...
I can compile the solution but get exceptions upon loading the compiled dlls.

[NathanKell]

@OliverPA no, that's what I mean--the interface receiver is in RF itself, so the extra dll isn't needed.

[OliverPA]

3 hours ago, NathanKell said:

@OliverPA no, that's what I mean--the interface receiver is in RF itself, so the extra dll isn't needed.

Alright, thanks for the clarification.

Unfortunately when I compile the dll it crashes on load so I can't do terribly much

However I have noticed that the dry mass in the right click menu of the tanks scales correctly whereas the actual part mass doesn't!
So ModuleFuelTanks.cs line 1007 produces a good result of >0 even when the actual part weight is <0.
Maybe it's worth looking at the if() starting line 970 which has only one code flow that does not lead to part.mass = mass, ie. line 985.

Also there always seems to be a difference between the dry mass in the right click menu and the part mass.
Take a Mk1 pod -> total mass of 850kg according to Engineer Report.
Add a FL-T100 tank with dry mass of 9.3kg -> total 860kg which is fine.
Scale the tank to 2.5m (= 2x) with a dry mass of 74kg (= 9.3*2^3) -> total vessel mass of 1362kg so +502kg instead of +74.

[Starwaster]

10 hours ago, OliverPA said:

Alright, thanks for the clarification.

Unfortunately when I compile the dll it crashes on load so I can't do terribly much

However I have noticed that the dry mass in the right click menu of the tanks scales correctly whereas the actual part mass doesn't!
So ModuleFuelTanks.cs line 1007 produces a good result of >0 even when the actual part weight is <0.
Maybe it's worth looking at the if() starting line 970 which has only one code flow that does not lead to part.mass = mass, ie. line 985.

Also there always seems to be a difference between the dry mass in the right click menu and the part mass.
Take a Mk1 pod -> total mass of 850kg according to Engineer Report.
Add a FL-T100 tank with dry mass of 9.3kg -> total 860kg which is fine.
Scale the tank to 2.5m (= 2x) with a dry mass of 74kg (= 9.3*2^3) -> total vessel mass of 1362kg so +502kg instead of +74.

There's an issue on Github about mass being set directly and apparently it's not supposed to be done at all anymore. I don't know if anyone is looking into that.

(edit: It was closed a long while back and didn't deal with tanks... but I still get email about it)

Edited July 19, 2016 by Starwaster

[NathanKell]

Yep, I'll fix that for the next update.

[coldsiemens]

Hi! I have a question about installing mod. In link from description configs are missing

"OR

Real Engines. It comes with Realism Overhaul, and converts engines that look like real engines into their real counterparts."

[Starwaster]

3 hours ago, coldsiemens said:

Hi! I have a question about installing mod. In link from description configs are missing

"OR

Real Engines. It comes with Realism Overhaul, and converts engines that look like real engines into their real counterparts."

If your question is about where to find engine configs, either install Realism Overhaul or Stockalike Configs (both links below)

 

[coldsiemens]

9 hours ago, Starwaster said:

If your question is about where to find engine configs, either install Realism Overhaul or Stockalike Configs (both links below)

 

Oh, my bad, i interpretate it wrong, don't see whis " It comes with Realism Overhaul ", ty for help

[delta wee]

There seems to be a bug where if you open a rocket in the VAB the button to auto-fill the tanks with the correct fuel amounts wont show up, even though engines are attached to the tanks. Reattaching the engines brings the button back.

Edited July 29, 2016 by delta wee

[Starwaster]

8 hours ago, delta wee said:

There seems to be a bug where if you open a rocket in the VAB the button to auto-fill the tanks with the correct fuel amounts wont show up, even though engines are attached to the tanks. Reattaching the engines brings the button back.

 

This report isn't sufficient to reproduce the problem. I just created a simple rocket, saved it and loaded it and the autofill button shows up just fine. Needs repro steps plus log. If you're not sure where your log is (and you do have one unless logging was explicitly disabled) then see below:

 

Edited July 29, 2016 by Starwaster

[delta wee]

2 hours ago, Starwaster said:

This report isn't sufficient to reproduce the problem. I just created a simple rocket, saved it and loaded it and the autofill button shows up just fine. Needs repro steps plus log. If you're not sure where your log is (and you do have one unless logging was explicitly disabled) then see below:

 

I think the bug is related to the FASA engines I was using. I was able to reproduce the bug by building a small rocket with a FASA engine, such as the RL10A-3, saving it, then loading it in the VAB. I've attached my log file. The auto-fill button also doesn't want to fill tanks with liquid hydrogen for FASA engines that are supposed to use liquid hydrogen.

log file

 

Edited July 29, 2016 by delta wee

[Starwaster]

3 hours ago, delta wee said:

I think the bug is related to the FASA engines I was using. I was able to reproduce the bug by building a small rocket with a FASA engine, such as the RL10A-3, saving it, then loading it in the VAB. I've attached my log file. The auto-fill button also doesn't want to fill tanks with liquid hydrogen for FASA engines that are supposed to use liquid hydrogen.

log file

 

Still not having any success and I don't see anything obvious in the logs directly relating to your issue. I see some issues with plugins not loading because they're not compatible: Gravity Turn can't be loaded. 

ModuleRCSFX also can't be loaded but that one is also unnecessary and should be deleted. Its functionality has been folded into the stock RCS module and it won't be updated. so you should uninstall that. (but as I said I don't *think* it's related to your issue)

I'll try poking at it again later. My RSS install isn't quite identical to yours; I don't have TweakScale installed.

How did you install RF, RO, RSS? If you did it manually, try uninstalling all of those and then use CKAN to install.

[delta wee]

1 hour ago, Starwaster said:

Still not having any success and I don't see anything obvious in the logs directly relating to your issue. I see some issues with plugins not loading because they're not compatible: Gravity Turn can't be loaded. 

ModuleRCSFX also can't be loaded but that one is also unnecessary and should be deleted. Its functionality has been folded into the stock RCS module and it won't be updated. so you should uninstall that. (but as I said I don't *think* it's related to your issue)

I'll try poking at it again later. My RSS install isn't quite identical to yours; I don't have TweakScale installed.

How did you install RF, RO, RSS? If you did it manually, try uninstalling all of those and then use CKAN to install.

I installed RSS manually, and RF and RO using CKAN. I'll try installing everything with CKAN to see what happens.

[delta wee]

22 hours ago, delta wee said:

I installed RSS manually, and RF and RO using CKAN. I'll try installing everything with CKAN to see what happens.

Installing everything with CKAN fixed my problem, thanks.