Fuel Cells Specs - Need Input

[zer0Kerbal]

Being released shortly is an updated parts pack that contains (amongst other parts) stack mounted fuel cells.

This discussion concerns those .1 (1/10) meter high stack fuel cells.

Each part contains the fuel cell, fuel tank (in units) and a batteries.

I want your input as to the game balance of each of the sizes, and where on the tech tree they should appear

Don't want to use Tweakscale on these due to the need for fine tuning game balance.

Fuselage versions (mk0-mk5) are included just in case a model/texture can be found/provided that will allow these to be created/included.

Thank you in advance!

the current plan is to introduce these as Monopropellent powered.

With the ODFC (On Demand Fuel Cells) patch that is included, these will expand to Liquid Fuel + Oxygen, Liquid Fuel + Intake Air, and Hydrogen + Oxygen (= EC + Water) modes. Doing this might also be possible with the latest B9.

Also any thoughts about if a patch to include a fuel cell in the stock pods and if yes, how much EC should it provide? Obviously the patch would not include the fuel or electricCharge.

thank you in advance!

For comparison

Stock Fuel Cell:

• cost:  750
• entry cost:  2250
• techRequired: Electrics
• mass:    0.05
• EC/s Produced:    1.5       
• EC storage: 50

Stock Fuel Cell Array:

• cost:  4500
• entry cost:  13500
• techRequired: Electrics
• mass:    0.24
• EC/s Produced:    18      
• EC storage: 300

 

Edited August 26, 2019 by zer0Kerbal
added stock comparison

[VoidSquid]

I'd like to have those fuel cells please  

Mind to create some radially attached ones as well?

But a question: why Monoprop based? The electrochemistry behind a LfOx fuel cell is well known (https://en.wikipedia.org/wiki/Fuel_cell), but what's 's your idea here how the electricity will be created?

Edited August 29, 2019 by VoidSquid

[Fraktal]

Turning a hamster wheel with high-pressure monoprop, perhaps?

[VoidSquid]

That I'd call a hamster wheel EC generator, not a fuel cell  

 

Edited August 29, 2019 by VoidSquid

[Fraktal]

It it looks like and quacks like Kerbin's equivalent of a duck...

Edited August 29, 2019 by Fraktal

[Zhetaan]

19 minutes ago, Fraktal said:

It it looks like and quacks like Kerbin's equivalent of a duck...

Then it's probably a decoy.  Keep watch for hunters.

Macabre pessimism aside, the hamster wheel is an EC generator, but the hamster itself is the fuel cell; it turns pellet fuel and oxygen into mechanical work and fertiliser--also entertainment, but I don't have the exact reaction for that.

@zer0Kerbal:

@VoidSquid is correct; the chemistry of fuel cells really doesn't work with monopropellant, so if realism is the goal, then that is not the solution.  The other solutions will work well, though of course you will want to adjust the efficiencies for the different fuel-oxidant mixtures.  For example, I would expect a pure-oxygen fuel cell to have far better efficiency than an air-driven model, though not necessarily five times better considering the heat production and the fact that the technology doesn't necessarily scale that way.

Pedantry notwithstanding, the point to remember about the difference between the solitary fuel cell and the array in stock is that the array produces (and consumes) twelve times the output (and input) as the single cell, but it only costs six times as much and masses 4.8 times as much.  That suggests to me that a significant portion of the mass and cost is in elements such as an essential control module of which you need only one, or in efficient packing that reduces the materials needed for the outer container.  The array also only has six times the battery capacity as the cell in despite of its twelvefold production, so perhaps the rule at work is that larger arrays save somewhat on the cost by reducing the included battery capacity.

Any way that you examine the problem, there appears to be an economy of scale at work that is more advanced than a simple first-degree curve.  All of this is information that you appear to have incorporated, but I don't quite follow your choice of scaling factor:  for example, your size .625 cell has essentially identical capability and mass as the stock cell plus an on-board fuel tank that the stock cell does not have, but at half of the cost.  Why?  Along similar lines, the size 3.75 cell has the same output as the stock array, but it has four times the mass, a bit over three times times the battery capacity, carries the equivalent of an FL-T300 fuel tank (if such existed), and only 150% the cost.  I understand that the fuel tank would add some cost and mass to the system, but as I said, I don't understand the scaling factor at work.

For another example, I would not expect your size 20 fuel cell to necessarily cost double or to have double the capability as the size 10, but rather expect it to be closer to four times as much, because--assuming that the height of the cells is the same--a twofold increase in diameter means a fourfold increase in cross-sectional area and corresponding volume.  This is why, for instance, the 1.25-metre battery is the Z-1k with 1,000 EC, but the corresponding 2.5-metre battery is the Z-4k with 4,000 EC.

I may be going into far too much detail and overthinking this, but I think that for a balance pass, I would choose to break the cells into their constituent components, scale the sizes and costs separately, and then add everything back together for a final value in order to preserve stock-like gameplay.  Doing this doesn't need to be difficult:  to wit, all KSP stock batteries have the same charge density, which is to say that they all have the same battery capacity per unit mass at 20,000 EC per tonne.  I therefore assume that all KSP battery technology, including that in the stock fuel cells, uses the same standard.  Similarly, all stock LFO tanks have a wet-to-dry mass ratio of 9, so that calculation is similarly easy:  simply calculate the size tank that you want for your on-board fuel supply and add its mass to the final product.  Monopropellant tanks do not follow that standard, but that's not insurmountable.

Of course, the fact of it is that I'm not the one making parts to expand the game, so there may be a lot more that goes into the determination that I simply do not know.  I think adding fuel cell capability is an interesting creative direction, so please do continue, and don't let anything I tell you be discouraging.

[VoidSquid]

1 hour ago, Zhetaan said:

I would expect a pure-oxygen fuel cell to have far better efficiency than an air-driven model

Simply: No  

"Problem" here: I studied chemistry and still remember some of that stuff... 

1 hour ago, Zhetaan said:

Pedantry notwithstanding

I'm living in free country and have the absolute right to be pedantic, harrrr!   

Now, instead of a several-page-post about the chemistry of fuel cell chemistry that nobody is interested in, what about you creating LfOx based fuel cells? I'd love to have some more variety than what stock KSP offers  

[Zhetaan]

1 hour ago, VoidSquid said:

Simply: No  

"Problem" here: I studied chemistry and still remember some of that stuff... 

You'll have to explain it, then.  If oxygen is the reactant, then it makes no sense to me that air, as a mixture of one-fifth the concentration, would have equal efficiency to the pure material.  If nothing else, nitrogen occupies space and ought to slow oxygen's ability to pass to and through the cathode.  Also, don't real fuel cells have a problem where carbon dioxide degrades them over time?  At the least, an air-breathing cell would need CO₂ scrubbers.

[VoidSquid]

Ok, I'll try then, several aspects here:

The basic chemical reaction is 2 H2 + O2 -> 2 H20 + Energy

This reaction does not occur in air, but usually in a liquid environment (today there are various other environments too, for example the polymer electrolyte membrane cell which works on a solid, not a liquid environment) (image from wikipedia):

Hence what is of interest here is the concentration of the gases in the electrolyte, which depends on several factors, among them the partial gas pressures (see https://en.wikipedia.org/wiki/Henry's_law for further details), but the electrolyte material plays a huge role here. Then you have the temperature (in general, the hotter, the faster the reaction, but the less efficient and less voltage), what semipermeable membrane materials you use, what catalyst, etc. etc... 

Regarding efficiency, there are different aspects here:

For starters, there is the galvanic series which basically tells you what max. voltage in theory you can get from any pair of chemical elements, for H2 and O2 it's 1.23 V at 25 °C, for practical applications it's usually only 0.5–1 V

The reaction itself delivers 286 kJ/mol, but only part of that can be obtained as electricity from a fuel cell, for the basic reaction 2 H2 + O2 -> 2 H2O it's, depending on the afore mentioned parameters, 40-60%

Another issue is that when you actually "pull" energy from the fuel cell, it's efficiently as well as the voltage lowers itself, part of this is the so called inner-resistance you have with every battery too.

About the CO2 you mentioned: yes, it lowers the efficiency, but even more important are any compounds/impurities that over time "poison" the catalyst, making it less efficient, even tiny traces of various specific compounds are highly poisenous, mentioning like sulfur compounds here as an example.

Got enough?  

[Guest]

On 8/26/2019 at 6:41 PM, zer0Kerbal said:

Also any thoughts about if a patch to include a fuel cell in the stock pods and if yes, how much EC should it provide? Obviously the patch would not include the fuel or electricCharge.

I think it would just add to the massive RMB menus mod-users already have to deal with on crewed parts, to add a feature that is only really useful in the very early game when you're making suborbital trajectories and don't have any batteries.

Still, UI issues aside it's not a bad idea and goes well with life support mods. For pod cells I would look at the default EC requirements for various life support mods and make them able to cover their crew capacity. For example USI-LS requires 0.01 EC/s per Kerbal.

Or I would dedicate a mass fraction of each pod to the generator, say 1%, and use the stock single fuel cell's mass per EC/s to decide how much it produces.

 

8 hours ago, VoidSquid said:

But a question: why Monoprop based? The electrochemistry behind a LfOx fuel cell is well known (https://en.wikipedia.org/wiki/Fuel_cell), but what's 's your idea here how the electricity will be created?

Monoprop could be passed over a catalyst to decompose it and release energy, as with rocket engines, but inside a closed system that extracts power from the increase in heat/pressure with a thermocouple/turbine. Granted this becomes some kind of combustion or gas turbine engine instead of a fuel cell, but I think it's fine to make it work the same as a fuel cell in gameplay terms.

Edited August 29, 2019 by Guest

[zer0Kerbal]

First:

Thank you

Second:

Love the pure fiestyness of the responses!

Continuing:

8 hours ago, VoidSquid said:

I'd like to have those fuel cells please  

Mind to create some radially attached ones as well?

But a question: why Monoprop based? The electrochemistry behind a LfOx fuel cell is well known (https://en.wikipedia.org/wiki/Fuel_cell), but what's 's your idea here how the electricity will be created?

They are coming - very shortly. For a VERY early look  - github. caveat emptor - I am still pushing through all the part.cfgs, and have asked for someone my graphically inclined to look at the textures.

For the base part, yes MP for development purposes.

For better or worse, am taking a meta-game approach to developing these parts for release. All the hard thinking is left to others, otherwise these parts would probably never get released. Now that might not please some, and that is fine. The target audience is most players. With that said am very willing to listen (and read) more in depth discussions - which is part of the reason for this thread.

It is easier to build the part.cfg's around MP then change to another fuel if that is what is desired. Since the two stock versions both use LFO, then that is probably how they will be released. Exception to this is the fuselage sized versions, which will use LF+IA (LiquidFuel + IntakeAir).

Also though, it is my (limited) understanding that MonoProp(Hydrazine) takes only one tank, and more potential energy can be stored per liter than LFO. Since space is limited, a MP driven fuel cell (and I now need to crunch numbers to verify) would pack the longest runtime into the smallest space.

I could include two (or more) versions of the parts, but then we are talking about a very large number of parts, and they all currently look identical in the editor parts list. I am also going to try to use B9 module switch so that you can select the fuel mode in the editor so only one part per size - with essentially 3(4) variants.

Radially attached. Yes, have considered that and will consider it. The biggest obstetrical is having a model/texture to use. There is this one part that should be released shortly, it works but haven't had the time to to the modernization pass.

What about including additional sizes of the stock Fuel Cell and Fuel Cell Array? I have an idea for another and will see what time permits.

If you install ODFC (being released hopefully in the next couple of days) there should be up to four total fuel modes. MonoProp; LiquidFuel + Oxidizer; LiquidFuel + Intake Air; (if you have Community Resource Pack installed) Hydrogen + Oxygen = EC + Water.

I am willing to add additional modes to the ODFC patches once it is released and if someone helps by providing the consumption/production/byproducts specifications (just need one - and a scale multiplier).

Am not concerned so much about 'realism - we are talking about kerbals here... rather keeping the specifications reasonably close to stock for game balance. Am currently

now a teaser to go with that spoiler:

Spoiler

7 hours ago, Fraktal said:

Turning a hamster wheel with high-pressure monoprop, perhaps?

am also considering adding partUpgrades (in a later release) to allow for the increase in efficiency.

Scaling is one of the issues am trying to tackle. Could just include a Tweakscale patch (in fact am trying to get the new improved consumption scaling to work in the tweakscale patch), and will look to that for some guidance.

Currently a spreadsheet is what is being used. Nothing fancy. Just plug in the EC/s production desired and it scales based upon the previous part. So if the previous (smaller) part produces x amount of EC, and this part is expected to produce Y then all other factors are directly scaled (multiplied) by Y/X. Simplistic, but wholly efficient and effective in its simplicity.

With that being said, I imagine the smallest size (stack) being the comprised of 50% by volume of the actual fuel cell, 45% by volume of fuel tankage, and 10% by volume of battery (or capacitors). this would gradually change as size (and the tech level) increased. At the largest size, only maybe 15% of the volume would be devoted to a fuel cell (or maybe 30% if two fuel cells are included in one part); leaving 30% battery and 45% for fuel storage and 5% for maybe something else. Am open to suggestions, in fact am soliciting them.

I do think I have to adjust the scaling and hand tune the parts much more. Instead of a straight scale, a cubic scale might be better. Thoughts?

First things first - is the EC/s produced by each size appropriate for its size?

Am also considering that if NFE (Near Future Electical) is installed, the battery be converted to a capacitor (stored charge) on the larger sizes in part or completely. This would also go for if KSEi (megajoules) is installed. Thoughts?

If I missed any points - please remind me. Very pleased with the conversation so far!

 

7 hours ago, VoidSquid said:

That I'd call a hamster wheel EC generator, not a fuel cell

but what happens if the hamster runs backwards?   and if a kerbal gets hungry enough ...

 

 

1 hour ago, Rocket Witch said:

I think it would just add to the massive RMB menus mod-users already have to deal with on crewed parts, to add a feature that is only really useful in the very early game when you're making suborbital trajectories and don't have any batteries.

Still, UI issues aside it's not a bad idea and goes well with life support mods. For pod cells I would look at the default EC requirements for various life support mods and make them able to cover their crew capacity. For example USI-LS requires 0.01 EC/s per Kerbal.

Or I would dedicate a mass fraction of each pod to the generator, say 1%, and use the stock single fuel cell's mass per EC/s to decide how much it produces.

combine these parts with On Demand Fuel Cells, and now you have the emergency back-up power supply, dark side supply, non-nuclear black and brown out protection that keeps kerbals alive, and warm.

I agree about PAW's, and will keep that in mind. Somebody might include a feature request on ODFC to be able to collapse the section (thought that feature was added) in the PAW menu and/or move the PAW menu controls to a separate toolbar window.

To keep things simple for now (can always add complexity) will keep the fuel cell small and single sized, and not include any battery or fuel, and keep it running on MP since pods usually have MP if they have any fuel. Thought is to go back and revisit once the parts are out.

[Gargamel]

After a bit of light reading :D, moved to Add-on Discussion. 

[zer0Kerbal]

that was, my last post - long.

 

Let's simplify the discussion (for now).

here are the sized planned to be included and their proposed EC/s production value. Remember, these are only .1 to .25m in stack height.

 

   
size    MaxEC produced
0.3125    0.45
0.625    0.75
1.25      1.5
1.875    4.45
2.5        8.85
3.75    18
5         36
7.5      50
10      75
20    125
    
mk2    5*
mk3    15/
    
Stock    
(R)FC    1.5
(R)FCA    18
    
Universal Storage    (payload wedge)
US    16
    
Universal Storage II    (payload wedge)
US2Alk    1.6
US2PEM    5
    
Bluedog Design Bureau    (integrated into Pod)
Gemini    1.5
Apollo(B2)    1.5
Apollo(LEM)    1.5
    
Stock Mining Enhancement (MiningExpansion)    
Sz0Generator    6
Mk2Generator    75
Size1Generator    50
    
RLA    
(R)Small FC    1.5
 
radial1    0.5*
radial2    4.5*
radial3    15*
radialArray1    9*
radialArray2    54*
radialArray3    170*

* placeholder - actual value not known yet

crewCabin    mk1-3pod    landerCabinSmall    mk1pod    mk1pod_v2    mk2LanderCabin    mk2LanderCabin_V2    mk3Cockpit_Shuttle    cupola
add Fuel Cell that produces 0.5 EC/s   

 

Edited August 30, 2019 by zer0Kerbal

[zer0Kerbal]

@Fraktal @Rocket Witch @VoidSquid @Zhetaan

Upon further consideration - agree - best way to do this is based upon EC per kg

so is this math correct?
 

Quote

 

Stock fuel cells take .0025 units of LFO to produce 1 Electric Charge; so 1 unit of LFO produces 400 Electric Charge.

Spoiler

math for blended fuel: (0.001125 Liquid Fuel + 0.1375 Oxidizer)

LFO has a density of 0.005 kg/unit so there are 200 units of LFO per kg.

Spoiler

math for blended fuel: ((0.005 *.45)+(0.005*0.55))

LFO therefor has an Electric Charge Density of 80,000 Electric Charge for one kg of LFO.

 

So if LFO has an Electric Charge Density of 80,000 Electric Charge per one kg of fuel, what should the Electric Charge Density should MonoPropellant? Hydrogen + Oxygen? Liquid Fuel + Intake Air have?

here is what I currently have:

• LFO = 80,000
• MP = 250,000 (because density of MP is 0.004 compared to 0.005 for LFO)
• HO = 5,999,001.94 7,332,113.49 (based upon US2 fuel cells) (gases are very low density)
• LFIA = 147,058.85 (should be higher? because intakeAir's low density)
• Batteries = 10 (20,000 per 2000 kg according to above) (something doesn't seem to be adding up)

I am temporarily holding off releasing On Demand Fuel Cells and the related patch (and Hot Beverages Irradiated(has lots of fuel cells) until am satisfied these numbers are solid or at least a few more days to allow for correction and commenting.

working on a goog spreadsheet to show math now.

Please commit there or here.

 

Edited September 2, 2019 by zer0Kerbal
clarity. so it doesn't read like a calculus textbook.

[Zhetaan]

On 9/1/2019 at 7:33 PM, zer0Kerbal said:

so is this math correct?

[snip]

(something doesn't seem to be adding up)

Your intuition is correct.

On 9/1/2019 at 7:33 PM, zer0Kerbal said:

LFO has a density of 0.005 kg/unit so there are 200 units of LFO per kg.

Remember that KSP uses metric tonnes as a base measure of mass, not kg.  The density of LFO is 5 kg / unit, or .005 tonnes / unit (LF and O share the same density for some reason).  A 'unit' in KSP used to be an arbitrary amount until the developers fixed it at one litre per unit and introduced density to make up the balance of the mass calculation.  Thus, the complete value is 5 kg / L for LFO, and a corresponding 4 kg / L for monopropellant.  In the resource definition files, this reads as the completely incongruous .005 tonnes / litre.

Your stock LFO-to-EC conversion is correct; one unit of LFO produces 400 electric charge.  Of course, that reduces the 'charge density' to 80 EC per kg, not 80,000.

 

I made a spreadsheet a few years ago to compare some modded RTGs to other power sources, but I included solar panels and fuel cells in the comparison.  I won't post the spreadsheet, but the essential idea that is relevant here is that in terms of actual power generated, the stock fuel cell array produces something in the rough neighbourhood of a Gigantor solar panel at Kerbin (it's actually about a third less), and the stock solitary cell produces something approximately equal to the OX-4L at Kerbin (or any of the other deployable OX-Stat solar panels; they're all equivalent in power production).  The reason that this is relevant is because you'll need to balance your fuel cells against the other power producers in the game:  you want something that is of clear use in specific situations (such as stock RTGs near Eeloo), but not something that is the clear winner in all situations (such as a .002 tonne RTG that produces 5,000 EC per minute indefinitely).

When converted to output per unit mass, the fuel cell array produces about 4,500 EC per minute per tonne and the singlet produces 1,800 per minute per tonne.  Note that this does not include fuel mass, mainly because there is no way to predict it.  The solar panels produce an average of 4,500 EC per minute per tonne at Kerbin (actual values vary between approximately 4,000 and 5,000; the Gigantor is a bit under 4,900) and a bit over 1,900 EC per minute per tonne at Duna.  The inclusion of fuel lowers the overall charge per tonne and imposes a time limit, but the point was to show that stock appears to envision fuel cells as primarily outer-system power supplies (solar panels at Dres average at 560 EC per minute per tonne and it only gets worse from there).

That being said, the solitary fuel cell produces 90 EC per minute and uses only .225 units of LFO per minute to do so.  If we were to assume a monopropellant-derived reaction to power a fuel cell that had similar production and consumption, then the stock Mk1 pod and its 10 units of monopropellant would produce 4000 EC over a period of nearly 45 minutes at maximum output.

I am inclined to think that this is likely quite good enough for a Mun landing, though I can see reasons to try something else.  I'd not be inclined to use such a part for Kerbin orbital operations; a single OX-STAT panel produces 21 EC per minute indefinitely, and for most operations, that's enough--even when accounting for the fact that that is a static panel and rarely produces its full output.

 

I don't quite follow your conversions for 'energy charge density'.  I see how you got 80,000 for LFO, but lower density should not result in greater power; it should only result in larger tankage.  The choice of reactants would affect power output, but that is not immediately related to the density.

Taking LFO as an example, the fuel cells readily convert it into EC at a rate of 400 EC / unit of LFO.  Since each unit is 5 kg, that makes an energy conversion potential of 80 EC / kg.  If we assume that monopropellant converts at the same rate (which is a total conceit), then that would yield 80 EC / kg * 4 kg / unit = 320 EC / unit.  To achieve the same power production, we'd need to run more units of monopropellant through the cell, which implies larger tanks.

Your idea of air-breathing fuel cells is intriguing.  I think that in that case, you'd still want an overall energy conversion potential of 400 EC per unit, but since you only need to carry the LF, the effective conversion potential is nearly 890 EC per unit (400 for LFO / .45 for needing to carry only LF)--but with the caveat that it only works in oxygen atmospheres.  There are interesting possibilities for intakes limiting the rate because of their ability to supply oxygen.

Edited September 3, 2019 by Zhetaan

[zer0Kerbal]

thank you.

am thinking.... will respond more after the aspirin kicks in.

the spreadsheet (available here through google.sheets (open for commenting)) reflects the corrections.

edit:

here is what I currently have after fixing formulas:

• LFO = 80 EC/kg
• MP = 250 EC/kg
• HO = 5,999 EC/kg
• LFIA = 147.06 EC/kg
• Batteries = 10 EC/kg
• solar panels/rtg = 80 EC/kg (estimate based upon above numbers)

initial thoughts after reading:

First step is to compare apples to apples - just the fuel cells.

Second step is to compare apples to plums and toss in the per minute calculations. including them now is like adding in volume to the discussion.

Fuel cells were used early on (and still are). Yes solar panels and RTG's are great - but that is an entire another discussion.

agree - spreadsheet formula is was off somewhere. Would appreciate if others would look and see what I missed. Think I found one issue. was inverting the density (1/density) then multiplying that by the unitEC. Now it is ( unitEC / (density *1000 ))

even if something isn't right - it is uniformly not right - so still can use numbers for a generalized comparison of efficiency.

LF+IA : seems about right - considering

MonoProp at 250 is way too high (believe the numbers came from RLA's MP fuel cells) - they need to drop to 200 if not 150 (accounting for variations in density)

HO - well, that is based upon Universal Storage 2's small fuel cell (which is less efficient than the medium) - will leave that for last.

 

making progress. any concrete (ie numbers) suggestions for these?

then we move on to how much EC per size

Edited September 4, 2019 by zer0Kerbal