[FES] Flywheel Energy Storage (and IPACS) - new battery types with SAS flavour

[riocrokite]

Kermangeddon Industries presents:

So simply speaking near-future mechanical batteries with SAS flavour. My initial aim was to provide a new type of batteries to enhance your experience with Near Future mods from Nertea. This mod is aimed mainly for career and science modes.

2012 Boeing presentation:

Flywheel Energy Storage compared to chemical batteries:

- Non-toxic and low maintenance

- can combine high power density (W/ kg) and energy density (W-Hr/ kg)

- Fast charge / discharge times compared to batteries

- Very long operating life time (exceeding 25 years)

- Superior operating temperature range

"A combination of advanced fiber technology and superconducting bearings enables the development of [...] extremely high energy-density, high-efficiency flywheel energy-storage system."

source: BOEING

Current NASA G2 flywheel, that will be installed on ISS, weighs roughly 200kg and holds about 1890 EC (525 W-Hr)

Source NASA

Balancing compared to stock batteries and NearFutureElectricals capacitors: (I assume 150 Wh/kg for stock batteries - source)

In total I plan 9 new parts:

3 form factors:

- 1.25m - specs in the table above,

- 2.5m that will have 4x mass and 4x EC

-2.5m IPACS (battery+SAS) that will have modest SAS capabilities at the expense of tad less storage and increased weight (but not much)

Those 3 forms will be offered in Graphite Composite, Fused Silica and Carbon Nanotube flavours which will differ in EC capacity at the expense of price and tech requirement.

TO DO:

- enhancing models?

- texturing

- side effects?

Edited March 31, 2015 by riocrokite

[FreeThinker]

As long as you use 2 fly wheels at the same time in counter wise direction, it might be realistic. If on the other hand you used only one, it should make your vessel spin out of control.

That would mean only the FES would work, the IPACT would probably destroy the vessel if used.

Edited March 31, 2015 by FreeThinker

[Nycidian]

Balancing compared to stock batteries and NearFutureElectricals capacitors: (I assume 150 Wh/kg for stock batteries - source)

I'm pretty sure your overestimating quite a bit

http://wiki.kerbalspaceprogram.com/wiki/Electric_charge

• Batteries are small and offer high storage capacities at the same time. All have the same capacity/mass ratio of 50 g/âš¡ or 20 âš¡/kg.
  

I'm pretty sure '20 ⚡/kg' is supposed to be W·h/kg

Also of note is right now with current technology flywheel batteries can reach energy densities of 100–130 W·h/kg the best technology being carbon fiber composite flywheels with magnetic bearing in a vacuum at the moment (from what I understand). These densities only reached in non commercial applications but most kerbal parts are supposed to be analogous to NASA level tech which is not really commercial.

http://en.wikipedia.org/wiki/Flywheel_energy_storage#Physical_characteristics

Any way I hope the information helps.

Edited March 31, 2015 by Nycidian

[riocrokite]

As long as you use 2 fly wheels at the same time in counter wise direction, it might be realistic. If on the other hand you used only one, it should make your vessel spin out of control.

That would mean only the FES would work, the IPACT would probably destroy the vessel if used.

Yah.

This means that 2.5m FES model is realistic since it comprises of 4 Z-axis flywheels so they could work in tandems spinning in counter wise direction. 2.5 IPACS has Z-axis flywheels and 1 X-axis and Y-axis (so will need to modify model a bit to include 2nd X-axis and Y-axis flywheel). Also I'll split 1.25 FES into 2 smaller ones so they spin in counter wise direction.

I'm pretty sure your overestimating quite a bit

http://wiki.kerbalspaceprogram.com/wiki/Electric_charge

I'm pretty sure '20 ⚡/kg' is supposed to be W·h/kg

Also of note is right now with current technology flywheel batteries can reach energy densities of 100–130 W·h/kg the best technology being carbon fiber composite flywheels with magnetic bearing in a vacuum at the moment (from what I understand). These densities only reached in non commercial applications but most kerbal parts are supposed to be analogous to NASA level tech which is not really commercial.

http://en.wikipedia.org/wiki/Flywheel_energy_storage#Physical_characteristics

Any way I hope the information helps.

Cheers.

Well there are two ways to consider this. Either we assume stock battery has 20 Wh/kg. Then 130Wh/kg would be equivalent to mid-tier flywheel (5-6x more energy density). So then I could rename flywheel names. Alternatively (as I did) we could assume that KSP stock batteries are equivalent to best technology we have today (LiPolymer batt.) so 150Wh/kg. Then we could assume that mid-tier and top-tier composite flywheels are connected to the future material science technology (see Boeing presentation) that has yet to be researched

The only thing is how to name 1st tier flywheel - graphite composite or carbon fiber composite?

- - - Updated - - -

I'm also thinking about adding some kind of disadvantage to those batteries to balance them against stock (additionaly to a high price for career and end-tier technology requirement for science mode). So for someone playing sandbox usage of them will be somewhat balanced as well.

1. First thing that came to my mind is usage of EC either constant or connected to spinspeed (so % of battery load).

1a. RL background for constant EC usage would be usage of some kind of hybrid system comprising of passive (no-maintenance) or electromagnetic bearings and active magnetic bearings (to overcome additional G-forces experienced during rocket motor burns).

2b. RL background for EC usage connected to battery load would be spinspeed that produces increased friction and heat. Nowadays this is one of a capacity limiting factor for NASA G2 flywheel.

3. EC usage connected to experienced G-force (RL background - active magnetic bearings activate to counteract excessive G-forces applied on flywheel)

4. Heat production and accumulation so one would need some kind of radiators or else battery capacities would be throttled as heat increases.

5. No surface attachment possible (to not disturb magnetic fields).

6. Easy to destroy (high G-forces experienced by too harsh landing might deform spinning wheel and damage battery).

What do you guys think?

Edited April 1, 2015 by riocrokite

[riocrokite]

some tests with Nertea's engines, big interplanetary tug for heavy weights, 2nd reusable stage, with vasmir engines. I wanted to achieve an energy buffer to let rocket motors burn for about 1000-1500 dV (typical burn). After that one can replenish EC using even small solar panels by utilizing long times between burns.

[funk]

Nice idea, but perhaps a little unrealistic. Flywheels have some issues regarding use in space and in general. I will try to explain:

1. Torque mentioned above -> gyroscopic effect

2. Mass - Stored energy is a function of mass and frequency (to be exact: inertia and angular speed). Frequency is the determining variable. In reality a flywheel needs to be fixed in a massive mount and needs to be shielded in case of a failure (crash). That means great increase in mass overall. If you assume 200-2000kg for the flywheel itself (5-10kWh), surrounding mass increases with frequency. You can have a "lightweight" flywheel rotating very fast, but will need a heavy housing.

3. Losses - besides magnetic losses in the engine and DC-Converters there are major losses caused by other parts needed. A flywheel causes friction with the surrounding medium. Ok... a vacuum pump isn't needed in space I guess. But it needs a bearing, which could be magnetic or hydraulic, others don't seem to be useful in space. So for a magnetic bearíng you'll have magnetic and eletrical losses. For a hydraulic one you'll need a pump (electrical) and some kind of periodic maintenance (generally flywheels are maintenance expensiv, once per year is a realistic value). Finally there are some minor losses caused by control system etc., but nowadays the attempts to realize an efficient flywheel consider even the smallest savings of energy losses.

4. Usecase is limited, mostly because of the losses. - Due to its good access time, high power potential and great number of cycles flywheels are used as a short time (seconds up to a few minutes!) energy storage device for balancing transient disturbances in energy systems. And they have to compete with capacitors or modern FACTS. Actual studies try to extend the usetime up to 12 hours, but I can attest, that efficiency drops heavily with time. ATM it is unknown, if it's possible at all.

That's said, I want to assure, that I don't want to belittle your approach. Perhaps you can use suggestions above for balancing or improvements.

Edit: I forgot to mention that cooling is necessary!

Edited April 1, 2015 by funk

[FreeThinker]

If I may make a suggestion:

The Flywheel should not behave like Ksp battery, which behave like an accumulator which can discharge all power in an instant except it does not lose power over time (like real capacitors do). Your fly wheel should only be allowed to convert Electric charge into Kinetic Energy and back to Electric Charge at a maximum rate. Preferable, it automatically stores excess Electric Charge and generate Electric Charge when it is low.

Edited April 1, 2015 by FreeThinker

[riocrokite]

Nice idea, but perhaps a little unrealistic. Flywheels have some issues regarding use in space and in general. I will try to explain:

1. Torque mentioned above -> gyroscopic effect

2. Mass - Stored energy is a function of mass and frequency (to be exact: inertia and angular speed). Frequency is the determining variable. In reality a flywheel needs to be fixed in a massive mount and needs to be shielded in case of a failure (crash). That means great increase in mass overall. If you assume 200-2000kg for the flywheel itself (5-10kWh), surrounding mass increases with frequency. You can have a "lightweight" flywheel rotating very fast, but will need a heavy housing.

3. Losses - besides magnetic losses in the engine and DC-Converters there are major losses caused by other parts needed. A flywheel causes friction with the surrounding medium. Ok... a vacuum pump isn't needed in space I guess. But it needs a bearing, which could be magnetic or hydraulic, others don't seem to be useful in space. So for a magnetic bearíng you'll have magnetic and eletrical losses. For a hydraulic one you'll need a pump (electrical) and some kind of periodic maintenance (generally flywheels are maintenance expensiv, once per year is a realistic value). Finally there are some minor losses caused by control system etc., but nowadays the attempts to realize an efficient flywheel consider even the smallest savings of energy losses.

4. Usecase is limited, mostly because of the losses. - Due to its good access time, high power potential and great number of cycles flywheels are used as a short time (seconds up to a few minutes!) energy storage device for balancing transient disturbances in energy systems. And they have to compete with capacitors or modern FACTS. Actual studies try to extend the usetime up to 12 hours, but I can attest, that efficiency drops heavily with time. ATM it is unknown, if it's possible at all.

That's said, I want to assure, that I don't want to belittle your approach. Perhaps you can use suggestions above for balancing or improvements.

Edit: I forgot to mention that cooling is necessary!

Well, all-in-all I think that flywheel batteries concept in KSP has much more realism than current chemical stock KSP batteries. Let's compare:

Stock KSP battery (chemical):

- unlimited lifespan and doesn't discharge when unused - in RL they have very limited lifespan and discharge slowly over time when unused - not realistic

- capacity doesn't change over time - in RL batteries have very limited charge/discharge numbers and their capacity decreases with each charge / recharge

- same battery effieciency no matter what load - in RL battery efficiency drop dramatically when overloaded (and then rest is released as heat) - and let's not forget that chemical batteries used in space are used mostly for low loads - not realistic

http://i.imgur.com/dnd6xxb.png

- very-wide / unlimited operating temperature range - in RL most high-capacity batteries have very limited operating temperature range - not realistic

- can discharge / recharge quickly - in RL chemical batteries have low power density -> are very limited when it comes to recharge/discharge rate, rapid discharge is domain of capacitors that have unfortunately low energy density - not realistic

- able to provide energy for different voltages with high power quality - in RL chemical batteries doesn't have this ability, not to mention limited ability when it comes to peak shaving, load leveling, reactive power support etc. so generally they wouldn't be able to sustain different electrical systems at the same time - not realistic

- doesn't require cooling - assuming they are Li-Polymer and are densely packed they should require cooling some form of cooling - not realistic

Now let's see where we land with flywheel batteries (mechanical) compared to KSP battery behaviour:

- RL practically maintenance free and cycle life times >25 years due to non-touch bearings - so in KSP you could use it for very long duration missions - realistic

- RL practically unlimited charge/recharge cycles and they doesn't affect capacity - see above - realistic

- RL ability so sustain heavy loads and high overall efficiency (90%) - so quite realistic

- RL wide temperature operating range - much wider than chemical batteries - realistic

- RL can discharge / recharge quickly - so it's in line with KSP battery behaviour - realistic

- RL able to provide energy for different voltages with high power quality - and they are perfect for load leveling peak shaving etc - realistic

- RL require cooling, yes, but even then using energy for cooling their discharge rate is comparable with chemical batteries when we use superconducting bearings (touchless) (and let's remember that primary battery usage is for microgravity) - so semi-realistic now, that's why I thought about implementing some small EC usage for those kind of batteries

proof:

Also with development of high temperature superconductors the cost decreases and low temperature point for cooling increases.

NASA uses HTSC:

http://large.stanford.edu/courses/2010/ph240/haefele1/

As for your point 1:

So the general research in flywheels is to maximize spin speed (w2) and minimize mass (I). So in effect flywheel mass to vehicle mass is usually very small to make undesired spin movement. Also we can use counterrotating flywheels in tandems so they cancel each others torque when changing rotation speed.

from B. Bolund - flywheel energy storage systems.

As for your 2nd point - mass and safety:

I think we are talking about two different types of flywheels, large steel ones and very small light composite ones. Composite flywheels doesn't break apart and damage everything as steel ones, composite material tend to micro-deform but is regarded much safer for eventual failure fallout. So no heavy casing. Besides composite flywheels are ultra-light and small, very low density, also composite shape is different than metal one (so no broad wheels rather hollow long narrow ones due to different optimal shape for max Ec between steel and composite ones).

I cited NASA G2 flywheel specifications for the whole system, not flywheel itself. The whole G2 system weighs 98.9kg. And that weight includes everything. G2's flywheel weighs 22.7 kg so only 23% of total system.

More on safety: G2 flywheel bearings can sustain (touchless) 3 times heavier flywheel (even if it's not spinning). So flywheel can safely operate under 3G accelerations. That's enough for me to use it in KSP. We can assume some kind of springs / dampeners for micro/noise accelerations.

3rd point - losses.

Most of it I explained earlier on. Also flywheel has on average higher efficiency than batteries used in space applications (90%).

4rd point - explanation same as 1st point.

----------

All in all, material cited above gives me strong scientific background to use it in KSP alongside stock batteries. IMHO the idea is much more realistic for KSP battery behaviour than stock chemical batteries.

- - - Updated - - -

If I may make a suggestion:

The Flywheel should not behave like Ksp battery, which behave like an accumulator which can discharge all power in an instant except it does not lose power over time (like real capacitors do). Your fly wheel should only be allowed to convert Electric charge into Kinetic Energy and back to Electric Charge at a maximum rate. Preferable, it automatically stores excess Electric Charge and generate Electric Charge when it is low.

Well I would argue that KSP stock battery should not behave like KSP battery

Aim of this mod is to keep things as simple as possible, one reason being that I'm not a coder, other being I don't think it is necessary to over-complicate this.

The proposed function you describe is what currently Nertea's capacitors do except they don't automatically charge/discharge (and they use more generic name 'stored energy').

I mean why create another type of resource (kinetic energy), when composite (NASA G2 and future) flywheels are able to store energy practically without losses in micro-g environment and discharge it very quickly when needed? So RL flywheel behaviour is pretty much equivalent of stock KSP battery?

The main restriction of flywheels I'm thinking of now is just that they will require small amounts of energy to operate. This will be explained by both need to cool super-conducting bearings and small kinetic energy looses over time. This solution is very elegant since it is enough realisti-sh for me and doesn't require any code to write

Edited April 2, 2015 by riocrokite

[FreeThinker]

Man I realy love this stuff, it's a nice opportunity to learn to know cutting edge technologies

I'm looking at the Boeing FlyWheel technology, it's quite impresive. But I wonder what happens if the bearing fail when this flywheel is at full speed (500.000 RPM)

- - - Updated - - -

Aim of this mod is to keep things as simple as possible, one reason being that I'm not a coder, other being I don't think it is necessary to over-complicate this.

Welll unless you want to superseed the Kerbal level of realism, you going to need a part which simulates this. As a first step you might want to use the resource converters found in Regolith with 2 conversion buttons, store and release. A second step would be to use a part specialised in energy storage. Notice I was planning to build a such a part as a generic storage of electrical power into chemical resources. Although ment for chemical properties, it should also be usefull for abstract etherial resources like Kenetic Energy. It would have to be automated, effectively charging when you have an energy surplus and supply electrical charge when you have a shortage. Edited April 2, 2015 by FreeThinker

[riocrokite]

Man I realy love this stuff, it's a nice opportunity to learn to know cutting edge technologies

I'm looking at the Boeing FlyWheel technology, it's quite impresive. But I wonder what happens if the bearing fail when this flywheel is at full speed (500.000 RPM)

- - - Updated - - -

Welll unless you want to superseed the Kerbal level of realism, you going to need a part which simulates this. As a first step you might want to use the resource converters found in Regolith with 2 conversion buttons, store and release. A second step would be to use a part specialised in energy storage. Notice I was planning to build a such a part as a generic storage of electrical power into chemical resources. Although ment for chemical properties, it should also be usefull for abstract etherial resources like Kenetic Energy. It would have to be automated, effectively charging when you have an energy surplus and supply electrical charge when you have a shortage.

As for fail I assume two types- overspin which won't be an option since momentum is controlled and throttled back if needed. 2nd type will be ofc excessive g-forces and I guess battery will simply go Kerbal-boom

As for charge/discharge efficiency and maximum powerload throughoutput - it isn't a problem for a flywheel so why not just leave it behaving like a battery? In the end I'm not seeing the difference whether I use EC or kinetic energy for storage. In the end excess EC ends up in energy storage and when there's a need it is released from storage. Only way to justify this would be to introduce cap on charge/discharge rate or specific usage for KSPI(?). However flywheels aren't powerload-capped unlike chemical batteries.

Edited April 2, 2015 by riocrokite

[funk]

I agree mostly with your statements, assuming less losses in kereality.

The picture in your post regarding efficiency over time is tbh a nice slide to show. But I have another trivial:

I don't know the RO energy densitity values of batteries, but if your values are related, then the storage amount won't be too high. So the benefits of your mod would be parts combining SAS with small EC storage. Though Nerteas caps are OP, too. If you like I can provide some studies from a research project I did in 2012 ...to be more reliable.

[FreeThinker]

However flywheels aren't powerload-capped unlike chemical batteries.

Regarding chemical batteries, we should realy have some Real Batteries Mod, which emplements precisely this.

- - - Updated - - -

I agree mostly with your statements, assuming less losses in kereality.

The picture in your post regarding efficiency over time is tbh a nice slide to show. But I have another trivial:

http://www.energybandgap.com/wp-content/uploads/2011/06/10.png

I don't know the RO energy densitity values of batteries, but if your values are related, then the storage amount won't be too high. So the benefits of your mod would be parts combining SAS with small EC storage. Though Nerteas caps are OP, too. If you like I can provide some studies from a research project I did in 2012 ...to be more reliable.

That diagram was made with old flywheel technology, which relied on mechanical bearings, the superconductive bearing are far superior

Edited April 2, 2015 by FreeThinker

[riocrokite]

Regarding chemical batteries, we should realy have some Real Batteries Mod, which emplements precisely this.

- - - Updated - - -

That diagram was made with old flywheel technology, which relied on mechanical bearings, the superconductive bearing are far superior

yah, diagram referred to metal/steel flywheels which is old tech. it is more or less consistent with different flywheels energy density from boeing presentation.

Actually a batteries mod later on would be cool, imagine Li-Air or Al-air batteries which have 20 times more capacity per kg than stock however they require intakeair resource to function lol

From the gameplay mechanics however I feel that x18 greater energy density than stock is a limit (or even it's too much since it overpowers TWR and dV of Nertea's engines). Will need to do some more tests. I wonder how this would fit KSPI?

edit: I've checked Nertea's caps, they might not be overpowered if we assume this is some kind of digital quantum battery which is comprised of billions of nanoscale capacitors and rely on quantum effects.

http://www.technologyreview.com/news/416852/a-quantum-leap-in-battery-design/

Edited April 2, 2015 by riocrokite

[FreeThinker]

From the gameplay mechanics however I feel that x18 greater energy density than stock is a limit (or even it's too much since it overpowers TWR and dV of Nertea's engines). Will need to do some more tests. I wonder how this would fit KSPI?

You do realise the Nertea's engines are highly overpowered from a realism point of view? To get anywhere near realistic (KSPI) power levels, you would have to multiply power consumption of Nertea engines by about 2000. Now you 18x Flywheel Energy Density Batteries no longer look so overpowered, rather the opposite. Why you could do is depending on wether NFT-E, KSPI is installed, use nerve the Flywheel to 18x density and use REAL WORLD powerlevels if KSPI is installed.

Edited April 2, 2015 by FreeThinker

[riocrokite]

You do realise the Nertea's engines are highly overpowered from a realism point of view? To get anywhere near realistic (KSPI) power levels, you would have to multiply power consumption of Nertea engines by about 2000. Now you 18x Flywheel Energy Density Batteries no longer look so overpowered, rather the opposite. Why you could do is depending on wether NFT-E, KSPI is installed, use nerve the Flywheel to 18x density and use REAL WORLD powerlevels if KSPI is installed.

Well, I'm referring to my batteries and to balance against stock and NearFuture mods, not RL (if it's semi-realistic and has some research backing it up it's a plus OFC). As for Nertea's engines I've never researched them. However from the game balancing point they were much more enjoyable for me than KSPI (only my personal experience) since end-game engines didn't have practically unlimited dV and high TWR. Earlier on I was fascinated by Interstellar mod however I've lost interested in it when I started using an engine that was also used by Scott Manley in a plane with reactor to fly wherever he wanted without fuel and dV consideration.

So yah I found nice scientific background for flywheel batteries but my ultimate goal is to balance them nicely not to broke and overpower high-end Nertea's engines (and maybe FTT mk2 nuclear engines). If those batteries are also useful and not overpowering KSPI at the same time it's an added bonus.

Edit: well when this mod is ready it can be easily patched to suit more whatever other want from it to do

Edited April 3, 2015 by riocrokite

[FreeThinker]

So yah I found nice scientific background for flywheel batteries but my ultimate goal is to balance them nicely not to broke and overpower high-end Nertea's engines (and maybe FTT mk2 nuclear engines). If those batteries are also useful and not overpowering KSPI at the same time it's an added bonus.

I suggest you use real values in the base parts and modify them yourself using a MM script to Kerbalise them. That's what I also did for KSPI. That way you parts will always be based on real data and addaps to it's environment in which it is used.

[riocrokite]

I suggest you use real values in the base parts and modify them yourself using a MM script to Kerbalise them. That's what I also did for KSPI. That way you parts will always be based on real data and addaps to it's environment in which it is used.

Interesting idea, will think about it

[FreeThinker]

Hey, I just noticed there are modules you can use to automaticly store energy into your flyheel and release it when needed.

check out [Plugin] [ODFC] On Demand Fuel Cells (v1.0 / KSP 0.90) it contains an Creative Commons Modules which can be configured to convert EC to Keneric power on demand.

[riocrokite]

Hey, I just noticed there are modules you can use to automaticly store energy into your flyheel and release it when needed.

check out [Plugin] [ODFC] On Demand Fuel Cells (v1.0 / KSP 0.90) it contains an Creative Commons Modules which can be configured to convert EC to Keneric power on demand.

cheers, I'll check it out next week.

[Svm420]

What made them OP? Based on your response in the NF thread? I forgot about this mod and seeing your post got me interested again.

[riocrokite]

What made them OP? Based on your response in the NF thread? I forgot about this mod and seeing your post got me interested again.

Most advanced model is able to store twice the energy of NFE capacitor per mass without its restrictions. So you can build very light spaceships that recharge during the voyage using small solar arrays/rtgs and dispose this energy for manouver nodes. Couple that with higher isp NFP engines and you get a lot lot of dV

[zer0Kerbal]

Did this ever get released?

and yes, ODFC would be perfect to convert KE<->EC