Factorio Forums

So I've encountered a problem with powering all those fancy laser turrets scattered all around when I have only 10 MW of maximum output. Basically there are literally HUNDREDS of laser turrets out there while my factory uses only about 8 MW for everything. And so after a while of getting creative, I've come up with this monster. Just look at it:

Oh boy, this is insane. The concept is simple: pipes can't handle more than 10 units of water and therefore putting water tanks in the beginning prove useless. So if I can't put them in the beginning, why can't I put them after every 10 steam engines? I put them and boom!, here we go. Water tanks are supposed to smoothen shortage of water after using 10 steam engines and also to store power as I mentioned here: https://forums.factorio.com/forum/vie ... 50&p=46183
In the screenshot you can see: 90 steam engines, 13 boilers, 16 water tanks, 1 pump. 90 steam engines have maximum output of 53.1 MW which is enough to power up nearly 89 laser turrets which basically means about 1335 damage/sec. 53.1 MW is like having 885 solar panels (it's goddamn 7965 tiles)! Do I even need to mention how MUCH space you can save by using this setup?

Speaking of sizes, let's compare this design with others by size:
90 steam engines: 90 * 15 = 1350 tiles
16 water tanks: 16 * 9 = 144 tiles
Not used space: 2*5*8*9 = 720 tiles
2214 tiles used.

Accumulators:
53.1 MW output can be get by using 177 accumulators.
177 accumulators take only 708 space excluding power lines. Three time as less as water capacitor but the problem is that its potential capacity is only 885 MJ.

Now compare it to water capacitor, which will work just like this guy:
16 water tanks are able to store 40000 units of water. According to the thread mentioned above, a unit of water heated to 100 degrees stores 85 kJ of energy. It means that water tanks alone provide 3 400 000 kJ (3.4 GJ, which is nearly 4 times as much as accumulators' capacity). But wait, there's more! Water is also stored in steam engines too. 90 steam engines hold 900 units which means 76500 kJ of extra energy. Summing it up altogether, we get 3 476 500 kJ of energy stored in that system. In order to get similar capacity there need to be 696 accumulators which will take 2781 tiles.
If we take a test and put 89 laser turrets to drain all that energy, we'll get next results:
1. For 177 accumulators we'll get only 250 damage inflicted.
According to Foreman, 177 accumulators will cost me:
1593 iron, 4830 crude oil (crude oil is usually precious resource)
2. For water capacitor we'll use only 2520 iron and that is FAR MORE common resource than oil.

So to sum it all up:
1. Water capacitors
Pros:
Extreme energy capacity
Power output can be easily changed as you please
Energy capacity can be changed independently from output (just put more water tanks next to existent ones)
Can be placed anywhere, but requires special shape
Cheap enough
Cons:
Energy output is NOT constant. It downs as it runs out of water.

It takes more space than accumulators.

2. Accumulators
Pros:
Can be placed anywhere in any shape
Configurable
Takes less space
Emits constant supply of power no matter what
Cons:
Сomparatively low energy density (5 MJ per 4 tiles (1.25 MJ a tile) is no match for 212.5 MJ per 9 tiles (23.6 MJ a tile))
Kinda expensive

Very nice. I used to think storing hot water is interesting but not very useful. But this setup is quite useful.

Related:
https://forums.factorio.com/forum/vie ... =10#p52712
https://forums.factorio.com/wiki/inde ... _hot_water
https://forums.factorio.com/forum/vie ... f=8&t=6665

I used that nearly always since last summer. It's a logical expansion of the early game steam-engine-setup. While I often see on youtube users creating their fourth boiler/steam-engine-setup I never use more than one steam-engine-setup. Only this.

It looks so:
- 3 pumps https://forums.factorio.com/wiki/inde ... shore_Pump
- 20-30 boilers. The problem is here: It is quite difficult to know the exact number of boilers you need, cause it depends on the length of the pipes, the layout of you liquid system, difference in level of liquids (=pressure) and many more. But I think with 30 you should be really on the safe side for a guaranteed 100 degree output. With 20 this might reach only 80 or less. Which is for that usage not even that bad!
- as much steam-engines, as possible. For the first 10 I look, that they are really in sequence. For the steam engines after that I try to make it as parallel, as possible, means: Many storage and steam-engines mixed together.
- as much storage-tanks as possible. I place them for example like above, or - I think that's better: lower side is for storage tanks, upper for the steam-engines.

In Detail: Between all blocks I leave some space (underground belts)

Yes, there is only one pipe for all that steam engines. This works of course only, if I use solar panels. But it saves me from doing much update: Not enough power? New storage tanks, BAM!, some pipes BAM!, connect steam-engines, FIXED!

The panels deliver the basic power, the steam-engines the top load. It is normally no problem to fill the storage tanks during day. If not, you know: More solar panels! But this setup gives you much time to do so... as seen above, the loading curve is really nice to protect you from a total power failure, and it flattens much more out then above, because the lower the pressure in the tanks, the more water will flow.

I think with v0.12 this will become the basic setup, cause I think we will be able to switch this on/off on demand.

- 20-30 boilers. The problem is here: It is quite difficult to know the exact number of boilers you need, cause it depends on the length of the pipes, the layout of you liquid system, difference in level of liquids (=pressure) and many more. But I think with 30 you should be really on the safe side for a guaranteed 100 degree output. With 20 this might reach only 80 or less. Which is for that usage not even that bad!

It doesn't matter, actually. Check this: https://forums.factorio.com/forum/vie ... 50&p=46183 (there are no pictures there anymore, but it doesn't matter either). Temperature doesn't really mean anything, because no energy is lost. If you transfer energy into water by heating, it won't get lost at all.
One boiler inserts 390 kJ of energy into water and steam engine is able to release 510 kJ of this energy. That's what it is. Experimentally I also found that on full usage one pump is able to provide hot water to 10 steam engines so 1 pump can give 5.1 MW. How I found everything else you can read in the link I gave. Also it doesn't matter how much water is flowing through steam engine, it will give the same amount of energy.

Sorry, but definitely need SOME more, if many tanks are empty. Which means, the flow is MUCH higher and you need all 3 pumps. But this is good, cause you can create 2-3 times more energy (not electric power! Hot water!) with this setup if empty, than with the conventional 1pump-14boiler-10steamengine setup. Try it out.

ssilk wrote:Related:
https://forums.factorio.com/forum/vie ... =10#p52712
https://forums.factorio.com/wiki/inde ... _hot_water
https://forums.factorio.com/forum/vie ... f=8&t=6665

I used that nearly always since last summer. It's a logical expansion of the early game steam-engine-setup. While I often see on youtube users creating their fourth boiler/steam-engine-setup I never use more than one steam-engine-setup. Only this.

It looks so:
- 3 pumps https://forums.factorio.com/wiki/inde ... shore_Pump
- 20-30 boilers. The problem is here: It is quite difficult to know the exact number of boilers you need, cause it depends on the length of the pipes, the layout of you liquid system, difference in level of liquids (=pressure) and many more. But I think with 30 you should be really on the safe side for a guaranteed 100 degree output. With 20 this might reach only 80 or less. Which is for that usage not even that bad!
- as much steam-engines, as possible. For the first 10 I look, that they are really in sequence. For the steam engines after that I try to make it as parallel, as possible, means: Many storage and steam-engines mixed together.
- as much storage-tanks as possible. I place them for example like above, or - I think that's better: lower side is for storage tanks, upper for the steam-engines.

In Detail:

Code: Select all

Storage Tank Area                Steam engines                           Boilers                Offshore Pumps

O O O O O O -- <- === === === === === === === === === === <-    <-bbbbbbbbbbbbbbbbbbbbbbbbb--   --<<<
O O O O O O -> -- === === === === === === === === === === 
O O O O O O -> -- === === === === === === === === === === 
O O O O O O -> -- === === === === === === === === === === 
O O O O O O -> -- === === === === === === === === === === 
O O O O O O -> -- === === === === === === === === === === 
...

Between all blocks I leave some space (underground belts)

Yes, there is only one pipe for all that steam engines. This works of course only, if I use solar panels. But it saves me from doing much update: Not enough power? New storage tanks, BAM!, some pipes BAM!, connect steam-engines, FIXED!

The panels deliver the basic power, the steam-engines the top load. It is normally no problem to fill the storage tanks during day. If not, you know: More solar panels! But this setup gives you much time to do so... as seen above, the loading curve is really nice to protect you from a total power failure, and it flattens much more out then above, because the lower the pressure in the tanks, the more water will flow.

I think with v0.12 this will become the basic setup, cause I think we will be able to switch this on/off on demand.

Being the cookie-cutter-guy that I am I created, inspired by this, a somewhat compact backup power facility which I currently use in my factory (images are cropped on the right):



In the center there is a two-line steam engine layout, which can be repeated to the left and right as often as desired. Arrows indicate intented water flow so the engines come before the tanks (not that is matters that much).
On the sides are two other blueprints shown that each add two additional tanks (to use the available space ) and then having a pipe going up.
This allows the addition of another row of dual-steam-engines, just like you did the one before. Rinse and repeat.

The water level in the tanks at the end goes down as previously described. Adding small pumps might help filling these also but when intended as a backup that's not necessary.


And here's a small example where it's in use:



The water has everywhere after the boilers a temperature of 100°C, which I have not expected but is really nice.

(As I'm not all too much into the Factorio fluid dynamic I wonder if the three water pumps at the beginning in combination with a single pipeline is an alright deign or do I need parallel pipes?)


Anyway. After cutting the water source and the other power sources, like solar fields, the example above can deliver initially around 25 MW.
As time passes this smoothes out as the tanks empty. On a 10-minute timescale it looks like this:



So after about 3 minutes the stored energy has been used up. There's plenty of room for improvements, like maybe using mentioned pumps, it's a first shot after some iterations.
I'm a little saddened my graph is not as pretty as OP's

On the left of the graph you can see the steam engines supporting the solar fields in fulfilling the power demand.


Edit: Addition:

This time I experiment with pumps. They are placed at every end of a steam engine row so they pump the hot water up to the top:



In the graph you can see energy is now available for nearly 10 minutes (I goofed up at the beginning by not creating enough demand, but the figure still stands).
Also nice to see where a steam engine row is starting to fail because the pumps drain all the water into the rows more north:




Edit: and another one:

This time like the "classic" design but with pumps on only one place, filling the steam engine and tank system as a whole:



Energy lasts only for around 5-7 minutes, but it tries to maintain the peak power output for as long as possible.
Because all engines empty the tanks at the same time they can run at full power but for not as long:




Edit: and the last one before falling to bed:

Adding side-pipes really helps saving time when filling the system as the water does not have to snail and snake through all parts:



Those can be part of the blueprints.

ssilk

The water capacitor design you're showing has a lot of potential, as hot water storage is extremely efficient, but it has a few flaws.
First off a few basic facts about the factorio fluid transport system.

-It runs on a "distance from source system", with offshore pumps and small pumps being considered the only "sources".
-The further a "sink" is from a "source", the lower the flow rate that can be maintained.
-Offshore pumps initiate a flow of 60 UOL (units of liquid) per second, small pumps pump 30 UOL/s
-A flow rate of 120 UOL can be maintained for up to 14 pipes (pipe to grounds each count as 1, allowing much further transfer)
-At maximum capacity, a single offshore pump can feed 10 engines (which use 6 UOL/s), and requires 14 boilers (1.4 boilers / engine)

For a quick example, if you were to place 2 offshore pumps directly pumping into a tank, it would fill at 120 UOL/s, if you put >14 pipes from the pumps to the tank, the flow decreases (to 90 UOL/s I believe). In your setup you have 3 offshore pumps feeding into a long line, such that the third pump is doing nothing for you.

My typical setup is to feed one offshore pump directly into 14 boilers, and then into 10 engines. Adding tanks and extra engines onto the end of this essentially just creates an overflow system, where the tanks will only fill when your first 10 engines are not running at capacity, and will empty when power demand is high. You can add in small pumps and pumping stations, along with multiple offshore pumps, but I find that if your power dips the pumps stop working and the whole system crashes, so it's better to use many, smaller setups, rather than building one giant one. Working with a single offshore pump in each system also ensures that your flow rate is only 60 UOL/s, which is maintainable over a very long distance, so you don't have to worry about line loss.

What I described is not thought for general usage, it is only usable together with solar power. The end-target is then "off course" to get rid of the whole steam-engines (turn them on, only when really needed).

I think I need to post some pics...

Cause I just have time yet, I show this setup as picture:
On the right: The three pumps. The third pump works only, if the storage tanks on the left are really empty. This can be seen through the glasses.

Above where the charater stands: 24 Boilers. Currently - with this level in the tanks - 23 boilers are running. They are not enough, if the storage tanks are really empty, but enough for this general usage. You see the temperature of the hovered storage tank on the right at 99 degrees.

Left: An (small) array of storage tanks.

This setup is just enough to reduce the total blackout at night an give the lasers somehow a "minimum of power". That time lasts in worst case just 5 to 15 seconds. Again: The target of this setup is
a) cheap power
b) some kind of emergency power / avoiding total blackouts
c) borowing time, if I forgot to build enough solar panels (or don't have the resources to build it yet)

You see, this setup creates 65 MW electric power with only about 1/3 to 1/2 of constant boiler power (power, not work!). https://forums.factorio.com/wiki/inde ... y_and_Work
This is later replaced with some intelligent sensor to be turned automatically on only when really needed.

So I decided to make something awesome. And I did it, I guess. I made custom script to collect data about filling tanks and stuff.
First I made simple setup of pump and storage tank. Here's full log of filling: http://pastebin.com/Try1UPg8
It's quite huge. There's a little error in the beginning since I had to keep my cursor on the tank to gather data.
Quick analysis shows that the actual filling speed of pumps to storage tanks 59.76 units per sec if we don't use pipes between a storage tank and a pump.
Also the log shows that we were wrong about 2000 units of water change in behavior. There is no significant (fourteen digits after the dot) change in filling speed after reaching 2000 units:

storage-tank had 1999.9791670038 of some liquid in at 64.783333333333
dA = 0.99601593625493
storage-tank had 2000.97518294 of some liquid in at 64.8
dA = 0.99601593625493
storage-tank had 2001.9711988763 of some liquid in at 64.816666666667
dA = 0.99601593625493

Actual change seem to appear upon reaching 2245 units of water:

storage-tank had 2244.9990873225 of some liquid in at 68.883333333333
dA = 0.99601593625493
storage-tank had 2245.9947381878 of some liquid in at 68.9
dA = 0.99565086528673
storage-tank had 2246.9885806005 of some liquid in at 68.916666666667
dA = 0.99384241271582

Seems like there is also some kind of QUANTIZE function which cuts the difference to 0 if it's less than 0.0001 which is quite obvious (the effect can be easily seen when accumulators discharge - at some point it drops to 0)

storage-tank had 2499.9745827787 of some liquid in at 107.93333333333
dA = 0.00010014781446444
storage-tank had 2499.9745827787 of some liquid in at 107.95
dA = 0
storage-tank had 2499.9745827787 of some liquid in at 107.96666666667
dA = 0
storage-tank had 2499.9745827787 of some liquid in at 107.98333333333
dA = 0

Because of that a tank never gets filled fully as well as pipes.

There is also some strange "waves" of difference in the beginning, I guess that has something to do with float rounding error.

How did you measure that? Do you have the data also as CSV?

How did you measure that? Do you have the data also as CSV?

I made custom script to collect data about filling tanks and stuff.

Here it is:
http://rghost.ru/6BmrzLZCv

Use console to control it
/c remote.call("calcs", "turn_debug") - switches the mod. It doesn't announce current status since I don't need it.
/c remote.call("calcs", "emit_data") - write collected data into ~/script-output/debug/data.log
/c remote.call("calcs", "clear_data") - clear current data

Here's what I came up with.
During the day the solar panels power the small pumps and prevent water from getting to the lonely steam engine, which in turn prevents any water from flowing to the engine array by deactivating the small pumps above it. At night the solar panels die, the engine is powered, and the whole array turns on, using the water stored in the day. Here it is at night.
Here's the electricity usage hooked up to a radar array for full consumption, it flickers a bit during dusk.

This thing has been so much fun to build, thanks for the inspiration ssilk. Here's my final design.
The left plateau shows array power generation when the tanks are empty, the right shows that when the tanks fill up the overflow is converted into additional power. (A few solar panels thrown in for day/night reference) The overview: The main switch system: Coincidentally I also discovered this really handy way of measuring flow rate perfectly, even in-line.
In the below picture simply divide the max power output in kW by 85 kW/UOL and you get UOL/s,
although for flow rates higher than 6 the setup needs to be bigger.

What I found i, when the water tanks and the engines are apart, the responsiveness isn't that good. Worse, when you have a separate tank and a separate engine area.
That's why the blueprint I posted some posts above is the one I use - tanks and engines are directly together and the engines don't have to zuzzle hot water through several tanks and pipes but have them immediately and, more importantly, sustaining on a high flow level.
Other than that, is works, too, yes.

That's why I included the pumps, every one of those engines is operating at max capacity at night.
Additionally, if you hook up your bases main power supply to the central electric pole (instead of the two solar panels) this setup becomes an on demand power supply, useful for bug attacks. When your power consumption exceeds production, the whole assembly turns on

So I realized that while building a water capacitor that only turns on at night is a neat idea, it really doesn't meet the design goals here.... I have another idea that I'm working on now that will hopefully work a bit better. Sorry, I didn't mean to kill this thread, I think I got a little over-excited.

I think the whole idea will get really, really cool, if we can turn on/off the steam engines automatically.

<waiting for 0.12>

Lupoviridae wrote:Additionally, if you hook up your bases main power supply to the central electric pole (instead of the two solar panels) this setup becomes an on demand power supply, useful for bug attacks. When your power consumption exceeds production, the whole assembly turns on

The best way to do that is to connect them both to an accumulator, since that will keep it as two power networks, and you won't lose power on the mini network unless the accumulator is empty, which means the main power network is under powered. Is that correct?

This is true - for now. Easier way we all hope to see added with 0.12 is detector that can activate or deactivate your steam engines depending on the main power grid power level.
All these beautiful - but complicated designs would become obsolete (but remain beautifully clever ^^)