Factorio Forums

Yeah, and at at the low end, steam control allows you to use the reactor ASAP, without neither wasting resources on ratioed heat exchangers nor on pre-Kovarex precious U-235 !
viewtopic.php?p=426275#p426275

Any day when you learn something new is a good day, or, still yet another different way to run out of power by Amarula

I am sharing this in the hopes it may help someone else not run into this problem; or at least you can get a good chuckle because I keep finding new ways to run out of power

When you are using steam tanks to control loading fuel into your reactors, the cycle goes something like this:

• Load fuel into reactor.
• Repeat as needed until steam tanks reach desired limit. Turbines are now operating on stored steam, and heat exchangers operating on stored heat from the reactors and heat pipes.
• Eventually the heat exchangers use up the heat in the heat pipes, and the steam turbines use up the steam in the tanks. The steam drops until it reaches the limit and that triggers reloading the reactors with fuel.
• Steam continues to deplete, until the reactors are again putting out enough heat, and steam levels start to rise.
• Fuel continues to be loaded until the steam tanks again reach the desired limit.

I tend to overbuild, from early experiences with running out of power. So now I build lots of power.
I am up to 11 power sectors, each with a 2x4 reactor core, and running my nuclear reactors this way has been working fine.

But as I have mentioned earlier in this thread, I just rebuilt all my reactors using the lovely design from disentius. All but two of the systems were shut down while I was considering whether to go steam tanks again, and after thinking about it, it really seemed the way to go, so I added the same amount of steam storage that I had been using previously.

Without considering any possible ramifications, I then started them up. All of them. At the same time (using the same master control system, controlled from inputs from my mall/home base/centre of operations).

So they all loaded fuel at the same time, pushing steam into their steam tanks at the same time, using the same steam tank limits, because they were all the same design. Which was fine.

Until the turbines used up the steam in the tanks. All at the same time. As soon as the first sector hit the limit, it relied on the other sectors picking up the slack, just for a minute or less, until its reactors were putting out heat again. Except that the next sector couldn't pick up the slack, because it had just run out of steam, which pushed even more load to the next sectors, which ran out of steam even faster. So for a minute or so, I was in a severe brown-out, with the only power coming from my starting coal powered unit of 40 steam engines.

And then the reactors recovered, and everything was fine... until 10 minutes or so later when it all happened all over again. Which tells me that previously my system of steam tanks was working only because of the random timing of starting up new reactor sectors, such they just happened to hit depletion at a time when all, or at least enough, of the other sectors were not depleted. In other words, they were working because of blind luck, and luck is not a robust control. The good thing is that I learned something today.

Now to figure out how to keep it from happening by design...

Reactor one = tank amount +1
Reactor two = tank amount +2
Reactor three = tank amount +3

which does hurt symmetry and blueprints, though.

You should in theory be able to fix this issue just by setting higher the steam level at which you reload fuel again ?

I ran into this issue with my very first second reactor design, that IIRC was literally using Steam = 0 as a reload condition. At which point coal picked up the slack, and shortly after that I had biter issues. Rinse and repeat every few dozen minutes or so...

Options I have come up with so far:

1. Changing cycle times.
   Say the current problematic situation has my reactors going through the cycle in 10 units. Changing the number of tanks, or the steam level, on a per sector basis, in effect changes the cycle time. So I might end up with one cycle time of 10, the next 9, the next 8... The problem is (the more cycle times, the further off the problem goes but), at some point the cycles line up again. For example 10 x 9 x 8 = 720, so at time 720 the 10-cycle AND the 9-cycle AND the 8-cycle all deplete at the same time. This mostly solves the problem, but has the sneaky effect of making it suddenly re-appear at lengthy intervals - kind of a mummy's curse
2. Shortening the cycle time.
   Lowering the steam limit for all the sectors in effect reduces the cycle time. So instead of taking 10 units, each sector takes 9 units to deplete. But they still all deplete at the same time. It does make things better, because the shorter the cycle, the higher the power level when things deplete so there is less to recover. Taken to the utmost, if the steam limit is set to the steam tank capacity, the reactors immediately refuel all the time, and there is no depletion.
3. Increasing steam storage.
   If there is enough steam on hand, the system will not brown-out before the sectors recover. The amount of steam needed depends on how much power is being used. Once you know how long it takes to reach from last refuel to depletion reaching the steam limit, you can estimate (or observe) how far the reactors have cooled down, which allows you to calculate (or observe) how long it will take to recover. With an estimate of how much power you need in the worst case, you can calculate how many turbines you need to generate that power, which allows you to calculate how much steam you need to have on hand to keep those turbines going for that length of time. Or of course you can just keep cramming in more storage until the problem goes away This is closest to a "real-world" solution but I admit it doesn't have a lot of appeal for me.
4. Changing start times.
   If each sector has the same cycle time, but they all start at different times, then the cycles will never line up. I really like this option, but I haven't figured out how to make it work. Yet.
   More to come

Honestly, the solution is probably in a different direction than where you’re looking.

How much time does it take, once the steam tanks are depleted, before you are making enough steam again? Lets say (its not, but for illustration purposes) it is 34 seconds.

At 34 seconds before the steam tanks run out, how much steam is left in the tanks?

Add 1000 steam (and more tanks, if necessary) and set your refueling to that steam level, rather than 0 steam.

You can do something other than 1000, that’s just a nice round buffer in case your timing is a little bit off. I don’t know how long it is taking your

astroshak wrote: Sat Jan 18, 2020 1:30 am Honestly, the solution is probably in a different direction than where you’re looking.

How much time does it take, once the steam tanks are depleted, before you are making enough steam again? Lets say (its not, but for illustration purposes) it is 34 seconds.

At 34 seconds before the steam tanks run out, how much steam is left in the tanks?

Add 1000 steam (and more tanks, if necessary) and set your refueling to that steam level, rather than 0 steam.

You can do something other than 1000, that’s just a nice round buffer in case your timing is a little bit off. I don’t know how long it is taking your

One annoyance with a tank based storage is how one reads the steam: if all the tanks are being read, you need to do desired average steam per tank * number of tanks and convert that to the energy window you want. For a "perfect" set-up, you could find exactly how long it should take to drain at maximum steam usage, and have that as the minimum above 0. The next limit is at most 200 seconds of reactor on-time. Adding a buffer helps a bit here: overbuilding the tanks, and shooting for say, 40-80% tank storage, and 150 seconds of reactor on-time, with enough tank steam left for the "un-accounted" 50 seconds. The circuit's a little complicated because you only want to insert fuel once per down-swing, and not on the way back up. The smaller the difference in power used vs stored, the smaller the steam window is: it could be as little as 1% over "enable" steam, if that's all the spare energy you're generating.

If someone hasn't made a spreadsheet for it, they should: for a 2xn reactor design, you know the max power output, convert that to steam, have a cell for desired steam over the minimum, put in a desired amount over the 200-second energy production, and output how many tanks and the steam read values. Another/set of calculations could be useful for recommended values when overbuilding. A big complexity here is at full-drain, there will be no change in the steam tanks, or worse, the engines are built over the ratio of heat exchangers, so you can be running at full power and still be draining steam.

I did something similar, here it is :
(it's pretty crude though - and don't forget about multiple sheets)

Amarula wrote: Thu Jan 16, 2020 4:06 pm All my nuclear plants are now upgraded (thank you again disentius); 10K fewer pipes and 5K fewer underground pipes, my UPS thanks you! No issues with heat exchangers out of water, no issues with turbines having no steam
Another related question: is it better to run more reactors than you really need so your steam turbines are running at say 50%, or to only power up what you need so that the turbines are pushing 95% capacity? (Not 100%, or I start seeing brownouts where I am not quite producing enough power).
The pros and cons that I see: fewer reactors operating means using less fuel (not a big concern at this point, thank you Kovarex); fewer reactors operating means a long lead time to warm up when I do need more power...

Off to figure out how to make a resetting timer using the circuit network, but that is a subject for a different thread!

The amount of power consumed is directly proportional to the amount of steam produced. Assuming you have enough storage capacity and the right number of heat exchangers and turbines so no reactor ever reaches 1000 degrees and wastes fuel then it is always better to run more reactors. Here is why:

Lets assume you have two 2x4 reactor complexes. You can fuel one of them 100% of the time or 2 of them 50% of the time. The amount of heat produced is the same, the amount of steam produced will be the same and therefore the amount of power produced will be the same. So no drawback there in either case.

But if you power one of the 2x4 reactors only partially things change. Reactors give a neighbour bonus only when fueled. So in a large 2xN array of reactors not fueling one of the middle reactors means you loose not only the 400% heat from the reactor but also 100% from each of the 3 neighbours for a total loss of 700% heat.

But what does 700% loss mean for the overall reactor? Lets assume you need the output of one reactor to power your base. A 2x4 reactor produces 2800% heat for 8 fuel. Each fuel lasts 200 seconds and you produce 28 times what you consume. 8 fuel will last 5600s giving you 5.143 fuel per hour. Turn of one of the middle reactors and it drops to 2100% heat for 7 fuel. Each fuel lasts 200 seconds and you produce 21 times what you consume. 7 fuel will last 4200s giving you 6 fuel per hour.

As you see the difference is nearly one fuel per hour. But lets go further and only power one reactor. It needs to run all the time and needs one fuel every 200 seconds. Or 18 per hour. More than 3 times what 8 reactors need for the same power.

mrvn wrote: Mon Jan 20, 2020 12:16 pm it is always better to run more reactors.

Very true, so if I am running a sector, all the reactors in the 2x4 array are running; all the inserters are wired together so when one loads, they all load. The question was if it was better to run all 11 of my sectors even if I didn't need that much power.

astroshak wrote: Sat Jan 18, 2020 1:30 am Honestly, the solution is probably in a different direction than where you’re looking.

You are describing my solution 3, to figure out how much steam I need to have on hand, and store that much. I do understand that it is the standard solution. I am a non-standard kind of person so it doesn't appeal to me - this doesn't make it a bad solution, just that quirky, strange, you have to be insane kind of ideas appeal to me more

Honktown wrote: Sat Jan 18, 2020 12:21 pm If someone hasn't made a spreadsheet for it, they should:

Yes! I love spreadsheets!
For me, it is pretty easy to figure out how many tanks you need, once you know how much steam you need. And once you know how much time you need the steam to last, it is pretty easy to figure out how much steam you need. But how do you figure out how much time you need?

Amarula wrote: Mon Jan 20, 2020 3:04 pm

Honktown wrote: Sat Jan 18, 2020 12:21 pm If someone hasn't made a spreadsheet for it, they should:

Yes! I love spreadsheets!
For me, it is pretty easy to figure out how many tanks you need, once you know how much steam you need. And once you know how much time you need the steam to last, it is pretty easy to figure out how much steam you need. But how do you figure out how much time you need?

The turn-on time would depend on reactor design. In theory, you could try to figure it out by how many branches and how they'd heat up (length and heat buffer effects, heat distribution from branches), then the heat transfer decay as exchangers come online, etc. Or just overbuild, and assume it'll take at least a minute or two (the lowest the heat pipes can go is 500C, then the exchangers stop drawing energy).

BlueTemplar wrote: Sat Jan 18, 2020 2:04 pm I did something similar, here it is :
(it's pretty crude though - and don't forget about multiple sheets)

Yay! thank you for sharing your spreadsheet! (Did I mention I love spreadsheets?)

Sadly I could not figure out what your spreadsheet is doing. Perhaps I should mention that I worked as an auditor, and much of my time was spent auditing spreadsheets

Such as it is, here is my start at a spreadsheet to calculate steam storage requirements.

Oh dear here I go again : have a good laugh at my latest doofus mistake - we learn best in moments of enjoyment!

I love the design that disentius shared for laying out a very simple, compact, and elegant 2x4 reactor. I love it so much, as mentioned earlier in this thread, I replaced all my nuclear sectors with this design. Then I ran into what I am calling cycle issues: all the reactors (well 9 of 11 sectors) all loaded fuel at pretty much the same time, creating a cycle where they all power down at the same time. So I thought I would throw in a few steam storage tanks to tide me over while I figured out a long term solution.

Since I started working with nuclear, my designs have followed the very general idea of a blob of heat exchangers, connected together with miles of pipe (mostly for water, but lots for steam too). The steam pipes form a kind of steam cloud (well it looks more like a plate of spagetti underneath the obscuring mist), and all the steam turbines are fed from this cloud of steam. And to add storage, I just throw a few steam tanks into the steam cloud. It didn't really matter where, just a few pumps pushing steam from the heat exchangers into the tanks, and more pumps sucking it out to feed the turbines, so I would drop in the tanks wherever I found room.

So doofus mistake of the day : I just threw in a few steam storage tanks on top of disentius' beautiful design. If you haven't done so, I invite you to take a minute to go back and look at the design, its on page 1. In particular, notice where the steam storage goes.

Oh yeah. Right.

There isn't any steam storage.

In fact, there isn't any room for any steam storage.

I clearly was not thinking when I pulled out a couple of turbines from the top row, and stuck in a couple of steam tanks (with pumps in and out to make nice pressurized tanks). But silly me there was no steam storage cloud, so the steam tanks were only feeding those couple of turbines (well 8 total, two heat exchangers feeding two tanks, each feeding two turbines, on both sides of the top reactor). Now mind you, 100K steam will last a long time when it only has to feed 8 turbines But 8 turbines out of 192 is only 4% of capacity. So by the time the steam tanks started losing steam, I was already in trouble

I really don't want to mess with disentius' gorgeous design, so I'm off to rework the fueling controls. I'm thinking a timer but we'll see how that works out...

mrvn wrote: Wed Jan 08, 2020 3:42 pm

astroshak wrote: Sat Jan 04, 2020 1:58 am You don’t need a mod for that.

If you use Landfill and fill in a lake (making sure not to fill in where you want your Offshore Pumps to go) and then build the reactor setup entirely on that Landfill .. you can include the Landfill in your BP. It WILL NOT put Landfill where dry land is, but by including the Landfill in the BP, you can make bots build everything including setting up the Landfill by plopping that BP down twice : once for the Landfill, the second time after the Landfill is placed, for the rest of it.

If you have any part of the BP you are saving over dry land, the BP will not consider saving Landfill there. It needs to entirely be on Landfill.

But the mod is much better. It takes your blueprint without landfill and adds landfill everywhere needed and only where needed.

If you build your reactor by hand and misplace landfill you can take a blueprint of the whole thing (without landfill) and export it to the library. Then load the last save and place the blueprint with landfill added. Voila. mistake gone.

Also if you later want a second reactor you can just copy&paste it to a new lake without running into problems where the old reactor didn't have lake.

I wrote the mod and this makes me happy. The reason I wrote it was two main use cases: Nuclear and trains. I've already saved more time using it myself than it took me to write it! It makes designs where offshore pumps directly feed the heat exchangers much easier to place if I may say so myself.

Amarula wrote: Mon Jan 20, 2020 3:34 pm

BlueTemplar wrote: Sat Jan 18, 2020 2:04 pm I did something similar, here it is :
(it's pretty crude though - and don't forget about multiple sheets)

Yay! thank you for sharing your spreadsheet! (Did I mention I love spreadsheets?)

Sadly I could not figure out what your spreadsheet is doing. Perhaps I should mention that I worked as an auditor, and much of my time was spent auditing spreadsheets

Such as it is, here is my start at a spreadsheet to calculate steam storage requirements.
Steam.ods

Sorry for not answering, Java "helpfully" decided to delete itself during an update, because it claimed to be "unused" - while LibreOffice did rely on it ! And since this was supposed to be just a "temporary" OS...

Amarula wrote: Mon Jan 20, 2020 2:49 pm

mrvn wrote: Mon Jan 20, 2020 12:16 pm it is always better to run more reactors.

Very true, so if I am running a sector, all the reactors in the 2x4 array are running; all the inserters are wired together so when one loads, they all load. The question was if it was better to run all 11 of my sectors even if I didn't need that much power.

Asked and answered.

Was there anything unclear in "Assuming you have enough storage capacity and the right number of heat exchangers and turbines so no reactor ever reaches 1000 degrees..."?

mrvn wrote: Thu Jan 23, 2020 10:40 am Was there anything unclear

Thank you for patience and willingness to help me out. No, your explanation was clear and understandable.