Comprehensive power management

Post all other topics which do not belong to any other category.
zOldBulldog
Smart Inserter
Smart Inserter
Posts: 1107
Joined: Sat Mar 17, 2018 1:20 pm
Contact:

Comprehensive power management

Post by zOldBulldog »

I have done steam, Steam as backup, solar with perfect ratios, nuclear that only runs when needed, Steam for spike backup, accumulators for spike backup, factories that shut down one at a time (lowest priority first) if power is insufficient, and of course alarms that let me know of problem conditions.

And yet, I still don't have a foolproof system that tells me when to increase capacity of something. I know it is possible, but Factorio does not make it easy.

My hope is that this thread will generate ideas that will permit me to create such a thing.

Here are my early thoughts:

A) Priority of power use:
- Steam. Top priority while wood is incinerated (to quickly empty wood storage chests), then shuts down if other power sources are available.
- Solar
- Nuclear, not feeding reactors unless needed, and with only enough nuclear plants being fueled as needed to ensure current demand is met.
- Steam from coal/solid/rocket fuel, whichever is abundant at the moment.

B) Emergency shutdowns, in order:
- Used only when power from all sources is insufficient to supply all factories without brownouts.
- Rocket production (accumulator charge below 20%)
- Science/Research production (charge below 18%)
- High volume malls, those that produce large quantities of specific construction supplies like modules, belts, etc. (charge below 15%)
- Volume production of circuits. (charge below 12%)
- Volume smelting lines. (charge below 10%)
- Low volume starter smelting and mall. (charge below 10%)

C) Solar:
The traditional approach of perfect ratio panels/accumulators is great but makes it hard to identify when it is time to expand capacity, often leading to outages. So I am thinking of being a bit paranoid.
- A primary bank of panels/accumulators at perfect ratio.
- A secondary and normally isolated backup bank of panels/accumulators, sized to about 10-20% of main bank. It only connects to the main network if the main network accumulator charge drops to 25% (at night) and disconnects once charge is back over 50% (morning) . Rings a one-time alarm as it connects. It should be possible to disable the alarm for the scenario when intentionally relying on Nuclear or Steam as primary power. Alarm ringing is a sign of needing to expand capacity.

D) Nuclear:
- Begins feeding reactors if accumulator charge drops below 23%.
- Activates one nuclear plant at a time. Most of my plants are 4-reactor 480MW.
- So, for a plant to activate the previous plant must be on AND accumulator charge must be below 23%. Deactivates once accumulators charge reach 100%.
- No alarms.

E) Steam final failsafe:
- This is not much capacity, just the now evolved/fully built early game steam plant, that operates as a wood incinerator most of the time, but activates with the most convenient fuel available as power source of last resource. It might just provide enough juice to supplement the other power sources until you upgrade other power capacity.
- Activates if the accumulator charge drops below 21% and steam in any backup tanks is low. Deactivates once accumulator charge is back to 100%.
- It of course has an alarm. We should never reach this point, but anything is possible.

Bauer
Filter Inserter
Filter Inserter
Posts: 285
Joined: Fri May 05, 2017 12:48 pm
Contact:

Re: Comprehensive power management

Post by Bauer »

We are talking late game, right?
Here are my 2-3 cents.
  • Steam Power
    If you use stack filter inserters connected to the requester chest and let the content signal set the filter of the inserter, it will output wood first. IMHO, steam power is not powerful enough to scale up beyond, let's say, 100 steam engines. This is what? 2% of the energy needed for a large base? No need to do anything to control it. I connect it to the grid and that's it. It's a good use of the starter base steam energy setup and a convenient sink of wood.
  • Solar
    I don't see your problem. If you set the power screen to 10 mins and look at the end of the day, you see immediately if you need more solar. While the accumulators are charged, the solar production is a straight line. After that, it dops below this level and becomes uneven. As long as this happens well before the end of the day, you're good.
    If you're base has high peak power (e.g. while moving 200 chests completely filled with green chips using 80,000 bots) you add an array of extra accumulators to your otherwise perfect 21:25 ratio. Same rule as above applies.
  • Nuclear
    You don't worry about UPS? No problem to build a 10 GW plant with a relatively small footprint.
    I actually operate about 17 GW of nuclear together with 20 GW solar. No need to balance or anything because both are basically for free (uran is not exactly sparse).
  • Emergency shutdown
    In early game for the steam plant, later for the nuclear plant, it is a best practise to operate the plant itself (inserters, pumps, etc., in early game the coal miners) with a separate power plant. As a result, a brown out doesn't lead to a black-out within a few minutes. Your factory runs a little slower. So what? Easy enough to detect with a single accumulator.
    I think that your reaction, i.e. to have a shut down plan for different factories is neat. But what you really need is more power! It is extremely unlikely that your consumption will go down again. Ever. Really. So get going and build more power production. That's the only way out. (It's not likely that this was the last biter wave for a while, either.)

Hannu
Filter Inserter
Filter Inserter
Posts: 720
Joined: Thu Apr 28, 2016 6:27 am
Contact:

Re: Comprehensive power management

Post by Hannu »

Nice systems, but I did not understand what is the actual problem? You get easily enough information from electric info screen. You can see from graphs if your factory runs at full power and see straight how much extra capacity (or deficiency in bad situation) you have. I build new powerplant or expand old when load exceeds 70-80 % of maximum production. At beginning I usually double my production and at later phase I add one 480-1000 GW coal or nuclear plant.

How about if you made a combinator system, which counts time when battery charge is less than 100 %. At 100 % counter would be reset to 0. If it counts up to for example 350 s (if I remember correctly the day cycle is little more than 400 s), it gives an alarm. Then you know that your production struggles to keep batteries charged during a day.

Your action limits seem quite low. They may work if you have very large accumulator buffer. I have buffer of about 10 s. I begin to run coal power at 96 % and coal powerplants work at full power at 56 %. I disable my green circuit factory (largest single consumer, over 1 GW) at 33 %, but I have never noticed it has happen. I made it as safety system when I modernized my coal powerplants, because they had production breaks and also lack of fuel due on insufficient production. Nuclear powerplants work at full power always (I do not like unrealistic steam buffer controls and fuel cells are practically free). I have 2.4 GW nuclear capacity, about 1 GW (peak power) solar and about 3 GW coal. Maximum usage is about 5 GW.

zOldBulldog
Smart Inserter
Smart Inserter
Posts: 1107
Joined: Sat Mar 17, 2018 1:20 pm
Contact:

Re: Comprehensive power management

Post by zOldBulldog »

@Hannu & Bauer,

I tend to get very focused on what I am working on, so I end up not monitoring the power graphs and get brownouts/blackouts if I don't have alerts. And it "feels" like a waste to waste resources significantly overproducing power just to avoid them.

I already used earlier versions of auto-plant shutdowns and they saved my bacon more than once, when I failed to notice a needed power expansion before building a large factory line of something. Prior to using the auto-shutdowns I had some cascade power crashes that took more work than I cared for to dig myself out of. I plan to keep using auto-shutdowns, but I would very much like to get the alerts *before* reaching that point.

@Bauer,

Early/late game:

- The final design would be for late but I need the warnings while I evolve from early to late.
- Circuitry and chests are not an issue, I get them very early.
- My typical evolution is:
--- Steam with up to one 20/40 plant gets me all critical research and low volume mall production of everything.
--- Build first Nuclear or Solar.
--- Building volume production of building supplies
--- Build volume smelters, circuits and science train outposts while adding power as needed. This is the stage that most requires alerts, as I go from consuming 30-40Mw to several GWs very quickly.

Solar is what I most want to use (planning to go with one of the Advanced Solar mods to make it more compact) but my past experience is that without taking special measures it was very easy to suddenly find myself in a significant power shortage unless I massively overbuilt... and a distraction could get me in a painful mess to dig myself out of. That is why I am trying to find ways to ensure an automatic backup and alert when I am nearing the time to build more.

As to Steam, I got that one figured out (including the burners you mention). It works like a beauty from burner stage to late game. (https://factorioprints.com/view/-Ld1hKDrgO-7i3vLPB2d) It can probably take some minor improvements (in particular on the Steam circuits, and maybe for chaining more than one of these plants so that only those needed run) but otherwise I am very happy with it. Its various features serve a purpose at different stages of the game and is cheap to build.

@Hannu,

I agree that the action limits seem low. I used to use much higher limits. The reason why I am looking at lower limits with backup power sources to fill in the behavior of Accumulators... with all accumulators in a network keeping the same charge level.

So all measurements have to be based on the same numbers, from when to kick in power sources to when to shut down things. And to compensate for that is that I am looking at the various power sources being off but kicking in as needed... with the associated alert to let me know it happened. So, I should really never even reach a factory line shutdown, since I would have been alerted and expanded long before. Those shutdowns are there just as a failsafe. Plus, as long as I have some Nuclear power (or even Steam) if I ever mess up and reach plant shutdowns, shutting down the less needed lines will reduce my consumption, and the most critical lines will keep running on the 1 or a couple Nuclear power plants, letting me get out of the problem gracefully (been there, done that already in the past).

Bauer
Filter Inserter
Filter Inserter
Posts: 285
Joined: Fri May 05, 2017 12:48 pm
Contact:

Re: Comprehensive power management

Post by Bauer »

With 0.17 I'm doing my 5th or 6th from scratch to mega-base run.

After more than 8000 hrs I have learnt how to do this efficiently. I think.
There are 3 phases.

1. Start-up base
  • Lot's of red, green, and blue science. Some yellow and purple for the most important stuff.
    The idea is to get all the tech needed for a mega-base. Takes 3-4 hours.
  • Energy: I start with a 1/20/40 steam-setup and double this at after green. After setting up oil I have a lof of solid fuel which is burned in another 2/40/80 steam plant.
2. Scale up
  • I build everything needed for the mega base. This is mainly red and blue chips for 1000's of modules, furnaces, bots, etc. This base does NOT use a lot of modules and beacons itself. Hence, steel smelters (16 red belts) with a low energy food-print are fine (thanks god we don't need to talk about the CO2 foodprint).
  • Even though I start to use bots, mainly construction bots, this phase is not very energy hungry. 200 steam engines are sufficient. This time, I started a nuclear plant during this phase as a preparation for phase 3 (hoping that the new fluid system doesn't kill my UPS). I have a tileable blueprint and start with 4 reactors only during this phase.
3. Mega base
  • Succesively, I scale up the production. Starting with a large furnace array with about 6000 electric furnaces. In order to allow other parts to keep up, I put speed modules in the existing lines and upgrade from yellow to blue belts. In the end, everything is rebuild. Setting up final designs alread in phase 2 costs too much time. I also build intermediate factories, e.g. for plastics. Something quick-and-dirty, no productivity, only speed modules for 12 blue belts of plastics.
  • This basis is going to use a lot of energy. Like 20-30 GW thanks to the extensive use of modules. Even though many parts of the fab are belt base, I still have 80,000 logi bots and 50,000 construction bots. When demolishing one of the quick-and-dirty bot-based temporary fabs, or when placing 100,000 concrete, the construction bots let the energy consumption peak badly.
    • This is why I find it neccessary to separate the energy plant's energy from the main-grid. I use a few solar panels and accumulators to supply my nuclear plant.
    • The nuclear plant is scaled up as needed. In fact, in most cases, I reacted to a brown out. There is buffer chest close by with the stuff needed to add 4 reactors. Since the energy system of the plant is separated from the main grid, there is no mayor harm in a brown out.
    • In an earlier save, I used the recursive blueprint mod to build a self-expanding solar array. Given a sufficient solar module and accumulator production, the solar plant would add a big solar blue print automatically when the accumulators would go below 15%.

User avatar
BlueTemplar
Smart Inserter
Smart Inserter
Posts: 2126
Joined: Fri Jun 08, 2018 2:16 pm
Contact:

Re: Comprehensive power management

Post by BlueTemplar »

Are you scaling so fast that the low steam warning isn't good enough for the job ? (Have you tried increasing the number of steam tanks & steam engines / turbines ?)

mrvn
Smart Inserter
Smart Inserter
Posts: 4050
Joined: Mon Sep 05, 2016 9:10 am
Contact:

Re: Comprehensive power management

Post by mrvn »

With nuclear you want to always run all reactors at the same time to get the neighbor bonus. Use heat or steam as buffer and only refuel when you have enough empty capacity to buffer one fuel worth of output. Add a timer for the refuel cycle and if you need to refuel more often than every 200-220s sound the warning to expand the reactor.

zOldBulldog
Smart Inserter
Smart Inserter
Posts: 1107
Joined: Sat Mar 17, 2018 1:20 pm
Contact:

Re: Comprehensive power management

Post by zOldBulldog »

BlueTemplar wrote:
Mon Apr 29, 2019 10:03 am
Are you scaling so fast that the low steam warning isn't good enough for the job ? (Have you tried increasing the number of steam tanks & steam engines / turbines ?)
Yes, unless I am busy rethinking my designs, I do tend to scale quite fast. So fast in fact that I am currently rethinking my early production to produce enough belts, inserters, electric smelters, assemblers, beacons and modules... and supply trains to deliver them whole I am placing them.

I do have several steam tanks on the one steam power plant that carries me to nuclear, but I use compact nuclear plants with one steam tank per reactor.

But I like where you are going. I might build one nuclear plant (the first one) with a lot of steam tanks and use its steam levels / depletion times as an indicator of having to add power, rigged to an alarm. If it proves reliable through all the stages at preventing brownout then I can get rid of all the other alarms and auto-shutdowns.
mrvn wrote:
Mon Apr 29, 2019 11:23 am
With nuclear you want to always run all reactors at the same time to get the neighbor bonus. Use heat or steam as buffer and only refuel when you have enough empty capacity to buffer one fuel worth of output. Add a timer for the refuel cycle and if you need to refuel more often than every 200-220s sound the warning to expand the reactor.
True, although I remember reading that with 4-reactor plants you only lose about 10% of the theoretical maximum. That makes a compact, tileable design that can be placed and built through the map screen quite reasonable. It can let you save space that would be occupied by steam tanks (one tank and a few circuits are enough), and can be circuit controlled to have only as many of these plant modules running as you really need at any given moment.

For a traditional Megabase the max neighbor bonus design you propose is clearly best.

But a base that is more about building and trying new designs, where power demands and expansion cycles can change drastically over time... flexibility and adaptation is king. In my case, I experienced massive growth at one time, followed by times of so little consumption that my nuclear plants would remain unfueled and running off steam stored in their one tank per reactor... for 30-60 IRL minutes. So - for my style of play - it is clearly more efficient to power up 4-reactor modules, one module at a time, as needed.

zOldBulldog
Smart Inserter
Smart Inserter
Posts: 1107
Joined: Sat Mar 17, 2018 1:20 pm
Contact:

Re: Comprehensive power management

Post by zOldBulldog »

General questions:

1) Do reactors always take the same amount of time to heat up from cold to max temp?

2) Heat pipes lose temperature with distance. Do regular pipes lose it too?

3) Is there a signal anywhere that gives you steam temperature?

DaleStan
Filter Inserter
Filter Inserter
Posts: 277
Joined: Mon Jul 09, 2018 2:40 am
Contact:

Re: Comprehensive power management

Post by DaleStan »

1) No. They take the same amount of energy. Note that the reactor is usually connected to a larger system, and energy may be provided from or taken by the larger system, meaning that the reactor may not need to provide all the energy necessary to heat it up, or that the reactor may need to provide extra energy to heat both it and the rest of the system.

2) No

3) Yes

Ultros
Fast Inserter
Fast Inserter
Posts: 128
Joined: Thu Apr 04, 2019 4:25 pm
Contact:

Re: Comprehensive power management

Post by Ultros »

DaleStan wrote:
Mon Apr 29, 2019 1:05 pm
3) Yes
Steam temperature is quite pointless though, since it depends on only one variable, what plant produced it. A boiler will produce 180°C steam and a nuclear plant will produce 500°C steam.

DaleStan
Filter Inserter
Filter Inserter
Posts: 277
Joined: Mon Jul 09, 2018 2:40 am
Contact:

Re: Comprehensive power management

Post by DaleStan »

"Temperature - rounded temperature inside a reactor"

EDIT: I managed not to notice the "steam" part of "steam temperature".
Last edited by DaleStan on Mon Apr 29, 2019 3:18 pm, edited 1 time in total.

Bauer
Filter Inserter
Filter Inserter
Posts: 285
Joined: Fri May 05, 2017 12:48 pm
Contact:

Re: Comprehensive power management

Post by Bauer »

zOldBulldog wrote:
Mon Apr 29, 2019 12:18 pm
I do have several steam tanks on the one steam power plant that carries me to nuclear, but I use compact nuclear plants with one steam tank per reactor.

But I like where you are going. I might build one nuclear plant (the first one) with a lot of steam tanks and use its steam levels / depletion times as an indicator of having to add power, rigged to an alarm. If it proves reliable through all the stages at preventing brownout then I can get rid of all the other alarms and auto-shutdowns.

(...)

For a traditional Megabase the max neighbor bonus design you propose is clearly best.

But a base that is more about building and trying new designs, where power demands and expansion cycles can change drastically over time... flexibility and adaptation is king. In my case, I experienced massive growth at one time, followed by times of so little consumption that my nuclear plants would remain unfueled and running off steam stored in their one tank per reactor... for 30-60 IRL minutes. So - for my style of play - it is clearly more efficient to power up 4-reactor modules, one module at a time, as needed.
Uranium is so abundantly available on the map that we don't have to actively manage it. However, it's a beautiful concept to insert the fuel cell only if the energy is needed. On the one hand, you're right neglecting the neighor bonus for the same reason (I just wouldn't call this "more efficient"). On the other hand, centralizing the nuclear power production makes a lot of sense from a mantainance point-of-view (production and re-use of old fuel cells should be in one location).

Don't forget the engery storage in the heat pipes and reactors. A reative control loop for the fuel cell insertion can only work well if the energy storage capacity is small, both in the heat-exchangers and the steam tanks.

If you have slightly more turbines than the heat-exchangers can deliver steam for, the time between the fuel cell insertion (e.g. at a steam level of 50%) and the turn around (=rising steam levels because you produce more energy than your fab needs) is a good indication for the total demand compared to your total capacity as long as the capacity of the tank is as small as possible, i.e. 1 tank per "row".

If you want to increase your energy storage capacity, you can add steam storage tanks after the turbines without a pump.

-- heat-exchangers -- pump -- indicator tank -- pump -- turbines -- additional storage tanks

These additional storage tanks are only for periods in which you produce less energy (with the reactors running 24/7) than you need. In order to cover peak demands. This implies that you build significantly more turbines than needed for the set of heat exchangers. I recommend this only if we are talking about peak consumption only. If the consumption doesn't (quickly enough) fall below your max production capacity you'll be in deep trouble. You can set an alarm on the empty (<500) indicator tank to see/hear that your demand > capacity.

mrvn
Smart Inserter
Smart Inserter
Posts: 4050
Joined: Mon Sep 05, 2016 9:10 am
Contact:

Re: Comprehensive power management

Post by mrvn »

zOldBulldog wrote:
Mon Apr 29, 2019 12:26 pm
General questions:

1) Do reactors always take the same amount of time to heat up from cold to max temp?

2) Heat pipes lose temperature with distance. Do regular pipes lose it too?

3) Is there a signal anywhere that gives you steam temperature?
1)
A reactor with fuel produces heat but never heats the reactor above 1000°C. Heat then dissipates to/from connected entities like other reactors or heat pipes. Reactors them self are quite good heat capacitors. When adjacent reactors have fuel the neutrinos from one affect the other causing extra heat build up, the neighbour bonus. So reactors clustered together heat up faster. Consuming heat from reactors, e.g. to drive steam turbines, slows them down.

So no. reactors don't always take the same amount of time to heat up.

2) To paraphrase: Temperature can neigther be created nor destroyed, it can only be transfered.

It takes fuel to create heat. The heat is then transfered from the reactor to a heat pipe, from heat pipe to heat pipe and eventually from heat pipe to heat exchanger to water/steam.

The amount of heat transfered through a heat pipe depends on the difference of temperature and a heat exchanger needs a certain amount of energy/s to produce steam at full speed. It only gets that energy when each previous heat tile is a certain amount hotter. That limits the maximum distance a heat exchanger can be from a reactor and still be efficient. It looks like the heat pipes loose heat with distance but that's because you are cooling it with the heat exchanger.

3) No. Steam temperature is always the same for each boiler or heat exchanger. Don't mix steam from different sources and you simply know what temp it is.

mrvn
Smart Inserter
Smart Inserter
Posts: 4050
Joined: Mon Sep 05, 2016 9:10 am
Contact:

Re: Comprehensive power management

Post by mrvn »

Bauer wrote:
Mon Apr 29, 2019 2:21 pm
zOldBulldog wrote:
Mon Apr 29, 2019 12:18 pm
I do have several steam tanks on the one steam power plant that carries me to nuclear, but I use compact nuclear plants with one steam tank per reactor.

But I like where you are going. I might build one nuclear plant (the first one) with a lot of steam tanks and use its steam levels / depletion times as an indicator of having to add power, rigged to an alarm. If it proves reliable through all the stages at preventing brownout then I can get rid of all the other alarms and auto-shutdowns.

(...)

For a traditional Megabase the max neighbor bonus design you propose is clearly best.

But a base that is more about building and trying new designs, where power demands and expansion cycles can change drastically over time... flexibility and adaptation is king. In my case, I experienced massive growth at one time, followed by times of so little consumption that my nuclear plants would remain unfueled and running off steam stored in their one tank per reactor... for 30-60 IRL minutes. So - for my style of play - it is clearly more efficient to power up 4-reactor modules, one module at a time, as needed.
Uranium is so abundantly available on the map that we don't have to actively manage it. However, it's a beautiful concept to insert the fuel cell only if the energy is needed. On the one hand, you're right neglecting the neighor bonus for the same reason (I just wouldn't call this "more efficient"). On the other hand, centralizing the nuclear power production makes a lot of sense from a mantainance point-of-view (production and re-use of old fuel cells should be in one location).

Don't forget the engery storage in the heat pipes and reactors. A reative control loop for the fuel cell insertion can only work well if the energy storage capacity is small, both in the heat-exchangers and the steam tanks.

If you have slightly more turbines than the heat-exchangers can deliver steam for, the time between the fuel cell insertion (e.g. at a steam level of 50%) and the turn around (=rising steam levels because you produce more energy than your fab needs) is a good indication for the total demand compared to your total capacity as long as the capacity of the tank is as small as possible, i.e. 1 tank per "row".

If you want to increase your energy storage capacity, you can add steam storage tanks after the turbines without a pump.

-- heat-exchangers -- pump -- indicator tank -- pump -- turbines -- additional storage tanks

These additional storage tanks are only for periods in which you produce less energy (with the reactors running 24/7) than you need. In order to cover peak demands. This implies that you build significantly more turbines than needed for the set of heat exchangers. I recommend this only if we are talking about peak consumption only. If the consumption doesn't (quickly enough) fall below your max production capacity you'll be in deep trouble. You can set an alarm on the empty (<500) indicator tank to see/hear that your demand > capacity.
The neighbor bonus is totally about efficiency. Building less than 4 reactors in a block is simply wasteful. But you seem to have the uranium to waste. Pesonally I have other uses for it / not enough centrifuges to waste fuel. And how will you know to build new reactors before you max them out if you don't measure the time between when you need to fuel them?

For this the control loop doesn't have to be very reactive. I had setups with huge steam tanks and a trigger at 5% steam that would then burn through 10 or more fuel cells before shutting down at 95% steam. You only need to achieve 2 things:

1) Insert fuel early enough so the reactor can heat up to produce more steam before the remaining steam runs out.
2) Don't insert fuel when you can't store the heat/steam that would produce.

The first is critical as you get brownouts. The second just wasteful.

Note that a reactor and heat pipes already have a huge storage capacity. It doesn't take much steam tanks to buffer steam as well, just at least one tank to be able to read the steam level at all since the temperature of the reactor can't be read.

Bauer
Filter Inserter
Filter Inserter
Posts: 285
Joined: Fri May 05, 2017 12:48 pm
Contact:

Re: Comprehensive power management

Post by Bauer »

too much uranium.jpg
too much uranium.jpg (10.87 KiB) Viewed 2155 times

I'm mining U ore at 3 k/min and I'm using 24 fuel cells/min and 5.6 nuclear fuel/min (for trains).
So, yes, I have plenty.

However, atm my reactor setup is 2x52 reactors (15.5 GW) that is pretty close to ideal and yes, I'm using a control loop to not waste fuel.

zOldBulldog
Smart Inserter
Smart Inserter
Posts: 1107
Joined: Sat Mar 17, 2018 1:20 pm
Contact:

Re: Comprehensive power management

Post by zOldBulldog »

mrvn wrote:
Mon Apr 29, 2019 2:34 pm
zOldBulldog wrote:
Mon Apr 29, 2019 12:26 pm
General questions:

1) Do reactors always take the same amount of time to heat up from cold to max temp?

2) Heat pipes lose temperature with distance. Do regular pipes lose it too?

3) Is there a signal anywhere that gives you steam temperature?
1)
A reactor with fuel produces heat but never heats the reactor above 1000°C. Heat then dissipates to/from connected entities like other reactors or heat pipes. Reactors them self are quite good heat capacitors. When adjacent reactors have fuel the neutrinos from one affect the other causing extra heat build up, the neighbour bonus. So reactors clustered together heat up faster. Consuming heat from reactors, e.g. to drive steam turbines, slows them down.

So no. reactors don't always take the same amount of time to heat up.

2) To paraphrase: Temperature can neigther be created nor destroyed, it can only be transfered.

It takes fuel to create heat. The heat is then transfered from the reactor to a heat pipe, from heat pipe to heat pipe and eventually from heat pipe to heat exchanger to water/steam.

The amount of heat transfered through a heat pipe depends on the difference of temperature and a heat exchanger needs a certain amount of energy/s to produce steam at full speed. It only gets that energy when each previous heat tile is a certain amount hotter. That limits the maximum distance a heat exchanger can be from a reactor and still be efficient. It looks like the heat pipes loose heat with distance but that's because you are cooling it with the heat exchanger.

3) No. Steam temperature is always the same for each boiler or heat exchanger. Don't mix steam from different sources and you simply know what temp it is.
1) Thanks, and I think I worded question (1) quite badly given everybody's responses. I probably should have said: In a specific setup (i.e.: 4 reactors, a specific layout of pipes, heat exchangers and turbines) and starting from a cold reactor, is it safe to assume they will take the same amount of time to reach operating temperature each time I fuel it?

2) Excellent. But now I am a bit confused, why is it that most people try to mimimize pipe distances from reactor to heat exchangers?

3) Again good. And again confused, if the setup is equally efficient from 500-1000, why are people so focused on heating things up to 999 degrees?

For reference, this is the 4-reactor setup I currently use (without any kind of alerts or chaining... yet). It is derived a BlakeMW's design (I am still working to understand all of the details of nuclear design). I mainly reorganized it to make it tileable, have good spots for roboports and fit my standard build-pad sizes. Fuel cells are handled by a few logistic bots. The setup is placed remotely through the map, built by construction bots with supplies delivered by train.
BulldogNuclearPower.jpg
BulldogNuclearPower.jpg (867.51 KiB) Viewed 2151 times
Full description at: https://factorioprints.com/view/-Lbz1RINCoczAIED4X9r
Edit: I just noticed the screenshot is out of date (missing wires to the bottom two tanks, but they should already be in the blueprint itself).

User avatar
BlueTemplar
Smart Inserter
Smart Inserter
Posts: 2126
Joined: Fri Jun 08, 2018 2:16 pm
Contact:

Re: Comprehensive power management

Post by BlueTemplar »

mrvn wrote:
Mon Apr 29, 2019 2:41 pm
Note that a reactor and heat pipes already have a huge storage capacity. It doesn't take much steam tanks to buffer steam as well, just at least one tank to be able to read the steam level at all since the temperature of the reactor can't be read.
Yeah, here's one design for a very different situation (DeathWorld Marathon) :
40MW_2x_Reactor.jpg
40MW_2x_Reactor.jpg (122.22 KiB) Viewed 2110 times
It's a 40MW, 200% efficiency design.*
It has just enough overall thermal storage to store the 8x2x2=32 GJ from the 2 fuel cells (even if the turbines are never working).
(Note that reading levels from all the 4 tanks is probably superfluous.)
Insertion is manual so far.

EDIT : Actually, looks like that trying to be efficient by using extra heat pipes might have been pointless :
1 Exchanger + 0.825.. Tank stores 5 times more than a Heat Pipe over a 200 seconds cycle (2.5 GJ over 0.5 GJ),
In expensive mode, 1 Exchanger + 0.825.. Tank costs as much copper as 5 Heat Pipes, but only 1.778.. times more iron !
(A reactor itself stores 10 times more, but costs 25 times more in copper and 50 times in steel only...)
(EDIT : And this is not even counting the overhead to heat the pipes - though it's negligible in the long run.)
This is a shame - looks like that game balance might be improved here...

*Thanks to the folks at freenode's #factorio that did most of the testing for the concept.
Last edited by BlueTemplar on Mon Apr 29, 2019 7:39 pm, edited 1 time in total.

User avatar
BlueTemplar
Smart Inserter
Smart Inserter
Posts: 2126
Joined: Fri Jun 08, 2018 2:16 pm
Contact:

Re: Comprehensive power management

Post by BlueTemplar »

mrvn wrote:
Mon Apr 29, 2019 2:34 pm
2) To paraphrase: Temperature can neigther be created nor destroyed, it can only be transfered.

It takes fuel to create heat. The heat is then transfered from the reactor to a heat pipe, from heat pipe to heat pipe and eventually from heat pipe to heat exchanger to water/steam.

The amount of heat transfered through a heat pipe depends on the difference of temperature and a heat exchanger needs a certain amount of energy/s to produce steam at full speed. It only gets that energy when each previous heat tile is a certain amount hotter. That limits the maximum distance a heat exchanger can be from a reactor and still be efficient. It looks like the heat pipes loose heat with distance but that's because you are cooling it with the heat exchanger.
Yeah, though there's no conservation of temperature, but only of energy :
For instance 1 Heat Pipe stores 1 MJ/°C, while 100L of Steam in a pipe store 200*100 = 0.02 MJ/°C.
If the Heat Pipe heats Steam (through a Heat Exchanger), the Heat Pipe going down 1°C will result in the 100L of Steam in the Pipe going up 1/0.02= 50 °C.
Physics Musings
zOldBulldog wrote:
Mon Apr 29, 2019 3:20 pm
1) Thanks, and I think I worded question (1) quite badly given everybody's responses. I probably should have said: In a specific setup (i.e.: 4 reactors, a specific layout of pipes, heat exchangers and turbines) and starting from a cold reactor, is it safe to assume they will take the same amount of time to reach operating temperature each time I fuel it?
You should rather consider starting from a 500°C-warm system : the "heat-fluid" elements can't have their temperature decrease under that temperature.
zOldBulldog wrote:
Mon Apr 29, 2019 3:20 pm
2) Excellent. But now I am a bit confused, why is it that most people try to mimimize pipe distances from reactor to heat exchangers?
I'd guess that if your heat pipes are too long, you can end up in the situation when the reactor cannot evacuate heat fast enough, and you end up wasting energy ?
zOldBulldog wrote:
Mon Apr 29, 2019 3:20 pm
3) Again good. And again confused, if the setup is equally efficient from 500-1000, why are people so focused on heating things up to 999 degrees?
You want to never reach 999 degrees (displayed) in the reactors, if you reached that temperature, you're likely wasting energy.
Conversely, if you never reach 998 degrees (displayed) in the reactors, then you probably over-built the heat evacuation & storage part, and could be more efficient.

Bauer
Filter Inserter
Filter Inserter
Posts: 285
Joined: Fri May 05, 2017 12:48 pm
Contact:

Re: Comprehensive power management

Post by Bauer »

mrvn wrote:
Mon Apr 29, 2019 2:41 pm
(..) And how will you know to build new reactors before you max them out if you don't measure the time between when you need to fuel them?

For this the control loop doesn't have to be very reactive. I had setups with huge steam tanks and a trigger at 5% steam that would then burn through 10 or more fuel cells before shutting down at 95% steam.
I insert the new fuel cell if the steam level falls below 50% (and only if the inserters removed the used cell).
My turbines are over-dimensioned by about 15%. If I use more steam than the heat exchangers create, the steam level will fall below 30%.
This is (was) my indication that I need to expand the power production.
In the meantime, I added more than 10 GW of solar. Now my indication is the time needed to charge the accumulators.

A very large steam storage leads to a very large eigenfrequency of the control loop. Or in your words, the respond time is 10 fuel cycles (=2000 seconds=33 minutes). IMHO, this is quite long.
However, as long as the total heat capacity of the reactors, heat pipes, and heat exchangers is not too large, this is not really a problem, because the respond time to a very high energy consumption will be realively swift.

What I don't understand is: What do you need a very large energy storage for? As long as the number of turbines is not much larger than the "ideal" ratio, you will not be able to get that stored energy out at a rate that is larger than your reactors can create energy.

How do you dimension the steam storage to the number of turbines (or the turbine/exchanger ratio) and why?

Post Reply

Return to “General discussion”

Who is online

Users browsing this forum: No registered users