Please explain to me how the electrical system works.

[adamwong246]

I've noticed some quirks and it's led me down a rabbit-hole. Maybe a dev can clue me in to who it works under the covers?

Part 1: Prologue

Like all bases, mine started with steam. At this point I was paying little attention to the grid, simply adding more steam capacity when the power went out.

Then I discovered solar+accumulators. I learned that by using solar, my production was way up during the day and I needed the right proportion of accumulators to get through the night. So I simply added more solar+accumulators as needed.

Then I discovered nuclear. Nuclear fuel burns for a long time and once the reaction starts, you have to let it finish. So it's possible to use nuclear fuel inefficiently, as I discovered when I realized my nuclear reactor was activating in the early dawn and running throughout the day- almost all of that fuel cell was wasted.

Then I discovered switches and signals. I realized I could use the signal from the accumulator to drive 3 sub-grids via switches. During the daytime, when energy was surplus, combinators would flip ON the switch to my secondary bus. At night time, the primary and secondary buses switched OFF to preserve my vast walls of laser turrets.

Part 2: The problem

By this point, I had 1 nuclear reactor, a surplus of nuclear fuel and enormous solar+accumulator farms. With multiple reactors, I was able to deactivate my steam plant to limit pollution. My system worked great for a long time- until I ran out of nuclear fuel. My solar resources weren't enough to keep my lasers up all night and the biters nearly got me. I quickly reenabled my steam plant while I sussed out my nuclear fuel problem.

Using what I had learned, I realized that I could use signals to trigger the steam-plant as an emergency backup generator but it's also where I ran into trouble- steam engines already exhibit this behavior. You don't need signals to activate steam energy, as it scales up and down automatically according to need. If your grid can absorb more power, the steam engine turns. Otherwise, it just halts.

Part 3: Factorio mechanics as I understand them

If you are moving an item off an assembler and onto a belt, and that belt is already full, the inserter will halt. This will halt the assembler and ad infinitum. The same logic applies to fluids and pipe, and I suspect it's how the electrical grid works as well. Power suppliers insert units of electricity and these units are quickly routed to a machine. But just like the other transport mechanisms, if theres no room for additional electricity, the "inserter" halts, where in this case the "inserter" is a steam engine. It may be technically impossible to change this behavior. If so, please tell me!

Part 4: Network priorities

from the wiki:

Electricity is provided on a priority basis. The demand for energy is satisfied by generators in following order:

Solar panels – Top priority; they always work at maximum performance available, unless they can cover all demand of the network, in which case they match demand.
Steam engines and Steam turbines – They match whatever demand solar panels cannot satisfy; note that Engines and Turbines do have the same priority, leftover demand is equally divided among both.
Accumulator – Last resort. They are only discharged when demand cannot be met by other means. They are also only charged when all demand is met, and there is yet more power available.
There may be situations where different behaviour is desired (such as solar panels combined with accumulators for night-and-day delivery), in which case clever use of a power switch and the circuit network is in order.

Part 5: A mod

The goal of this screed is to explain my idea for a mod which removes these safe-guards, requiring the player to REALLY manage their grid. Electricity that is over-produced will simply be wasted. Boilers which are not disabled will continue to spew pollution, regardless of the electrical need. And accumulators now need to be controlled, switched between charging and discharging via signals. I would like full control of the electrical grid, without handrails.

But the real point of this post is to ascertain if such a mod is even possible
And if so, how difficult would it be?
Lastly, is my understanding of the grid mechanics accurate?

[Serenity]

So it's possible to use nuclear fuel inefficiently

With Kovarex enrichment it's really not an issue and you can just let reactors run all the time. A single centrifuge can run around 52 reactors while using a minuscule amount of uranium:
https://kirkmcdonald.github.io/calc.htm ... 3:p3:p3:p3

If you want to be efficient you can buffer steam, measure the steam level in the tanks and only insert new fuel cells if the steam level is too low

You don't need signals to activate steam energy, as it scales up and down automatically according to need. If your grid can absorb more power, the steam engine turns. Otherwise, it just halts.

That's not true. As you've read accumulators have the lowest priority. You do need to switch steam power if you want to use it only when the accumulators are empty. Once you get a significant amount of solar coal power is supposed to be a backup. If you don't use combinators they will both be used together

For solar + coal backup you want solar -> accumulators -> coal

[foamy]

That's not true. As you've read accumulators have the lowest priority. You do need to switch steam power if you want to use it only when the accumulators are empty. Once you get a significant amount of solar coal power is supposed to be a backup. If you don't use combinators they will both be used together

For solar + coal backup you want solar -> accumulators -> coal

I prefer to switch it on when the solar can't keep up, which is somewhat more complex; you need to be able to gauge the accumulator discharge rate and see if you're approaching maximum discharge (i.e. brownout), among other bits.

[astroshak]

There is/was a mod that contained “low priority steam generators” that were intended to come on when the regular power generation was not sufficient - for say, the power demand of a bunch of lasers suddenly having to fire. These things were on the same priority as accumulators, and the mod even warned you, that they broke accumulators. I never actually used the things personally, however.

I do not know if it is possible to add in new tiers of priority. It was apparently not possible back then; I do not know if such functionality has been added as I do not write mods.

As someone else mentions, the Kovarex process basically provides you with effectively unlimited 235 (once you amass enough to start it). This renders fuel economy concerns about reactors moot, as one 235 gets you 2000 seconds of reactor warming time. A small patch of uranium can supply your entire reactor power needs for months of IRL gaming.

That said, there IS a way of conserving nuclear fuel. Its not really necessary, thanks to Kovarex basically giving you unlimited 235 from your stockpile of 238, but it is definitely possible. Simply store the extra steam in Storage Tanks. It takes about 14 tanks to store the steam from a 1x2 reactor setup, and for every 1x2 you tack on (making it a 2x2, then a 3x2, then a 4x2, etc) add an additional 26 storage tanks. Monitor the level in those tanks, and when it starts to get low (but not low enough that you’ll run out before the power plant produces more) that is your trigger to control the reactor fuel input and spent fuel output Inserters. The setup for this is a Green Wire connects the two Inserters, with the fuel input Inserter set to Stack Size 1 and to Enable when Spent Fuel Cell > 0. The spent fuel cell output Inserter is set to Read Hand Contents (Hold), and to Enable when Steam < whatever amount you determined. Connect this Inserter to the steam storage tank(s) with the Red Wire.

Once you get it up and running, which does include manually feeding the first Fuel Cell to the reactors, it is automatic and does NOT waste nuclear fuel.

What it does not do, though, is what you wanted TC : allow you to control it manually. ALL of the electricity generating devices, modded and vanilla, that I have encountered automatically throttle the steam used based upon the electrical demand. This mirrors real life, as well, where steam to the turbines (or fuel to the engine) is throttled based upon the speed of the rotor in the generator. As electrical demand increases, it increases the drag on the rotor, which attempts to slow the turbine (or engine). The machine senses the speed reduction and throttles the input up to maintain the speed, thus using more steam/fuel to provide more electrical power. As electrical demand decreases, the reverse happens : drag on the rotor decreases, the machine speeds up, and gets automatically throttled down to maintain speed.

That's not true. As you've read accumulators have the lowest priority. You do need to switch steam power if you want to use it only when the accumulators are empty. Once you get a significant amount of solar coal power is supposed to be a backup. If you don't use combinators they will both be used together

For solar + coal backup you want solar -> accumulators -> coal

I prefer to switch it on when the solar can't keep up, which is somewhat more complex; you need to be able to gauge the accumulator discharge rate and see if you're approaching maximum discharge (i.e. brownout), among other bits.

Actually, you can do that without combinators. Simply put a single Accumulator down near the Offshore Pumps providing water to your coal/wood/solid fuel power plant, and wire it to those Offshore Pumps. Set them to operate when A is less than some desired amount. If you have enough power generation capacity by boiler, your Offshore Pumps will rapidly cycle on and off when A gets down that low. Add in a SR-Latch (which is what you could want Combinators for) to prevent the rapid cycling, but its not strictly speaking a necessity.

[starlinvf]

Quick side bar on the OP's post.

The Electrical networks are not flow-based like fluids. Each network is a flat resource pool, and is evenly distributed across all machines bound to it. There can also be multiple networks on the map, each essentially having a relatively flat overhead cost to process. Networks can also be merged, like when adding a power pole to connect things, or power switches enabling. The way its been explained to me; you want to avoid creating multiple power networks at larger scales (between the overhead cost and unpowered machines being more UPS expensive than idle machines), and avoid power switches flapping/cycling as merging networks is comparatively expensive to process.

Thats why you'll see a lot of suggestions to Control fluid sources to stop a Production Node's input rather then using power switches, as thats computationally cheaper and effectively does the same thing. But as with everything in Factroio, there are fringe cases where different solutions could be more desirable for different design demands.

There is some situations where it takes multiple ticks for a machine to respond to triggers; but for the most part this is inconsequential until you start scaling into megabases with UPS problems, or are working with designs using very delicate timing set ups.



I can't speak to the difficulty of making a mod like that....but the idea sounds plausible. But one flaw in the overall idea is that its giving up scaling nature of Steam Gens, and turning the whole thing into "stepping" values; which I think might end up being counter productive in practice because of how it massively complicates precision power production. AMs and Inserters cause huge spikes/dips in power draw on tick by tick basis as they activate. Big enough with late game modules that you have multiple megawatts difference between ticks. I speculate that this on its own would make managing it with Circuit networks a huge pain in the ass, unless you purposely over-produce power.... which defeats the purpose of circuit controls in the first place.

To counter that, the quick and dirty solution would be to have all machines draw max power at Idle, giving you a fixed target to work with based on whats bound to that network. So in short... all power production nodes operate at 100% output, while all factories draw max power at all times.

Now thinking about it.... a second major flaw with the overall idea is that it defeats the purpose Accumulators of you have to manually switch them on and off. Even with dynamic power draw in play, the response time alone makes it impossible to manage correctly. You either completely over engineer power production to avoid brown out; or you're forced to use lagging circuit logic to respond to brown outs.... which already has a tendency to spiral into blackouts, because all your fuel production needs electricity to self-sustain.

[adamwong246]

It's unfortunate but I feel as if I am pulling in a different direction than the community.

Massive megabases ARE really cool but that's really not the draw, for me personally. I don't really want to push UPS with a planet-sized factorio- I'd rather have a smaller, but more technically complex base. You can always build bigger and eventually overload the game, so I'm not sure why so many people think pushing the limits is the true end-game?

As an aside, why are signals such the problem child? There so freakin cool but most people steer me away from them.

[Koub]

It's unfortunate but I feel as if I am pulling in a different direction than the community.

Massive megabases ARE really cool but that's really not the draw, for me personally. I don't really want to push UPS with a planet-sized factorio- I'd rather have a smaller, but more technically complex base. You can always build bigger and eventually overload the game, so I'm not sure why so many people think pushing the limits is the true end-game?

As an aside, why are signals such the problem child? There so freakin cool but most people steer me away from them.

You should ignore others' opinions if they are not relevant for your favourite playstyle. Factorio's moddability will allow you to craft the very specific gameplay experience you desire. There are mods for almost anything and everything.

[starlinvf]

It's unfortunate but I feel as if I am pulling in a different direction than the community.

Massive megabases ARE really cool but that's really not the draw, for me personally. I don't really want to push UPS with a planet-sized factorio- I'd rather have a smaller, but more technically complex base. You can always build bigger and eventually overload the game, so I'm not sure why so many people think pushing the limits is the true end-game?

As an aside, why are signals such the problem child? There so freakin cool but most people steer me away from them.

Both can be answered by "communal knowledge" as an aggregate. The nature of the game is VLSI, Adaptation and Optimization. So naturally when people ask questions, its in usually in service of trying to solve a problem, rather then create new ones. However..... lack of understanding is what creates new problems, when the person thinks they're creating a solution. As a result, when people try to describe a problem, the first few answers you get are going to be a relatively simple solution, with an explanation as to why you'd use this solution.

Secondly, you have a solution that you're trying to invent a problem for... and thats something that needs to be more carefully explained, and demands a greater justification as to why anyone should care in the first place. Is this something that helps the game as a whole? Or is this just to satisfy your OCD? That latter one happens way too often, so people are getting tired of it.

As to why UPS/SPM is such a big deal around here? Its an easy to grasp benchmark to represent a myriad of things the game is doing, as a means to calculate efficiency and performance. The game's primary target is whats needed to accomplish a rocket launch. However, because "everything scales", highly efficient code performance has multiple benefits across the board for almost everyone who plays. Better baseline performance of vanilla means mods have more wiggle room on performance cost. Better performance scaling with mods means more mods can run in parallel without ruining the play experience. Problems identified at very large scales can also lead to better methodology that works at all scales. Better overall performance of existing code also means more overhead to add new features that cost against performance budget.


if you've ever closely followed the Dev life cycle of a Modern MMO, you'll quickly find most live on the edge of their performance budget on a nearly constant basis. As more content and features are added, they engine needs to be continually refined to lower the cost or expand the overhead capacity to keep being able to expand the game over time. If you've ever heard of the term Accretion in continuous development, and long life titles, this is what they're talking about. However... this is finite. Theres only so much you can revamp and modify within an existing architecture, before you reach the hard limits of what that architecture is good for. Thus, how well does your design scale?


Of course that doesn't stop you from making the mod anyway. Or running whatever mods you want. Its just that if you're gonna involve the community for any reason, you'll to provide enough reason for them to care enough to invest the effort into the discussion at hand.

[starlinvf]

It's unfortunate but I feel as if I am pulling in a different direction than the community.

Massive megabases ARE really cool but that's really not the draw, for me personally. I don't really want to push UPS with a planet-sized factorio- I'd rather have a smaller, but more technically complex base. You can always build bigger and eventually overload the game, so I'm not sure why so many people think pushing the limits is the true end-game?

As an aside, why are signals such the problem child? There so freakin cool but most people steer me away from them.

You should ignore others' opinions if they are not relevant for your favourite playstyle. Factorio's moddability will allow you to craft the very specific gameplay experience you desire. There are mods for almost anything and everything.

I would disagree if you're gonna involve others in something that asks for more then a casual effort. "A solution in need of a problem" would require more effort in explaining to a group thats spent an significant amount of time to finding and sharing the optimal set of solutions to existing problems. Or in this case.... why a problem should be more complicated then it already is.

[Serenity]

so I'm not sure why so many people think pushing the limits is the true end-game?

Yeah it's a somewhat silly obsession that can distract from the game and from having fun. Performance is important, but people do a lot of premature optimization. They will tell you do not do this and that and worry about tiny details. When in reality much of it only matters in specific circumstances.

There is nothing wrong with building smaller bases and starting over when you get bored. Moreover, these days you can build a multiple thousand SPM base and not necessarily worry too much about performance. The game is optimized to such a degree that a base was cutting edge and caused issues years ago runs just fine today.

[adamwong246]

I'd like to pull this conversation back to the question at hand.

But the real point of this post is to ascertain if such a mod is even possible
And if so, how difficult would it be?
Lastly, is my understanding of the grid mechanics accurate?

[starlinvf]

I'd like to pull this conversation back to the question at hand.

But the real point of this post is to ascertain if such a mod is even possible
And if so, how difficult would it be?
Lastly, is my understanding of the grid mechanics accurate?

Wouldn't it more productive to talk to other modders instead, then?



viewforum.php?f=82

[foamy]

Actually, you can do that without combinators. Simply put a single Accumulator down near the Offshore Pumps providing water to your coal/wood/solid fuel power plant, and wire it to those Offshore Pumps. Set them to operate when A is less than some desired amount. If you have enough power generation capacity by boiler, your Offshore Pumps will rapidly cycle on and off when A gets down that low. Add in a SR-Latch (which is what you could want Combinators for) to prevent the rapid cycling, but its not strictly speaking a necessity.

I didn't think offshore pumps could be wired? But either way, no, that's not sufficient for brownout prevention. That might save you from blackouts -- though it's arguable as to how effective it is at such -- but it doesn't allow you to detect when your accumulator banks are being overwhelmed. To detect an impeding brownout before it impacts your machines you need to track accumulator discharge rates, and that means combinators.

[astroshak]

It is as effective as the fossil fuel power plant can be. If you have 2 GW of solar power, need 2.4 GW of power thanks to expanding some production using beacons and modules, but still have a 36 MW fossil fuel plant set up, then no, using the Accumulator wired to the Offshore Pump won’t save you from a brownout or blackout.

Ultimately, thats why I stopped doing that at all. In order to keep things functioning smoothly, I had to choose to either keep developing main power or backup power. And I really cannot see any point in continuing to develop backup power when keeping main power well above demand is simply the better choice. I see the 36/72 MW fossil fuel plant scenario, as backup to a multi-GW solar or nuclear power plant, as being the most likely thing. Those few MW (relative to the GW main power plant) are essentially meaningless.

It may be useful while still getting your first few solar BP’s up and running, but once you have a few hundred MW of solar or nuclear, having backup power generation is a waste of time and effort that would be better spent on expanding primary power generation capacity.

Detect an impending brownout? If you’re using solar, either you have the power or you don’t. Either you have enough solar panels and accumulators, or you don’t. If you don’t, you’re not just going to have brownout, it’ll be a blackout come nightfall. Nuclear? Yea, there you can have a brownout, but simply extending the 2xN reactor setup by 1 pair of reactors (which should get you an additional 320 MW of power) is most likely going to dwarf the contribution of your fossil fuel plant.

For the brief window where it is useful, simply hooking an accumulator to the offshore pumps is sufficient. Later on, its insufficient unless you’re spending an inordinate amount of time maintaining and expanding the fossil fuel power plant.

[foamy]

Detect an impending brownout? If you’re using solar, either you have the power or you don’t. Either you have enough solar panels and accumulators, or you don’t. If you don’t, you’re not just going to have brownout, it’ll be a blackout come nightfall. Nuclear? Yea, there you can have a brownout, but simply extending the 2xN reactor setup by 1 pair of reactors (which should get you an additional 320 MW of power) is most likely going to dwarf the contribution of your fossil fuel plant.

I think we may have a definition collision here. When I say 'brownout' I mean the state of having power supplied to the grid, but it being insufficient for all draws, resulting in things like inserter slowdowns and whatnot. This can occur with any kind of power supply, including solar; it simply means that there's not enough power for all the loads put on it.

Solar is less vulnerable to brownouts, yes. This is because since accumulators can discharge in approximately 16.7s, and the night is ~41.7s long, a pure solar setup capable of lasting through the night at average power draw will be able to supply a minimum of roughly 2.5x that in spike loading. Likewise, in the day, a spike load can draw from both the panels and the accumulators. However, it's still possible to do, and I've done it.

But it's actually solar you most want the early warnings with! A large-scale solar setup with a backup that trips only on accumulator level is unlikely to have that backup engage in time to accomplish anything. You want to plug your backup into the system early if you pick up warning signs that you're over-drawing your accumulators. This requires some circuit trickery; the simplest one is to have it trip if the accumulators ever discharge above ~30% of their maximum rate, say. More complexly, you can do some forecasting of where the accumulators will end up if the backups are not engaged and turn them on if that's below your comfort level.

A couple of other points: Not all backup power systems are created equal. For a start, you can use nuclear as the backup to a solar main system (and a boiler system to backup the nuclear plant's various bits, if you wish). A single quad-core reactor can drive 480MW continuously and way more if you include tanks and additional turbines, and makes a spectacularly compact and useful backup plant if you've expanded your factory a bit too much. Secondly, you appear to be thinking in terms of quite small boiler setups; a basic 1:20:40 setup for the 36MW you quoted. I will usually build out to six times that, and then convert the engines into turbines and add tanks when I do the conversion to backup power generation. That pulls the generation capability up into the 432MW range until the tanks run out, and that'll take a while: six turbines, being 1/2 supplied from boilers and 1/2 supplied from tanks, will draw 180/s from the tanks; every tank, therefore, buys you about 138s of time to fix the problem. And it's not particularly difficult to build more of those if you really want to.

[astroshak]

Detect an impending brownout? If you’re using solar, either you have the power or you don’t. Either you have enough solar panels and accumulators, or you don’t. If you don’t, you’re not just going to have brownout, it’ll be a blackout come nightfall. Nuclear? Yea, there you can have a brownout, but simply extending the 2xN reactor setup by 1 pair of reactors (which should get you an additional 320 MW of power) is most likely going to dwarf the contribution of your fossil fuel plant.

I think we may have a definition collision here. When I say 'brownout' I mean the state of having power supplied to the grid, but it being insufficient for all draws, resulting in things like inserter slowdowns and whatnot. This can occur with any kind of power supply, including solar; it simply means that there's not enough power for all the loads put on it.

Solar is less vulnerable to brownouts, yes. This is because since accumulators can discharge in approximately 16.7s, and the night is ~41.7s long, a pure solar setup capable of lasting through the night at average power draw will be able to supply a minimum of roughly 2.5x that in spike loading. Likewise, in the day, a spike load can draw from both the panels and the accumulators. However, it's still possible to do, and I've done it.

But it's actually solar you most want the early warnings with! A large-scale solar setup with a backup that trips only on accumulator level is unlikely to have that backup engage in time to accomplish anything. You want to plug your backup into the system early if you pick up warning signs that you're over-drawing your accumulators. This requires some circuit trickery; the simplest one is to have it trip if the accumulators ever discharge above ~30% of their maximum rate, say. More complexly, you can do some forecasting of where the accumulators will end up if the backups are not engaged and turn them on if that's below your comfort level.

A couple of other points: Not all backup power systems are created equal. For a start, you can use nuclear as the backup to a solar main system (and a boiler system to backup the nuclear plant's various bits, if you wish). A single quad-core reactor can drive 480MW continuously and way more if you include tanks and additional turbines, and makes a spectacularly compact and useful backup plant if you've expanded your factory a bit too much. Secondly, you appear to be thinking in terms of quite small boiler setups; a basic 1:20:40 setup for the 36MW you quoted. I will usually build out to six times that, and then convert the engines into turbines and add tanks when I do the conversion to backup power generation. That pulls the generation capability up into the 432MW range until the tanks run out, and that'll take a while: six turbines, being 1/2 supplied from boilers and 1/2 supplied from tanks, will draw 180/s from the tanks; every tank, therefore, buys you about 138s of time to fix the problem. And it's not particularly difficult to build more of those if you really want to.

We have the same definition of a brownout. If you have a brownout during the day, with solar, you’re going to have a blackout after night falls.

I too have done the 6x 1/20/40 array before; I usually start with two of those arrays, building them up as needed, one on each side of a yellow belt of coal that I’ll upgrade to solid fuel. I once had a full blue belt of solid fuel being eaten by a bunch of those arrays before, when I did not go nuclear or solar that game. So what? Anecdotes like that are meaningless. Its not usually necessary to expand fossil fuel power plants beyond 1-2 arrays like that unless sticking with them past the point where you can start making solar, or nuclear (without Kovarex backup). Obviously the bigger you build the sooner you’ll need to upgrade or expand.

And again, rather than build up the backup I’d rather use the resources to better build up the main power. Eventually that 432 MW of power won’t mean much, will it? Not the duration it lasts, but the amount of power itself. Very large factories use GW of power; less than half a GW will provide a low power brownout but not much else.

Ultimately, you need to provide power for all your machinery. You can do it via several means, fossil fuels, solar, or nuclear (or additional means if playing modded). All a backup plant means is you are splitting your power plant efforts between two sources, one of which is designed to come on if the other is somehow unable to provide sufficient power. If you don’t let the primary plant become unable to provide sufficient power, then all your efforts at upgrading and improving the backup are wasted time, effort, and material. All it takes for that is just the occasional monitoring of your power plant statistics by clicking on a power pole, and expanding the primary plant as needed. No need for a secondary power plant, just a larger primary power plant.

[Yoyobuae]

But it's actually solar you most want the early warnings with! A large-scale solar setup with a backup that trips only on accumulator level is unlikely to have that backup engage in time to accomplish anything. You want to plug your backup into the system early if you pick up warning signs that you're over-drawing your accumulators. This requires some circuit trickery; the simplest one is to have it trip if the accumulators ever discharge above ~30% of their maximum rate, say. More complexly, you can do some forecasting of where the accumulators will end up if the backups are not engaged and turn them on if that's below your comfort level.

The easiest way to simulate the charge/discharge rate of an accumulator across the day is to use create a mini-electric network with a a solar panel, an accumulator and 42kW worth of power consumption. The charge level of the accumulator will then always be the minimum required to keep a solar/accumulator power network running thru the day:

Screenshot_2021-02-03_00-40-04.png (303.88 KiB) Viewed 5342 times

Screenshot_2021-02-03_00-39-59.png (64.56 KiB) Viewed 5342 times

You can compare the signal from that accumulator with the signal from an accumulator in the actual electric network. If the accumulator in the actual electric network is at a lower charge level then it's necessary to inject power into the accumulators to survive the day. This kind of check works not only at night when accumulators reach their minimum level but at any point during the day.

[foamy]

But it's actually solar you most want the early warnings with! A large-scale solar setup with a backup that trips only on accumulator level is unlikely to have that backup engage in time to accomplish anything. You want to plug your backup into the system early if you pick up warning signs that you're over-drawing your accumulators. This requires some circuit trickery; the simplest one is to have it trip if the accumulators ever discharge above ~30% of their maximum rate, say. More complexly, you can do some forecasting of where the accumulators will end up if the backups are not engaged and turn them on if that's below your comfort level.

The easiest way to simulate the charge/discharge rate of an accumulator across the day is to use create a mini-electric network with a a solar panel, an accumulator and 42kW worth of power consumption. The charge level of the accumulator will then always be the minimum required to keep a solar/accumulator power network running thru the day:

Screenshot_2021-02-03_00-40-04.png

Screenshot_2021-02-03_00-39-59.png

You can compare the signal from that accumulator with the signal from an accumulator in the actual electric network. If the accumulator in the actual electric network is at a lower charge level then it's necessary to inject power into the accumulators to survive the day. This kind of check works not only at night when accumulators reach their minimum level but at any point during the day.

You'd need to match the ratio of your minigrid to the mains for precise control with that, wouldn't you? It's a really slick idea, though.

[Yoyobuae]

You'd need to match the ratio of your minigrid to the mains for precise control with that, wouldn't you? It's a really slick idea, though.

To have some safety margin it's easy enough to adjust the circuit signal afterwards, so no need to modify the mini grid.

For non-perfect ratio solar/accumulator setups the mini grid would need to be adjusted to the same ratio. Same with modded solar panels/accumulators which don't have the same ratio as vanilla.