Increase Accumulator Maximum Output Rate

[AntiElitz]

Accumulators currently have very limited use outside of Fulgora. There are situations where they would be ideal, such as covering the large power spike during a rocket launch on Gleba, but their maximum output rate is too low.

To handle that spike, it can take over 30 accumulators, even though the total stored capacity would already be enough. Because of this, it is usually cheaper and more compact to expand power generation instead. A single steam engine replaces about three accumulators, and one turbine replaces around twenty in output rate.

Raising the discharge rate would make accumulators useful for handling short-term power spikes without affecting their normal role in day–night power storage, since total capacity would remain unchanged.

There is also the issue that capacity and max input/output rate scale differently with quality, leading to inconsistent behavior between tiers.

[Tertius]

They scale with quality. A legendary accumulator has 750 kW power rate.

[AntiElitz]

That does not change anything. First of all that is not enough, secondly the moment you have legendary accumulators you are using fusion everywhere already and wouldn't care. This usecase for spikes is usually about the stage where you start building up your planets.

[crimsonarmy]

To better illustrate the problem assuming you are using some sort of 500 degree steam power (heating tower, nuclear, etc.):
one tank and a steam turbine (normal quality) can hold ~2.5GJ (~80 legendary accumulators) and has a max output of ~5.8MW (~8 legendary accumulators)

[Hurkyl]

To play devils advocate for the specific example in the OP, you could use efficiency modules (maybe with beacons) in your rocket silo....

But I generally agree, accumulators seem extremely underwhelming, only seeming useful for buffering solar power specifically, and being a hook for circuit logic.

[AntiElitz]

To put some numbers to it, I’d suggest increasing the accumulator max input / max output
from: 300 / 390 / 480 / 570 / 750 kW
to: 500 / 1000 / 1500 / 2000 / 3000 kW
for common / uncommon / rare / epic / legendary.

Right now, capacity and input/output rate scale differently with quality. This change keeps total capacity unchanged but makes each accumulator able to fully charge or discharge in 10 seconds, regardless of quality.

[Hurkyl]

For something specific that high discharge rate would get in the way of....

On the Steam forums, someone developed an uninterruptable power supply: they added logic to a solar farm that ensured it would never deliver more power than it was rated for, so that during the night your factory can only brownout and will never blackout.

This relied on the slower discharge rate of the accumulator; IIRC during high power usage, the network would break apart into seven segments and rotate between them to throttle the rate of power delivery.

With a large increase to the discharge rate, it would inflate the number of needed segments to a (likely) unmanageable level.

[AntiElitz]

For something specific that high discharge rate would get in the way of....

On the Steam forums, someone developed an uninterruptable power supply: they added logic to a solar farm that ensured it would never deliver more power than it was rated for, so that during the night your factory can only brownout and will never blackout.

This relied on the slower discharge rate of the accumulator; IIRC during high power usage, the network would break apart into seven segments and rotate between them to throttle the rate of power delivery.

With a large increase to the discharge rate, it would inflate the number of needed segments to a (likely) unmanageable level.

https://xkcd.com/1172/

[aka13]

+1. The accumulators for a very long time feel kinda useless, unless you are spamming them with solar, and caching steam is cheapper/better/faster/stronger.
I don't play too seriously in the last year, so I don't have an opinion on the number, but even twice the current value wouldn't be too much, imo.

[Hurkyl]

For something specific that high discharge rate would get in the way of....

On the Steam forums, someone developed an uninterruptable power supply: they added logic to a solar farm that ensured it would never deliver more power than it was rated for, so that during the night your factory can only brownout and will never blackout.

This relied on the slower discharge rate of the accumulator; IIRC during high power usage, the network would break apart into seven segments and rotate between them to throttle the rate of power delivery.

With a large increase to the discharge rate, it would inflate the number of needed segments to a (likely) unmanageable level.

https://xkcd.com/1172/

Is there not value in collecting all of the ways the current behavior has advantages over the proposed change, so that the idea can be accepted (or not) accurately on its merits? Or is it just supposed to be accepted blindly?

[crimsonarmy]

For something specific that high discharge rate would get in the way of....

On the Steam forums, someone developed an uninterruptable power supply: they added logic to a solar farm that ensured it would never deliver more power than it was rated for, so that during the night your factory can only brownout and will never blackout.

This relied on the slower discharge rate of the accumulator; IIRC during high power usage, the network would break apart into seven segments and rotate between them to throttle the rate of power delivery.

With a large increase to the discharge rate, it would inflate the number of needed segments to a (likely) unmanageable level.

https://xkcd.com/1172/

Is there not value in collecting all of the ways the current behavior has advantages over the proposed change, so that the idea can be accepted (or not) accurately on its merits? Or is it just supposed to be accepted blindly?

I think the point is that it is such a minor use case.

[Hurkyl]

For something specific that high discharge rate would get in the way of....

On the Steam forums, someone developed an uninterruptable power supply: they added logic to a solar farm that ensured it would never deliver more power than it was rated for, so that during the night your factory can only brownout and will never blackout.

This relied on the slower discharge rate of the accumulator; IIRC during high power usage, the network would break apart into seven segments and rotate between them to throttle the rate of power delivery.

With a large increase to the discharge rate, it would inflate the number of needed segments to a (likely) unmanageable level.

https://xkcd.com/1172/

Is there not value in collecting all of the ways the current behavior has advantages over the proposed change, so that the idea can be accepted (or not) accurately on its merits? Or is it just supposed to be accepted blindly?

I think the point is that it is such a minor use case.

Never said otherwise and wouldn't disagree if someone did argue that limiting power discharge is a use niche application better discarded for other niche use cases.

But mockery of the very idea of bringing up examples....

[eugenekay]

From the entity definition:

Code: Select all

energy_source =
{
  type = "electric",
  buffer_capacity = "5MJ",
  usage_priority = "tertiary",
  input_flow_limit = "300kW",
  output_flow_limit = "300kW"
},

5 Mega-Joules / 300 kilowatts * 60 Ticks per Second = 1000 Ticks at full drain

I think that the Accumulator's current math is absolutely flawless, and strongly oppose any change on the grounds that it will upset the Powers of Ten.

[Hurkyl]

Spite is a powerful motivator, so I thought to actually look more carefully at the rocket silo example.

A common quality accumulator stores 16.7 seconds of charge.

In my tests, I prepped by loaded a rocket silo up with two full rockets and charging the accumulators. Then I disconnected the power network, turned off the rocket part ingredients feed (and cleared the inventory), and loaded the silo up with items.

I tested a few configurations:

• No modules: consumed 75 MJ at 4 MW peak
• Prod 1 modules: consumed 193 MJ at 10.37 MW peak
• Prod 3 modules with one speed 3 beacon (on a separate power network): 463 MJ at 25.14 MW peak

In all cases, the energy consumed is around 18.7 seconds of peak power. So if you have enough accumulators to satisfy the energy demand, you already have sufficient discharge rate.

So the problem observed in the OP, insofar as the rocket silo is concerned, seems limited specifically to quality accumulators, where the storage capacity grows much more quickly than the (dis)charging rate.

---

For completeness, I'll add that if you never have to cover both spikes -- i.e. both the rocket reveal and the rocket launch are spread out in time to allow accumulators to recharge in-between -- then having just enough accumulators to cover the energy storage will indeed leave you lacking in discharge rate.

[Hurkyl]

Addendum: I was doing the "launch once when you have two rockets ready" because it was an easy way to eliminate rocket part production from the test.

But other use cases have different energy profiles. (the following for no modules involved)

If you start without a rocket ready, the "reveal a new rocket, launch, then close the silo" operation takes 92 MJ of energy, or 23 seconds of peak power. (the actual procedure takes 34 seconds, but isn't drawing peak power the whole time)

If you start with two rockets ready, do two launches, and let the silo close, it takes 120 MJ of energy, or 30 seconds of peak power. (spread out over around 56 seconds)

[AntiElitz]

From the entity definition:

Code: Select all

energy_source =
{
  type = "electric",
  buffer_capacity = "5MJ",
  usage_priority = "tertiary",
  input_flow_limit = "300kW",
  output_flow_limit = "300kW"
},

5 Mega-Joules / 300 kilowatts * 60 Ticks per Second = 1000 Ticks at full drain

I think that the Accumulator's current math is absolutely flawless, and strongly oppose any change on the grounds that it will upset the Powers of Ten.

That is not true for quality accumulators anymore as capacity scales different than the charging rates currently.

[AntiElitz]

Silos are just one example. It basically is equally true for power spiked on space platforms and other peaking environments

Spite is a powerful motivator, so I thought to actually look more carefully at the rocket silo example.

A common quality accumulator stores 16.7 seconds of charge.

In my tests, I prepped by loaded a rocket silo up with two full rockets and charging the accumulators. Then I disconnected the power network, turned off the rocket part ingredients feed (and cleared the inventory), and loaded the silo up with items.

I tested a few configurations:

• No modules: consumed 75 MJ at 4 MW peak
• Prod 1 modules: consumed 193 MJ at 10.37 MW peak
• Prod 3 modules with one speed 3 beacon (on a separate power network): 463 MJ at 25.14 MW peak

In all cases, the energy consumed is around 18.7 seconds of peak power. So if you have enough accumulators to satisfy the energy demand, you already have sufficient discharge rate.

So the problem observed in the OP, insofar as the rocket silo is concerned, seems limited specifically to quality accumulators, where the storage capacity grows much more quickly than the (dis)charging rate.

---

For completeness, I'll add that if you never have to cover both spikes -- i.e. both the rocket reveal and the rocket launch are spread out in time to allow accumulators to recharge in-between -- then having just enough accumulators to cover the energy storage will indeed leave you lacking in discharge rate.

[mmmPI]

If you increase the maximum output rate, it's still possible for players to gate their accumulators behind power switch if they want to mitigate the drain on them to artificially recreate a lower discharge rate, the reverse isn't possible, to me that's in favor of the balance change proposed.

Also currently on Vulcanus the higer the tier of accumulator, the more you waste, because the night is so short they don't have time to discharge, such change would make using quality accumulator more efficient and not less, that's another "buff" outside the use for silo spike.

[Hurkyl]

Huh, I just learned the power throttling application is even on the Factorio wiki.

I feel like there are two separate balance concerns here:

• There's the issue that accumulators seem balanced just for buffering solar and are woefully underpowered in comparison to buffering steam and overprovisioning on turbines/engines
• For quality specifically, the fact capacity grows much faster than (dis)charge rate throws its statistics out of whack

The excess capacity, I think, makes them better for buffering solar/lightning, but it seems only useful in that purpose?