Will AC power distribution become obsolete?

[Marco]

Sure. But what non-solar renewable can’t be generated synchronously?

Tidal too, any time you can't let the input power accumulate for a bit (water level, steam pressure, temperature, coal pile) and it has to be used immediately.

Though with just wind and solar together making AC awkward it's already a pretty bad look, so much for simplicity and efficiency.

[Someone]

In theory people should be paying *now* for the inertia that the older mechanical systems connected to the AC power grid provide because that stability is required, just like they should be paying a premium for dispatchable power, which does not include solar or wind, because it is available at any time.  This premium comes from solar and wind *not* being able to provide baseload or dispatchable power, unless backed up with expensive battery and inverter systems.

In practice they are not required to, so solar and wind sources are receiving a subsidy by making dispatchable and baseload systems less economical to operate, even though they are required for grid stability.

Spinning generators are not required. Notably the Australian market does price stability and pay to operators who can provide it (be that flywheels or batteries) while at the same time prohibiting wind/solar from enabling that capability. Agree it should be part of the pricing/costing but that needs to be in a fair market where any technological solution that delivers the functionality/resource can participate equally.

That is right.  I argued earlier that inverters can be used to provide grid stability, but they must be designed that way.  A large battery bank with an inverter which can both source and sink power can provide grid stability up to its power limit, which is exactly what Australia has done.

Now contrast that with other areas that are installing solar and wind without installing additional grid stability and then must limit the contributions of solar and wind.  The cost of the inverters and batteries is considerable.

You've glossed right over the important bit....
solar and wind (where they use DC linked inverters) can provide inertia/frequency control using the inverter with some appropriate software.

As a specific example in Australia, there are wind farms that have this feature already in the controls but it is prohibited from being used at an administrative (albeit technically informed) level.

Wind and solar do not necessarily need additional inertia sources to be installed, they can provide it themselves if required. Without additional hardware.

But not infinitely as the clocks in the small power sources will drift! At some point the small power sources will need to be re-synchronised to the large, synchronised power sources (= power plants). If the large power plants are no longer providing a low enough impedance to do this, there will have to be alternative ways to synchronise the grid. This is a basic clock synchronisation problem. And no, GPS is not suitable for this purpose as it is way too vulnerable / unreliable.

The long term drift is not an issue, inertia is only short term (flywheels have no inherent timing sync either).

[David Hess]

You've glossed right over the important bit....
solar and wind (where they use DC linked inverters) can provide inertia/frequency control using the inverter with some appropriate software.

As a specific example in Australia, there are wind farms that have this feature already in the controls but it is prohibited from being used at an administrative (albeit technically informed) level.

Wind and solar do not necessarily need additional inertia sources to be installed, they can provide it themselves if required. Without additional hardware.

The inertial sources used to generate power are 4-quadrant, so can source or sink power.  A solar inverter cannot sink anything, and inverter based wind generation cannot sink power either.  These must rely on other sinks like resistive loads or inertial sources.

Inverters should be able to produce VARs enabling power factor correction, but the circulating currents subtract from the inverter's power capability.

[Marco]

The inertial sources used to generate power are 4-quadrant, so can source or sink power.  A solar inverter cannot sink anything

Near zero crossing it can't do anything, but a GW plant stopping any output is going to drop the voltage across most of the phase on any line near it, without sinking anything.

For the large PV farms you could add a small stage to feed back power to the DC capacitors, it only needs to sink a bit of power near the zero crossing.

[IanB]

For the large PV farms you could add a small stage to feed back power to the DC capacitors, it only needs to sink a bit of power near the zero crossing.

The behavior of a generator connected to a large grid is that the power flow depends on the phase angle between the generator and the grid. Ordinarily the generator is locked in phase to the grid. If the grid starts to speed up and the generator does not want to follow, then a phase angle will start to open up between the generator and the grid. This will result in the generator having to sink ever increasing amounts of power to maintain its own frequency (this is what inertia means). Either this sunk power will need to be dissipated through large resistor banks, or it will need to be stored in large battery banks. If neither resistor banks nor battery banks are available, the generator will have to follow the grid, and there is nothing software can do about this. The converse is if the grid starts to slow down, then the generator will have to deliver ever increasing amounts of power to maintain its own frequency, and if the required power delivery exceeds the generator's capacity, then again it will have to slow down and follow the grid frequency.

The only way software can influence this is by simultaneously commanding all generators delivering power to the grid to cut back on their power output in order to balance the load with the consumers. This can work over longer time scales, but it doesn't help so much with short time instability, which is where mechanical inertia helps.

Also, if increased power delivery is required, then software can only command this from generators that have additional capacity to deliver (meaning dispatchable or baseload generators). You can't command a solar farm to deliver more power than it is currently receiving from the sun.

[Marco]

The behavior of a generator connected to a large grid is that the power flow depends on the phase angle between the generator and the grid. Ordinarily the generator is locked in phase to the grid.

It only has to sink for a few degrees near the zero crossing, the load is still overwhelmingly resistive. The capacitors can easily take that, the inverter just needs to be bidirectional (though the sink part can be a fraction of the power). Or it could just let the sine distort a bit and try to force it to the desired phase/shape away from the zero crossing (where instantaneous power output is always positive except in failure modes).

If the grid starts to speed up and the generator does not want to follow, then a phase angle will start to open up between the generator and the grid. This will result in the generator having to sink ever increasing amounts of power to maintain its own frequency

It doesn't have to replicate the way a generator could blow itself up in a failure mode. The tug of war between generators is temporary and net zero power.

[David Hess]

IanB covered what I would have said so there is no need for me to repeat.

It only has to sink for a few degrees near the zero crossing, the load is still overwhelmingly resistive. The capacitors can easily take that, the inverter just needs to be bidirectional (though the sink part can be a fraction of the power). Or it could just let the sine distort a bit and try to force it to the desired phase/shape away from the zero crossing (where instantaneous power output is always positive except in failure modes).

The input capacitors are only large enough to decouple the inverter from the source.  A battery storage site like the one Tesla built in Australia can do it at scale.  Solar farms cannot, and wind farms probably cannot either.

Inverters should be able to produce VAR power because the circulating current alternates between cycles, but the inertia of an alternator can last for orders of magnitude longer.

If the grid starts to speed up and the generator does not want to follow, then a phase angle will start to open up between the generator and the grid. This will result in the generator having to sink ever increasing amounts of power to maintain its own frequency

It doesn't have to replicate the way a generator could blow itself up in a failure mode. The tug of war between generators is temporary and net zero power.

The generator pushes against the inertia of its prime mover, like a water turbine.  The energy is not dumped into the generator, really polyphase alternator, except for losses.

[Someone]

You've glossed right over the important bit....
solar and wind (where they use DC linked inverters) can provide inertia/frequency control using the inverter with some appropriate software.

As a specific example in Australia, there are wind farms that have this feature already in the controls but it is prohibited from being used at an administrative (albeit technically informed) level.

Wind and solar do not necessarily need additional inertia sources to be installed, they can provide it themselves if required. Without additional hardware.

The inertial sources used to generate power are 4-quadrant, so can source or sink power.  A solar inverter cannot sink anything, and inverter based wind generation cannot sink power either.  These must rely on other sinks like resistive loads or inertial sources.

Inverters should be able to produce VARs enabling power factor correction, but the circulating currents subtract from the inverter's power capability.

The inverters used in many (most?) DC-grid applications are 4 quadrant, and have significant DC bus capacitance (which could be increased if there was incentive to add more energy storage). How many times does this need repeating? That is completely adequate/capable/practical for producing inertia.

[nctnico]

You've glossed right over the important bit....
solar and wind (where they use DC linked inverters) can provide inertia/frequency control using the inverter with some appropriate software.

As a specific example in Australia, there are wind farms that have this feature already in the controls but it is prohibited from being used at an administrative (albeit technically informed) level.

Wind and solar do not necessarily need additional inertia sources to be installed, they can provide it themselves if required. Without additional hardware.

The inertial sources used to generate power are 4-quadrant, so can source or sink power.  A solar inverter cannot sink anything, and inverter based wind generation cannot sink power either.  These must rely on other sinks like resistive loads or inertial sources.

I'm not sure that is entirely true. In the end you want an inverter to supply to the grid while being in sync with the grid. So what is needed is to drive the clock of the inverter accurate enough so that the grid stays at the right phase + frequency. A high quality clock like a ovenized oscillator should be able to have a long enough hold-over to guarantee that for a couple of hours. I don't see the need for an inverter to be a sink as well. In my mind the inverter checks if the impedance of the grid is low enough and if not, it goes into holdover mode to ensure all the inverters together are not creating a feedback loop which starts to oscillate.

[David Hess]

The inverters used in many (most?) DC-grid applications are 4 quadrant, and have significant DC bus capacitance (which could be increased if there was incentive to add more energy storage). How many times does this need repeating? That is completely adequate/capable/practical for producing inertia.

The inverter's output structure supports 4-quadrant operation, but there is not enough input capacitance to matter beyond circulating current over one cycle.  The input capacitance is only high enough to meet ripple current and decoupling requirements for the DC source.

[Marco]

The input capacitors are only large enough to decouple the inverter from the source.

Which is a lot more energy than needed to cover a couple degrees of sinking near the zero crossing. When the instantaneous input power is constant and the instantaneous output power is sin(t)^2 at 50/60 Hz, you have to store power for more than a couple degrees.

Inverters should be able to produce VAR power

VAR power is net zero across 360 degrees.

The generator pushes against the inertia of its prime mover, like a water turbine.  The energy is not dumped into the generator, really polyphase alternator, except for losses.

If it has to "sink ever increasing amounts of power" it will become a motor instead of generator and generally something will break. Even pumped hydro won't just switch from generation to pumping that easily.

By the time you are more than 90 degrees out of sync, it's time to call it a day.

Anyway here's one from Google :
https://www.sciencedirect.com/science/article/abs/pii/S1364032113001925

[Someone]

The inverters used in many (most?) DC-grid applications are 4 quadrant, and have significant DC bus capacitance (which could be increased if there was incentive to add more energy storage). How many times does this need repeating? That is completely adequate/capable/practical for producing inertia.

The inverter's output structure supports 4-quadrant operation, but there is not enough input capacitance to matter beyond circulating current over one cycle.  The input capacitance is only high enough to meet ripple current and decoupling requirements for the DC source.

The amount of energy that can be absorbed by the capacitors is a quantifiable, the amount of virtual inertia is quantifiable. That could persist over arbitrary time scale depending on the power demand, just as any energy storage such as spinning mass. It's not a yes/no zero/infinity thing.

Although a single inverter (with the capacitance as currently deployed) can only provide a relatively small amount of virtual inertia energy compared to a spinning mass, that does not make it a failure.

So we have at least whatever the installed capacity of inverters is (given lots of those are on variable/unreliable energy sources, there is already a multiplier above the load average).

The additional cost is practically nil as its almost entirely software that is in the way at the moment. You could increase the peak/surge capacity of the inverters and add some more bus capacitance if even more inertia is required above the current availability.

this is going off into the sort of silliness/disingenuous/sealioning that would produce the question:
"but a residential solar install is only 5kW, how could that deliver the xxGW the grid currently demands?"

H-bridge inverters are 4 quadrant devices and as implemented in grid tied DC inverters already have some storage and virtual inertia capacity.

It's not zero and it's not infinity. Currently it's almost completely "disabled" at a software level so is sitting under-utilised with what is already there. More energy storage can be added if there is economic incentive. There is this ongoing/general problem that grid operators enjoy the profit from various markets that distributed generators are unable/prohibited from accessing, not because it is impractical or dangerous, but historically because distributed control systems and micropayments/accounting were uneconomic. Those are no longer justifiable barriers.

Seems like a whole pile of anti renewable mouthpieces in here.

[johansen]

https://www.eia.gov/todayinenergy/detail.php?id=3990

In 1 second the frequency rises .1hz and already generation is being taken offline, frequency continues to rise for 2 more seconds until generation matches load and the frequency begins to fall.

Its relatively easy to program an inverter to not be able to change its frequency that quickly, and reduce the power output in accordance with how quickly the frequency climbs, but slowly over a 10? second time frame, return to full power output regardless of frequency (within limits)

I suspect that such a simple algorithm would provide a lot of virtual inertia. Not enough to hold the grid up, but make it easier on the grid operators to give them more time to shut down generation when the frequency rises because the load is falling. This comes at very small decreased revenue by the solar farms. yes someone has to pay.

Using the solar farms to provide reactive vars.. only helps if the utility operators command it to. when the power demand changes, the power factor can go either way, leading or lagging.. you don't know this ahead of time.

IanB's idea that generators sink power is not correct. what happens is the prime generators are driven by machines that produce torque, the generator will always spin faster if you decrease the load, and they will always follow the grid frequency and slow down when the power demand increases. The grid frequency rises when power in is greater than power out. What we need is more inertia in the system, and its a massive amount. Tf you read the prior link I posted, the UK wants 140 GJ of inertia. sounds like a lot, but when you are trying to balance out a system that needs to stay within 60.1 to 59.9hz, the energy miss match between what is pushed into and pulled out of the system between those frequencies is only 1/600th of 140 GJ. for a .1 second per second frequency increase or decrease. that's only 233MW miss match will result in a .1 hz frequency change per second.

I strongly suspect that the 40 ton flywheel mentioned by the prior link i posted, is storing 900MJ in the total inertia of the rotor at 1800 rpm for reasonable, 4 foot diameter, 40 tons worth of steel. it sounds like a lot but it really isn't, which explains why they need something like 100 of those added to the grid over the next decade.

Also I don't understand why Marco is talking about the zero crossing. It is true that most small solar farms are dumb inverters that follow the grid and wait .05 milliseconds each half cycle so they have anti-islanding capacity, but it doesn't matter much. The stored energy in the inductance of the power grid and the capacitors needed to offset that inductance, is pretty significant and it balances out most of the harmonics.

Consider the fact that 1GW sized transformers are inductive to the tune of 20%, and all of that inductance has to be canceled out with capacitors. It makes for a pretty good sized LC filter.

[IanB]

IanB's idea that generators sink power is not correct.

It's not that they actually do sink power. It's just that they would sink power if you somehow tried to force them to stay at the same frequency (lagging the grid), if the grid is advancing ahead of them. When means they cannot do that, they have to follow the grid frequency. Earlier in the thread there was the implication that using software, you could force a grid-tied inverter to stay at 60 Hz even if the grid is running at 60.1 Hz. This defies the laws of physics.

What can happen is what you said, namely to reduce the total power input into the grid if the total grid load decreases, thereby causing the grid to reduce its frequency. The ideal way to do this is to have all generation sources controlled centrally, and to coordinate how much power is being fed into the grid to match demand. This is easy(er) when the grid is powered by relative few large large generation sources. It is much harder when the grid is powered by multitudes of micro-sources.

[Someone]

Earlier in the thread there was the implication that using software, you could force a grid-tied inverter to stay at 60 Hz even if the grid is running at 60.1 Hz.

Where in the thread?

What can happen is what you said, namely to reduce the total power input into the grid if the total grid load decreases, thereby causing the grid to reduce its frequency. The ideal way to do this is to have all generation sources controlled centrally, and to coordinate how much power is being fed into the grid to match demand. This is easy(er) when the grid is powered by relative few large large generation sources. It is much harder when the grid is powered by multitudes of micro-sources.

Frequency is a very strong signal across the grid as to aggregate balance and is the current primary method for controlling power on short term basis, there is no need for (and a current lack of) centralised control on timeframes of seconds or even minutes.

Which an inverter with a crystal oscillator (and no central control) can already participate in by following the frequency cues, as well as adding virtual inertia to slow the changes.

[IanB]

Where in the thread?

Here:

Which an inverter with a crystal oscillator (and no central control) can already participate in by following the frequency cues, as well as adding virtual inertia to slow the changes.

At this point, I do not know what to say. The grid frequency is "an immovable object". A grid-tied inverter can do nothing about it, crystal oscillator or not, unless it is big enough to be "an unstoppable force". This is simple physics.

[Someone]

Earlier in the thread there was the implication that using software, you could force a grid-tied inverter to stay at 60 Hz even if the grid is running at 60.1 Hz.

Where in the thread?

Here:

Which an inverter with a crystal oscillator (and no central control) can already participate in by following the frequency cues, as well as adding virtual inertia to slow the changes.

At this point, I do not know what to say. The grid frequency is "an immovable object". A grid-tied inverter can do nothing about it, crystal oscillator or not, unless it is big enough to be "an unstoppable force". This is simple physics.

You're using infinity. There is nothing which can provide an immovable object, and your hypothetical immovable object is the aggregate of many "less movable" things.

Why is it so hard to imagine that many smaller energy storage devices can add up to the same as a larger energy storage device?

Also note the quote you chose to use says... follow the frequency. So your straw man is just that, an imagined argument that no-one was pushing and you added.

Existing plant providing frequency control keeps some power above and below its current operating point in reserve to push the frequency back to ideal. Pretty much any generator can offer to do that, inertia being how fast that frequency resists change an entirely orthogonal matter, something inverters are also capable of but are currently prohibited from doing in Australia (and other jurisdictions).

[Someone]

Apparently an "engineering" forum has attracted a significant user base who cannot understand the nuances of power vs energy, speed vs torque, phase vs frequency, etc.

[tautech]

Apparently an "engineering" forum has attracted a significant user base who cannot understand the nuances of power vs energy, speed vs torque, phase vs frequency, etc.

While our grid frequencies are defacto standards we must also consider now they are set/fixed and managed/modified.

Imagine a Pelton wheel turbine which could spin out of control with or without a load so its speed needs be closely controlled simply by governing nozzle flow and linking its control to grid frequency or some local frequency standard which in effect becomes a speed governor.

Upscale this to multi turbine hydro or steam, gas or whatever which throughout the globe serves as our baseload generation and where these current sources are all on the same grid some frequency/speed control becomes necessary.
Until we have a HV DC link to another location, state, island, country, all the AC needs is be near perfectly synced which acts as the standard frequency for all grid tied inverters to sync to.

[johansen]

IanB's idea that generators sink power is not correct.

It's not that they actually do sink power. It's just that they would sink power if you somehow tried to force them to stay at the same frequency (lagging the grid), if the grid is advancing ahead of them. When means they cannot do that, they have to follow the grid frequency.

talking in circles to justify the prior incorrect statements is a waste of everyone's time.

[johansen]

Until we have a HV DC link to another location, state, island, country, all the AC needs is be near perfectly synced which acts as the standard frequency for all grid tied inverters to sync to.

all that does is make the problem more global. it doesn't fix the problem.

[tautech]

Until we have a HV DC link to another location, state, island, country, all the AC needs is be near perfectly synced which acts as the standard frequency for all grid tied inverters to sync to.

all that does is make the problem more global. it doesn't fix the problem.

HV DC links breaks grid tied frequency unless it's bidirectional.

[Someone]

While our grid frequencies are defacto standards we must also consider now they are set/fixed and managed/modified.

Imagine a Pelton wheel turbine which could spin out of control with or without a load so its speed needs be closely controlled simply by governing nozzle flow and linking its control to grid frequency or some local frequency standard which in effect becomes a speed governor.

Upscale this to multi turbine hydro or steam, gas or whatever which throughout the globe serves as our baseload generation and where these current sources are all on the same grid some frequency/speed control becomes necessary.
Until we have a HV DC link to another location, state, island, country, all the AC needs is be near perfectly synced which acts as the standard frequency for all grid tied inverters to sync to.

This is more of the extreme exaggerations like needing super accurate synchronised clocks during offline.

The grid frequency is not ppm level sensitive, 1% change is a major event, 0.1% is significant, 0.01% useful for day to day control (100ppm and well within crystal oscillator accuracy before even using something "controversial" like a PLL). On the other end generators do not explode with a few percent change, those excursions are disconnected from the grid and needed otherwise to protect themselves in the absence of the grid (in case of a disconnection event).

Inertia is mostly about keeping the already distributed control system of generators and loads stable by limiting the rate of change (bandwidth).

[IanB]

Apparently an "engineering" forum has attracted a significant user base who cannot understand the nuances of power vs energy, speed vs torque, phase vs frequency, etc.

And it seems to me that you are a member of that user base, so, sadly, we cannot make any progress   

[Someone]

Apparently an "engineering" forum has attracted a significant user base who cannot understand the nuances of power vs energy, speed vs torque, phase vs frequency, etc.

And it seems to me that you are a member of that user base, so, sadly, we cannot make any progress   

Where so? More baseless claims despite being able to quote any part of the conversation which might back this up.