By Robert Bradley Jr. — May 6, 2025
“The challenge with rising renewables: As power systems rely more on asynchronous generation, frequency changes can occur much faster, increasing the risk of grid instability.” (- Dave Edwards_
Dave Edwards post begins: “As an engineer I like to understand how systems operate, this is useful for fault diagnosis, especially when a system fails and a Root Cause is needed, though maybe more than one contributing factor. So let’s Talk “Inertia” 🙋♂️Everyone’s Saying It, But Who Really Understands It 🤷♂️”
The floor is his:
“Inertia” is getting a lot of airtime in power system discussions lately, but what is it, really, and why is it so critical for grid stability on an electrical power generation system.
More importantly, which generation technologies actually provide useful inertia⁉️
Image created using Chapt GPT 4o
❓What is inertia?
Think of inertia a bit like us seasoned engineers on LinkedIn: “It’s the ability of an object (engineer) to resist change.”
For rotating objects (like generator rotors), this is called rotational inertia or moment of inertia.
👉 Two key factors influence it:
• Mass
• Mass distribution (i.e., the object’s shape)
Example: A solid disk and a ring with the same mass and radius have different moments of inertia.
• Solid disk: Mass is closer to the center = lower inertia
• Ring: Mass is farther from the axis = higher inertia, result, the ring is harder to start or stop spinning.
👉 Why does this matter in power systems? The spinning parts of generators store kinetic energy, which helps stabilise grid frequency during sudden events, like a generator or interconnector tripping offline.
👉 What actually provides useful inertia? Synchronous generators (coal, gas, hydro, nuclear):
• Spin in sync with grid frequency (e.g UK: 50Hz, US: 60Hz)
• Have large rotating masses, provide physical inertia.
• Help slow down frequency changes, giving time for systems to react, either demand or generation.
Asynchronous generation (most wind turbines, all solar PV):
• Magnetically decoupled from grid.
• Has little or no rotating mass.
• Provide minimal to no physical inertia.
💡The challenge with rising renewables: As power systems rely more on asynchronous generation, frequency changes can occur much faster, increasing the risk of grid instability.
To counter this, we need –
• 🔋 Fast battery response
• ⚙️ Synthetic inertia
• 🧠 Grid-forming inverters
• ⚡️ Synchronous Compensators
or just plain old conventional generation like nuclear or gas.
👉 Bottom line:
As we increase renewable penetration, understanding and managing inertia becomes critical.
It’s not just a buzzword, it’s foundational to keeping the lights on 💡
Hopefully if you didn’t know, you now know what inertia is in the context of the grid and why important for grid stability.
Asynchronous generators produce DC. They DC needs to be converted to AC to power the grid. Why not connect the DC asynchronous generators to the grid through motor generator sets with big flywheels? They may be less efficient than inverters, but they would have inertia.
Because they will cause massive (pun intended) problems when that DC stops being generated for any reason. Like a cloud, or a breeze dropping. Starting them up again would take a lot of power, which probably wouldn’t be available for a long time.
Basically, inertia works both ways…
I guess to put it simplistically by my non-engineering thinking:
Hard to stop, good; hard to start, bad.
you nailed it!
Why not connect them to the grid through a battery providing synthetic inertia? Doing that solves more problems including benefiting from renewables that are generating more energy than required on the grid and provision of peak load energy.
As battery prices drop, this solution looks more and more favourable.
You didn’t comprehend the point of the post.
Why is that? I’m going to guess you dont know what synthetic inertia is.
All hail toolmans synthetic inertia god that only he understands.
It sort of sounds plausible, if you don’t know anything about physics, which he doesn’t.
I aspire to being able to understand radiation in terms of boards, holes and springs like real physicists such as yourself.
I tried to give you dumbed-down descriptions suitable to your level of education. (It wasn’t easy.) I could immediately tell that you had never studied physics formally because you rambled on about “net work” and “net power”, then insulted me while I was trying to teach you that those are false conceptions. Then you helpfully contradicted yourself by showing us that the definitions of “energy”, “work”, and “power” did not, in fact, include the word “net” anywhere. Who taught you your physics? Or your logic, for that matter?
I insulted you? Lol. I am patient but you have helped me find my limits so thanks for that.
I too am glad that we found the limits of your understanding of both physics and logic. Acknowledging those limits will help you look smarter. You’re welcome!
“I aspire”
Says all I need to know.
Sophisticated power electronics aka big complicated inverters theoretically quickly respond to frequency deviations. These inverters can be programmed to increase their output or reduce it based on the frequency of the grid, and in theory becoming an inertial supply.
So where the UK could/should be using Nuclear+Gas+Hydro+Pumped Storage+Other Odds and Sods to generate electricity reliably, Hydro and pumped to cope with sudden peaks. We now have Nuclear+Gas+Onshore Wind+Off Shore Wind*+Hydro+Pumped Storage+Solar PV*+ Domestic Solar*+Interconnectors*+Batteries*+Drax+Odds and Sods.
*=DC Sources. (The UK interconnectors are DC as opposed to the ones in Spain which were AC. I think that is because the connections are in excess of 30 miles under the sea and DC connections have lower losses)
All the DC connections have to be converted to AC using inverters. But some if not most of those inverters now have Synthetic Inertia Inverters to prevent a Spainish Blackout. This would be a difficult sell in “Dragons’ Den” (Shark Tank). The more complex a system the more likely it is to fail. An unneccessarily complex system is also more expensive to build and maintain.
If somebody knows the reason for DC underwater and can explain it simply then please do. I think it is to do with the properties of water, relative permittivity and relative permeability which are unique? But that’s something from 55+ years ago.
It’s not so much about the water it’s about the distance.
Your losses in a DC cable are I squared R your losses on an AC cable are I squared R + Inductive losses + capacitive loses.
So why do we have high voltage AC lines at all .. well one word “corrosion”. Try running your retic system on DC and see what happens at every terminal and join 🙂
There is a distance at which the corrosion prevention cost drops away compared to the power loss cost.
It’s probably worth a read of
https://www.inmr.com/10th-anniversary-of-worlds-highest-hvdc-line/
It explains the choices they face in selector insulator and cable design.
Yes, DC HV transmission lines have lower losses. The disadvantage is that the terminal stations are so costly that they are only being used for point to point long distance lines. AC HV lines can be tapped into at any point along the way for distribution at a reasonable cost. That is why they are more common and were the standard.
Leon,
The reason AC is used for high voltage is to get the voltage up high through transformers. DC can’t be easily transformed up or down to hundreds of thousands of volts and being sent as three phase power is another huge advantage. Why high voltage? because electrical energy is very diffuse and as a force reduces with the inverse square of the distance so to send power over long distances requires very high voltage.
Part of the problem is phase differences even if frequencies are maintained. Large inertial generators not only “lock” in frequency but phase also.
Think about a subdivision with 100 houses feeding the grid from solar and inverters. Which signal does your inverter follow? Your neighbors? The one two blocks away? Maybe it just oscillates trying to keep up as all 100 just move around with different phases.
Yes, too many variables to solve that equation, and the grid is more complex.
They all add their energy to the grid at exactly the same frequency and hence phase as the grid frequency that is supplied by say the gas turbine that is also on the grid. Roof top solar inverters don’t have their own frequency that they need to synchronize with. They are utterly bound by the existing grid frequency. The need that gas turbine to work.
Some have a mode whereby they can supply their own frequency so a house can operate off grid but it doesn’t apply when on grid.
“They all add their energy to the grid at exactly the same frequency “
“They are utterly bound by the existing grid frequency.”
And what happens when the existing grid frequency generator goes away? Like it did in Spain? All those inverters free run – they *will* have their own frequency. Which one does *your* inverter sync to?
They stop putting their energy into the grid because they turn off,
…almost immediately sense that there is no stable grid frequency, go out of spec and turn off their output.
This is absolutely by design. When a segment of the grid goes down, they’re designed to stop outputting to the grid.
And you end up with a 1.6 billion euro blackout in Spain.
AC creates an varying electric field even with shielding that affects water life.
That is one factor.
AC also requires periodic step up transformers, that HVDC does not. That adds complexity and very real maintenance challenges.
That is one factor.
The transformers are not cheap making an AC connection more expensive. That is partially offset by having to generate very high DC voltages.
That is one factor.
That is a short list.
As LdB stated, underwater cables have capacitive losses in the form of a dielectric loss and I_squared_R loss from the current needed to charge/discharge the capacitance. The other reason is to allow an asynchronous connection between two power grids.
The problem with “quickly respond to frequency deviations” is sensing where the frequency deviations are occurring. Transformers have a non-zero leakage inductance, so a sudden change in load will change the phase shift across the transformer and thus appear to be a deviation in frequency with respect to the other side of the transformer.
Up to a point, of course. Which probably explains why there aren’t more of them in use. Battery electric vehicles were supposed to become cheaper, but it seems that when government subsidies dry up, so do BEV buyers.
In my worthless opinion, both wind and solar power produce more energy during the day, where certainly more domestic power is required at night. A 10KWh battery will run a domestic oven, toaster, jug, and moderate size air conditioner for an hour or two before it is flat.
And of course, it can burst into a fierce fire unexpectedly. Of course, you will have this battery attached to your building, probably with a roof over it. The more batteries the more expense, the more space, the more complicated management software, the greater risk of fire . . .
That’s why I declined the kind offer of the government a few years ago, to provide solar panels at no cost (apart from ongoing maintenance). I’m happy for the grid operators to look after the technical aspect details, while I pay a base charge plus a unit cost. To each his own.
“As battery prices drop, this solution looks more and more favourable.”
Numbers please. Real people spend real money.
From here we have the trend
Lithium ion batteries that I can buy have not fallen very much since about 2012.
I question the data on that graph
Feel free to find data that contradicts it. There are many many sources that agree but I’ve not found any that suggest otherwise.
The other thing to note is that efficiency has improved over time too. So for example the 18650 cell has considerably improved its capacity since it was introduced.
Economies of scale are part of that. Demand increases, production increases and manufacturing efficiencies improve. ECON 101.
I doubt the trend continues. I expect it to flatten, then rise. Why? Geopolitical effects on supply chains. Depletion of critical resources, such as rare earth elements.
Also note that chart has prices adjusted for inflation without providing the inflation data. As such, that is not the price you pay.
As you say, Economics 101 and production has massively increased so why doubt the fact their price has come down?
There’s tons of scope for continued reduction in price.
Your chart ends in 2018. Surely if the trend has continued the cost must have gone to zero in the last 7 years.
Undoubtedly they have gone past zero, and manufacturers are now paying us to take their batteries.
Funny how those mythical price drops never made it to the consumer.
They did. When was the first time you bought a lithium battery and how much did you pay for it? My memory is that “rechargeable batteries”, NiCad the the time, were extremely expensive in the 90s but very affordable now.
Synthetic inertia isn’t the same as real inertia. Synthetic inertia is limited by battery charge and discharge rates, electronic component current limitations, and most importantly the software and response time used to implement it. Some sources indicate that generator frequency must be held to within a 0.3 Hz. limit and that a half Hz. (8msec) error will force a dropout. Synthetic inertia must react much more quickly than that. Real inertia is instant, and not limited by power ratings of the components, particularly the batteries, and most of the solar sources have no batteries at all. Real spinning inertia might even be cheaper and certainly has less risk of fires than synthetic inertia.
Its effectiveness is limited by the mass. Response is limited by the rate of change of the prime mover and that is inevitably much, much slower than battery response. Rotating mass has some benefits but so does synthetic inertia.
You missed the point. Mass inertia has zero response time. Synthetic inertia has a non-zero response time. Mass inertia can provide time for fusing elements to work. Synthetic inertia cannot match this. Thus damage from things like high current faults can cause more damage with synthetic inertia than with mass inertia.
The response time of the driver has to do with response over longer periods of time. And in many cases it isn’t the response time of the driver that is in control, it is once again the inertia of the mass. If it takes a long time to slow it also takes a long time speed up because of the inertia. Batteries have the problem that if the charging system goes away how do they recharge. A coal or gas driver can operate independently. If you provide coal or gas drivers to recharge the batteries then why have the batteries to begin with?
Batteries also have voltage delay turn on times.
Depending on age and load, these can be consequential.
Having tested older batteries, some of those took minutes.
You are trying to substitute software for real multiple 1500 tonne turbo generator sets rotating at 1500 or 1800 rpm. That constitutes real inertia.
How many bits of software to the tonne?
Faster response can have its own unintended consequences.
Response is instantaneous, as soon as the terminal phase starts moving away from the rotor phase, power will start to flow. The damper windings will provide additional response to a change in relative frequency. The nature of a synchronous condenser is that it provides better inherent support for voltage regulation than static capacitors.
The time taken to measure and respond from a battery is tiny.
And for normal operation this is just fine but when the load change is larger, then the energy in the mass can’t maintain frequency for long and battery response is effectively immediate compared to any other generation such as gas turbines which are much slower.
So thy have their pros and cons.
How tiny is the response time? A large short circuit fault can have catastrophic results in milliseconds. If the response doesn’t start *before* the short circuit is removed the damage will compound. The response from a spinning mass starts instantaneously. So tell us just how fast “tiny” is for a battery?
“ when the load change is larger,”
It isn’t the gas turbine that is slow. It is the spinning up of the physical mass that is slow. For the same reason that it is hard to slow the spinning mass down – inertia.
Nor is the issue the load change from the work day to evening. The issue in Spain wasn’t people coming home from work and turning on air conditioners.The biggest issue is when there is a fault on the grid that generates huge current demands. Spinning mass responds instantaneously. Non-spinning mass doesn’t.
I don’t think you realize how truly ugly the voltage and current waveforms can get in a fault condition, which can complicate the response needed. The windings on a large synchronous machine will take a lot of abuse for a short time.
This is why its difficult to account for all the fault conditions and why practical experience is going to be the best teacher. Nobody said synthetic inertia was easy.
TTTM,
Rubbish, the moment the frequency reduces in a grid fed by large AC thermal power stations the governors on the turbines sense the drop and provide more power to keep the voltage stable. In a 50HZ system those 800 tonne rotors are spinning at 3000 r.p.m. and in a 60 HZ system at 3600 r.p.m. Considerable numbers of those generators in a power station have enormous overall mass, ten generators would have over 8000 tons of spinning mass. Batteries cannot provide under when the system is under significant load. Try pushing a battery powered lawn mower into longer than normal grass and it stalls. You also miss the point about solar power in residential areas as a street or road has to balance to power according to which phase each home comes off. You can’t provide solar power back to say just A phase and leave B phase and C phase out of phase.
Once phases get well out of sync in big systems, generator drives can get sheared. Another point about solar and wind apart from destruction by severe weather is thermal stress. No form of machinery or metal handles continual stop/start, cold hot conditions.
But that’s too slow for stability. For example AEMO defines response to maintain stabilisation as “slow” for good reason. AEMO FCAS definitions from here where:
Opening the governor doesn’t instantly create more steam for a steam turbine and that is a slow process. Coal heating steam is typically “slow” (10s of seconds), gas turbines might be “fast” (ie 6 seconds or less) and for Hydro generation its typically “very fast” (ie 1s)
Sure, if you have enough of them compared to the size of the load change then they’ll cope just fine.
This is just a bad analogy. A Tesla model S plaid will drag off almost every muscle car because its batteries can output an enormous amount of power. Batteries can absolutely do the job and are doing so in many grid battery installations already.
That’s not the way it works. solar inverters dont change the frequency or phase when they add their energy.
Arguing that you cant add energy on one phase doesn’t work either because the opposite is obviously not true. Homes typically are allocated one of the three phases to draw from and that’s unbalanced except in aggregate where it averages out.
All true but it certainly can be good enough, which is why people are adding batteries to the grid.
50-60Hz is an eternity for fast switching transistors.
Synthetic inertia may not be the same, but it is good enough.
Its also very expensive. Like everything to do with renewable energy.
good enough is not enough. High current faults cause big damage. A few milliseconds extra because of response time from synthetic inertia can cause significant additional damage compared to mass inertia.
That’s why equipment is protected by circuit breakers. Large enough faults are going to be a problem whatever generation method is used.
How fast do you suppose those circuit breakers operate? How much infrastructure can be destroyed in that amount of time?
You ever have lightning get into your house over the service line? I have. The circuit breaker at the service pole didn’t operate nearly quickly enough to protect my TV’s, ham radio, and several appliances. Neither did the circuit breakers in my fuse box. We had to call the fire dept to use their heat detection equipment to insure there was no fire inside the walls. Took out stuff in the neighbor’s houses as well. Yet the substation never reacted to isolate the neighborhood feed. This had to have generated a huge current spike down the line but the spike didn’t last long enough for the protective equipment to react.
However if the lightning had caused a fault (as opposed to frying electronics) then that would have potentially caused the house fuse/circuit breaker to trip or substation to trip if the problem remained.
Are you trying to liken lightning to a fault that conventional generators can handle but batteries providing synthetic inertia cannot?
They is big expensive tools and you have to know where best to put them when you switch from large hub and spoke synchronous generation to asynchronous spaghetti and meat balls if you get the drift-
SA synchronous condensers installed – Energy Magazine
There are lots of ways to skin a cat. Those synchronous condensers were installed years ago when batteries were less economically viable and grid forming was less well understood.
Here is a question for you. Let’s say there is a grid forming inverter with 100 MW of power available, putting the proper voltage. What makes the hundreds or thousands of async inverters down the line lock onto that one isolated transmission, especially if they are on the distant side of say two or three transformers?
A much better solution is for a separate timing network that every inverter supplying power to the grid must use to insure proper synchronization all throughout the grid. It would simplify inverter design at the same time.
I have the answer to that, but I doubt that you would understand it.
What makes the asynchs lock on is because that is exactly what they are designed to do.
What makes using a master clock impossible is that falling frequency is the way the generators signals to the load it cant deliver the power. Along with falling voltage. Remove that and the grid gets totally unstable. Renewables would try and drive, overcurrent and switch off.
Think in terms not of frequency or inertia but in terms of energy storage. If you have an energy supply whose power is strictly limited and you try and draw more power, it collapses.
If it is less strictly limited, the voltage sags.
They are designed to “lock” on to a frequency on the grid. That frequency is the input to a phase lock loop that insured the frequency matches. Due to filtering, there is no guarantee that phase will also controlled. With hundreds of async inverters generating the correct frequency but not necessarily all in phase, what do you think will happen?
Spain
Spain’s solar inverters dropped off the grid for one of two reasons.
Either the large generators dropped off so there was no grid frequency left for the solar inverters to work with. Or the grid frequency went so far out of spec the inverters dropped off by design.
You said: “They don’t try to synchronize, they use the grid as their reference to add the energy.”
“Either the large generators dropped off so there was no grid frequency left for the solar inverters to work with.”
Which is exactly the situation Jim Gorman laid out: “What makes the hundreds or thousands of async inverters down the line lock onto that one isolated transmission, especially if they are on the distant side of say two or three transformers?”
You basically answered that this wouldn’t happen. Now you are saying that it *did* happen.
They don’t try to synchronize, they use the grid as their reference to add the energy.
Try thousands of solar inverters on typical grids these days. Tens of thousands even. What happens is what they’re meant to do, they add their energy according to the frequency of the grid and don’t effect that frequency.
Except when: ““Either the large generators dropped off so there was no grid frequency left for the solar inverters to work with.””
In which case they detect it (because there is no stable frequency) and quickly drop off by design.
You seem to be lacking an accurate definition of synchronize.
A simple definition of “synchronise” is:
But…synchronised implies independently doing it. At least it does to me. In the case of an inverter, the signal isn’t “synchronised to”, it is better described as “defined by”.
You could say one car is synchronised with another if they travel along a road together at a constant distance from each other. But you wouldn’t say they were synchronised if one car was towing the other. Synchronised isn’t the right term in that case.
But yes, its a matter of interpretation and I’m sure people have different opinions.
Nit pick. This would add an additional network of sync signals that would need to be maintained. If that additional network is powered from the grid then when the grid fails so would the sync network. Repeaters for the sync signals would incur transmission delays from node to node thus causing sync problems over long distances. You might simplify inverter design but would incur other parallel complications.
Power flow through a transmission line depends on the phase difference between the ends of the line. Since power flows will change with changing demand and generation, the relative phase between endpoints will need to change. This happens automatically with rotating generation.
How much battery, at what cost, at what risk, and how much material investment? Maybe you can do it, but is it practical at the necessary scale? Why do it?
To move us away from fossil fuels which are depleting. Not to move us away from fossil fuels for reasons of climate change, that is a non reason IMO. Unfortunately the world thinks climate is the main reason and so the rate at which we’re transitioning is dictated by that wrong reason.
That must certainly take place at some point, but to force it using unproven and inadequate “replacements” is unacceptable. My question of why addresses only that aspect of said transition.
Because, like the rotary inverter it costs a lot more.
Renewable energy is already fiendishly expensive and a voracious consumer of resources. Its hard top situation the lie of cheap renewables when you need a battery AND a gas power stations to make them work
By the way, this is precisely what batteries are being added to grids for. They certainly cant cover the loss of solar power overnight.
For 4 hours.
The build is expensive. Running them, not so much.
We should expect that as fossil fuels deplete in the future and the supply reduces, running costs are going to increasingly matter.
We should expect by the time that future arrives we will have engineered sensible solutions and made an economical transition to those new solutions.
What we are witnessing is too much, too fast, with too little technical analysis, analysis of alternatives, and analysis of unintended consequences.
What alternatives? Nuclear? Nuclear energy is reasonably well understood but repeatedly hasn’t been selected. You can complain about the reasons it wasn’t selected but you cant alter the fact.
What we are witnessing is a much needed correction on the stability side of the transition as opposed to the generation side.
The description of the above is not quite correct. The reason dispatchable generation has “inertia” is because the generators are pushing a specific frequency into the grid – this arises from the tuned rotational speed of said generators. A lower grid frequency would have no effect on the rotation of the dispatchable generators because they are effectively independent in the frequency domain.
Inverters, on the other hand, whether battery based or solar PV/wind based, must get their reference frequency from somewhere. In theory, you can put an independent clock on a circuit and use this to generate frequency. But obviously, the existing inverters don’t do that.
And note that this independent clock is not trivial because all power going to an IC is coming from the outside/grid. Whether it is AC frequency variation or DC voltage variation (or other issues like temperature), these can all affect semiconductor clocks. Couple this with the general difficulty that semiconductors have with high power, well, I am no expert but it seems obvious to me why inverters don’t do things “right”.
All true, but we do still have vacuum tubes and those are less power constrained and will be available until we shut down AM and FM radio stations.
There is a reason we don’t use vacuum tubes anywhere, anymore, if there is a solid state silicon semiconductor that can do the job instead: they break frequently and unpredictably.
But more importantly: vacuum tubes don’t make for good clocks – they still have the same power supply variance issues.
You need something like an atomic clock or any other clock whose performance is not tied to input power supply.
In the case of dispatchable fossil fuel, hydro or nuclear – the frequency is an outcome of mechanical engineering.
Vacuum tubes are still a big part of guitar amplifiers and this won’t be changing.
As battery prices drop? As demand for batteries increases, prices increase. Part of that is supply chain.
Batteries providing synthetic inertia last how long? Maybe 4 hours is enough, but until the proper analysis is completed, one cannot claim synthetic inertia is a panacea that solves the ultimate question of life, the universe, and everything.
Then there are failure. The more complex, the more failure modes, and reliability goes down. Once a proper validated FMEA and MTBF are completed, again it is not “42.”
So, your assertion that it solves more problems is vastly unproven.
Spain was the first large scale demonstration project and it failed miserably.
Until a proper demonstration project, with all the incumbent cost, performance, reliability, and failure modes and effects, is completed successfully, I would shy away from claiming it will be successful.
I am an engineer and a scientist. Skepticism is key. Verification is vital.
You can’t “demonstrate” at the scales needed to learn the lessons that running renewable grids will teach us. Instead you transition towards it and deal with issues along the way.
1.6 billion euros and 6 people dead in this “transition.”
Did you look into the actual deaths?
Three people believed to have died because they used a generator and died of CO poisoning. Another died in a fire thought to be started by a candle.
The power outage had a role but wasn’t “causal”
Asynchronous isn’t DC. All wind turbines produce AC. They can be configured to change their AC Asynchronous output to DC then using an inverter change it to AC to suit the grid. Asynchronous just means that the AC produced by the alternator isn’t the same frequency as the grid.
True but irrelevant. Renewables represent a fluctuating source of energy with no inherent storage.
Whether its AC or DC is not the point. It’s easy enough to switch between the two.
The key point is no energy storage beyond maybe a smoothing capacitor.
Cos it cost a lot more. And is not as efficient
End of.
Rotary converters have been around since before transistors were invented.
That’s an M-G set, the term “rotary converter” applies to a machine with a stationary field and an armature with slip rings on one end and a commutator on the other end.
Wind and solar aren’t really “generators”. They are energy collectors.
Not precisely. They are energy converters.
Given EM radiation is sinusoidal, SV is AC to DC conversion.
Given wind is not AC (very low frequency thus roughly DC for the analogy) WTG is DC to AC conversion.
STORY TIP
The CENTENNIAL GLEISSBERG CYCLE: If you’ve been enjoying the auroras of Solar Cycle 25, we’ve got good news. The next few solar cycles could be even more intense-a result of the little-known phenomenon called the “Centennial Gleissberg Cycle”
The respective paper
https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2024SW004238?campaign=woletoc
Here we go
Miliband sets out proposals for solar canopies above car parkshttps://www.timesandstar.co.uk/news/national/25142846.miliband-sets-proposals-solar-canopies-car-parks/?callback=in&code=YJJLODK0N2UTZDU2MC0ZZDJJLWI0NDUTMDG1NTNJOTVJMZUY&state=9cff66b84e5b4061bfc5f649b92cdcc8
Great idea. This way, when the panels self-immolate, they can destroy an entire car park of cars along with themselves, thereby maximizing the destruction of property. Splendid.
If the cars are EVs, the immolation will be even more spectacular.
We have parking lots with the illuminators powered by solar voltaic charged batteries atop the lamp poles. And excelling niche application.
Funny, Nissan produced one, more than 10 years ago. It didn’t catch on (maybe it caught fire instead 😄).
While inertia is very important that is not the whole story.
Inertia is a damper and slows rate of change of frequency caused both by demand and supply variation.
The grid needs to be balanced with respect to demand and supply, they must match instantaneously. Renewables cannot modulate output to match demand and keep frequency within the necessary strict limits.Solar for example, due to it’s output increasing as the sun gets higher often at the time when demand is falling, not a good situation, or vice versa.
The U.K. now relies entirely on gas to perform this balancing which is done automatically using speed governors. At times the grid operator will identify from experience, that the gas generation will have too big a task to keep within limits which is where storage is useful as it can provide power very quickly, but is limited in time, i.e. it’s capcity is relatively small and it also needs to have what it stores replaced.
Synchronous dispatchable generation is essential to a grid and is not easily replicated artificially.
Why any country goes against natural engineering and physical limits and risk grid securiyty just to get more renewables on a grid is beyond my comprehension?
Most UK BESS have an operating time of 1.5 to 2 hours. As you say when the UK underwent an extended period of dunkelflaute for 5 days last November we relied upon large amounts of gas fired generation.
This is not correct: Renewable absolutely CAN modulate to meet supply and demand. The way they would do it is the same as the way everyone else does it: turn off when there is too little demand.
The only reason the renewables don’t do this is because of bad contracts that force favoritism of them over dispatchables.
And the reason why these contracts exist is because otherwise, the renewables would not be profitable.
The way intermittent generation *should* work is that they should be *THE FIRST* to be turned off when there is a curtailment situation, not the last. Solar PV before wind, wind before fossil fuel, fossil fuel before nuclear.
In an ideal world, the “renewable” generating stations would have some sort of local energy storage such as batteries to enable them to vary their output (decrease AND increase) in response to changing demand regardless of the wind speed or amount of sunlight.
In an ideal world those thousands of batteries would have no maintenance costs. Telephone companies found out that trying to provide backup batteries at remote carrier (like for subscriber line carrier systems) sites all over the country cost a lot more for maintenance than they originally anticipated. You can’t just stick the battery in a housing and assume it will last for five years – or at least you shouldn’t if you plan for high reliability.
Your model is incomplete. You are referencing very short term variation.
In the real world, wind and solar have not just daily, weekly and monthly variation but seasonal and yearly variation. Europe has had at least 2 situations in the past 5 years where wind output was 30% below projections for 3 months or more.
Needless to say, this is an ENORMOUS storage requirement.
And note this is further complicated by the over-supply requirements for Net Zero: if you have to have 4x or 5x wind and/or solar PV to replace 1x dispatchable, then you WILL have situations where the massive generation overcapacity is going to exceed demand. Right now, dispatchable is being modulated to compensate; what I am saying is that this is exactly backwards.
Proven technology.
If it ain’t broken, don’t fix it is one engineering saying.
The right tool for the right job.
— Attributed to Scotty
I’ve been in discussion with a person on YT, regarding inertia. I raised the issue, that traditional coal/gas/nuclear have large turbines and generator rotors. That have a lot of mass.
I then questioned whether SMRs could provide the same inertia, being considerably smaller.
As I understand it, if there’s a change in demand, yes the frequency can drift, but drawing additional current, will have a feedback effect on the generator and try to slow it down. But due to the mass, the effect is minimal, before the system stabilises again.
Would an SMR, be more prone to the same problem as wind/solar?
Is my understanding correct, or am I thinking it the wrong way round?
Comments very much appreciated by anyone, who actually works in the power generation industry.
“Small” modular nuclear reactors are still very large, typically in the 100 MW range. And they still spin large rotary generators with a lot of inertia.
It would be a little more difficult to keep grid stability with 10 of these 100 MW generators than with one 1 GW generator, but still much easier than with large amounts of “inertia-less” generation.
Thanks Ed, so my thoughts of a network of SMRs was sort of on the right track, but it wouldn’t be as . problematical.
There isn’t an easy answer to your question. Load variations can range from small to large, in both directions – heaving load to almost no load and almost no load to very heavy load. The amount of time this variation takes can range from a very short period (think a big short circuit) to a very long period (think working day to evening).
SMR’s take time to react, spinning mass doesn’t. The impact of the reaction time depends on the size of the variation and the time period over which it occurs.
If it was easy to analyze and provide for amelioration of the impacts the grid wouldn’t be very complicated. So is life.
Thanks Tim. It seems like we’re going to have to experience a period of ‘suck it and see’.😀
An excellent question.
There is also the issue of deployment.
For example, an array of co-located 10MW SMRs could be designed to operate as a single 100MW nuclear generator.
So the question devolves to single, isolated SMRs.
Still, an excellent question.
Thanks Sparta, my question assures me, that I’m in good company, with WUWT team. 👍
And how much is this going to cost?
An arm and a leg.
…and another leg.
and both arms
And 16 virgins, plus all the first born across the land.
Come along Eric, what is this ‘cost’ that you mention. The woke world of virtue signals demands climate action without reference to costs.
How they imagine they will survive, once all of the resources of society have been consumed chasing pointless inefficient power provisions? Well that is something that never enters debate either.
Dontchaknow wind and solar are “free?”
/s
I can give you the costs for the Western Australian grid of nominal 5900MW output
5 large battery systems each costing over $2-3B each so lets call it $10B and no-one knows if it will work yet. The ongoing maintenance cost and staff is classified.
East coast Australia is at 19500 MW so using WA numbers you need 3.3 times WA so you are going to need 16 or 17 Mega batteries at around $33 to $51B initial cost.
Is anyone like me got that NBN feeling because that is about the costs they initially quoted wasn’t it. Unfortunately we can’t sell or privatize the NBN because it isn’t commercially viable and I do wonder the resale value of Mega Batteries.
You need to put in a payer to the powers above that those batteries give us a few years at that cost. TimTheToolman has a much cheaper plan but as yet he hasn’t lodged a tender but I am a believer.
To minimize the cost, simply elect the last option, then disconnect all of the worse-than-useless asynchronous generation from the grid.
Problem solved AND cost minimized!
Exactly.
But Renewables are Cheap!
Not when you add batteries, uber long transmission lines that are seldom used to capacity, gas backup power stations, short lifetimes and high maintenance costs. Plus the high amount of energy needed to build the whole shebang in the first place.
And the cost to fix the unintended consequences.
How much did the black out cost? I heard it was 10s of billions.
Australia only produces 30 million tons of CO2 from power generation so it might actually be cheaper to buy Carbon Credits which are like $25 a ton. So that is $750M per year which is probably cheaper than the whole renewables gig.:-)
Story Tip
Danish firm shelves huge UK windfarm project over rising costsØrsted cancels fourth stage of Hornsea project off Yorkshire coast, which was expected to include 180 giant turbines
It’s in the Guardian but not in the BBC
https://www.theguardian.com/business/2025/may/07/danish-firm-shelves-huge-uk-windfarm-project-over-rising-costs
Can’t make a decent profit despite all the Taxpayers money we throw at these things
Earlier this year Orsted said it was reducing its investment programme to 2030 by 25%.
Equinor have also scaled down investment in unreliables and low carbon solutions because they are not seen as providing enough profitability. They had earlier also pulled out of hydrogen projects in Germany as there was no market.
Expect others to follow.
attention, any electrical engineers well versed in the generation field.
a poster above says that all wind turbines produce AC. i have no reason to doubt this. so for clarification, can someone answer the following, specific to the raw output of the generator, at points of generation.
nuclear generators…AC or DC
coal generators…AC or DC
nat gas generators… AC or DC
hydro generators… AC or DC
wind turbines…AC or DC
portable non inverter (conventional) gen-sets…AC or DC
portable inverter gen-sets…AC or DC
i think clarification on raw output will be beneficial to future discussions .
thanks.
Non-engineer, but I think I understand the basics for a short answer.
Inverters convert DC to AC and vice versa.
Anything that spins, coal, gas, nuclear, hydro, produces AC, this includes wind turbines.
DC is solar plus batteries and supplies that have been converted from AC via inverters, usually for long distance transport.
On the subject of wind turbines, the fact that they produce AC is more or less irrelevant, because they lack the physical mass of a steam/gas turbine, and spin far slower, thus they lack the inertia to stabilize the grid frequency, which is the real engineering problem with wind/solar.
They can and have been arranged to be synchronous, but that only works over a narrow part of the turbines output
To maximise output they rectify to DC d then up convert to the right voltage and frequency to match what they find on the grid.
Isn’t that the point?
They can synchronize with the grid, but they need something else to tell them what to synchronize to.
There are both AC and DC spinning generators. The thing is the AC versions are cheaper and more efficient. The generator in your car is actually an alternator that produces AC. That AC is then converted to DC so it can be used to power your cars electronics.
For this reason, the direct output from wind mills is AC, however the frequency being generated is directly related to the speed of the blades. To get around this, and to fix the problem of synchronizing. That AC is converted to DC, then converted back to AC that is grid synchronized.
Leo Smith nailed it.
Just a minor detail:
WTG have a tough time maintaining a constant frequency in variable winds.
As such they convert the AD to DC for output. There could be exceptions. I doubt one person has studied all the tens of thousands implemented.
Typo “AC to DC”
Wind turbines can apparently be DC or AC generators so you’d expect DC generation when they’re coupled with batteries and in the case of Hornsdale wind farm in South Australia the Tesla battery gets a big earn providing Frequency Control Ancilliary Services (FCAS)-
Neoen – Tesla batteries at Hornsdale Power Reserve | news.com.au — Australia’s leading news site
Well at least it did earn a motzah as an early adopter with the fickles grid but it has competition now-
SA synchronous condensers installed – Energy Magazine
It’s all expensive stuff and there’s been a couple of occasions on a weekend day like Spain in ideal conditions where the fickles have provided 100% of demand and naturally the fanbois are applauding.
It all adds up in the power bills but if you check the NEM grid output at night and select the Fuel Mix you find it’s 61% fossil fuels at present which can reach 85% plus with aircon demand soon and EVs are only selling 6% of sales at present-
AEMO | NEM data dashboard
What happens when EVs are supposed to be 100% by whenever we can manage it and owners want to charge them at home at night?
Oops… I meant 71% fossil fuels although it’s 72% now
DC generators (coils and magnets) are more expensive than AC generators.
Cheaper to add the inverter to get DC.
More specifically wind turbines are AC-DC-AC: the issue is the frequency of the initial AC conversion is a function of how fast the blades are turning, i.e. the wind speed, and thus cannot be directly connected to the utility (the grid).
Everything else in your list does generate AC electrical power at the correct frequency, 50 or 60 Hz (don’t know about the last one, might be something intended for off-grid power).
Another issue is that AC alternators are smaller and cheaper. That’s why your ICE car has an alternator not a generator.
A wind turbine does not produce DC like a battery. The design of a DC generator does what is known as rectification, that is, it cuts off one-half cycle of an AC waveform. Depending on the number of separate windings on the armature one may generate a fairly high frequency that has a higher average “DC” component. However a DC generator prevents direct connection to the grid.
thanks jim and others. the whole point of me asking is to identify all the points of rectification and inversion in the different schemes.
example: do the generators at the hoover dam spin and produce AC from start to finished product at the end user? if so, is frequency matching performed by regulating rpm vs other sync mechanisms?
AC. Frequency by regulating RPM. Primary source that others synch to.
It is possible to perform slight variations of RPM to adjust the phase, but I do not know if that is done.
Generators – whether wind or hydro or whatever – can either generate DC electricity or convert DC electricity to AC. A DC to AC generator is an inverter.
So yes, the generators comprising wind turbines are actually the inverters themselves and wind turbines should be generating AC.
But note that because wind speeds are variable, a wind turbine is not a source of frequency “inertia” because its rotational speed –> frequency is not consistent. You would have the same situation if a hydroelectric generator operated on a flowing river as opposed to a catch basin: river flow variations would also cause frequency variations whereas dams use sluice gates to maintain a consistent flow of water through the generation system.
And so you can see how a large percentage of wind power, going into a grid, can be a source of frequency instability because a widespread change in wind speed would result in a widespread input of changed frequency input AC onto the grid.
Now I am sure the wind turbines have some control to try and prevent this type of situation, but I am pretty sure that this control is fairly narrow and not a priority given the types of contracts and lack of accountability they enjoy.
Even easier.
Inertia means that rotating masses contain energy that can be used in an emergency to hold the grid up for a few seconds. Making it resilient.
Renewable energy has no reserves whatsoever.Making it brittle.
It doesn’t bend. It breaks
True, but I don’t believe that a windmill has no inertia. We just don’t know how to use it efficiently.
Of course a WTG has inertia, but it is not the same as grid inertia.
Can it be used? Interesting question.
Could clear statement of issue….thanks.
Suggest: 1. give a time frame to compare 2min, 30 min….more. and to allow sensing and switching equipment to operate on a transmission grid. How long does it take a grid to compensate for “green” effects?
Also, eliminate the “we need”…insert each “green” insufficient inertia facility must provide physical or monetary compensation for leaching stability from the grid.
Thanks….
Sparta Nova 4 said something in the previous inertia article which is well worth repeating here:
Let’s expand upon Sparta Nova’s remarks. From the part of my career in nuclear doing quality assurance, I was involved in doing root cause analysis for some number of adverse events of various kinds, and in formulating corrective actions for solving the underlying issues and problems which spawned those events.
These events involved not only failures in physical systems, but also the failures of employees, managers, organizations, and management systems to do the jobs assigned to them.
In nuclear, one hard lesson learned in the 1980’s was that it isn’t enough to look at a single adverse event as an individual occurrence separate and apart from the physical and organizational systems that the adverse event occurred in.
If one is going to be highly proactive in avoiding further similar events, the lower level problems and issues must be integrated into a larger picture. Doing this often reveals larger cross-cutting problems and issues which would not have become visible had that larger picture not been created and analyzed.
Being highly proactive in preventing emerging issues and problems from occurring carries risks for those involved in identifying the larger cross-cutting issues inside of complex physical and organizational systems.
For example, in the 1980’s, the NRC was directed to become much more proactive in identifying emerging issues in the nuclear construction industry and to work closely with that industry in an effort to prevent the serious failures in nuclear quality assurance which were occurring in that era.
Being highly proactive sometimes requires questioning the capability of management — something which is not easily done.
If an investigator is wrong in his analysis of the larger picture, the impact on that person’s reputation is a disaster. On the other hand, if his analysis is right and proactive measures are taken to solve the lower level issues and problems, adverse incidents which might have occurred don’t occur.
Anyway, what the NRC found in the 1980’s was that it had given construction licenses to power utilities which simply weren’t capable of managing their nuclear projects nearly as well as those projects needed to be managed.
Here in the year 2025, what we see now is the complete refusal of those promoting wind and solar to perform an end-to-end analysis of the many technical and programmatic issues facing the transition away from conventional fossil-fueled power, and to come clean in acknowledging how much further work needs to be done to make their transition a reality.
Your experience in nuclear is splendid.
Rockets and payloads are much smaller, less complex systems.
I did not address the personnel issues due to the nature of my work. I hinted at it, though.
Each one has resulted in corrective actions. Our number one issue is people working around very large “bombs” and how to keep them safe. The cause and corrective actions in each event is documented and distributed.
Keep up the good work, sir.
I’d say that in the past, too much focus has been on building more and more subsidised, income producing, renewable energy generation and not enough focus has been placed on grid stability. But I also see grid stability getting much more focus these days so the engineers are winning their battles with the pointy haired bosses…
It is starting, to be sure. FYI, my boos does not have a pointy beard, but I do. 🙂
“…or just plain old conventional generation like nuclear or gas.”
Why is coal missing? Coal is significantly cheaper/quicker to build and operate than nuclear which is the most expensive form of conventional generation.
We build nuclear for purposes of energy security and reliability, not because nuclear is the cheapest way to generate electricity.
Going with nuclear is a public policy decision, for which we are obliged to pay a premium over what coal-fired or gas-fired generation costs.
IIRC, the capital cost of Vogtle 3 & 4 has added roughly 10% to the consumer price of electricity in the state of Georgia. In return for that price increase, power customers in Georgia got a reliable supply of 2400 MW of electricity for the next eighty years.
Georgia’s politicians made a conscious decision to support Vogtle 3 & 4 regardless of its cost overruns, doing so as a means of installing an absolutely reliable 2400 MW of long-term capacity for the state.
In the Era of Trumpsonian Energy Policy, the largest obstacle to new-build coal-fired capacity is stiff competition from new-build gas-fired generation.
Here in the US, if a power utility can gain access to a reliable long-term supply of natural gas, the advantages of new-build gas-fired over new-build coal-fired in terms of environmental compliance costs and operational costs is enough to give the nod to new-build gas.
Where possible, is it not cheaper to ship gas via pipeline, than it is to overland ship coal by train and truch?
As an engineer who understands inertia, risk analyses, and reliability, I have one simple question: Why? Why in this 21st century modern civilization are we doing this (solar, wind, batteries, etc) to ourselves when we’ve thrived on reliable, low risk, high inertia, dispatchable, rampable thermal generation to get us here? We have NO NEED to add this risk, cost, and land and resource devouring crap to our Grid.
i agree 100%.
+10
Old ways seem more reliable than new ways because:
1) Most of the non-computerized engineering technology is 20th century. Ideas that did not work are gone and forgotten. Flaws in ideas that can be made to work are safe and familiar.
2) Manufacturing likes making more of the old thing. Even if the new thing is better, they don’t know how to make it yet. They might not want to know.
3) Engineering management knows it will take twice as long to design as business management’s spreadsheet predicts plus sales already promised the customer they could pull the schedule in and signed the agreement for early delivery. Big changes are risky.
4) The standard warranty for power products that sometimes sit in a brackish swamp is 20 years. Made sense for giant magnets wrapped in copper inside powder-coated steel boxes. Does not make sense for circuit boards with tiny electrical chips
Point is:
Technological capability changes faster than engineering and manufacturing wants to change. As examples consider older industries like auto… count the ratio of GM, Toyota and Kia on the highway next drive. How did that happen? What does that look like in the long warranty, high capital cost, aging workforce, low population growth rate of utilities in the USA?
Moore’s Law still applies.
It’s a real balancing act with extremely thight margins. This is Europe right now.
Based on a sample of the discussion, synthetic inertia could work but is expensive and requires a bunch of very flammable batteries all charged up all the time. Otherwise, I don’t see batteries as being part of the solution since they require inverters and have no inertia on their own. Plus the cost. The cost of inertia is already built into a traditional power plant.
Australia spec for grid forming inverters. Of course they add yet another cost to the ‘cheap’ renewables and hence reduce the rent seeker’s profits. https://aemo.com.au/-/media/files/initiatives/primary-frequency-response/2023/gfm-voluntary-spec.pdf?la=en&hash=F8D999025BBC565E86F3B0E19E40A08E
Important stuff, we need lots of important stuff.
Dave Edwards explains grid inertia brilliantly—clear, concise, and insightful. A must-read for understanding renewable energy’s impact on power stability.
learn more : https://www.inymbus.com/
I have read : “National Grid [UK NESO] said that its current policy was to ensure that system inertia was always above 140 GJ. “
140 GJ is about 39 MWhrs. If this energy is provided by thermal generators can anyone tell me what this means in terms of the actual generator capacity in MW? Is it dependent upon the speed of reaction required?
Apologies if this is a silly question. I am not an engineer, let alone an electrical engineer. I want to know what thermal generating capacity is required to be online/operating to achieve this amount of “system imertia”.