Mains frequency

There is a lot more to this than just the inverters. It is an enormous control system problem, with many many parameters, some of which are unknown. An Inverter connected to a windmill must take all the available output and get it into the grid cycle by cycle. The blade angle has a control system to adjust the blade angle, but this is much too slow to deal with wind gusts immediately. If the grid load does not match the blade speed, they will speed up, but slowly due to the huge inertia. Gas turbines are also not known for their immediate response to more fuel (heat) and must be synchronous anyway. The same goes for steam turbines, but here the limiting factor is probably rotor inertia. There are also time delays due to distribution systems and consumers loads, particularly if these are large rotating machines. There is a huge problem with batteries, some of it due to the required inverters, as well as the chemical problem of short extremely large charge and discharge cycles, perhaps even at the inverter switching frequency, perhaps 20-50 kHz. Overall large amounts of wind power are a pretty bad idea, unless we have nice constant winds, which are somewhat unusual at many of the wind sites. Wind companies are not very interested in further development unless government makes large subsidies, overall costs are higher than expected and in many cases revenue lower. The person who solves this lot is quite likely to end up fairly rich, so who is volunteering?

Or build big battery banks around the country.  Spare power from renewables can be stored.  Then when the generations drops too low, the batteries can top the grid up for long enough to bring backup generators on line.

None of this is possible with wind or solar generators and static inverters. They supply into grid whatever the wind or solar source can supply. There is no question of briefly supplying a bit more in response to a falling frequency.

It might be possible. Say an inverter, in normal circumstances, shoved only say 95% of the available power into the grid - sending the remainder into some local dump (resistive heater say). Then when a bit of extra power is required it could increase output (up to the amount normally dumped) pretty much instantly(*) - whether the need for more power is communicated by falling frequency or by some other method.   For wind, you probably don't even need to generate the extra electricity and dump it - but just tweak the angle of the blades slightly to catch a little less wind.


Certainly at the moment, where the primary aim is to produce as much as possible from renewables, we probably don't want to adopt such an approach - so keep using inverter designs that try to export every mW that can - but if we were in the situation where almost all our power was from renewables and so most of the time we had a bit of excess generating capacity lying around, it might be workable.


(*) - I can't claim the approach as a new invention - I've seems something along those lines in a old (probably 1920s) generator powered by a water wheel - with some very dodgy looking open frame resistances!


   - Andy.

I do not believe that any centralised clock would help.

The problem is that inverters have no inertia and can not help stabilise the grid frequency.

A steam turbine has considerable inertia, and in case of a frequency drop will "try" to limit the drop in frequency. As the turbine slows, the governor detects this and supplies more steam, the resulting drop in boiler pressure will result in the fuel supply being increased.

The stored rotational energy is supplied in milliseconds, and the energy stored by the steam in the boiler is applied within a second or so, the chemical energy stored in fuel bunker can be applied at an increased rate within a few minutes.


None of this is possible with wind or solar generators and static inverters. They supply into grid whatever the wind or solar source can supply. There is no question of briefly supplying a bit more in response to a falling frequency.


Short term grid storage, mechanical flywheels, or batteries for example will be needed, trial installations are already underway.

I've often thought about the impact of more and more invertor derived AC on the grid, and less and less conventional synchronous generation on the grid


As we move more to the former, and less of the latter one has to wonder if at some point we need to introduce a central or regional frequency clock to kick all the invertors back into synch


The man who owns that clock, owns the world !!


Regards


OMS

It is interesting - though I must say have not seen a mains derived electronic clock for years - quartz crystals are pence in bulk.

Mechanical generators slow down and generate a  lower frequency as the load current rises, and we can use that to sychronise a generator into phase -lock with an existing supply - you get it nearly synchronised so the difference frequency is sub Hz, and close the breaker, and the current that flows either adds or subtracts some torque, and in turn speeds up or slows the rotor until things settle to exactly the same frequency, and a voltage and phase offset that represents the fraction of the load handled by that one generator relative to the average of the rest.

There is an assumption in the way that the rules of the grid are written that all generators do this, but of course inverter derived things like solar farms, and most wind turbines (certainly those at the end of DC links...) have no such characteristic, the voltage droops as more current is taken, but the zero crossings occur where ever the electronics tells them it is time to invert the DC to do the other half cycle - Imagine if you will a big battery and a reversing switch operated at 100 times a second (as a full cycle is 2 reversals.)

It is possible in software to alter the timing of the zero crossings, either in a phase shift way, i.e. a one time shift and then a constant period again, or in a frequency shift way (all cycles have the new period.) and right now we go to some effort to make these inverter sources mimic the behaviour of a traditional generator, but it is not clear if as we may soon approach a situation with more kit on the grid emulating rotating generators than actually are, we wish to continue.

It may end up like the phone system with electronics at one end pretending to be a mechanical exchange, and stepping up to a really awkward 50V and 80V ringing   voltage, while at the other end we have phones containing electronics emulating a carbon microphone and a moving iron earpiece stepping down from this really awkward voltage to handle audio at lower levels internally. If we were starting again, this is not the system we would design.

Traditionally with comms systems inefficiency like this is a secondary consideration after compatibility and usability,   it is not clear that it should be for power system like a national grid.

A lot of people think only old style clocks with synchronous motors are affected by the mains frequency. This is not the case, and it is surprising how many new electronic clocks, including alarm clocks, use a calculation based on the mains frequency rather than a highly accurate chrystal oscillator. I guess this is because the calculation results in a cheaper clock for the achieved accuracy. 


Regards,


Alan.

I suspect that the frequency display on the dynamic demand site is not working correctly.

For some weeks it has been displaying a generally lower frequency than that displayed from other sources.


The DD meter is much quicker acting than most other sources, it is virtually real time rather than being smoothed or averaged over seconds or tens of seconds. Therefore on ANY ONE occasion, the DD meter might well be accurately displaying information that is not shown elsewhere.

However, the frequency shown on Dynamic demand is too much lower, too often, and for too long to be compatible with actuality.


Extreme load, sudden and unexpected increases in load, or sudden losses of generating/import capacity can all cause a drop in frequency. Such situations are usually handled fairly quickly with the frequency recovering in minutes, or very exceptionally in a few hours.


The national grid aim to keep the frequency within normal operational limits of 49.8 cycles up to 50.2 cycles, brief excursions beyond these limits occur fairly regularly.

The legal limits are rather wider at from 49.5 cycles up to 50.5 cycles, averaged over some short time. Excursions outside these legal limits are rare.


Any severe drop in frequency triggers automatic starting of OCGT plant and calls for extra output from other plant. If that does not relieve matters, then areas are automatically blacked out, a rare event that last occurred in August 2019, and previously about ten years before that.

Thanks all for your replies looking at the dynamic demand meter is a little misleading as it looks like the frequency is closer to the lower limit than it really is  49.5 to 49.7 cycles is only a tiny bit of movement which makes it look like things are worse than they are. I agree that newspapers seem to say that we're all doomed to sitting in the dark each winter which hasn't happened yet except when there's been a major fault somwhere  and that's rare of course dramatic headlines sell help them sell these daily rags  so that's that