# Consistency of the mains frequency throughout the UK

I'm working with a number of frequency meters which are reading the mains frequency in the UK from different locations. The problem is that I haven't been able to find information on whether there are local fluctuations in the grid frequency due to the geographical location, the local grid or the environment in which the meter is placed. While I expect that there is some level of local fluctuation I have no idea of how significant it is (i.e. is it in the mHz, uHz or nHz range).

I'm aware of the discrepancies I would expect to see due to the varying accuracy and precision of the individual meters. The tolerance of the meters I'm using is +/- 1mHz so if the local fluctuation was in the uHz or nHz range I’d consider it to be negligible.

My overall goal is to determine if I have a reading on one meter how close will the other meters be reading.

Edit: Just to be explicit my question boils down to: When measuring the mains frequency in the UK from multiple locations what, if any, variance exists between measurements if we were to disregard measruement error? And what is the cause of said variance?

• The real-time graph of power line frequency across the UK national grid is here. Local differences beyond 20 milliHertz from grid control will supposedly trigger corrective controls in mainland Europe (source) so the UK probably follows the same rule. Local fluctuation between subscribers within a single grid source group (say one power station) is typically marginal if any. – Anindo Ghosh Jan 9 '14 at 14:29
• It is difficult to determine what is being asked here. Can you rephrase the question? – posipiet Jan 9 '14 at 14:34
• @posipiet He is asking what is the expected deviation of the mains frequency – alexan_e Jan 9 '14 at 15:06
• @posipiet Sorry about that, I'm a devil for asking questions without actually asking a question. – Peadar Doyle Jan 9 '14 at 18:10

National Grid (NGET) "Guidance Notes - Synchronous Generating Units" Issue 12 says

## Target Frequency

If a generating unit is required to have a Frequency control device as specified in CC6.3.7 it must also have the facility to accept and act upon Target Frequency instruction from National Grid. In order to adjust electric clock time the Grid Operator may instruct settings of 49.95Hz or 50.05Hz. However, under exceptional circumstances, the instructed settings could be outside this range. A minimum setting range from 49.90Hz to 50.10Hz is required.

I suspect Andy's answer is correct concerning local fluctuations within parts of the UK connected to the national grid.

See Wikipedia on Synchronisation

An AC generator cannot deliver power to an electrical grid unless it is running at the same frequency as the network. If two segments of a grid are disconnected, they cannot exchange AC power again until they are brought back into exact synchronization.

NGET Frequency Management includes several interesting graphs that may be relevant.

So, if a calamity occurs, you may briefly find that one end of the country is 0.05 Hz adrift from the other end for a fraction of a second. End users might need expensive and well calibrated meters to accurately detect that.

• The last graph is very interesting, it's the first evidence I've seen that the frequency is not a "one true value" all the time. – Peadar Doyle Jan 9 '14 at 18:50

I'm sticking my neck out here.

The easiest approximation is this: the power grid is totally frequency locked and any local movement measured will in fact be representative of the grid as a whole. Getting into the detail means looking at transient effects as generators come on line - there will be a tiny bounce in frequency due to the "machine" changing from a motor (on the way up to synchronous grid speed) to a generator but this is really just a transient.

Any measurement you might make that looks like a different frequency on different parts of the same grid is probably an error between the two meters used.

• I agree, I read an article about the UK government recording mains hum so they could correlate it to voice or video recordings for time stamping and to ensure they had not been altered. They asserted that this method worked nation wide for exactly the reason you're stating. – Samuel Jan 9 '14 at 17:56
• @Samuel I came across said article in my quest for answers. Very interesting read. – Peadar Doyle Jan 9 '14 at 18:22
• @Andy aka Thanks for the info. Mind sticking your neck out again and take an educated guess as to whether this would happen locally if large, say 250kW, motors are being started/stopped. – Peadar Doyle Jan 9 '14 at 18:26
• @PeadarDoyle This is definitely sticking my neck out (LOL). I can't see it making a difference at all. I can see it taking the supply voltage down 10% and maybe tripping the odd breaker or glitching a PC so it crashes but I see no cause and effect mechanism that could fight the grid control systems and win (even a little). It would take a concerted effort on a national level for this to happen I'd imagine. – Andy aka Jan 9 '14 at 18:46
• @Andyaka Thanks once again. I had a colleague insisting that local frequency changes occur due to local effects but the more I dig it appears this isn't the case. – Peadar Doyle Jan 10 '14 at 15:09

Another simplified way of looking at it is ... the entire "grid" is essentially the same enormous piece of wire, or fuseboard bus, that everyone around the country taps into. It has the same signal on it everywhere. There may be a slight propagation delay from one end to the other, but in a country barely 1000km long and with signal propagation speeds north of 150,000km/s, and the sync being continually reinforced at rather smaller jumps along its length, that's pretty much immaterial.

If something is unable to comply with the group-mandated frequency, it's either forcibly brought back into line by the other components (as mentioned, an underperforming generator will be, in possibly stuttering fashion, driven as a synchronous AC motor (much like that in your vacuum cleaner) by all the others, until its rising speed and their falling speed are back in sync) or, if that cannot be efficiently achieved, the appropriate set of safety breakers snap open and it's unceremoniously ejected from the network.

Thus anything that shows a significant difference from the rest of the grid network for more than a second or two, with properly calibrated meters attached to both items ... most likely has been disconnected and ISN'T on the grid any more.