OT: 'Photon Farming' in California

[Tom Gardner]

On 10/08/19 13:52, Martin Brown wrote:

On 10/08/2019 08:05, Tom Gardner wrote:
On 04/08/19 15:41, Tom Gardner wrote:
On 04/08/19 15:01, Rick C wrote:
[snip]

Yup, the UK is stuffed.  Better get used to it.

Agreed, the UK energy industry is screwed up beyond belief.

Political/economic ideology is the core reason.

And it happened - a concrete demonstration of how perilously
close to the limit the UK power system is.

And in mid summer when you might expect things to be safe.

I hope this will finally get through to Rick C that EV
infrastructure isn't as simple as it is in his neighbourhood.
But I doubt it

Yesterday *large*[1] parts of the UK had power cut because two
plants went offline within two minutes - one gas plant, one
wind plant. Yup, wind power does suddenly stop.

[1] There blackouts across the Midlands, the South East,
South West, North West and north east of England, and
Wales.

It seems likely that the initial problem which was the first gas turbine
generator dropping out wasn't dealt with in the two minutes before a second
independent failure of a wind turbine farm. The two being offline together then
took the mains frequency sufficiently far out of bounds that some inverters
stopped as well. Cascade failure followed.

So it seems. Gridwatch shows down to 49Hz when the limit
is 49.5Hz.

I look forward to seeing more in the trade press,
e.g. the IET and comp.risks


> Dinorwig can start in under 20s so someone wasn't paying attention...

Maybe it was otherwise occupied?

<waves arms>
Maybe the velocity of the cascade was both too fast for humans,
and too slow for the automated system on the other side of the
country to kick in.
</waves arms>

The thing that they have failed to model adequately for grid management is that
conventional turbine generators have huge synchronised spinning rotors that
offer considerable inertia to changes in grid frequency.

I pretty sure they know that and have modelled it It may well
be that load shedding occurs to prevent excessive stress being
put on the remaining rotors.

By comparison semiconductor inverters seem to offer almost none and give up
completely if they find df/ft too large or f outside accepted bounds. I don't
see why this needs to be the case provided that the devices are protected from
overload - a couple of percent frequency variation shouldn't take a decent power
transformer into saturation.

It was astonishing how widespread the UK power cuts were for a fault that was
concentrated in the SE near Cambridgeshire - they lost Newcastle Airport nearly
300 miles away as a direct result.

Yes indeed.

I think it has got some politicians attention. Whether they will
continue to think about it is questionable, especially in the
face of the disaster of their own making.

NHS hospitals that beancounters had put on the load shedding cheapest tariff
found themselves without any power at all - and the odd one discovered their
emergency backup generators didn't work either.

Classic.


> Lucky it happened at 5pm when most scheduled operations were over.

A few months ago my daughter was waiting in the anteroom all day
and night, and the operation finally happened at 7:30pm.

There appeared to be parts of the London underground that lacked emergency
lighting too (or if they had it then it failed too work).

It was also rather discouraging how few of the major systems recovered
gracefully from a sudden loss of power when power *was* restored after about an
hour. The trains are still a complete mess today.

That's unsurprising. You can't have a train from X; you have
to have a train to X followed by a train from X

All you have to do is look at the compass points to realise
how widespread that was.

That wide geographic spread is a bit weird - the protection system is supposed
to limit the contagion and shed the right sorts of load first and nearby. We
were not cut off but others in major cities were.

https://www.bbc.co.uk/news/uk-49302996

Offgen is looking into it. Don't hold your breath...

Yes, curious minds do want to know.

 

[Rick C]

On Saturday, August 10, 2019 at 11:50:05 AM UTC-4, Tom Gardner wrote:

On 10/08/19 16:41, piglet wrote:

If so that may not be quite so stupid, although I can't see why the cutout had
to be set so close to the regulatory limit, most loads would still be perfectly
content at 47Hz for instance.

You have to cater for those that wouldn't be happy
and have been designed with the stated limits in mind.
But that is only hand waving.

I would speculate it is more about the concern that a significant aberration is happening and it is simply best to go offline rather than continue to draw power.

--

Rick C.

+-++ Get 1,000 miles of free Supercharging
+-++ Tesla referral code - https://ts.la/richard11209

 

[Tom Gardner]

On 10/08/19 16:27, Rick C wrote:

On Saturday, August 10, 2019 at 3:05:45 AM UTC-4, Tom Gardner wrote:

Yesterday *large*[1] parts of the UK had power cut because two
plants went offline within two minutes - one gas plant, one
wind plant. Yup, wind power does suddenly stop.

Yup, wind power stops suddenly, but not because the wind stopped blowing.

"Energy watchdog Ofgem has demanded an urgent report from National Grid into how the failure happened as parts of the UK were battered by strong winds and heavy rain."

“The first generator to disconnect was a gas fired plant at Little Barford at 16:58. Two minutes later Hornsea Offshore wind farm seems to have disconnected."

No one said the wind stopped blowing. lol

I think that comes under the category of "no additional information".

 

[piglet]

On 10/08/2019 16:47, Tom Gardner wrote:

On 10/08/19 16:39, Rick C wrote:
On Saturday, August 10, 2019 at 11:03:30 AM UTC-4, Tom Gardner wrote:
On 10/08/19 15:52, Tom Gardner wrote:
On 10/08/19 13:52, Martin Brown wrote:
On 10/08/2019 08:05, Tom Gardner wrote:
On 04/08/19 15:41, Tom Gardner wrote:
On 04/08/19 15:01, Rick C wrote:
[snip]

Yup, the UK is stuffed.  Better get used to it.

Agreed, the UK energy industry is screwed up beyond belief.

Political/economic ideology is the core reason.

And it happened - a concrete demonstration of how perilously
close to the limit the UK power system is.

And in mid summer when you might expect things to be safe.

I hope this will finally get through to Rick C that EV
infrastructure isn't as simple as it is in his neighbourhood.
But I doubt it

Yesterday *large*[1] parts of the UK had power cut because two
plants went offline within two minutes - one gas plant, one
wind plant. Yup, wind power does suddenly stop.

[1] There blackouts across the Midlands, the South East,
South West, North West and north east of England, and
Wales.

It seems likely that the initial problem which was the first gas
turbine
generator dropping out wasn't dealt with in the two minutes before
a second
independent failure of a wind turbine farm. The two being offline
together
then took the mains frequency sufficiently far out of bounds that some
inverters stopped as well. Cascade failure followed.

So it seems. Gridwatch shows down to 49Hz when the limit
is 49.5Hz.

I look forward to seeing more in the trade press,
e.g. the IET and comp.risks


Dinorwig can start in under 20s so someone wasn't paying attention...

Maybe it was otherwise occupied?

waves arms
Maybe the velocity of the cascade was both too fast for humans,
and too slow for the automated system on the other side of the
country to kick in.
/waves arms


The thing that they have failed to model adequately for grid
management is
that conventional turbine generators have huge synchronised
spinning rotors
that offer considerable inertia to changes in grid frequency.

I pretty sure they know that and have modelled it It may well
be that load shedding occurs to prevent excessive stress being
put on the remaining rotors.


By comparison semiconductor inverters seem to offer almost none and
give up
completely if they find df/ft too large or f outside accepted
bounds. I don't
see why this needs to be the case provided that the devices are
protected from
overload - a couple of percent frequency variation shouldn't take a
decent
power transformer into saturation.

It was astonishing how widespread the UK power cuts were for a
fault that was
concentrated in the SE near Cambridgeshire - they lost Newcastle
Airport
nearly 300 miles away as a direct result.

Yes indeed.

I think it has got some politicians attention. Whether they will
continue to think about it is questionable, especially in the
face of the disaster of their own making.


NHS hospitals that beancounters had put on the load shedding
cheapest tariff
found themselves without any power at all - and the odd one
discovered their
emergency backup generators didn't work either.

Classic.


Lucky it happened at 5pm when most scheduled operations were over.

A few months ago my daughter was waiting in the anteroom all day
and night, and the operation finally happened at 7:30pm.


There appeared to be parts of the London underground that lacked
emergency
lighting too (or if they had it then it failed too work).

It was also rather discouraging how few of the major systems recovered
gracefully from a sudden loss of power when power *was* restored
after about
an hour. The trains are still a complete mess today.

That's unsurprising. You can't have a train from X; you have
to have a train to X followed by a train from X


All you have to do is look at the compass points to realise
how widespread that was.

That wide geographic spread is a bit weird - the protection system
is supposed
to limit the contagion and shed the right sorts of load first and
nearby. We
were not cut off but others in major cities were.

https://www.bbc.co.uk/news/uk-49302996

Offgen is looking into it. Don't hold your breath...

Yes, curious minds do want to know.

Another aspect is becoming apparent. The homes affected appear
to be in a fractal patchwork, rather than across a whole area.

For example, I didn't notice anything, but 4 and 10 miles away:
"the company’s website shows 194 homes in Keynsham and 25 in
Nailsea are still [yesterday] without electricity."

Why so few homes, and there not elsewhere?

Can't say about the UK, but in the US areas are often provided with
power from more than one source by redundant paths.  If one fails,
they draw more power from another.  At times this can result in an
overload of any single circuit.  Not surprising failures were
sporadic.  When you stress a system to it's max you find weak spots
first, where ever they are.


The UK is better than the US in that respect.

The problem wasn't the grid, it was insufficient generating
capacity.

I've previously mentioned that if two plants were to have
unscheduled downtime in winter, we would be in trouble - and
it isn't even winter

This is a concrete demonstration that I wasn't being alarmist.

Yes, very strange. Apparently demand was "only" 30GW which is way lower
than winter peak capacity - I guess a lot of generating capacity was off
line for summer maintenance works. Left too fragile for unexpected dual
failure.

piglet

 

[Tom Gardner]

On 10/08/19 16:27, Rick C wrote:

On Saturday, August 10, 2019 at 3:05:45 AM UTC-4, Tom Gardner wrote:

Yesterday *large*[1] parts of the UK had power cut because two
plants went offline within two minutes - one gas plant, one
wind plant. Yup, wind power does suddenly stop.

Yup, wind power stops suddenly, but not because the wind stopped blowing.

"Energy watchdog Ofgem has demanded an urgent report from National Grid into how the failure happened as parts of the UK were battered by strong winds and heavy rain."

“The first generator to disconnect was a gas fired plant at Little Barford at 16:58. Two minutes later Hornsea Offshore wind farm seems to have disconnected."

No one said the wind stopped blowing. lol

Actually, the wind /does/ stop blowing for *days* at a time.
All it needs is a blocking high pressure over the UK

 

[Rick C]

On Saturday, August 10, 2019 at 11:57:07 AM UTC-4, Tom Gardner wrote:

On 10/08/19 16:27, Rick C wrote:
On Saturday, August 10, 2019 at 3:05:45 AM UTC-4, Tom Gardner wrote:

Yesterday *large*[1] parts of the UK had power cut because two
plants went offline within two minutes - one gas plant, one
wind plant. Yup, wind power does suddenly stop.

Yup, wind power stops suddenly, but not because the wind stopped blowing.

"Energy watchdog Ofgem has demanded an urgent report from National Grid into how the failure happened as parts of the UK were battered by strong winds and heavy rain."

“The first generator to disconnect was a gas fired plant at Little Barford at 16:58. Two minutes later Hornsea Offshore wind farm seems to have disconnected."

No one said the wind stopped blowing. lol

I think that comes under the category of "no additional information".

Lol!

--

Rick C.

+--+ Get 1,000 miles of free Supercharging
+--+ Tesla referral code - https://ts.la/richard11209

 

[Rick C]

On Saturday, August 10, 2019 at 11:58:18 AM UTC-4, Tom Gardner wrote:

On 10/08/19 16:27, Rick C wrote:
On Saturday, August 10, 2019 at 3:05:45 AM UTC-4, Tom Gardner wrote:

Yesterday *large*[1] parts of the UK had power cut because two
plants went offline within two minutes - one gas plant, one
wind plant. Yup, wind power does suddenly stop.

Yup, wind power stops suddenly, but not because the wind stopped blowing.

"Energy watchdog Ofgem has demanded an urgent report from National Grid into how the failure happened as parts of the UK were battered by strong winds and heavy rain."

“The first generator to disconnect was a gas fired plant at Little Barford at 16:58. Two minutes later Hornsea Offshore wind farm seems to have disconnected."

No one said the wind stopped blowing. lol

Actually, the wind /does/ stop blowing for *days* at a time.
All it needs is a blocking high pressure over the UK

And does that happen suddenly? I think the power from a wind farm will vary much more slowly than a-ok to offline in two minutes.

I believe I read they are investigating if the first plant failure caused the wind farm to go offline.

--

Rick C.

++-- Get 1,000 miles of free Supercharging
++-- Tesla referral code - https://ts.la/richard11209

 

On Sat, 10 Aug 2019 13:52:03 +0100, Martin Brown
<'''newspam'''@nezumi.demon.co.uk> wrote:

There blackouts across the Midlands, the South East,
South West, North West and north east of England, and
Wales.

It seems likely that the initial problem which was the first gas turbine
generator dropping out wasn't dealt with in the two minutes before a
second independent failure of a wind turbine farm. The two being offline
together then took the mains frequency sufficiently far out of bounds
that some inverters stopped as well. Cascade failure followed.

It also may be that they had a bit bad luck.

Good network management requires that the N-1 rule is followed. The
network should survive the loss of _largest_ production unit. In
practice during the first minutes the spinning reserve of other power
plants are used (.i.e. overloading them). During that time quick
start emergency gas turbines are started, which then takes over from
the spinning reserve. When the emergency gas turbines are running, the
power of existing power plants are added e.g. by burning more coal to
get more steam or other slower starting power plants are starting.
When these are running, the emergency gas turbines can be shut down
and the N-1 criterion is restored, waiting for the loss of the next
largest production unit. These emergency gas turbines are typically
running for less than an hour.

However, as long as the emergency gas turbines are used, the network
doesn't tolerate the loss of the next largest unit.

In this case, there was the second fault while the emergency gas
turbines were running. Of course if two medium size faults occur
nearly simultaneous, there should not be a problem as long as the
combined lost power is less than the largest unit.

Now both units were quite big, the Little Barford 727 MW and Hornsea
Project 1 of 1200 MW, however it is unclear how much of that capacity
had already been installed.

Little Barford is a typical combi power plant with two gas turbines,
exhaust goes through a boiler, running a single steam turbine. Even if
one turbine/generator failed, the remaining turbines should still be
in production, so it should be a bad case, if all 727 MW are lost at
once for some reason

While one wind turbine may loose production due to loss of wind, this
happens much slower in a wind farm with some geographical extent.

If the whole wind farm was lost at once, it is unlikely that it was
due to loss of wind, which can be predicted quite well in advance.
Since this is an offshore farm, power must be transferred to shore
with cables. A total wind farm failure sounds like a problem in this
cable or other farm network components, not in the turbines itself.

Then there is the question, was the wind farm loss truly independent
or caused by the gas turbine fault, since it occurred during spinning
reserve period. How much semiconductors were involved in the link to
shore, which might have tripped due to overload.

 

[Tom Gardner]

On 10/08/19 16:41, Martin Brown wrote:

On 10/08/2019 16:03, Tom Gardner wrote:

Another aspect is becoming apparent. The homes affected appear
to be in a fractal patchwork, rather than across a whole area.

For example, I didn't notice anything, but 4 and 10 miles away:
"the company’s website shows 194 homes in Keynsham and 25 in
Nailsea are still [yesterday] without electricity."

Why so few homes, and there not elsewhere?

There was independent damage to some local infrastructure by lightning strikes
in addition to the main event. A village not far from me took a direct hit last
night. We just got a to see a spectacular noisy display.

I didn't see anything.

There was a truly spectacular lightning storm 15 miles
away a couple of weeks ago. Around a strike a second
for >>10 minutes, i.e. until I got bored.

I think they got most of the load shedding stuff back on within a couple of
hours. Most times they get us back on within half a day unless there is major
damage affecting larger towns and cities in which case we wait.

Yes, and I think it was a few hours this time too.

But then I remember revising for exams by candlelight,
back in the 70s

Most of us have other ways of heating and lighting our homes since the usual
mode of failure is severe winter storm tree breaks the line.

I've now got a pocket-money second-hand UPS which
may or may not work; I've yet to have a chance to
look at it. Principle use case: keeping freezer cold
after a no-deal brexit.

 

[Rick C]

On Saturday, August 10, 2019 at 11:47:04 AM UTC-4, Tom Gardner wrote:

On 10/08/19 16:39, Rick C wrote:

Can't say about the UK, but in the US areas are often provided with power from more than one source by redundant paths. If one fails, they draw more power from another. At times this can result in an overload of any single circuit. Not surprising failures were sporadic. When you stress a system to it's max you find weak spots first, where ever they are.


The UK is better than the US in that respect.

The problem wasn't the grid, it was insufficient generating
capacity.

No one said the initial problem was the grid. The point is as events unfolded parts of the grid were stressed including both generation and transmission.

I've previously mentioned that if two plants were to have
unscheduled downtime in winter, we would be in trouble - and
it isn't even winter

This is a concrete demonstration that I wasn't being alarmist.

Slightly different problem. Simply being down may not have caused this problem. Everyone seems to focus on the fact that they went down within two minutes of one another. That clearly precluded any adjustments to generation and distribution before problems started happening.

--

Rick C.

+-+- Get 1,000 miles of free Supercharging
+-+- Tesla referral code - https://ts.la/richard11209

 

[Tom Gardner]

On 10/08/19 16:41, piglet wrote:

On 10/08/2019 16:03, Tom Gardner wrote:

Another aspect is becoming apparent. The homes affected appear
to be in a fractal patchwork, rather than across a whole area.

For example, I didn't notice anything, but 4 and 10 miles away:
"the company’s website shows 194 homes in Keynsham and 25 in
Nailsea are still [yesterday] without electricity."

Why so few homes, and there not elsewhere?

Perhaps different switch centers have differing under-frequency-lockout
sensitivity and responsiveness delay?

Nailsea population 15k, but only 25 houses?
Keynsham population 16k, but only 194 houses?

And they are nowhere near the plant with downtime.

If so that may not be quite so stupid, although I can't see why the cutout had
to be set so close to the regulatory limit, most loads would still be perfectly
content at 47Hz for instance.

You have to cater for those that wouldn't be happy
and have been designed with the stated limits in mind.
But that is only hand waving.

 

[Tom Gardner]

On 10/08/19 16:39, Rick C wrote:

On Saturday, August 10, 2019 at 11:03:30 AM UTC-4, Tom Gardner wrote:
On 10/08/19 15:52, Tom Gardner wrote:
On 10/08/19 13:52, Martin Brown wrote:
On 10/08/2019 08:05, Tom Gardner wrote:
On 04/08/19 15:41, Tom Gardner wrote:
On 04/08/19 15:01, Rick C wrote:
[snip]

Yup, the UK is stuffed.  Better get used to it.

Agreed, the UK energy industry is screwed up beyond belief.

Political/economic ideology is the core reason.

And it happened - a concrete demonstration of how perilously
close to the limit the UK power system is.

And in mid summer when you might expect things to be safe.

I hope this will finally get through to Rick C that EV
infrastructure isn't as simple as it is in his neighbourhood.
But I doubt it

Yesterday *large*[1] parts of the UK had power cut because two
plants went offline within two minutes - one gas plant, one
wind plant. Yup, wind power does suddenly stop.

[1] There blackouts across the Midlands, the South East,
South West, North West and north east of England, and
Wales.

It seems likely that the initial problem which was the first gas turbine
generator dropping out wasn't dealt with in the two minutes before a second
independent failure of a wind turbine farm. The two being offline together
then took the mains frequency sufficiently far out of bounds that some
inverters stopped as well. Cascade failure followed.

So it seems. Gridwatch shows down to 49Hz when the limit
is 49.5Hz.

I look forward to seeing more in the trade press,
e.g. the IET and comp.risks


Dinorwig can start in under 20s so someone wasn't paying attention...

Maybe it was otherwise occupied?

waves arms
Maybe the velocity of the cascade was both too fast for humans,
and too slow for the automated system on the other side of the
country to kick in.
/waves arms


The thing that they have failed to model adequately for grid management is
that conventional turbine generators have huge synchronised spinning rotors
that offer considerable inertia to changes in grid frequency.

I pretty sure they know that and have modelled it It may well
be that load shedding occurs to prevent excessive stress being
put on the remaining rotors.


By comparison semiconductor inverters seem to offer almost none and give up
completely if they find df/ft too large or f outside accepted bounds. I don't
see why this needs to be the case provided that the devices are protected from
overload - a couple of percent frequency variation shouldn't take a decent
power transformer into saturation.

It was astonishing how widespread the UK power cuts were for a fault that was
concentrated in the SE near Cambridgeshire - they lost Newcastle Airport
nearly 300 miles away as a direct result.

Yes indeed.

I think it has got some politicians attention. Whether they will
continue to think about it is questionable, especially in the
face of the disaster of their own making.


NHS hospitals that beancounters had put on the load shedding cheapest tariff
found themselves without any power at all - and the odd one discovered their
emergency backup generators didn't work either.

Classic.


Lucky it happened at 5pm when most scheduled operations were over.

A few months ago my daughter was waiting in the anteroom all day
and night, and the operation finally happened at 7:30pm.


There appeared to be parts of the London underground that lacked emergency
lighting too (or if they had it then it failed too work).

It was also rather discouraging how few of the major systems recovered
gracefully from a sudden loss of power when power *was* restored after about
an hour. The trains are still a complete mess today.

That's unsurprising. You can't have a train from X; you have
to have a train to X followed by a train from X


All you have to do is look at the compass points to realise
how widespread that was.

That wide geographic spread is a bit weird - the protection system is supposed
to limit the contagion and shed the right sorts of load first and nearby. We
were not cut off but others in major cities were.

https://www.bbc.co.uk/news/uk-49302996

Offgen is looking into it. Don't hold your breath...

Yes, curious minds do want to know.

Another aspect is becoming apparent. The homes affected appear
to be in a fractal patchwork, rather than across a whole area.

For example, I didn't notice anything, but 4 and 10 miles away:
"the company’s website shows 194 homes in Keynsham and 25 in
Nailsea are still [yesterday] without electricity."

Why so few homes, and there not elsewhere?

Can't say about the UK, but in the US areas are often provided with power from more than one source by redundant paths. If one fails, they draw more power from another. At times this can result in an overload of any single circuit. Not surprising failures were sporadic. When you stress a system to it's max you find weak spots first, where ever they are.

The UK is better than the US in that respect.

The problem wasn't the grid, it was insufficient generating
capacity.

I've previously mentioned that if two plants were to have
unscheduled downtime in winter, we would be in trouble - and
it isn't even winter

This is a concrete demonstration that I wasn't being alarmist.

 

[Tom Gardner]

On 10/08/19 17:03, upsidedown@downunder.com wrote:

On Sat, 10 Aug 2019 13:52:03 +0100, Martin Brown
'''newspam'''@nezumi.demon.co.uk> wrote:

[1] There blackouts across the Midlands, the South East,
South West, North West and north east of England, and
Wales.

It seems likely that the initial problem which was the first gas turbine
generator dropping out wasn't dealt with in the two minutes before a
second independent failure of a wind turbine farm. The two being offline
together then took the mains frequency sufficiently far out of bounds
that some inverters stopped as well. Cascade failure followed.

It also may be that they had a bit bad luck.

Good network management requires that the N-1 rule is followed.

Yes, well, that's the idea

The
network should survive the loss of _largest_ production unit. In
practice during the first minutes the spinning reserve of other power
plants are used (.i.e. overloading them). During that time quick
start emergency gas turbines are started, which then takes over from
the spinning reserve. When the emergency gas turbines are running, the
power of existing power plants are added e.g. by burning more coal to
get more steam or other slower starting power plants are starting.
When these are running, the emergency gas turbines can be shut down
and the N-1 criterion is restored, waiting for the loss of the next
largest production unit. These emergency gas turbines are typically
running for less than an hour.

However, as long as the emergency gas turbines are used, the network
doesn't tolerate the loss of the next largest unit.

In this case, there was the second fault while the emergency gas
turbines were running. Of course if two medium size faults occur
nearly simultaneous, there should not be a problem as long as the
combined lost power is less than the largest unit.

I will be very interested to hear
- why the *second* fault was so close in time
- what the actual faults were

Considering the latter can lead to movie-plot theories.

Now both units were quite big, the Little Barford 727 MW and Hornsea
Project 1 of 1200 MW, however it is unclear how much of that capacity
had already been installed.

Little Barford is a typical combi power plant with two gas turbines,
exhaust goes through a boiler, running a single steam turbine. Even if
one turbine/generator failed, the remaining turbines should still be
in production, so it should be a bad case, if all 727 MW are lost at
once for some reason

While one wind turbine may loose production due to loss of wind, this
happens much slower in a wind farm with some geographical extent.

If the whole wind farm was lost at once, it is unlikely that it was
due to loss of wind, which can be predicted quite well in advance.
Since this is an offshore farm, power must be transferred to shore
with cables. A total wind farm failure sounds like a problem in this
cable or other farm network components, not in the turbines itself.

Then there is the question, was the wind farm loss truly independent
or caused by the gas turbine fault, since it occurred during spinning
reserve period. How much semiconductors were involved in the link to
shore, which might have tripped due to overload.

Good questions.

 

[Tom Gardner]

On 10/08/19 17:12, piglet wrote:

On 10/08/2019 16:50, Tom Gardner wrote:
On 10/08/19 16:41, piglet wrote:

Perhaps different switch centers have differing under-frequency-lockout
sensitivity and responsiveness delay?

Nailsea population 15k, but only 25 houses?
Keynsham population 16k, but only 194 houses?

And they are nowhere near the plant with downtime.


But with a national grid and frequency being the same everywhere the substations
with the more sensitive cutouts don't have to be any place special at all. So
you'd almost expect random locations to trip?

Where the outage is marginal, that is plausible.
I can think of other possibilities, but they would
only be uninformed speculation.

 

[Tom Gardner]

On 10/08/19 16:56, Rick C wrote:

On Saturday, August 10, 2019 at 11:50:05 AM UTC-4, Tom Gardner wrote:
On 10/08/19 16:41, piglet wrote:

If so that may not be quite so stupid, although I can't see why the cutout had
to be set so close to the regulatory limit, most loads would still be perfectly
content at 47Hz for instance.

You have to cater for those that wouldn't be happy
and have been designed with the stated limits in mind.
But that is only hand waving.

I would speculate it is more about the concern that a significant aberration is happening and it is simply best to go offline rather than continue to draw power.

That's plausible for this event, and I will be interested to
see if it was the case.

 

[Tom Gardner]

On 10/08/19 16:58, piglet wrote:

On 10/08/2019 16:47, Tom Gardner wrote:
On 10/08/19 16:39, Rick C wrote:
On Saturday, August 10, 2019 at 11:03:30 AM UTC-4, Tom Gardner wrote:
On 10/08/19 15:52, Tom Gardner wrote:
On 10/08/19 13:52, Martin Brown wrote:
On 10/08/2019 08:05, Tom Gardner wrote:
On 04/08/19 15:41, Tom Gardner wrote:
On 04/08/19 15:01, Rick C wrote:
[snip]

Yup, the UK is stuffed.  Better get used to it.

Agreed, the UK energy industry is screwed up beyond belief.

Political/economic ideology is the core reason.

And it happened - a concrete demonstration of how perilously
close to the limit the UK power system is.

And in mid summer when you might expect things to be safe.

I hope this will finally get through to Rick C that EV
infrastructure isn't as simple as it is in his neighbourhood.
But I doubt it

Yesterday *large*[1] parts of the UK had power cut because two
plants went offline within two minutes - one gas plant, one
wind plant. Yup, wind power does suddenly stop.

[1] There blackouts across the Midlands, the South East,
South West, North West and north east of England, and
Wales.

It seems likely that the initial problem which was the first gas turbine
generator dropping out wasn't dealt with in the two minutes before a second
independent failure of a wind turbine farm. The two being offline together
then took the mains frequency sufficiently far out of bounds that some
inverters stopped as well. Cascade failure followed.

So it seems. Gridwatch shows down to 49Hz when the limit
is 49.5Hz.

I look forward to seeing more in the trade press,
e.g. the IET and comp.risks


Dinorwig can start in under 20s so someone wasn't paying attention...

Maybe it was otherwise occupied?

waves arms
Maybe the velocity of the cascade was both too fast for humans,
and too slow for the automated system on the other side of the
country to kick in.
/waves arms


The thing that they have failed to model adequately for grid management is
that conventional turbine generators have huge synchronised spinning rotors
that offer considerable inertia to changes in grid frequency.

I pretty sure they know that and have modelled it It may well
be that load shedding occurs to prevent excessive stress being
put on the remaining rotors.


By comparison semiconductor inverters seem to offer almost none and give up
completely if they find df/ft too large or f outside accepted bounds. I don't
see why this needs to be the case provided that the devices are protected
from
overload - a couple of percent frequency variation shouldn't take a decent
power transformer into saturation.

It was astonishing how widespread the UK power cuts were for a fault that was
concentrated in the SE near Cambridgeshire - they lost Newcastle Airport
nearly 300 miles away as a direct result.

Yes indeed.

I think it has got some politicians attention. Whether they will
continue to think about it is questionable, especially in the
face of the disaster of their own making.


NHS hospitals that beancounters had put on the load shedding cheapest tariff
found themselves without any power at all - and the odd one discovered their
emergency backup generators didn't work either.

Classic.


Lucky it happened at 5pm when most scheduled operations were over.

A few months ago my daughter was waiting in the anteroom all day
and night, and the operation finally happened at 7:30pm.


There appeared to be parts of the London underground that lacked emergency
lighting too (or if they had it then it failed too work).

It was also rather discouraging how few of the major systems recovered
gracefully from a sudden loss of power when power *was* restored after about
an hour. The trains are still a complete mess today.

That's unsurprising. You can't have a train from X; you have
to have a train to X followed by a train from X


All you have to do is look at the compass points to realise
how widespread that was.

That wide geographic spread is a bit weird - the protection system is
supposed
to limit the contagion and shed the right sorts of load first and nearby. We
were not cut off but others in major cities were.

https://www.bbc.co.uk/news/uk-49302996

Offgen is looking into it. Don't hold your breath...

Yes, curious minds do want to know.

Another aspect is becoming apparent. The homes affected appear
to be in a fractal patchwork, rather than across a whole area.

For example, I didn't notice anything, but 4 and 10 miles away:
"the company’s website shows 194 homes in Keynsham and 25 in
Nailsea are still [yesterday] without electricity."

Why so few homes, and there not elsewhere?

Can't say about the UK, but in the US areas are often provided with power
from more than one source by redundant paths.  If one fails, they draw more
power from another.  At times this can result in an overload of any single
circuit.  Not surprising failures were sporadic.  When you stress a system to
it's max you find weak spots first, where ever they are.


The UK is better than the US in that respect.

The problem wasn't the grid, it was insufficient generating
capacity.

I've previously mentioned that if two plants were to have
unscheduled downtime in winter, we would be in trouble - and
it isn't even winter

This is a concrete demonstration that I wasn't being alarmist.

Yes, very strange. Apparently demand was "only" 30GW which is way lower than
winter peak capacity - I guess a lot of generating capacity was off line for
summer maintenance works. Left too fragile for unexpected dual failure.

That would be my /guess/.

 

[piglet]

On 10/08/2019 16:50, Tom Gardner wrote:

On 10/08/19 16:41, piglet wrote:

Perhaps different switch centers have differing
under-frequency-lockout sensitivity and responsiveness delay?

Nailsea population 15k, but only 25 houses?
Keynsham population 16k, but only 194 houses?

And they are nowhere near the plant with downtime.

But with a national grid and frequency being the same everywhere the
substations with the more sensitive cutouts don't have to be any place
special at all. So you'd almost expect random locations to trip?

piglet

 

[Tom Gardner]

On 10/08/19 17:02, Rick C wrote:

On Saturday, August 10, 2019 at 11:58:18 AM UTC-4, Tom Gardner wrote:
On 10/08/19 16:27, Rick C wrote:
On Saturday, August 10, 2019 at 3:05:45 AM UTC-4, Tom Gardner wrote:

Yesterday *large*[1] parts of the UK had power cut because two
plants went offline within two minutes - one gas plant, one
wind plant. Yup, wind power does suddenly stop.

Yup, wind power stops suddenly, but not because the wind stopped blowing.

"Energy watchdog Ofgem has demanded an urgent report from National Grid into how the failure happened as parts of the UK were battered by strong winds and heavy rain."

“The first generator to disconnect was a gas fired plant at Little Barford at 16:58. Two minutes later Hornsea Offshore wind farm seems to have disconnected."

No one said the wind stopped blowing. lol

Actually, the wind /does/ stop blowing for *days* at a time.
All it needs is a blocking high pressure over the UK

And does that happen suddenly?

No. There several days warning, and it decreases slowly.

Nonetheless, there must be sufficient conventional plant
that is normally idle to take up the slack. The question is
who pays for the "normally idle" plant.

The green zealots say "don't add the costs to the cost of
green electricity.

The accountants say "only generating 2 days/year means the
electricity is uneconomic - shut the plant down". That's
just happened with another coal plant owned by RWE.

.... and the result is too little margin in the available
generation capacity.


> I think the power from a wind farm will vary much more slowly than a-ok to offline in two minutes.

I would be very interested if that wasn't the case


> I believe I read they are investigating if the first plant failure caused the wind farm to go offline.

They would be negligent if they weren't doing that.

 

[Bill Sloman]

On Sunday, August 11, 2019 at 2:03:40 AM UTC+10, upsid...@downunder.com wrote:

On Sat, 10 Aug 2019 13:52:03 +0100, Martin Brown
'''newspam'''@nezumi.demon.co.uk> wrote:

[1] There blackouts across the Midlands, the South East,
South West, North West and north east of England, and
Wales.

It seems likely that the initial problem which was the first gas turbine
generator dropping out wasn't dealt with in the two minutes before a
second independent failure of a wind turbine farm. The two being offline
together then took the mains frequency sufficiently far out of bounds
that some inverters stopped as well. Cascade failure followed.

It also may be that they had a bit bad luck.

Good network management requires that the N-1 rule is followed. The
network should survive the loss of _largest_ production unit. In
practice during the first minutes the spinning reserve of other power
plants are used (.i.e. overloading them). During that time quick
start emergency gas turbines are started, which then takes over from
the spinning reserve. When the emergency gas turbines are running, the
power of existing power plants are added e.g. by burning more coal to
get more steam or other slower starting power plants are starting.
When these are running, the emergency gas turbines can be shut down
and the N-1 criterion is restored, waiting for the loss of the next
largest production unit. These emergency gas turbines are typically
running for less than an hour.

However, as long as the emergency gas turbines are used, the network
doesn't tolerate the loss of the next largest unit.

When South Australia bought its big Telsa battery, with the fast inverter built in, it more or less immediately demonstrated the convience of having a really fast-acting voltage and frequency correcting device.

https://reneweconomy.com.au/how-the-tesla-big-battery-kept-the-lights-on-in-south-australia-20393/

Dinowig is fast, but not that fast.

<snip>

--
Bill Sloman, Sydney

 

On Sat, 10 Aug 2019 17:17:09 -0700 (PDT), Bill Sloman
<bill.sloman@ieee.org> wrote:

On Sunday, August 11, 2019 at 2:03:40 AM UTC+10, upsid...@downunder.com wrote:
On Sat, 10 Aug 2019 13:52:03 +0100, Martin Brown
'''newspam'''@nezumi.demon.co.uk> wrote:

[1] There blackouts across the Midlands, the South East,
South West, North West and north east of England, and
Wales.

It seems likely that the initial problem which was the first gas turbine
generator dropping out wasn't dealt with in the two minutes before a
second independent failure of a wind turbine farm. The two being offline
together then took the mains frequency sufficiently far out of bounds
that some inverters stopped as well. Cascade failure followed.

It also may be that they had a bit bad luck.

Good network management requires that the N-1 rule is followed. The
network should survive the loss of _largest_ production unit. In
practice during the first minutes the spinning reserve of other power
plants are used (.i.e. overloading them). During that time quick
start emergency gas turbines are started, which then takes over from
the spinning reserve. When the emergency gas turbines are running, the
power of existing power plants are added e.g. by burning more coal to
get more steam or other slower starting power plants are starting.
When these are running, the emergency gas turbines can be shut down
and the N-1 criterion is restored, waiting for the loss of the next
largest production unit. These emergency gas turbines are typically
running for less than an hour.

However, as long as the emergency gas turbines are used, the network
doesn't tolerate the loss of the next largest unit.

When South Australia bought its big Telsa battery, with the fast inverter built in, it more or less immediately demonstrated the convience of having a really fast-acting voltage and frequency correcting device.

https://reneweconomy.com.au/how-the-tesla-big-battery-kept-the-lights-on-in-south-australia-20393/

Dinowig is fast, but not that fast.

snip

Originally the idea with the N-1 rule was that after the failure of
the largest production unit, the remaining production units can be
slightly overloaded for a while until new production capacity has ben
dispatched (spinning reserve).

This assumption is true with _BIG_ synchronous generators with can be
overloaded by 5 - 10 % for a minute or two without overheating too
much. Thus if a 1000 MW unit is lost the total grid on-line capacity
needs to be 10-20 GW, each of which can be overloaded by 5 - 10 %,

After this minute or two, fast starting emergency gas turbines or
diesels must be running so that the big generators can be scaled down
to nominal capacity.

However, a lot of renewable sources are inverter connected. These
inverters have no overload capacity or at most a few second capability
before overheating.

About a decade ago there was a huge wind energy boom in Germany due to
ample subsidies, so that every farmer wanted a wind turbine on their
field, no matter how bad the wind conditions are. These could also
sell all production to the net, this running at 100 % load as long as
there is wind.

In a fault condition. when all remaining production were asked to
provide the spinning reserve, i.e overloading of generators. Since
many wind turbines had practically no overload capability, some were
tripped more or less immediately before the emergency gas turbines had
been started. With loss of some wind turbine production the remaining
capacity was asked to produce more overload power. More and more wind
turbines overloaded and tripped. The wind turbine tripping rippled
through the network.

There are at least two solutions to this problem.

1.) Allow inverter based system to generate only 90 % of rated power
on a day to day bases, thus when asked to provide the spinning
reserve equivalence, the output can be increased to 100 %, which
doesn't cause overload tripping. Greens would cry loudly foul, when
they are not able to sell all their production at all time.

2.) Provide a battery backup system as in Fairbanks Alaska or South
Australia. The combined output power from these units should be as
big as the largest unit in the network, say 1000 MW. The batteries
should be able to provide that power for a few minutes (replacing
spinning reserve) before emergency gas turbines have started. If the
combined battery is sufficiently large, say 1000 MWh, it could also
replace the emergency gas turbines.

Anyway the cost of either of these methods should be included in the
cost of renewable sources.