Posted by Don Kelly on April 9, 2009, 3:44 am
Tim Jackson wrote:
> daestrom wrote:
>>
>>> The grid collectively controls the frequency, which controls the
>>> speed of all the generators, each individual generator controls its
>>> output current by adjusting the input power. When any generator or
>>> load changes its current, lest say a load increase, first the grid
>>> voltages fall slightly which increases the mechanical load presented
>>> by each of the generators, so they all slow down a bit, which in
>>> turn signals their local control systems to increase their power
>>> input, if possible.
>>>
>>
>> Except that the majority of the generators on the grid are not
>> 'regulating' units. Their governors are used only during starting
>> and synchronizing to the grid. Then they are run up out of the way
>> so that they have no affect on the steam-valves, fuel-line,
>> wicket-gate, whatever.
>>
>> For example, base-load coal/nuc plants typically control the amount
>> of load they carry either manually or a remote signal from the system
>> operator. If the grid frequency drops a tenth of a cycle or rises,
>> the load the unit carries will remain the same. This is because the
>> speed change will not impact the steam-valve / boiler controls. But
>> if the system operator *asks* for more load, then either the local
>> operator or the automated system will ramp up fuel flow and open the
>> steam valves accordingly.
>>
>> Only a fraction of the plants under a system operator's control are
>> set for regulation. They will automatically adjust the amount of
>> load they carry based on the frequency of the grid. These are the
>> only units that will respond to a grid load change. The system
>> operator will monitor how much load the regulating units are carrying
>> and if they reach one end or the other of their range, he will call
>> up a base load unit and have them pick up/shed load in order that the
>> grid frequency will change enough that the regulating units move back
>> towards their mid-range point.
>>
>> Similarly, many remote-operated peaking units are of the 'fixed load'
>> type. The operator makes a call or even just pushes a button and a
>> gas turbine unit or even diesel will startup, automatically synch to
>> the grid and load to a preset amount. Push the button and voila,
>> another 4 MW of generation is on-line. Push another button and it
>> unloads, disconnects, cools down and shuts down back to standby.
>> Expensive MW, but cheap to install.
>>
>> But the regulating unit governors don't have to be extremely fast.
>> Because *all* the motors and units are tied together, a sudden change
>> in load is largely made up for by the huge amount of inertia of all
>> the machines. To accelerate several hundred generating units and
>> thousands of AC motors takes quite a lot of load change. The
>> regulating unit governors have a few seconds to respond before grid
>> frequency changes enough to be trouble.
>>
>>> Bringing two grids together to enable a power transfer is harder
>>> because you've got to move the whole loaded grid into phase by
>>> fiddling the target frequencies and get it to stay there long enough
>>> to lock it up. Then two *target* frequencies have to be moved apart
>>> until the desired flow is achieved, one grid "pushing" the other.
>>> This does not mean the actual frequencies move apart, obviously
>>> that's not physically possible, it means that any given actual
>>> frequency produces a stronger signal to the exporting grid to
>>> generate power than it does the importing one.
>>>
>>
>> Indeed. System operators tried to tie the east and west US grids
>> together a few times in the later part of the 20th century (I forget
>> what decade). Even when you get them tied together, the tiniest
>> disturbance on one side or the other would cause *large* amounts of
>> power to flow across the tie and it would quickly re-open. Like
>> connecting two elephants together in a parade with a tiny thread. As
>> long as nobody changes how they elephants are moving, the thread just
>> hangs there. But if one gets the idea to turn left, that thread
>> parts instantly.
>>
>> And that's one of the beauties of HVDC. The two grids don't have to
>> be synchronized to get power transfer.
>>
>> daestrom
>>
> OK, now we've established that we agree on how the grid works, maybe
> we can apply that to the issue of "dynamic demand". Obviously the
> intent is that by dropping loads on peaks of demand, we can avoid
> having to bring on those expensive pushbutton megawatts.
> But on the other hand it also has issues of regulation. The more units
> you have trying to control the grid, the harder it is to predict the
> stability of the control system. The effective "gain" of load
> shedding appliances would be unpredictable and very hard to legislate
> for. So I suspect that if you put a lot of these things on line, you'd
> end up with an unstable network. That's why I was equivocal about the
> idea in the first place.
> Same sort of problems happen if you have too big a % of wind power,
> worse really because that itself varies unpredictably, so designing a
> base-load system to work with it is not for the faint hearted.
> I wonder if similar considerations are why our UK power companies are
> reluctant to allow private tie-in cogeneration. They say the grid
> can't cope with it. On a local scale I can't see any power
> transmission issues as long as it remains a minority and there are
> some rules about voltage regulation, (ie you can't export if your line
> voltage is at the high end - indicating a lack of local demand) but I
> do wonder if they are concerned about potential instabilities of the
> system.
> Tim Jackson
Instability, in the sense of the common usage of the term power system
instability" would not be a problem. It is the sudden large changes in
transmission capability that are problematic- and these are due to
faults on the system- not load changes but splitting the system so that
some areas have excess load and others have a deficiency of load and
inadequate transfer capability to correct this. With a healthy
system, even if everybody shuts off their water heaters at 5PM, there
should be no problem-as there is sufficient transfer between units to
keep them in synchronism-i.e. stable. Frequency will drop or rise
until governors and other regulatory mechanisms take over and restore
frequency and a desirable load distribution. The "two elephant case"
that Daestrom mentioned was a particularly weak tie (insufficient
transfer to keep the two systems together but weak enough to not cause
instability in either system- the string broke without being noticed by
either elephant- and was known to be inadequate -at that it held for
about 30 minutes if I recall correctly.
The tie in generation is a problem for two reasons- one is a safety
factor but that is not an issue with proper equipment and the other may
be financial- the utility sees that it has to subsidize the private
producer in many situations- they don't want to pay for the transmission
of someone else's energy nor do they want to pay peak or even average
rates for generation that occurs off- peak (on peak the person with the
generator may have little or nothing to spare). Most utilities have
come to some fair compromise with regard to this.
As for wind power, you will always have to have hot or spinning capacity
on line, ready to take over. This can take over quite quickly- again due
to the coordination of the droops of the governors on machines on line.
A system that can cope with the sudden loss of a 500MVA machine can cope
with the slower loss of several 10MW wind units without blinking.
Note also that on the recent- turn off the lights for an hour day, some
regions had load drops of 5 to 10% and coped very nicely.
--
Don Kelly
dhky@shawcross.ca
remove X to reply
Posted by daestrom on April 10, 2009, 12:31 am
<snip>
> The tie in generation is a problem for two reasons- one is a safety factor
> but that is not an issue with proper equipment and the other may be
> financial- the utility sees that it has to subsidize the private producer
> in many situations- they don't want to pay for the transmission of someone
> else's energy nor do they want to pay peak or even average rates for
> generation that occurs off- peak (on peak the person with the generator
> may have little or nothing to spare). Most utilities have come to some
> fair compromise with regard to this.
With somewhat larger co-gen installations around here, another issue is
re-calculating the available fault currents. Putting a small unit of just a
MW or so on the far end of a line *could* require that customer protection
equipment be upgraded.
The extreme case would be if you put in a co-gen and now your neighbor's
service entrance fault current increases to the point where the typical
service entrance equipment can't interrupt it. He needs to upgrade his
service for his own protection, but it's because of the co-gen you
installed. Who's to pay?
daestrom
Posted by Don Kelly on April 10, 2009, 1:27 am
daestrom wrote:
> <snip>
>>
>> The tie in generation is a problem for two reasons- one is a safety
>> factor but that is not an issue with proper equipment and the other
>> may be financial- the utility sees that it has to subsidize the
>> private producer in many situations- they don't want to pay for the
>> transmission of someone else's energy nor do they want to pay peak or
>> even average rates for generation that occurs off- peak (on peak the
>> person with the generator may have little or nothing to spare). Most
>> utilities have come to some fair compromise with regard to this.
>>
> With somewhat larger co-gen installations around here, another issue
> is re-calculating the available fault currents. Putting a small unit
> of just a MW or so on the far end of a line *could* require that
> customer protection equipment be upgraded.
> The extreme case would be if you put in a co-gen and now your
> neighbor's service entrance fault current increases to the point where
> the typical service entrance equipment can't interrupt it. He needs
> to upgrade his service for his own protection, but it's because of the
> co-gen you installed. Who's to pay?
> daestrom
Unless the neighbor's load is large , I would assume that the capability
of his protection would be based on an "infinite bus" behind the
transformer supplying him. In that case there would be no requirement.
In cases of industrial co-generation (10MW is in that category) in an
industrial area, that *may* not apply but in that case it would be the
utility side protection that has to be enhanced.
In the "pa and ma" low level co-generation- where the fuss arises- this
would not be a concern unless the co-generation source is on the same
side of the transformer as the neighbor.
--
Don Kelly
dhky@shawcross.ca
remove X to reply
Posted by Alistair Gunn on April 5, 2009, 6:52 am
In alt.energy.renewable Eeyore twisted the electrons to say:
> Tim Jackson wrote:
> > synchronous electric clocks stay in time, In the short term the
> > frequency drifts by anything up to a Hz or so.
> Utter nonsense. Except in 3rd world countries maybe.
Clearly the National Grid thinks that the UK is a third world country
then?
"System frequency will therefore vary around the 50 Hz target and
National Grid has statutory obligations to maintain the frequency within
+/- 0.5Hz around this level. However, National Grid normally operates
within more stringent 'operational limits' which are set at +/- 0.2Hz."
--
These opinions might not even be mine ...
Let alone connected with my employer ...
Posted by Ken on April 5, 2009, 7:07 am
wrote:
>> The mains frequency is very rock solid. It HAS to be for the gris to work.
>> You're talking crap.
>> Graham
>
>
> The instantaneous mains frequency isn't rock solid. The control system
> keeps the long term *average* mains frequency rock solid so that
> synchronous electric clocks stay in time, In the short term the
> frequency drifts by anything up to a Hz or so.
>
> The main reason is the time it takes mechanical generators to respond to
> changes in overall load.
>
> See, for example,
> http://en.wikipedia.org/wiki/Mains_frequency
Very true. When the load is heavy, the frequency goes down.
Here in Sweden the variation normally is 49.85 Hz to 50.15 Hz.
Now in the morning it's around 49.87 Hz.
XML site map
>>
>>> The grid collectively controls the frequency, which controls the
>>> speed of all the generators, each individual generator controls its
>>> output current by adjusting the input power. When any generator or
>>> load changes its current, lest say a load increase, first the grid
>>> voltages fall slightly which increases the mechanical load presented
>>> by each of the generators, so they all slow down a bit, which in
>>> turn signals their local control systems to increase their power
>>> input, if possible.
>>>
>>
>> Except that the majority of the generators on the grid are not
>> 'regulating' units. Their governors are used only during starting
>> and synchronizing to the grid. Then they are run up out of the way
>> so that they have no affect on the steam-valves, fuel-line,
>> wicket-gate, whatever.
>>
>> For example, base-load coal/nuc plants typically control the amount
>> of load they carry either manually or a remote signal from the system
>> operator. If the grid frequency drops a tenth of a cycle or rises,
>> the load the unit carries will remain the same. This is because the
>> speed change will not impact the steam-valve / boiler controls. But
>> if the system operator *asks* for more load, then either the local
>> operator or the automated system will ramp up fuel flow and open the
>> steam valves accordingly.
>>
>> Only a fraction of the plants under a system operator's control are
>> set for regulation. They will automatically adjust the amount of
>> load they carry based on the frequency of the grid. These are the
>> only units that will respond to a grid load change. The system
>> operator will monitor how much load the regulating units are carrying
>> and if they reach one end or the other of their range, he will call
>> up a base load unit and have them pick up/shed load in order that the
>> grid frequency will change enough that the regulating units move back
>> towards their mid-range point.
>>
>> Similarly, many remote-operated peaking units are of the 'fixed load'
>> type. The operator makes a call or even just pushes a button and a
>> gas turbine unit or even diesel will startup, automatically synch to
>> the grid and load to a preset amount. Push the button and voila,
>> another 4 MW of generation is on-line. Push another button and it
>> unloads, disconnects, cools down and shuts down back to standby.
>> Expensive MW, but cheap to install.
>>
>> But the regulating unit governors don't have to be extremely fast.
>> Because *all* the motors and units are tied together, a sudden change
>> in load is largely made up for by the huge amount of inertia of all
>> the machines. To accelerate several hundred generating units and
>> thousands of AC motors takes quite a lot of load change. The
>> regulating unit governors have a few seconds to respond before grid
>> frequency changes enough to be trouble.
>>
>>> Bringing two grids together to enable a power transfer is harder
>>> because you've got to move the whole loaded grid into phase by
>>> fiddling the target frequencies and get it to stay there long enough
>>> to lock it up. Then two *target* frequencies have to be moved apart
>>> until the desired flow is achieved, one grid "pushing" the other.
>>> This does not mean the actual frequencies move apart, obviously
>>> that's not physically possible, it means that any given actual
>>> frequency produces a stronger signal to the exporting grid to
>>> generate power than it does the importing one.
>>>
>>
>> Indeed. System operators tried to tie the east and west US grids
>> together a few times in the later part of the 20th century (I forget
>> what decade). Even when you get them tied together, the tiniest
>> disturbance on one side or the other would cause *large* amounts of
>> power to flow across the tie and it would quickly re-open. Like
>> connecting two elephants together in a parade with a tiny thread. As
>> long as nobody changes how they elephants are moving, the thread just
>> hangs there. But if one gets the idea to turn left, that thread
>> parts instantly.
>>
>> And that's one of the beauties of HVDC. The two grids don't have to
>> be synchronized to get power transfer.
>>
>> daestrom
>>
> OK, now we've established that we agree on how the grid works, maybe
> we can apply that to the issue of "dynamic demand". Obviously the
> intent is that by dropping loads on peaks of demand, we can avoid
> having to bring on those expensive pushbutton megawatts.
> But on the other hand it also has issues of regulation. The more units
> you have trying to control the grid, the harder it is to predict the
> stability of the control system. The effective "gain" of load
> shedding appliances would be unpredictable and very hard to legislate
> for. So I suspect that if you put a lot of these things on line, you'd
> end up with an unstable network. That's why I was equivocal about the
> idea in the first place.
> Same sort of problems happen if you have too big a % of wind power,
> worse really because that itself varies unpredictably, so designing a
> base-load system to work with it is not for the faint hearted.
> I wonder if similar considerations are why our UK power companies are
> reluctant to allow private tie-in cogeneration. They say the grid
> can't cope with it. On a local scale I can't see any power
> transmission issues as long as it remains a minority and there are
> some rules about voltage regulation, (ie you can't export if your line
> voltage is at the high end - indicating a lack of local demand) but I
> do wonder if they are concerned about potential instabilities of the
> system.
> Tim Jackson