Frequency, Load Demand Control

[powertrainee]

I have an interest of looking into demand-side management, specifically, the control of specific loads dependent on grid frequency. I am looking for resources on possibily statistical analysis done on grid frequency data and algorithms which initiate "triggers" the control loads.

From what I have been able to search, there is very little out there. I am hoping some of you can help me in finding some papers, books or other articles.

Any help would be greatly appreciated!

 

[waross]

Only very very small grids have any appreciable frequency excursions. Most grids have a rock solid frequency.
I did see an interesting hydro generation installation.
1> The size of the installation was chosen as 50 KW on the basis that the government department funding the project required a full engineering study on installations larger than 50 KW. The actual load was several times that.
2> The frequency was to be controlled by adding load. There was a large water tank with about 50 KW of electric immersion heaters installed something like an inside out porcupine. As the frequency rose above 60 Hz the heaters were to be turned on in steps.
They also got government funding to build a laundromat to run off the free hot water that the control scheme would produce.
3> The plant was so undersized that the control heaters never turned on. The laundromat didn't get any hot water
4> They brought in some diesel generators to carry the load.
5> The seasonal creek that they had chosen without making local inquiries ran dry.
My tax dollars at work.
Is this what you had in mind?

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[powertrainee]

In a way, but not to that extent

I am looking at it more as a way to "remedy" the incorporation of intermittent renewable generation. The two obvious control schemes (or triggers) that come to mind are absolute frequency (greater than or less than 60Hz) and derivative frequency (increasing and decreasing trends), both used to indicate that load should be controlled in some way. Hopefully looking to find some papers/books that went into some analysis in this topic.

Again, all help is appreciated!

 

[wayne440]

As a government employee, I completely appreciate that 50kw installation. I'd love to hear the specifics of that story.

 

[byrdj]

I would look at NERC requirements

 

[powertrainee]

Is there a database online of frequency readings (hopefully somewhere in the Eastern interconnection) over let's say, a monthly (or longer) timeline?

 

[waross]

A sample taken at one second intervals may look like this:
60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-61,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-59,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-61,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-59,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-61,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-59,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-61,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-60,-59,-60,-60.
You may get the occasional 59 Hz that will probably be offset by a 61Hz correction. Depending on the design of the trigger of the counter that is monitoring the frequency, A phase shift of a few degrees may cause the counter to respond to 59.9 Cycles in a second as 59.0 cycles in that particular second.
That notwithstanding, frequency excursions may occur.
An example:
A large mine-mill was built at the north end of a large island. The island was fed by a combination of hydro-electric plants and a High Voltage DC submarine cable.
The utility was concerned that if the capacity of the DC cable was lost, the mine load may drag down the frequency of the whole island.
To that end, the required the mine to install an under frequency relay to trip the main power supply to the mine in the event that the grid frequency dropped.
The mine also installed their own under frequency relay to trip their main power.
In the event that the frequency dropped, if the mine tripped themselves off and by so doing, allowed the grid to recover, they could go back into curtailled production after a phone call to the utility load dispatch center to determine how much capacity was available, if any.
If their relay failed to trip their power or if even after the trip, the system frequency continued to decay, the utility relay would trip. The procedure for restoring power after the utility relay tripped was much more onerous and time consuming. As I remember the relay settings were either 56Hz and 58Hz or 58Hz and 59Hz. I don't know if any of the relays ever did trip, other than for power failures.

An instance of a utility frequency that did and does vary. The little island utility that I used to be involved with.
This was a diesel powered plant with a total load that peaked at about 1.2 MW. The normal frequency range was a 3% drop from no load to full load. However, with operator intervention every 15 minutes, The frequency was usually within 1/2 Hz. With a droop of 3%, a block load of 20% of the set capacity would result in a frequency drop of 20% 0f 3% or 0.6% That translates to a frequency drop of 0.36Hz.
Compared to the North American grids frequency excursions to 59.64 Hz are pretty sloppy.


Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[wayne440]

There is a program on the right top of this page

http://gridwise.pnl.gov/technologies/transactive_controls.stm

that displays one plant's frequency in near real time.

I have no association with said page, etc.

 

[Compositepro]

My understanding is that the grid is controlled to have the same number of cycles every day. This keeps clocks that do cycle counting very accurate (digital or synchronous motor).

 

[byrdj]

again, look at NERC documentation and the studies they are doing

http://www.nerc.com/

The utility generation control systems I worked on respond propertianly to system frequency deviations, however there is some inherent deadband with the mechanism that change the load. while NERC seems to be pushing utilities for increased propertianly response and tighter deadbands, I was wondering what the effect of these new generation schemes would have.

another soure of information would be the Texas Grid

http://www.ercot.com/

I'm not as fimialr with what they are doing with the daily control of system frequency, but I recall these stations have posted procedures for frequency dictated load shedding

As for the grid being 60.00 hertzs most of the time, I have never seen the grid that stabile for any period of time. I judge the quaility of the mechanical governors and valve actuators based upon how much they are constainly correcting for the microHertz deviations

 

[mgtrp]

If you're interested in utility under-frequency load-shedding schemes, some information can be found in summary form here:

http://www.systemoperator.co.nz/f3210,36010779/AUFLS_Summary_Card.pdf

Or the full report:

http://www.systemoperator.co.nz/f3210,36010943/AUFLS_Technical_Report.pdf

For what it's worth, I've never really seen the NZ grid frequency out by more than about +/-0.02 Hz, although this is just me glancing at meters in power stations/substations/system control centres/etc rather than anything particularly analytical. I would expect various grids in North America to be even more stable, given how much larger they likely are. The grid(s) that I work with now might see deviation by as much as +/-0.3 Hz on a "system normal" situation, and much more during system events (for those of you involved with protection of more important systems: you'd be surprised at how many relays have issues caused by off-nominal or rapidly changing system frequencies!).

I would expect that, other than on an isolated grid, you won't get sufficient difference in frequency to use this as a trigger for a straight DSM system that is just working to shave peak loads. You might be able to create some form of emergency load-shedding scheme, but I suspect that the cost of this may be prohibitive, and utilities are unlikely to want to hand control of such a critical system over to the general public. I could see making changes to how the scheme operates being somewhat of a nightmare as well.

What I have seen used effectively are ripple control systems (or similar), which allow the utility to switch off certain loads (e.g., residential hot water heaters) during peak periods by briefly adding high frequency signals over the fundamental system frequency to turn on or off the loads.

The utility that I work for now used to operate a scheme whereby certain customers were able to receive reduced rates on electricity in exchange for our SCADA operator being able to interrupt these loads if we had a shortfall of renewable energy. These loads were typically heaters, with the customers having some form of fossil fuel-based backup. This scheme finished up in the past year or so, as we no longer have an excess of hydro generation available.

 

[ScottyUK]

I've seen a couple of very severe swings on the UK grid, one as a result of a very large station tripping off the system (the one near Selby ) which took the system down to about 48Hz, and one where a main interconnector opened and the north of the country with most of the generation almost separated from the south. The latter must have been interesting for grid control, the former was certainly interesting for us because the gas turbines went into system reserve operation to try to pull the frequency back up, burning up hot parts life at about 20x normal rate.

One thing to remember is that a base-loaded station can only deload to control system frequency, unless it is capable of over-firing the prime mover to exceed baseload rating.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[cswilson]

Compositepro -- This is not really relevant to the OP, but back in the 1960s, my father made some inquiries as to how accurate the accumulated time value from counting cycles was. What he found out is that there could be a noticeable deviation across a day, but that every night in the middle of the night, the grid operators would deliberately vary the frequency to compensate for the accumulated deviation. So there would be no long-term (multi-day) deviations, but depending on the accuracy required, you could not necessarily count on extremely high accuracy during a day.

This was for the eastern US grid, and I believe the ultimate grid control was based near Niagara Falls.

 

[PeterWedo]

Think they missed your point. If your talking about providing Ancillary Service (demand load shedding) to the power grid during an event e.g. large generator fault. Then look at dynamicdemand web site. Note I been looking at this for a while and there very little out there. Some trials in US with air-con and heaters.