Islanded Generating Station 6

[Mbrooke]

How does an Islanded 600-2000MW generating station behave? What does typical critical clearing time look like? What is the voltage and frequency divergence during and after a fault for various fault scenarios? How is active and reactive power dispatched/controlled? How does generation behave during trip and reclose events? And how do you "block load"?

 

[waross]

I should add that when the island becomes islanded, the plant may easily control one set as a swing set for closer frequency control.
It may need only a switch closure and a set point adjustment to convert an islanded set to a swing set.

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

 

[Mbrooke]

Question. What is a droop curve? And why can't all the gens be swing sets?

 

[waross]

Droop:> Off-set equal to the proportional band.
Thus a 5% droop is a 5% proportional band with a 5% offset.
A 60 Hz governor with 5% proportional band will drop 5% under load.
Thus 60 Hz at no load and 57 Hz at full load.
Now add 5% off-set or 3 Hz offset and the set will run at 63 Hz at no load and 60 Hz at full load.
In a grid, the swing set is a PI (Proportional plus Integer) control.
If the load increases the frequency will drop slightly and the swing set will increase its output to correct the frequency back to 60 Hz.
If the load on the swing set is dropping load dispatch will instruct other operators to reduce their output.
This can be done in two ways.
1. Take sets off-line.
2. Reduce the base loading of the sets.
If the load is growing on the swing set, load dispatch will instruct other operators to increase their out-put
Base loading. The load that a set accepts may be varied from zero to full load by setting the droop governor at 60 Hz for zero load up to 63 Hz for full load.
Set the governor to 61.5 Hz for 50% load.
If the swing set is lost, or if part of the system is islanded, the droop sets will work well together in droop, but with slightly greater frequency deviations.
The frequency deviations will generally not be anywhere near the full 5%. The frequency change will be proportional to the percentage change in the loading when a part of the system is islanded.
The initial frequency devition will be even less if the swing set is lost but nothing is islanded.
Did I explain this properly CR?
Only one set may be run in PI mode.
If there is more than one set in PI mode (swing set) they will fight each other. It has been a while since I considered this but as I remember, one or the other swing set will hog the load and the other will be unloaded until the first swing set is swamped.
It is not good to run a large generating set at no load.


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

 

[Mbrooke]

Thats a lot to digest. Thanks! Can the governor be manual? What if you had no swing set- just "sped" or "slowed all 10 gens to match the load as it changed? Do you black start the system by turning on one gen set- load it- then start and sync the others- the pick up more load?

 

[waross]

What if you had no swing set- just "sped" or "slowed all 10 gens to match the load as it changed?

Droop control is a basic feature of a diesel governor.
With all parallel sets at the same droop setting the sets will share the load in proportion to their capacity.
Droop control nicely limits the over speed and allows service to continue at no more than a 5% increase of frequency in the event that the swing set drops offline.
Other feature such as integral or reset will be added on top of droop control.
For example for the swing set. The first response to a load change will be a droop response.
This will cause a deviation from 60 Hz.
The integral or reset function will detect the error from the 60 Hz set point and bias the set point to bring the frequency back to 60 Hz.
On a large system the load typically does not change that fast.
When a load change is considered as a percentage of the total load and over the time constant of the integral function, the actual frequency deviations are generally much less than the 5% droop.
In the event that a generator was fully loaded, and lost all of the external load and the swing set connection, the frequency of the station service transformer would rise to 63 Hz.
This is a very worst case and is still acceptable.
Our very small plant, total 2.2 MW, had five diesels. They ran in droop with no swing set.
While the governors responded to load changes by following the droop curve, the operators manually trimmed the governor set points to keep the frequency at 60 Hz. It was rare that the frequency would be off by more than 1% or 2%.
Block loading.
We had two out-going circuits.
Our system could not take block loading of a complete circuit.
Our system went online by manually closing in each circuit one phase at a time.
This was a special case.
A larger system would be expected to have enough outgoing circuits that circuits could be switched in sequentially so that the block loading was within the capability of the generator(s).
Most systems of this size or smaller will be privately owned and the load will be under the control of the owners.
Consider an industrial plant on a self owned, islanded set.
Typically, after an outage, the large motors will be offline and will be restarted sequentially.
Our worst block loading was residential circuits.
As well as the transformer inrush, almost all of the refrigerators, freezers and A/C units would be drawing starting surges.

Extra reading.
PID controllers.
Droop control is the P of PID.
The swing set uses the PI of PID.
When you get your head around basic Proportional control the addition of integral will come easier.
A hint: Set loading. Consider a set in parallel with an infinite grid. The grid controls the frequency.
With 5% droop and the governor controlling at 60 Hz, the set will be unloaded.
As the control point is advanced, the set tries harder to raise the infinite grid frequency by opening the throttle and picking up more load.
When the governor is controlling at 105% or 63 Hz, the throttle will be 100% open and the set will be at full load but the frequency will be at the frequency of the infinite grid.
The error correction of the swing set helps the grid to appear infinite. (After the short correction period.)

To put a set online, the frequency is slightly above the grid frequency (In the order of 60.1 Hz to 60.15 Hz.)and the set synchronized and closed in.
The set will be supporting just enough load to prevent a reverse current trip.
The governor setting is then increased until the set picks up the required load.


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

 

[Mbrooke]

My bad. I'm thinking steam. 10 250,000kw gensets in parallel. Diesel sets stop at around 3,250kw if I have it right.

Think in the old days where people manually controlled steam generators. A coal or nuclear plant. Hydro could also be used in this example.


BTW- great reply!


Black loading would be about 200-300MW- perhaps 600MW with one circuit out- in the system I have in mind.

 

[waross]

Think in the old days where people manually controlled steam generators.

How old would that be?

Centrifugal governors were invented by Christiaan Huygens and used to regulate the distance and pressure between millstones in windmills in the 17th century.[1][2] In 1788, James Watt adapted one to control his steam engine where it regulates the admission of steam into the cylinder

Running flat out!
Balls to the walls!
Not horses and not smut.
If a generator lost its load and went over speed, The governor flyballs would rise up closer to horizontal and would describe a larger arc.
This condition was often described as flat out or balls to the walls.
Droop control of steam engines has been around for a lot longer than generators.
Governors are generic devices that are applied to much more than generation.

Terminology: Old control theory texts used the term "Reset".
Newer texts refer to "Integral".
When reset or integral control is enabled on a generator is is referred to as isochronous mode.

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

 

[Mbrooke]

Before computers. 3 mile Island was all relay logic if I'm being told right. Still then, in the 50s there were no semiconductors. Everything was EM. Power plants used resistors for the excitation.

 

[crshears]

Hey all, back from eight days off but under an altered schedule whereby I'm now part of a "bio-group" or cohort that always works together, with absolutely no interchanging of staff happening. Dayshift operations are from the main control centre, nights from the back-up, allowing thorough / "deep" cleaning [ whatever that means ] of each of the control centres while they are unoccupied; no other staff whatever are being permitted site entry.

Did I explain this properly CR?

Explained very well indeed, Bill.

Before composing the following, I searched through my previous posts to see if I had already expanded on this topic before, but I couldn't find where I had, so here goes.

I know that on our grid there are multiple generating station that can operate as the "swing" set; all on-line units at that location are loaded and/or unloaded together via AGC [Automatic Generation Control], and loading of the rest of the maneuverable generation in the province is directed by the IESO so as to maintain the proper load/generation balance while respecting market objectives, system element loading constraints, et cetera et cetera.

One other wrinkle is added to this when an otherwise separate grid becomes part of an interconnection, viz., the vectorial total of all of that grid's interconnecting circuits is summed and is factored in to the operation of AGC; Tie Line Bias and Area Control Error [ ACE ] are two of the terms often encountered in this regard.

In a nutshell, if AGC detects a decline in interconnection frequency combined with a net outflow from our grid to the rest of the interconnection, the swing plant or plants will contribute to the frequency correction to a lesser degree, as the fact that the net outflow has increased from schedule indicates that the frequency decline is due to a generation deficiency OUTSIDE of our grid, which other entities must address.

A frequency decline combined with a net reduction of outflow on the other hand [ we are almost always in net export ] indicates a generation deficiency WITHIN our grid, to which AGC will respond to a much larger degree so as to both correct the frequency decline and restore tie line flows to schedule.


CR

"As iron sharpens iron, so one person sharpens another." [Proverbs 27:17, NIV]

 

[crshears]

Before computers. 3 mile Island was all relay logic if I'm being told right. Still then, in the 50s there were no semiconductors. Everything was EM. Power plants used resistors for the excitation.

"Resistors for the excitation"...

Except in the smallest units, this generally caused far too much power loss, with the energy wasted as heat.

"In the 50s there were no semiconductors"...

True; but rotary AVRs such as Amplidynes or Rototrols were generally employed in this type of service. Magnetic Amplifiers also enjoyed a brief time in the sun, and although one rarely hears of the latter nowadays, I strongly suspect many of them remain in service around the world.

I have no knowledge of whether thermionic tube amplifiers were ever employed in excitation control systems...

CR

"As iron sharpens iron, so one person sharpens another." [Proverbs 27:17, NIV]

 

[Hoxton]

Brown Boveri made the resistor type of automatic voltage regulator. I'm trying to remember the name. Was it a ‘pile’ type or ‘moving element’?

https://www.eeeguide.com/brown-boveri-regulator/

I assume that the response was better than an amplidyne.

I do not know of an AVR incorporating thermionic valves, but we came across a synchroniser based on valves (tubes) some years ago. We were replacing an existing panel and had the schematics of the existing equipment. We were intrigued by reference to ‘warm up time’ on the synchroniser. Investigation revealed it had valves, which were only energised when synchronising was required. A timer delayed synchronising until the valves were warmed up!

I worked on a large island near India with a hydro based generation system. They ran most of the hydro units at full load, keeping one set on variable load as frequency controller. When that was at full load, they would start another one.

 

[Mbrooke]

Have power plants always been automatic excitation or has excitation ever been adjusted manually?


What about large sets in the 20-60s- what did they use besides resistors if they weren't present? How great are the losses in resistive excitation systems? What about DC generators for the excitation?

 

[crshears]

This generating station

https://www.opg.com/powering-ontario/our-generation/hydro/

https://www.google.com/maps/@44.186...4!1sGRMDUMR7yeqrwLaz6Wspvw!2e0!7i13312!8i6656

has four units, 1 MVA each, IIRC, and two separate excitation sources; one is an MG set and the other a water-turbine-driven D C generator. There is one common excitation supply bus, and the field strengths of the individual units are controlled by large resistors. If memory serves, re-starting of this plant requires in-person attendance and local operation of all devices, remote operation of the wicket gates [ for water control purposes and local trash rack cleaning ] being the only remotely controllable element.

To the best of my recollection this size and vintage of plant would be the only sort where such primitive excitation control would be tolerated; anything in the > 10 MVA total plant output range would have to be just a tad more sophisticated.

CR

"As iron sharpens iron, so one person sharpens another." [Proverbs 27:17, NIV]

 

[Mbrooke]

anything in the > 10 MVA total plant output range would have to be just a tad more sophisticated

Maybe I'm wrong, but I've seen prints to gens in the 125MW range with resistors. Technically- if it works for 1MVA, it should work for 250MWs.


I like the green backup generator BTW.

 

[crshears]

You mean the turbine exciter? My understanding was that it was always the preferred source, with the MG set serving as the backup since it would consume part of the station's production.

CR

"As iron sharpens iron, so one person sharpens another." [Proverbs 27:17, NIV]

 

[Mbrooke]

For the Excitation. I'm curious how they did it in the old days without AVR and automatic governors assuming thats how they did it.

 

[crshears]

Oddly enough, back in the day even the smallest units out there generally incorporated governors, reason being that they were often the only supply to varying local loads, and later on ran as part of what were essentially micro-grids [mini-grids?] with < 10 MW of total connectable generation, rendering governors indispensable. It was only years later that grid capacities measured in gigawatts turned these previous mainstay units into mere fleas on the camel's back, with the predictable result that rebuilding or even maintaining these governors was no longer economically justifiable.

The inevitable fallout of this reality was that these beautiful cast governors with their brass accessories, as beautiful and elegant as they were functional, were unceremoniously ripped out and replaced with just plain ugly gate actuators, built up out of plate steel, to which were affixed gear boxes, a normal duty AC drive motor and an emergency gate closing actuator consisting of a large DC truck engine starting motor [ often whatever was cheap and available ], and a truck battery or two with a trickle charger.

Depending on the load supplied, AVRs weren't necessarily fitted because the station was attended 24/7, and it would be the operator's duty to tweak the settings of the field resistors in real time as the load changes dictated.



CR

"As iron sharpens iron, so one person sharpens another." [Proverbs 27:17, NIV]

 

[Mbrooke]

That makes sense now, I was wondering about that.

What are the governor like in coal plants? I've heard this debate about running boilers with the steam valves fully open. What is that about?

 

[crshears]

The older steam plants I know of, whether fossil fuelled or nuclear, all employed a hydraulic governor system; the flyweights were connected to a pilot valve which fed or bled pressurized oil to or from the pilot oil [ called sensitive oil by some turbine manufacturers ] system; the pilot valve itself was constructed in such a way that the greater the pressure in the pilot oil system, the greater the bias pressure developed against the flyweight assembly tending to drive the weights apart. Therefore when the turbine was under governor control, as the turbine speed dropped, the flyweights would move in, opening the pilot valve and admitting more oil into the pilot oil system; the increase in pressure partially opposed the action of the flyweights and arrested further movement until there was a further drop in turbine shaft speed. This configuration provided a built-in proportional action, and hence stability, to the entire governor system. The percentage speed droop could be readily varied by an adjustment knob which varied the bell crank length of a portion of this system.

Each steam valve to the turbine was provided with a pilot actuator; the actuator's job was to admit power [ or "relay" ] oil to the spring-opposed hydraulic servomotor which operated the steam valve itself. Power oil pressure was therefore directly proportional to pilot oil pressure...I'm pretty sure I still have some governor oil system drawings in my archives at home; if I can manage to find them, I'll scan & post.

Offset was provided by transition of the pilot valve's outer sleeve which embodied the oil admission and release ports; there was a small electric motor, wired up to a pistol-grip spring-return-to-centre switch in the control room which enabled the unit operator to match turbo-generator output frequency to the system for synchronizing purposes, then adjust unit loading once on line.

I remember lots more, but I'll stop here.

I have next to no clue about today's modern electronic governors.

CR

"As iron sharpens iron, so one person sharpens another." [Proverbs 27:17, NIV]

 

[Mbrooke]

I can't thank you enough for this post. I can't lol.


How did it respond to faults where the gen is speeding up then slowing down? Would it be sufficient in Island mode?

The pistol grip switch makes a lot of sense- clears half the picture up for me- I wondered how they manipulated the governor as to sync to the grid.

Don't at all be worried about electronic governors. I'm not as interested in those, in fact I find them "boring". I'm fascinated by the past- by hydraulics and relay logic. In part because doing anything without semiconductors is inconceivable to me.