power grid synchronization

[urlkrueger]

A few years ago I toured the Grand Coulee dam and powerhouse in Eastern Washington state, USA. What impressed me most was standing down in the generator/pump room surrounded by tons upon tons of concrete and the whole vibrating from the unimaginable forces at work.  It was a bit scary but gave me an idea of the tremendous amount of power produced.

Now I live near another smaller dam which along with Grand Coulee are both connected to the Bonneville Power grid.  What I don't understand is how they synchronize the phase of the AC power from these and other generators before attaching them to the grid.

Anybody know how this is done?

[BradC]

A few years ago I toured the Grand Coulee dam and powerhouse in Eastern Washington state, USA. What impressed me most was standing down in the generator/pump room surrounded by tons upon tons of concrete and the whole vibrating from the unimaginable forces at work.  It was a bit scary but gave me an idea of the tremendous amount of power produced.

Now I live near another smaller dam which along with Grand Coulee are both connected to the Bonneville Power grid.  What I don't understand is how they synchronize the phase of the AC power from these and other generators before attaching them to the grid.

Anybody know how this is done?

The simple answer is by varying the speed of the alternator. The complex answer is by varying the speed of the alternator

Many years ago I saw a great photo of what was left of an old powerhouse in the North of Western Australia after one of the machines was accidentally synced up 180 degrees out of phase. It leapt off its mounts and out the side of the powerhouse.

[Jr460]

In the very old days.  A set of 3 lights, connected from Gen phase A to Grid phase A, Gen phase B to Grid phase B, Gen phase C to grid phase C.  When all three lights are out and stay out you have matched phase and voltage.

Bring unit up to speed, then adjust slight fast or slow to get the blinking pattern to slow down and almost stop.  Time your movement of the breaker handle so that it is closed just as the 3 lights are out.

Newer than that, something called a syncro-scope, looked like a clock with just one big hand.  I think it had two drive coils, the hand spun around and slowed down when you got the speeds right.   Phase is a match when it is point up, an 12 o'clock.

Newer stuff, like in the past 40-50 years, some more complex automatic sync.   If you think about it, it is just a big PLL, the error signal drives the generator speed controls.

[StillTrying]

"Anybody know how this is done?"

With 3 light bulbs.   en.wikipedia.org/wiki/Synchronization_(alternating_current)

[Jr460]

The simple answer is by varying the speed of the alternator. The complex answer is by varying the speed of the alternator

Many years ago I saw a great photo of what was left of an old powerhouse in the North of Western Australia after one of the machines was accidentally synced up 180 degrees out of phase. It leapt off its mounts and out the side of the powerhouse.

Tthe GE manuals said, at least for the big stuff, 100 to 1000+ Megawats,  If you sync more than 20 degree out, you have to shut it down and do a complete inspection.  That inspection is going to take at least 2 weeks even if everything is OK.

[duak]

You might remember hearing about something that went wrong a few years ago at a Russian dam: https://en.wikipedia.org/wiki/2009_Sayano%E2%80%93Shushenskaya_power_station_accident

It was nothing to do with synchronization but it points out what can happen when large scale ordinary forces are accidently unleashed.  I was impressed by the engineering that tames these forces after touring dams and powerhouses.  Even small ones, of only a few MW, are pretty serious things.

Anyway, once the generators are synchronized, the turbine gates are adjusted to vary the current and its phase relative to the network voltage to control the power delivered.  The generator's voltage is controlled by varying the excitation ie., the strength of the rotor's magnetic field.

[calexanian]

Once upon a time they simply just carried the speed by variation in water flow with a phase meter and closed the breakers when they were close enough. The next way they did it was to turn the generator by a secondary motor, get it up to speed and close the breaker with no load or no water where it will find its way into synchro in less than a few cycles, then ass the water (Or steam depending on the energy source) and now of course its done digitally.

My info is many decades old and I am sure they have new ways but the great western DC link line used to use massive thyratrons as the switches and massive pulse forming networks to make the 60hz and they would start by triggering the thyratrons from the AC signal so  they would be already in phase and the pulse shaping network would do the rest. Like a giant ferroresonant transformer being driven from a square wave. I am sure its all real time digital control now though,.

[Halcyon]

Fascinating. Some of that gear looks pretty old, but I bet it's designed for ultra-high reliability.

[sibeen]

To add to what duak stated; to control the amount of power (watts) delivered to the grid you control the phase relationship, to control the amount of VARs delivered you control the voltage.

[urlkrueger]

So, what everybody seems to have described is manual methods like "Transistor Man" in "Art of Electronics" but with some automation.

What I was expecting was more along the lines of atomic clocks at each site synchronized to a universal time standard with automation to ensure that frequency was correct and that 0^o phase occurred at specific clock ticks.

And then there is the problem of how do you handle it when the steel mill turns on their electric arc furnace and your generator slows down due to the change in load.

Way beyond my abilities.  I think I'll stick with the idea of "Generator Man" turning a knob.

Thanks...

[station240]

I think the modern grids sync via fibre optic data links.

[Jr460]

So, what everybody seems to have described is manual methods like "Transistor Man" in "Art of Electronics" but with some automation.

What I was expecting was more along the lines of atomic clocks at each site synchronized to a universal time standard with automation to ensure that frequency was correct and that 0^o phase occurred at specific clock ticks.

And then there is the problem of how do you handle it when the steel mill turns on their electric arc furnace and your generator slows down due to the change in load.

Way beyond my abilities.  I think I'll stick with the idea of "Generator Man" turning a knob.

Thanks...

Can't happen that way.

It doesn't matter that the phase is at zero at some point in the grid at a point in time.   The grid a set of transmission lines and phase zero at one point is not phase zero at all points.   

So if you pick a point, say your house, the phase at the power is different and it will change as changes are made to the grid.

No one cares what the absolute phase is, only that when you close that break on the generator that things match.

Now if you are talking about the overall line frequency, well that varies and utilities keep a running total of the error and correct for it at low load times, so that a clock with a sync motor will stay correct over a 24 period.

So yes, when the steel mill turns on, or the bigger jump in load occurs each morning as people wake up around the same time and turn on light, the whole gird slows down a tad, and the turbines push harder and bring things back to the correct rate.   If you watch the RPM design of a large unit, it will be at 3600 RPM all the time (most didn't show more digits).   That means it is putting out 60Hz, and with the limited digits, that means the freq is off by at most about 1/3600, or % 0.02 if my math is right.

Remember this has been done with 100+ years without digital or fiber links.


The only place they care about a good frequency reference is in the downtown control center where they keep track of slow/fast and send load corrections to the different units.  At each plant you don't care.  If the grid is .00001 Hz slow when you bring the unit online, that is what you sync to, not 60.000000000000Hz.

[IanB]

So, what everybody seems to have described is manual methods like "Transistor Man" in "Art of Electronics" but with some automation.

What I was expecting was more along the lines of atomic clocks at each site synchronized to a universal time standard with automation to ensure that frequency was correct and that 0^o phase occurred at specific clock ticks.

When generators are connected to the grid, they are automatically locked in sync and cannot shift. So there is no need to synchronize them while they are running. The only time when synchronization is needed is when first connecting a generator that is not already connected. And since the bad effects come from any relative difference between the generator and the grid the synchronization tools here try to adjust the relative difference to zero before closing the contacts.

In this context absolute phase is not useful, only relative speed and relative phase.

And then there is the problem of how do you handle it when the steel mill turns on their electric arc furnace and your generator slows down due to the change in load.

That is easy. If the generator tries to slow down you feed more power into it.

The secondary question (which maybe you did or didn't ask) is: how is the grid as a whole maintained at the right frequency and phase? Here an atomic clock may be useful. When the whole grid slows down you have to feed more power into all the generators at once to speed it up. If the whole grid lags behind, you have to speed it up until it gets back where it should be. (You can speed it up by making all the generators go faster until they catch up.) Somewhere there is a central control room that coordinates the whole grid to keep the voltage, frequency and phase on target.

[T3sl4co1l]

It's 50/60Hz man, you could sync it with NTP for all that matters.  Don't need atomic clocks for something this old fashioned!

Tim

[floobydust]

I worked in a multi 485MW power plant, what they did was spin up the generator and give the field excitation a "bump" which would force the rotor to lock on phase.
Some guys just gave very weak field excitation until the synchroscope looked good, then they switched in.

If the control room operator was in a hurry to go on-line, they'd just hit it and the entire building would literally jump which they switched in.

[dmills]

The trick is that grid frequency is proportional to the integral WRT time of the net power (Power input - power used), to raise the frequency you just turn up the throttles at one or more prime movers and the whole thing slowly winds up to a higher frequency, same thing if you are generating more power then is being used, the frequency ramps up until the excursion causes units to throttle back reducing the power input until the frequency is back where it should be.

It is the setup of the governors on the prime movers that determines which units load up when.

For example (In a 50Hz system), if you have some massive hydro or nuke or something plant that is very economical as a base load plant, you maybe set the governor to 50.1Hz with little droop so that as long as the frequency is below that the thing will be running flat out. On the other hand if your plant is a open cycle gas turbine or a Tesla battery or something (Expensive to run under load, or of limited duration, but fast to bring up), you set the governor to 49.9Hz with full output occurring at say 49.8Hz, that way your plant stays at idle until the grid frequency drops to well below normal at which point the throttles start opening to apply torque to the rotating machine, and you start feeding power into the grid.
Interestingly all the machines on the grid are turning at the same speed (number of poles dependent), what varies based on the governor settings is the torque produced by the prime movers. 
   
When considering three phase machines, you can consider the system as a phasor between the rotating field in the stator (due to the grid connection) and the rotor, with the difference being an angle either positive or negative depending on whither the machine is generating or motoring, steady state this angle can be seen as a torque on the rotor.

An interesting case is a power station with a transmission line to the grid protected by a recloser at each end, if the recloser trips there is then a very finite time for the recloser to close again before the rotor has gotten so far ahead of the (now missing) stator field that the phase error will cause damage, activation of such reclosers is by all accounts something everyone at the power plant tends to notice.

Regards, Dan.