Synchronous Generators - an overview of construction and operation

I've posted quite a bit about other aspects of power plants - posts about gas turbines, steam turbines, the steam cycle, nuclear reactors, and done a few posts about different or interesting places where I've worked in the past.

Oddly enough, I've never posted about electrical generators before, except for historical electrical progress and small household emergency generators.  It seems a bit odd that I've overlooked this for so many years, because the entire reason power plants are built is to spin the generator!  The machines we will be discussing are actually an alternators, because the output is alternating current.  In the business, we tend to use the term "generator" more often than "alternator", so please bear that in mind while reading.



So... let's just jump into it!  The simplest generator is a simple bar magnet, spinning around, with a coil (also called a winding) on each side, connected together and to an electrical load.  In this case, the electrical load is a light bulb. 

As the bar magnet sweeps past the coils to the left and right, its associated magnetic field passes through the coils, and sweeps electrons along inside the wires.   Electrons have a negative charge, so where there are excess electrons, we develop a negative voltage.  The end of the coil where electrons have been swept away from develops a positive voltage.  This voltage difference then causes electrical current (electrons again) to flow through the light bulb, causing it to light up.   Maximum voltage will be present when magnet is horizontal.  If you make a graph of the voltage over time, the output voltage will form a sine wave, passing through zero when the bar magnet is vertical, 90 degrees away from the coils, and at maximum or minimum when the north or south pole is in line with the coils.

Below is an example of magnet position vs. electrical output.  In this image, the electrical coils would need to be located at the top and bottom of the circle for the maximum and minimum voltage output to match the drawing up above.

A utility-size generator is considerably more complex than what is shown in these diagrams.  The simplified diagrams above describe the construction and output of single-phase generators.  Industrial generators are three-phase, meaning there are three pairs of windings, spaced 120 degrees apart (both physically and electrically).

Three phase alternators are used in order to eliminate periods where no power is being produced - which happens when the sine wave passes through zero.  A three-phase system also reduces the number of conductors required to send an equivalent amount of power to the load.

If you look at a three-phase voltage sine wave (below), you can see that the power is more consistent.  At no point does the system voltage (and therefore power output) ever drop to zero as happens with a single phase generator.  Interestingly though, the sum of the three voltages always equals zero.  Pick any point on the graph and draw a vertical line.  The positive values will be exactly cancelled by the negative values.

A simple three-phase generator animation and output voltage graph looks something like this:

All of the simple alternators above use a rotating bar magnet to create the output voltage.  In the real world, we cannot use a bar magnet.  A bar magnet won't work, due to the fact that it has a fixed magnetic field strength, and it therefore cannot be used to adjust generator output voltage.  We need a way to adjust the rotating magnetic field strength, so that we can raise and lower generator voltage.  It's important to be able to adjust the voltage for several reasons:

• To match with grid voltage to connect our generator to the grid
• To compensate for increasing and decreasing electrical loads
• To increase or decrease reactive power sent to the grid (discussed this later in the post)

Instead of using a fixed strength bar magnet, we use a large electrical winding, and then apply a large  adjustable DC current through it.  In this way we develop an intense but variable magnetic field on the rotor, and by adjusting this DC current flow, we can vary the output voltage of the generator.  This rotating winding is called the field winding, because it adjusts the magnetic field that sweeps through the generator and induces the output voltage onto the stator windings.

In the diagram above, we see that the adjustable DC Excitation Voltage ("Vf") increases or decreases the Excitation Current "If" (aka "Field Winding Current"), and this in turn strengthens or weakens the rotating magnetic field, shown as the dashed green lines.  This adjustable magnetic field then increases or decreases the output voltage of the generator.   By making small adjustments to the DC excitation voltage, we can get quite a large change in the generator AC output voltage.  This makes the exciter and generator a DC to AC voltage amplifier of sorts.

Voltage control gets more complex yet - the ability to adjust generator output voltage is only part of the picture.  There must be an automatic feedback loop to increase the output voltage when electrical load is increased, and to reduce voltage when electrical load drops off.  This is analogous to a speed governor, but for voltage.

Below is a simple block diagram of how an automatic voltage regulator adjusts the amount of excitation current that is fed to the field winding.  There are a couple of different types of exciters, and the voltage regulation schemes differ slightly with each.  The voltage regulator described below would be for a brush and slip-ring type, known as a static exciter.

The "Sensing Circuit" block represents three voltage transformers - one for each electrical phase.  These voltage transformers reduce the high-voltage output of the generator (which is typically thousands of volts) to just a few volts, so that the signal can be safely used for metering and control.  The "Reference Circuit" block is the desired voltage setpoint, which is set and adjusted by the operator.  The "Comparison Circuit" block compares the "Sensing" and "Reference" blocks, and sends an error signal to the "Amplification Circuit".  The "Amplification Circuit" takes a small (usually digital) control signal an electronic board and sends a much larger output signal, which is required to drive power electronics in the following block.  The output of the Amplification Circuit is what controls the block labeled "Signal Output Circuit".  Within the "Signal Output Circuit" are large power-switching electronics - thyristors.  Thyristors are used to convert a separately-supplied AC voltage to DC, and then used to regulate the amount of DC current that is supplied to the field winding.  The block labeled "Feedback Circuit" uses a signal from the exciter output, which is sent as negative feedback to the Amplifier Circuit.  This is done to prevent over/undershooting the desired setpoint following a step voltage change, such as might happen during when a large electrical load is turned on or off.

The block diagram above does a decent job of explaining the very basic operation of a static exciter, but they are really quite complex.  They are required by federal law to have very fast response times to electrical disturbances.  They contain a crowbar circuit for de-excitation when the generator field is de-energized.  Exciters on larger machines in the Western US are also required to have active harmonic dampening.  The portion within the voltage regulator that provides the dampening is called a Power System Stabilizer (PSS).

Below is a bit more detailed block diagram of an Automatic Voltage Regulator

Below, inside the cabinet of a small brushless exciter.  Most modern machines have dual exciters and static excitation, with one exciter loaded and the other exciter ready to take over in the event of an electronics failure.

Below, a display on the front of an exciter cabinet. This machine uses a brushless exciter with a permanent magnet generator, so this is a smallish generator.  We are applying 37 volts DC to achieve 6.5 amps DC current to a pilot exciter.  The pilot exciter will amplify this current and supply the field winding with much greater than than 6.5 amps.  Generator voltage is 14650 volts at 3430 Amps, and the Power output is 83.56 Megawatts.

Cool eh?  There is a short Wiki article on electrical exciters here:

Let's take a break from electrical theory and regulators, and have a look at the physical construction of a few  generators.

Below is a Stator or Armature winding of a three-phase, two pole turbo-generator.  The armature is where the output voltage is produced.  This is the stationary part of the machine - in this image the rotor has been removed.  If this were assembled, In the middle would be a rotor, holding the field winding, which would be driven at 3600 RPM by a gas or steam turbine.

The armature windings pass down the center between high-silica laminated steel punchings.  The steel concentrates magnetic flux, making the generator more efficient.  The steel is laminated and insulated to minimize circulating electrical and magnetic currents, which causie self-heating. 

Although generators are highly efficient, there are large amounts of electrical current flowing throughout the machine.  The currents are large in both the field (rotor winding) and in the armature (stator windings).  As a result of these high current flows, a great deal of heat is created.  This heat is due to resistance heating, and this heat is trapped within the enclosed space of the generator.

Heat within the generator will damage the electrical insulation and lead to electrical faults, and so all generators are manufactured with heat removal systems.  The rotor and stator electrical windings are made with slots, so that air can circulate around and through them.  Large and mid-sized generators are filled with pure hydrogen gas, because hydrogen is superior to air at conducting heat.  When hydrogen cooling is used, the generator shell is also a pressure vessel.  Pure Hydrogen gas is pressurized to increase its density, which further improves heat removal.

Very large generators have cooling water tubes embedded in the stator windings to remove heat, in addition to being equipped with hydrogen coolers.  The Hydrogen is circulated by fans mounted on each end of the rotor shaft inside the generator.  This gas circulates in a closed loop between the electrical windings and hydrogen coolers, which are typically mounted at the bottom of the generator shell. 

Below, a mid-size generator, with five hydrogen-to-water coolers at the bottom

Below is a turbo-generator rotor.  At the far end is a coupling flange to connect the rotor to the turbine.  At either end of the winding section are fans, which circulate hydrogen for cooling.  Just outboard of the shipping straps are shiny silvery sections.  These are called "retaining rings".  I will get back to them.

Moving further to the right, we have a journal bearing wrapped in cloth to protect the surface.  Next we have two shiny slip rings, where DC excitation current is supplied to the field winding for voltage control.  Lastly, we have a smaller fan to blow filtered cool air continuously across the slip rings.  Carbon brushes ride on the slip rings, and the fan cools them and blows carbon dust out of the enclosure.

Below is a brush rigging assembly.  This is how excitation (or field) current is transferred from the stationary excitation source onto the rotating field winding.  A large number of brushes are pressed against the rotating slip rings (seen in the image above) by spring pressure.  A lot of brushes are needed, because while carbon conducts electricity, it has more electrical resistance than copper does.  Therefore much more contact area is necessary to get the job done.  The increased electrical resistance also generates quite a bit of heat, which is why there is a dedicated brush cooling fan at the right side end of the rotor above.   The brush cooling system is open-circuit, meaning that the air does not circulate.  It is drawn in, filtered, and exhausted continuously.   The carbon brushes wear down over time, and carbon dust would tend to build up and cause electrical shorts and fouling issues issues if the air were to be circulated through a cooler.  For this reason the fan simply blows the loose carbon dust out of the enclosure.  It's a little grimy where this duct exhausts to.

Below is a generator rotor undergoing servicing.  The silvery retaining rings that I mentioned on the rotor above have been removed in this photo.  Retaining rings are one of the most physically stressed pieces of equipment in a power plant.  They have also been responsible for numerous highly destructive accidents.  Retaining rings are high tensile strength alloy cups that hold the ends of the field winding in place.  Between the retaining rings, where the windings are straight, they are held in place by wedges, and these wedges get tighter the faster the rotor spins.  The wedges hold these straight sections of copper in place quite well.  However, at the ends of the rotor, there is a large mass of copper that is not wedged in place, and that's what the retaining rings are for.  However the centrifugal force at 3600 RPM (60 rotations per second), with a huge mass of copper trying to unwind, is immense.  When a retaining ring fails, the generator is usually destroyed.  If it is hydrogen-cooled it will also be engulfed in fire.  Frequently, adjacent equipment is also damaged by flying pieces of the ring after it exits the generator.  Over the decades, a great deal of research has gone into improving the strength of the retaining rings.  The metallurgy of the rings will be dependent on whether the generator is cooled by hydrogen or air.

Below, the retaining ring goes around the bare copper section, and the straight segments pass through slots cut into the rotor, and wedged into place.

Below, a bare rotor with the field winding removed.  This shows the machined slots that the copper windings are held in.

Operation:
When run up to 3600 RPM, our turbo generator is almost ready to start making power.  We apply DC current to the field winding (rotor), and the armature (stator) begins producing output voltage at 13.8 KV, 18KV, 21KV or even more.

These voltages are fairly high, but not high enough to use for a transmission line (which typically run at 115KV, 230KV or 500KV) , so the voltage must be matched through a main transformer, sometimes called a Generator Step-up (GSU) transformer.  The GSU may step the output voltage up from 18KV to 230KV for instance, so that the generator output can be sent to the power transmission system.

When we close the output breaker of the generator, we are going to synchronize our generator with hundreds of other generators on the electrical grid.  In fact, you can think of all the other generators on the grid as one infinitely large generator that is much larger than you are.  When you synch your generator to the grid, you are coupling your generator BOTH ELECTRICALLY AND MAGNETICALLY to something that is enormous.  This is important, because I said it in caps.

So you match the generator output voltage with the grid voltage by adjusting excitation current, and adjust your generator speed so that your frequency is slightly higher than grid frequency.  That is done so that when you shut the generator breaker, the generator will be putting out a little bit of electrical load.  If the machine were at a slightly lower frequency than the grid, when you closed the generator breaker, the grid would be driving the generator (and turbine) as a motor.  This is bad, and a protective relay will put an end to it, and you will get in trouble for tripping the unit.

Once you get the generator in phase with the grid, you close the generator breaker, and you are putting a couple of Megawatts of Real Power on the grid.  Good stuff.  The guy in the video below discusses this process quite well.

To increase power output, you apply more fuel to the gas turbine, or apply more steam to your steam turbine.  Does your generator speed up?  NO!  Your stator is electrically coupled to an infinite-sized generator, and your rotor is magnetically coupled to the stator.  You are not able to speed your generator up while it is coupled to an infinite-sized one.

So to explain further, up until the generator breaker is shut, increasing fuel to the gas turbine will make the machine speed up.  Once the breaker is shut, adding more fuel will not speed the machine up, but instead it will make more Real Power, measured in Megawatts.

What happens electrically inside the generator is that the linked magnetic fields between the stator winding and the field winding stretch as more fuel is added.  As more torque is applied to the generator shaft, the magnetic field of the rotor moves out ahead of the magnetic field of the stator, but never gets away from it.  This is called the "torque angle" or "slip angle".  Note that the rotor is out ahead of the stator by a few degrees, but it is always rotating at the same speed, the rotor simply leads it.  There is a gross misconception about this, even among some power plant guys. 

 

I find this to be useful to help understand the relationship between the rotor and stator magnetic fields.  No, the donkey is not going any faster than the millstone - but he is slightly out ahead of it, and pulling it.  The leather straps on his harness stretch as he works harder.  That is all.

If you somehow manage to get the rotor to slip out of synch, that's bad.  A protective relay will put an end to it, and you will get in trouble for tripping the unit.  

You can accomplish this feat by under-exciting the generator - that is, making the rotor's magnetic field so weak that it slips out ahead of the stator's magnetic field.  Exciters have limiters to prevent you from goofing up this badly, so hopefully this scenario won't happen.

That said, I have seen a field slip when I was in the Navy.  We had a large and small generator connected together.  We were trying to synch to shore power.  For some reason, pier voltage was really high on that day.  The operator kept raising voltage on the bigger generator to match voltage on the pier.  The bigger generator dominates ships voltage.  He should have adjusted the smaller generator up as well, but neglected it.  Eventually he raised voltage on the large generator so high that the small generaror didn't have enough magnetic coupling between the rotor and stator, and it lost synch.  The lights got bright, then out, then bright, then out.  This went on for several seconds before they went off for good.  Warships don't have protective relays, so this went on for a little bit before high fault currents finally took out both generators.   Fortunately no damage occurred.

All generators have limits on how hard they can be run.  The limits are based on localized electrical heating of different components inside the generator.  Below is what is called a "reactive capability curve"  sometimes called a "generator operating curve". 

 You will notice the left-right axis is Megawatts.  That is Real Power, delivered when your gas or steam turbine is cranked up and spinning that generator as hard as it can, with maximum torque angle.  You should be able to churn out 200 Megawatts and still be within the most limiting curve - that is, staying to the left of the curve that says 15 psi of hydrogen pressure in the generator.  Obviously if you have more hydrogen density to remove heat, the generator can be run harder yet.  270 MW at 30 psi, and 320 MW at 45 psi of hydrogen pressure.  Impressive!

The reactive capability curve is actually composed of three different curves.  The three different curves are connected together, but you can clearly see different arcs on the top, right and bottom.  Each of the three curves is a limit based on avoiding damaging heat in to specific components inside of the generator.  

I'll start by explaining the easiest one.  The right side curve is based on too much current heating the armature (stator) windings of the generator.  This could happen because you are making too much power - pushing too many amps of electricity to the generator output.  Amps generate heat, and too many amps generate too much heat!  You really can't do this however, because nobody builds a power plant where the turbine output isn't precisely matched to the generator output.  Doesn't happen.  Why purchase a 500 MW turbine and connect it to a 400 MW generator.  That would be a waste of money spent on the turbine, so builders match the machine's output ratings.

To explain the top and bottom curves, I'll have to explain our electrical and magnetic relationship to the rest of the world with respect to Alternating Current.

The other axis that we need to worry about?  The vertical axis of positive and negative Megavars.  What are Megavars, you ask?  Sucks to be me right now...

A VAR is a volt-amp reactive.  Up to this point, I've been talking about real power.  Real Power is power that can perform work.  In AC systems, we also deal with Reactive Power.  This is a form of power that provides voltage in the system, but does no physical work.  Just like real power, reactive power is shared between generators though, and it is quite important.  The relationship of real to reactive power is called "power factor" and those are the straight lines radiating out from Zero on the generator reactive capability curve.  If the generator is making zero reactive power, and all real power, the power factor is 1.0 or "Unity". 

Let me 'splain something about the electrical grid.  Nearly everything on the grid uses electricity, and everything that uses electricity causes voltage to drop.  Turn on your vacuum and watch what the lights do.  They dim, because voltage drops.  Multiply that voltage drop by a million vacuums, refrigerators, large industrial air-conditioning units, water well pumps, etc...  you get the idea.  Everything drags grid voltage down.  Only one thing can actually increase grid voltage (with minor exceptions), and that is a generator.

And as noted earlier, increasing excitation current on the rotor winding increases generator voltage - unless that generator is synchronized electrically and magnetically to the grid.  In that case, the raising and lowering excitation current raises and lowers Reactive Power, measured in Megavars.

Megavars are what help support grid voltage.  It may take a lot of Megavars (MVAR from now on) to maintain grid voltage on a day where the grid is heavily loaded, and it may take negative MVAR to keep voltage from going too high on a lightly loaded grid in the middle of the night.  MVAR is shared between generators. Because you are electrically and magnetically tied to an infinite generator, raising and lowering MVAR doesn't have a huge effect on the grid.  However it is important to carry a share of the voltage support by trying to keep the electrical grid at the specified voltage, SO LONG AS YOU DON'T VIOLATE THE REACTIVE CAPABILITY CURVE OF YOUR GENERATOR.

Back now to the other two limiting curves on the reactive capability curve.  

 

The top one is based on not overheating the windings of the rotor with high current, by trying to provide too much voltage support.  The combination of many Megawatts and many MVAR obviously is more limiting than a few Megawatts and a lot of MVAR, because field current is a combination of both.  The exciter circuitry has an OverExcitation Limiter (OEL) that will prevent you from reaching this curve and overheating the field winding.

The bottom curve is due to stator end heating.  When the machine is operating highly under-excited, the magnetic flux tends to concentrate at either end of the stator, and therefore the current will concentrate there as well, leading to localized heating.  The exciter has an UnderExcitation Limiter (UEL) to prevent reaching this curve, as well as to prevent loss of synch, which occurs slightly below the heating curve.

Below is a generic reactive capability curve with notations.  The "steady state liability limit" curve is  the point where you might find yourself slipping a pole.

Back in the old days, you used to have to manually plot this yourself on one of the above charts, however modern computing power can plot it for you in real-time.  Pretty cool, huh?

Note in the above picture the generator is putting out 125 MW of real power and about 75 MVAR of reactive power, and yet the generator is nowhere near any kind of limit.  Carrying a large reactive load increases electrical currents inside the generator, and so of course it increases heat.  However the generator coolers are sized for this heating - or else the curve would be much more limiting.  There is another misconception among many, many power plant guys that carrying reactive load decreases efficiency.  It does not do that.  If anything, it's easier to push real power onto the grid at higher voltages.

For those unaware, it would be highly abnormal to run a generator anywhere near any of the limits.  Nevertheless anyone running a grid-connected generator should understand its limits and what they are based on.