Generator synchronization

My first advice would be: FORGET IT.

You can sync them by just starting them both and placing them in parallel. The one will slow and the other will speed. (I suspect that you may have to try to do this more than one. If you close when they are far out of sync you will pop the breakers on both machines.

But without an adjustable governor and some way of monitoring how much power each machine is delivering, it will be difficult to share the load.

The "classic" way of syncing two machines was to place a lamp across the contacts of the switch that, when closed, will put the machines in parallel. You put a 3/4 load on the first running machine and bright the second machine up to speed but slight slow. You "inch up the speed" of the new machine and when it is SLIGHTLY faster you thown the switch when the bulb is out.

| I am interested in hearing about simple methods of synchronizing two | consumer grade standby generators under varying loads.

Most consumer grade standby generators are not designed for such things. They vary from mechanically speed regulated (can be plus or minus a couple Hz and who cares normally), to electronically. But few have any interconnect to syncronize. You're better off buying a larger generator.

Small gennys below ~12.5kw are engine speed governed. Above that line there are usually controls that are speed independent. So if you below the magic number, buy a bigger gen or split your loads.

No. You have no control of phase. One will feed the other if out of phase, and blow it out. Even if the are EXACTLY on frequency, which they will not be, they float.

DC is a little different you still get one will try to pull all the load, and you can get surging which will blow one of them out. . But You can add series resistance, I used welding rods one time, (as a series resistor) and got the mismatch in current to within 20%(14 volts at

200 amps)

--------- Ideally-yes- in practice- no- there always will be small but real differences in the speed regulation of the two machines- the net result is that the "faster" machine will try to grab all the load--trouble. In particular, for constant speed settings on both machines, even if the load was equally shared, any change in load would result in one trying to run the other as a motor as well as carry the useful load. In utilities, the governors (and yes, there have been and still are, very good mechanical governors) are set to provide a set speed droop as load increases which helps control sharing by providing a definite intersection between the two speed- torque curves (or frequency-power). Attempts to raise or lower speed on a unit will modify this sharing. With no droop, there is no control of sharing of load.

If you were to try to run two small units in parallel, then it would be best to set one to constant speed and adjust the no load speed of the other (which would have a speed droop) to determine its share of the load. Any load changes would be handled by the "fixed" speed unit. Mechanical governors would be adequate but capability to adjust the "no-load" speed and the droops would be needed. Is it worth the hassle for small machines?- not likely.

-------- Utilities have two main controls on the machines- speed and voltage. They do not have direct control of phase. They survive.

Adjust the speed of the incoming machine to match the on-line machine as try to get the differential voltage as low as possible- there will always be some small phase and voltage magnitude difference but, done properly there will be no more than a slight "bump' Of course if you try to synch when 90 or 180 degrees out of phase- then there will be major problems. It's not a problem to synchronise but it may be a problem with load sharing due to different speed-torque characteristics.

Only in a perfect world, and as we all know, our world isn't perfect.

For two or more AC machines to share the load well, the governors on the machines (either electronic or mechanical) *must* have a characteristic called 'droop'. This is the phenomenon where the governor output (fuel control/throttle) only rises if the speed is below a setpoint. And in order to get more fuel, the speed further below setpoint. Speed droop is often measured as a percentage. If the unit is run isolated from all other generaters and has its load varied from no-load to full-load, the amount of speed reduction needed to have the governor go from cut-off to full throttle is measured. This speed reduction (in RPM) is divided by the no-load speed (also in RPM) and expressed as a percentage. Speed droops of small to medium sized machines often runs 3 to 7 % (i.e. @5%, the speed of an 1800 RPM, unloaded generator will drop to 1710 when fully loaded)

If you take two AC machines with the same droop characteristic (let us assume 5%), they will share the load quite nicely, even if one is much larger than the other and running at a different speed (with different number of poles in their respective generators). For example, say you have a 10kW, 1800 RPM machine carrying 2kW operating in parallel with a 100kW,

3600 RPM machine carrying 20kW. Now suddenly switch on a 55kW load. Both machines will begin to slow down and their governors will increase their fuel flow. But look closely at what happens. The small machine slows down about 2.5% to about 1755 RPM. The governor of that machine will increase fuel flow by 50% of rating and the generator is now carrying 7kW. Similarly, the large machine slows down about 2.5% to about 3510 RPM. The governor of *that* machine will increase fuel flow by 50% of rating and the generator is now carrying 70kW.

So, with the same droop characteristic, these two machines were originally loaded to 20% of their capacity, and after switching on a large load they are both loaded to 70% of their capacity. And it all came about because the two governors have similar characteristics. These machines do *not* need any cross-connections for this to work. Similarly, a sudden reduction in load will come off the two machines nicely and neither will reverse power the other unit.

The only problem is that frequency went from 60hz original, to 58.5hz. This can be corrected by adjusting *both* governor setpoints upward at the same time. If you adjust just one of the governor setpoints upward, it will sense that engine speed is further away from the setpoint and increase fuel flow to the engine. This will increase the load on the generator. As the load is 'picked up' by this machine, the opposite machine will see a reduction in load and start to speed up slightly. So the second machine's governor will sense speed closer to the setpoint and decrease the fuel to its engine and its generator load will decrease.

I've paralleled hundreds of generators over the years using both synchroscope and lights (three phase lights can be fun as they can be wired to 'rotate' :-). Almost universally, the 'incoming' machine is run slightly faster than the system it is being connected to. This helps to ensure that the moment it connects, its governor will see a slight drop in speed and increase fuel flow. This is important as most generators intended for parallel operation have reverse-power protection and in order to avoid false tripping of the unit, it is best to pick up some load immediately. Even extremely *large* units (1200MVA), when first connecting in are run slightly fast (after all, if the speed is exactly matched, the synrchoscope doesn't rotate at all).

Now, that all said, there is yet a different situation with some of the small portable units available. Some of the small portable units are actually DC generators with electronic inverters. Since the machine is actually DC, all this previous talk about governors and droop is not applicable. What is important when connecting inverters together is how the inverter electronics are designed. Some manufacturers (Honda I believe is one) have a special cable to connect between the units. This cable actually connects between the *inverters* so the two *inverters* that are inside the units can communicate and share load between them.

daestrom P.S. DC generators can be made to share their loads in a similar manner by designing the voltage regulators to have a 'droop'. Or just the machine's inherent voltage regulation, if properly designed, will do it as well.

Thanks for the explanation. I didn't know about Honda having a cross connect. I will look into that feature.

daestrom wrote:

| Utilities have two main controls on the machines- speed and voltage. | They do not have direct control of phase. | They survive.

Why not? It can be done.

The phase locking is automatic, very difficult to do by hand anyway.

|> | Utilities have two main controls on the machines- speed and voltage. |> | They do not have direct control of phase. |> | They survive. |>

|> Why not? It can be done. |>

| | The phase locking is automatic, very difficult to do by hand anyway.

I believe that what everyone is saying is that it isn't necessary to drive the prime mover to specifically be in phase sync, but rather, if you switch in the added generator at just about phase sync where the slight speed difference was slowly taking things in and out of phase sync, that the load variations due to circulating currents when the phase varies just a bit will drag them together within a certain degree of sync. I just worry that with 3 or more generators, this can be a chaotic system without a lot of dampened control of the prime mover power source.

I've always wondered how they do that with hydroelectric. You can't just change the water flow levels that fast.

I can't speak for current practice, but some years ago I was involved with a large industrial that generated about 175MW, no machine larger than 25 MW, all steam. We generally ran isolated from any utility, but could connect if necessary. Usually, we connected to help them out, but anyway.........

We had a few machines with special Woodward governors. At any time, one of these machines machine functioned as our speed master. All the other machines were controlled based on load. Keeping the speed master fully loaded kept the frequency nearly constant, the governors on the other machines were trimmed to keep the load divided as conditions required, always keeping the master fully loaded. This pretty much made frequency regulation a non-issue, other than the master, correct load-division assured correct frequency.

We had both Synchroscopes and lights, and generally cut-in the incoming machine as we let it coast down from a slight (1-2 Hz) overspeed. There were no provisions for automatic synchronization, the lights and Synchroscope were it. There were protective relays to double-check the operators' intentions. We later modified these to permit operation of 2 machines as synchronous condensers, but we had to make special provision to allow a small steam flow to cool the machine.

I have no clue how it is done today, but the "master machine" was pretty much standard control practice at the time.

Wicket gates control the water flow to the turbine, driven by a large hydraulic cylinder. I don't have exact numbers but at Manitoba Hydro's "Limestone" generating station the units are about 130 MW each, and the hydraulic cylinder looked to be about a 14 or 15 inch bore - governor oil systems run about 1000 to 2000 PSI.

Bill

Synchronism is not "Phase locking" which implies a set phase angle difference - preferrably 0 on the initial connection. This is the objective. However the synchronising process involves more than phase ocking -relative speed and differential voltage. Certainly, if conditions are not exact- 0 phase difference, 0 speed difference and 0 voltage difference at the time of connection- there will be a bump - The machines are more forgiving than you imply. A utility would have no load control on its generators in a phase lock situation but I assume that you are referring only to the instant of connection. As far as synchronising by hand, this is not a difficult process. It is easier to do it with a synchroscope than with a set of lights but in either case it is not difficult. In fact a good operator can synchronise an incoming machine to the system faster than an automatic control because he can anticipate and allow for reaction time etc. The automatic synchronisers dither around to try to get things "just right" This was touched on by others where the incoming machine is a bit "fast".

As for Phil's comment- yes, a synchroscope does indicate phase (not exactly a precision instrument) as well as relative speed. It does not indicate voltage differences. Is a precision circuit needed- no. Is common sense needed- Yes.

---------- Having 2 or more generators on line with one incoming machine (and you are synchronising one at a time) will make things easier. The incoming machine is then relatively smaller and has less impact on the system. Hydor machines are not a problem- yes it can take time to adjust speed but the process of nearly matching speed and switching as it approaches the best point works well. Still I have seen a 60MW machine come up from start, on remote, automatic, control and pick up load within 1-2 minutes (most of which were due to the synchroniser dithering).

| | Wicket gates control the water flow to the turbine, driven by a large | hydraulic cylinder. I don't have exact numbers but at Manitoba Hydro's | "Limestone" generating station the units are about 130 MW each, and the | hydraulic cylinder looked to be about a 14 or 15 inch bore - governor | oil systems run about 1000 to 2000 PSI.

But how fast can it react?

--------------- On starting the machine and governor are set to come up to a "speed no load" setting. This will be quite close to the final speed and only small speed "twitches" are needed to bring the speed and phase close enough. Fast reaction is not necessary. Certainly, for large changes in power output, once synchronised, large gate changes are needed. No-load to full load may take 0.5 seconds for a steam turbine and several seconds to a couple of minutes depending on the machine and its turbine/penstock system for a hydro unit. I know of a 300ft head plant where the first unit had its gate speed limited (2 minutes to fully open or close) as faster operation could lead to problems with the penstock (collapse possible on sudden load increase) while an older unit of the same size could operate much faster because of a surge tower. On the other hand, Alcan's Kemano (Pelton wheel- high head ) plant (160MVA units) supplying an aluminum pot line is required to, almost instantaneously, pick up a 100MW pot line (all or nothing). Here the gates were opened fully and deflector vanes were used for control. Load on, deflectors moved aside so full blast on the wheel. Not efficient for normal control so conventional- governors and gates used for that.

Whether the machine is hydro or not- sometimes the slowest synchronising is when the speed is almost bang on but the phase is out - then the phase may change in the order of a couple of degrees a second, or seem to dither around as the system or incoming frequency wobble a bit