It's just a limitation of the software or algorithm you are using. It's done to simplify the calculations. In most cases, this corresponds fairly well to reality if you have only one utility source.
The software I use (EasyPower) allows any number of swing sources for power flow analysis. There can be convergence issues, but normally it's fixable. Swing sources can be specified with a specific power angle.
I've used SKM and it requires at least one swing bus (generator or utility source) but the upper limit is much higher than 1. The fewer swing busses the easier it is for the calculations to converge if a solution in possible with that swing bus. Computer power is generally sufficient now that you should be able to set everything that should be a swing bus as a swing bus and running the calculations. It may be necessary, occasionally, to fix certain sources to control the nature of the solution found.
You can have some PV generators, if that is realistic, but there has to be someplace, the swing bus, that can make up the difference in watts and vars between load and the scheduled generation.
The term 'Swing' generator is usually used just in software to determine a source the voltage of which does not change during the analysis. As mentioned, pretty much every software available allows more than one swing bus connection.
In the real world, the theoretical connection with no voltage change is not possible.
As mentioned already, the swing bus is required in the mathematical solution to make up the difference in watts and vars.
In the real world, where you have local generators in parallel with a power grid, the utility sources function essentially as swing buses (within reason), providing whatever power and vars are required or absorbing any excess from the local generators.
DavidBeach: I thought that PV generator will also make up for VAR's (by keeping voltage constant on it's termianals)I also thought vast majority of gens out there are PowerVoltage (PV) with scheduled Power export and scheduled Voltage bus. What do you mean with "if it is realistic"?
DanDel - thank you, but are you saying we model different types of gens just for mathematical calculation (they don't exist in reality?)
dpc:
So, I understand "spinning reserve" generation is to make up for lack of generation (usualy gas turbines, for fast demand changes), and every generator "type" with voltage regulator (AVR)can "consume" excess of VAR's to their machine limitations, not just swing type gens. Is "swing" generator (IEEE says biggest gen in utility studies or utility source for industrial study) strogest source of scheduled generation or the one who deliver less or more Watts/VARs quickly enough when needed?. So is it fair to say that this is not strongest source but one with quickest
dS/dt?
"If it is realistic" depends on the system being modeled. A standby system with two generators and associated load would have the two as swing generators and you wouldn't have any PV generators. On the other hand, if you have a system where you have multiple sources each contributing a certain load then you could have several PV generators. A PV generator will supply vars, within its capability, at its scheduled watt output, but will it have the capability to supply (or absorb) a sufficient quantity of vars? There can also be baseload machines running in PQ mode supplying a scheduled amount of watts and vars. The swing bus/swing source is where the software can find the necessary watts and vars necessary to make it all work out.
Unless you can model the entirety of your system, all generators and all loads, you will have to make some compromises and you have to know when you are close enough.
The utility connection that always gets modeled as a swing bus isn't really, load will change the voltage, but not by much and not enough to make worth the complication of modeling those affects. Part of engineering is knowing when you are close enough to the answer. If you have line models that get you to within a percent or two of the correct phase impedance (never mind your zero sequence impedance) then getting the contributions from each source nailed down to the 1/10 of a percent is false precision and a waste of time because you are now at 1/10% +/- 2%; whoopee!
If your maximum load change is small compared to the system feeding it, you can treat the source as a swing source with instantaneous response, not exactly true but close enough. On the other hand, if you are modeling 1.8MW of load being fed from two 1MW generators you will need to see all the system effects. If the software places the generator impedance after the voltage control point (the right place for fault studies but the wrong place for load flow) you may have to adjust the voltage of your swing source to overcome the voltage drop of its internal impedance.
I'd start any study with all sources as swing, but with a rated capacity, and see where the numbers fall compared to how you think the system is going to operate. If the result is close to how it should be, then you might find it could operate with reasonable stability. If you have to use PV and PQ sources to force significant changes in load flow you might want to reexamine your expectations or your model.
Yeah, if you've got local generation running in parallel with the utility and the operating mode is fixed watt and var output go ahead and model it that way; but any source that has some freedom of operation should start out as a swing source and then be modified if/as needed.
pajce, the term 'model' is typically used when entering the mathematical parameters of an electrical device into engineering software, so, yes, we do 'model different types of gens just for mathematical calculation'.
A 'swing' bus is the bus in the system model whose voltage does not change. This is necessary for the calculation (it is usually the Utility connection bus).
In ETAP, you can also model a generator as 'Voltage Control', 'VAR Control', PF Control', etc., based on the software (SKM has SB, PV, & PQ options). But you must have at least one swing bus in any case. Making the generator anything but a swing bus, the generator voltage will change during the analyses.
Thanks guys. I think I can recap now that "swing" generator is no different then any other generator (in reality, not in mathematical program analysis), this is what I would normally expect. There is only one principle of work of the generator.
