Estimating the source inductance of a 3ph. generator

[MagicSmoker]

I'm working on a rectifier for a drive system that will be supplied by a 625KVA diesel genset. The existing rectifier being used has the usual 3% impedance line reactor as normally specified when a drive is supplied by the mains and this strikes me as both counterproductive (more voltage drop) and superfluous, as even a 625KVA synchronous generator will exhibit a much higher source impedance than the mains. I know that to determine the interrupt capability of the breakers required in the event of a short one uses the subtransient reactance, X"d, to determine the initial current spike (lasting ~ a half to one full cycle), but for the purposes of reducing harmonic distortion and improving rectifier power factor, which reactance value is appropriate? Using the synchronous reactance, Xd, gives what seems like an absurdly high value of inductive reactance (e.g. - 1.2mH). Using the armature time constant, and knowing the resistance of the stator windings, gives an eminently plausible value of 70.3uH. Finally, using X"d gives a somewhat lower than expected, though still plausible, inductance of 38.5uH.

For reference, the following specifications for this 480VAC generator are known:

KVA base rating for reactance values = 825
Xd = 2.80 PU
X'd = 0.13 PU
X"d = 0.09 PU
Ta = 0.019 seconds (armature time constant)
Stator winding resistance = 3.7 milliohms

 

[LiteYear]

According to the reference that 7anoter4 just posted[1], the most appropriate value is the sub-transient. This seems to make sense, since the distortion due to the non-linearity of a rectifier is definitely within a cycle (ie. at multiples of the mains frequency), so in the region of sub-transients.

http://www.idc-online.com/technical...al_engineering/ANALYSIS OF SUBSYNCHRONOUS.pdf

By the way, I tend to agree with your premise that additional line reactance would be counter-productive. Ultimately the reactance is normally there to reduce harmonics for other loads on the line side. If you're running off a generator then you want what you can get from it.

[1] over in

http://www.eng-tips.com/viewthread.cfm?qid=339052

 

[MagicSmoker]

Thanks for the response LiteYear. I was leaning towards the subtransient reactance as the one most appropriate, but I was having a hard time grasping this on an intuitive level. Poking around on IEEE Xplore netted me a few promising papers on the matter (by Bonwick, dating June 1973 and January 1980) and at least one Master's level thesis...