Ok so chances are those are diode front end VFDs which typically have some pretty awful line performance. A line reactor blocks triplen harmonics but those aren’t the harmonics tge VFDs create so that is a waste of time. Essentially you have been given bad advice. Current harmonics are HARMLESS unless it’s affecting the transformer. Look at it this way. You have a 225 kVA transformer. If the load is purely sinusoidal then the usual rating system works. But it’s not. With harmonics there are times where you are drawing more current than simple sine wave math predicts. If the transformer load is below the rated current then it won’t distort the output. If it’s above the rated current voltage starts to drop and the more you exceed the rating the more the output sags. If this happens you get voltage harmonics, which is NOT harmless and what we need to avoid.
Where there are issues is two things. First off the transformer short circuit rating is pretty close to what I suspect the VFDs are rated for. So unless they have fast fuse protection or something like that you’ll get a lot of VFD failures where something as simple as a motor failure causes a VFD to fail. This is where a 3% line reactor or splitting them up among multiple transformers can make a difference and leads to what is probably your issue.
So I don’t know the ratings and how it’s timed but I’m estimating 21 kVA each? 400 x 30 x 1.732 is 21 kVA. At 30 drives that’s 630 kVA. I’m not so concerned about operational issues when it’s running as you said that seems OK. But when all those drives initially power up you have every drive trying to precharge simultaneously. I have serious concerns you’ll get all kinds of boot up glitches as the line voltage takes a dump when this occurs, which is what you describe. A simple test is hook up your meter on alligator clips and watch line voltage (phase to phase) looking for minimum voltage. I’ll bet it’s dipping low causing momentary power losses causing malfunctions in the electronics (think intermittent power failures to 30 tiny computers). A better strategy is to power them up in stages. Often machine builders neglect this or feel you should just leave it powered up 24x7 and reboot a few as needed. A 1500 kVA transformer would do much better but again cause even worse short circuit issues. So you can use contactors and possibly power things up in stages with timers/PLC logic or get more smaller transformers.
Right, I was reading about a device that delays start up, as an alternate to a line reactor.
Speaking empirically, we seem to get two kinds of errors - one where we can press reset on the drive and the drive resets and we're good to go. And the other where the drive seems to trip, resetting it will not restore and the only option is to power down, power up.
I have monitored THD-V as well - they seem stable between 2-4%.
I have been told that the transformer is undersized by one electrical engineer (out of three). I have been told that that we do have fast fuse protection on the components but I haven't verified.
In any given week we have call it 12-20 dips in the range of 80-90% power, I'll upload a couple of representative ITEC curves.
I didn't understand your comment about a 1500 KVA transformer causing bigger short circuit issues. Could you elaborate?
I have seen peak power at 1200 Amps or so, with 200 ms power at 700 Amps or so.