Limiting inrush current 2

[fsck]

Was reading <

http://www.eng-tips.com/viewthread.cfm?qid=272481&page=1>

with interest.

Has anyone dealt with transformer startup surges & inverter sources?

This installation shall have a 112.5 kva 3ph 480:208/120 transformer at the house, about 1000 ft. away from the utility building (feed).

The utility building shall also house PV-fed grid-tie inverters, and battery-fed backup inverters/generator.

The BU inverters won't be able to source enough current for the transformer's inrush. (The inverter manufacturer stated:

30 Minutes = 6500W
5 Minutes = 7200W
1 minute = 8400W
100 Milliseconds = 150A

(per unit @120V; there will be one/phase.) They don't appear to have any kind of soft-start.

Other people have proposed schemes such as having an oversized generator that starts first, handles the inrush, then shifts the load to the BU inverters, and shuts down again until needed for battery recharge. I'm not enthralled with such, for several reasons.

I spoke to an retired friend who designed & built large [100KW-several MW] UPS's and the transformers within them. (He is the only person I've ever met who really understands both low and high-power transformer magnetics, especially ferromagnetics; vs. reading the textbook, as I did years ago. He holds patents in the area.)

He reminded me the surge is really because you hit the xfmr with voltage at a random point in the cycle, as compared to the residual flux remaining from the point of interruption; if only you knew exactly when and reenergized then, there's almost no bump. (They did this for some of their schemes.)

Custom low-inrush transformers came up, but if I followed his thinking, they have lower efficiency by a significant amount, and far higher price tags; and still may not be low enough for the inverters.

His alternative was to put resistance in series with the inverter output, sized to keep the worst-case current within the inverter ratings, start it, and after the transformer is ...realigned? [not sure what word is best here..]; shunt the resistance out, and then reconnect the loads. This would take milliseconds, but he wouldn't hazard a guess of how many without a lot more data on the transformer in question. (But way below tens of seconds; likely less than a second. His first thought was the pickup time alone of that shunting contactor would be enough, even without an added delay.)

The break in power is acceptable; any scheme I can see that prevents that has offsetting disadvantages. (We can put the non-laptop computers/TV's on small UPS's, etc.)

If possibly viable, I'd propose paying him for a few hours consulting; it would be way less money than any other approach, such as upsizing the inverters.

I welcome comments on this topic.

 

[jghrist]

I suspect that energizing a transformer from a current-limited source like an inverter would eliminate inrush and there won't be a problem. I don't have any experience in this, but I know that inrush depends on the stiffness of the source. Energizing a transformer on a system with low fault current available (high source impedance, weak source) will have lower inrush.

 

[itsmoked]

fsck; Everything you've discussed in regards to your 'friend' is 'right on'. I suggest you proceed with his help.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[ThePunisher]

fsck, shunting the resistance out of an already energized resistance may cause a transient voltage spike across the contactors. This may damage the contacts after several operations.

The series resistance will decrease the total X/R ratio which will decrease the DC component of the inrush as well as increase the rate of decay of the inrush. But the biggest factor it does is that it creates a voltage drop that it will be like starting the transformer at a lower voltage to decrease the inrush current.

There are other stuff that determines your inrush magnitude like the switching the transformer at zero period of the voltage sinusoid and remamence. Switching the transformer with a secondary load of high power factor will damp the inrush a little bit.

 

[itsmoked]

I'm not seeing the damage with shunting..

That is a form of soft starting which doesn't usually thrash contactors any worse than direct starting. Which brings up the point of using a soft stater?

Keith Cress
kcress -

http://www.flaminsystems.com

 

[waross]

Look at this thread with particular attention to the posts by Skogsgurra.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[cherry2000]

FSCK..I did some study on Xmer inrush which depends much more on factors like remnant flux from earlier energization, core dimensions etc. It is an inherent nature of that particular Xmer and can be worse when the switching takes place at Voltage zero. X/R ratio is not that important (unlike in motors where X/R ratio is important). Now the inverter will shut off on overcurrent when the limits prescribed are crossed. Here only the first 100msec are important, since the inrush lasts for 3 to 5 cycles, and dies down. Your friend's suggestions are correct. Intentional insertion of resistance in the circuit will help in dampening the inrush. Even the supply cable of 1000ft can have an effect on the inrush. For actual calculation, one needs to model the system in MATLAB and run a study to see what value of resistance helps in keeping the inrush within the 150A, for the 100msec limit.

 

[itsmoked]

Bill; Your link is as good as my spelling of, "starter"..

Keith Cress
kcress -

http://www.flaminsystems.com

 

[waross]

Look at this thread with particular attention to the posts by Skogsgurra.

thread238-271076


Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[fsck]

waross:
Reading your reference....Very interesting. Are you proposing that this might be a better approach than Lee's?

ThePunisher:
Not sure I follow. By the time we shunt across the resistors, the current should be down to IR where I == the magnetizing current, and R the limiters.

I vaguely recall magnetizing current was ~~1-2% FLA, or here 3-6 amps. To limit inverter load to say 6KW, that's 2.4 ohms of series R; so we're shorting out 14 volts...The contactor will be rated at oh 100A.

We'd actually discussed measuring each drop, and inhibiting shunting until the voltage drop is small. (One of my fears was what happens if one of the resistors opened? I'm now thinking of several elements in parallel.)

Yes, if we had a smart way to connect each phase to the transformer load at zero-crossing; we would avoid as big a surge. But I don't know of such that's affordable.

Cherry 2000:
As for modeling, ISTM the first step is to see if the transformer vender will be forthcoming on data.





(The whole picture is slightly more complex. The grid-tie inverters [not part of the start-up issue] are 277V out. But [grrr..] it appears the inverter manufacturer only offers 120V out in its backup unit models sold in the US. So we will also need a far smaller [~15KVA] 120-277 step-up at the shed between the BU inverters and the 1000' cable input. It will also supply 120v in the shed for misc. use.)

BTW, for laughs we also briefly looked at just putting the stepdown at the shed, and running 208/120:200A the 1000 feet. You can imagine the conductor size needed for *that* approach....(The owner's name is Bill, but *not* Gates.)

I know it seems inelegant, but won't the <50 A initializing current assure success eventually? [Where eventually is "less than hundreds of cycles.."]

One question: Is there any reason to look at a custom low-inrush transformer? (Even they were not low-enough for our needs, or I'd have likely never posted.) Such comes at big initial [$$] and ongoing [efficiency] costs; and I don't see any advantage to same here.

 

[fsck]

After talking with the electrician; he thinks the AHJ may have kittens over a locally built current limiter; even when inside a switchboard.

So I'm seeking a reduced voltage motor starter that will accomplish the same thing. Of course, all I can find are the new SCR-based ones that won't do what we need; grrr..

 

[ScottyUK]

You could consider a 'point-on-wave' switch which would consistently achieve the optimal switching. Not sure if they are available in very small sizes, but maybe a Google of the phrase would help?


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[fsck]

So it's gotten more interesting. First, resistive-type starters seem to be as easy to find as a creampuff '65 Convair van.

The Point on Wave stuff looks to be, well expensive....

But I've found a bunch of work across the pond. It appears the EU has a regulation on startup surges, and combined with the popularity of toroids; there's work on Transformer Start Relays [which are arguably misnamed.] such as Avoiding of transformer inrushcurrent peaks and TSR.

I'm trying to gather more data on that approach.
For example, exactly how do they limit instantaneous current?