24 kA A.C. Source? 9

[Voltswagen]

For UL testing we need to equip our lab with a power source capable of pushing at least 24 kiloAmperes at any low voltage for nine seconds through our 500-MCM product under test, which has a typical impedance around 250 micro-ohms.

Some adjustability of output will be needed, as heating of the sample quickly reduces current going through it, and that minimum of 24 kA must be maintained for the 9-second duration.

The 9-second pulse would need to be repeated a few times in a few hours, so a 9-seconds-on, 30-minutes-off duty cycle or better is fairly important.

Are there any good places to head with this project, beyond ganging up multiple series loading transformers to a variac or several? Bosses are frantic to get started and just bought an 8 kA loading transformer on the assumption that for just 9 seconds, it can handle triple the rated input voltage.

I smell smoke already-- not just the kilobucks already spent, either-- and expect that we'll need at _least_ two more such transformers in series.

Can we feed them from one variac?

Thanks for any guidance.

--Voltswagen
('03 Civic Hybrid)

 

[ScottyUK]

VW,

Assuming that the 8kA transformer delivers 8kA into your load when supplied with rated voltage, you can not just increase the transformer voltage by a factor of three. The transformer core will saturate long before you reach three times rated voltage, and the transformer will draw a colossal current from your supply, either tripping the protective device or damaging your transformer.

The assumption that the transformer can handle 24kA, which is virtually a fault-level current, for nine seconds is questionable. You need to discuss this with the transformer manufacturer - the thermal limitations may be ok, but the mechanical stresses may be too high to be acceptable in a repetitive process such as the one you are describing.

 

[DougMSOE]

First, make sure that your source to feed the xfmrs is capable of supplying the load.
Second, the Z% should be quite low. Arc furnace transformers work very well but again the Z% and the source Z comes into play.
Other methods are there. Rotating short circuit generators are available in US and Europe.
Good Luck.

 

[RAMConsult]

Achieving these short circuit currents in a safe manner is a non-trivial task; therefore I suggest you check the following links to see how others do it:

http://powersearch.cpri.res.in/LABS/sc.htm

http://powersearch.cpri.res.in/stds/1500mva.htm

http://www.ruscan.com/e_products/e_test_short.html

http://www.kinectrics.com/en/serviceline/HighCurrent.html

If you happen to have a spare AC generator lying around that can supply that level of short circuit current you can try and duplicate what these laboratories do.

 

[busbar]

Long shot — location of work unknown, but in St Louis, MO, USA there is “…Paul P. Gubany Center for High Power Technology… supports tests up to 300 kA and 640 MVA for fault evaluation.”

http://www.bussmann.com/services/gubany/

 

[DanDel]

There are LV circuit breaker test sets which can routinely deliver more than 60kA for several seconds. The problem is the impedance of your load. The impedance of a closed C/B contact is typically at the micro-ohm level, and the output of the test set is usually 15V maximum. What length of cable are you using?

 

[busbar]

If I hit the right keys, about 1 volt is needed for compliance, but for some, nine seconds is close to an eternity for 24kA. It is stated that 250µΩ is impedance and not merely resistance, so the lead/bus makeup and extreme reactive requirement must empirically be very familiar. Carefully defined “impedance matching” is crucial.

Besides Gubaney Lab, I have heard that Ferraz-Shawmut has some in-house high-current capabilities, but I do not know in which country. A possible solution might be a NETA-member firm {netaworld.org} that could arrange for the testing—for some have onsite primary-current injection breaker-test capabilities and operators/technicians which can do this testing like apes in a coma.

Unstated is whether the current must be at a particular line voltage that would be needed for effectively driving an interrupting overcurrent device.

I would carefully study the proposed thermal time-current capability of various test equipment, and define a performance spec before any work is authorized. An example of test equipment might be

http://www.megger.com/common/documents/DDA3000_DS_En_V01.pdf,

but multiple series/parallel sets may be needed for the specified time duty and current amplitude.

Another possibility is finding the right person at Burndy, T&B or Penn-Union that has contracted their IEEE Std 837 tests on grounding components. AB Chance claims to do in-house ASTM-type fault-current testing on their line-maintenance products.

 

[electricuwe]

The solution to your problem is quite simple, it's called Ohm's law:

To provide a current of 24 kA in a load of 250 µOhm Impedance you need a voltage of 6 V. A transformer designed for 8 kA continous current should be able to deliver three times the current for 9 seconds without damage.

The only problem is to built a low inductance arrangement so that inductive voltage drop will not increase the circuit impedance.

 

[ScottyUK]

electricuwe,

What you say is correct in a simplistic sense, but do you see a problem with this sentence copied from the original post?

[blue]Bosses are frantic to get started and just bought an 8 kA loading transformer on the assumption that for just 9 seconds, it can handle triple the rated input voltage.[/blue]

What happens when an iron-cored transformer is fed with three times rated voltage? Overfluxing of the core? Supply circuit breaker tripping? Transformer primary winding burning out? Fire?

Ohm's Law is simple, but there are some real-world issues to address too.

 

[busbar]

Yes indeed — 6 volts compliance and not 0.6V.

 

[jbartos]

Suggestion: Visit

http://www.thomasregister.com

and type Transformers: Custom under Product or Service, which will return 224 companies to approach to for a suitable custom made transformer 6VAC, 60Hz?, ~150kVA, 24kA for 9 seconds secondary ratings. The secondary conductor size will be in thousands of MCM, however, there will be very few turns.
Considering a large amount of money being charged for UL testing, it should be a good investment. Maybe, the 8kA transformer would be possible to trade in.

 

[Voltswagen]

Many thanks for all your responses and the benefit of your diverse thoughts on this project.

I probably misused the term Impedance in my post, assuming it meant merely alternating current resistance, and should have simply stated the typical resistance of the sample under test, roughly 250 micro-ohms and usually less than that.

Most samples are composed of a clamp or connector joined to one or two pieces of stranded copper or aluminum conductor usually not longer than 12 inches. These are typically clamped into aluminum terminal lugs bolted to heavy copper busbars 24 inches long, connected to the power source.

Over several years I have brought various, often unconventional challenges to the great team at the High Power Test Lab of Ferraz-Shawmut in Newburyport, MA, and they have always come through with a way to do what we needed. (Recently they showed us how much damage just a few cycles' worth of 20 kA current can do to our products at 60 Hz., compared with the 8 or 10 microsecond rise time of lightning surge tests they have performed for us many times.) We have simply reached the point where in-house testing to the straightforward, conventional UL and other specs is the only way our product development will progress at a reasonable pace, and the value of being able to perform prompt, informal, but decisive trials on prototypes is inestimable.

More than one posted response hints at generators, which we have discussed in the past for both test-lab service and staying in business during mains outages. Surely we would still need transformer(s) between generator and lab; does the generator approach simplify achieving the current values required in our testing?

We keep our test leads as short as possible, usually bridging the test sample between massive copper bars that provide only about two feet of distance (and heat sink) from the primary current injection unit, Oden AT:

http://www.programma.se/odenat.html

we have been using for several years. Even so, we seldom can get more than 10 kA through our samples with it, despite its 21-kA max output rating, which is apparently more theoretical than actually achievable. For this reason we do not bother trying to calibrate it higher than 10 kA. I will contact the manufacturer in Sweden about the slim chance of adding more power modules to the three we already have, but have little Hope that it's a viable option.

Busbar's link to Megger's pair of circuit breaker test sets is intriguing and just might be adequate. Attractive is that the entire system design has been worked out so we are not likely to damage anything from failure to select and connect compatible components, and energize them properly. Our Oden AT is on its second fine-adjustment rheostat, though; we were urged to keep that set at zero and use only the coarse adjustments (SCRs?) during short-term high current tests.

I'll call for a quote on the bigger Megger, already bracing for sticker shock. Big question is how the +6 kA max continuous current (50 percent duty cycle, 30 min. ON; 30 min. OFF) prorates at 9 seconds if its max current through a breaker is 60 kA and through a short circuit, 100 kA. It just might be beast enough for our needs. Input current is 350 amps.

Electricuwe mentioned:
> The only problem is to built a low inductance arrangement so that inductive voltage drop will not increase the circuit impedance.

Aside from superconducting busbars, twisting together lead cables of different polarity to reduce voltage drop and keeping them as short and heavy-gauge as possible, how else might inductive losses be minimized?

Seriously; are there alternatives to solid copper busbars, like carbon, that might be less of a bottleneck between the innards of the power source and the test sample? We made heavy flexible braided copper replacement leads for the Oden; is there any benefit from high surface area conductors like that as opposed to solid bar stock, beyond flexibility and perhaps better convective cooling?

Thanks again for the helpful and thought-provoking posts, every single one. We welcome any additional links, leads, questions and comments at any time and will summarize here any conclusions reached.

 

[electricuwe]

ScottyUK,

please consider the following: For getting a current three times the nominal current from a transformer it is not neccessary to apply three times the rated voltage !

And if this should happen by accident a circuit breaker or a fuse should clear this fault without problem.

 

[busbar]

Volts — As a rough example, I am familiar with test transformers with 2-turn, 3 x 0.375-inch secondaries. They are nominally rated with a secondary open-circuit voltage of 10, and a continuous short-circuit current of 2kA, short-time 20KA at about 175 pounds. [Clearly there is some specsmanship involved, for the transformer probably could not serve 2kA at 10 volts.] Secondary-loop reactance can be murder for anticipated high currents. It seems like multiple series/parallel units could be feasible.

http://www.phenixtech.com/products/Circuit_Breaker/brochures/30205-HC16C-HC30C-HC100C.pdf

http://www.phenixtech.com/products/Heat_Cycling_Test_Sets/brochures/heat_90402.pdf

Also, depending on time and budget, custom-made transformers with your own in-house control gear might work.

 

[ScottyUK]

Hmm electricuwe,

For a given load which of a defined impedance, getting 3x current through it does require 3x voltage. Or is Mr. Ohm wrong? He will be disappointed after all these years!

Of course if the transformer voltage is variable in the region below rated primary voltage, or the load is variable, then you are absolutely correct. I had assumed that the load was of a fixed value, through which the transformer noted was able to drive 8kA. In this case, I am correct.

It all comes down to interpretation of the information presented, which is what makes Eng-Tips so interesting!

 

[jbartos]

Suggestion: Visit

http://www.isomatic.co.uk/welds.htm

for welding transformers up to 67kA

 

[Voltswagen]

The helpfulness and scope of your posts are overwhelming and we appreciate every one.

Semantics are 90 percent of searching so the links from busbar to sites dealing in circuit breaker test sets opened up an entire area that we would not have found in our searches for power source, power supply, etc. I might soon fly to Dallas where Megger has one plugged in so we can verify that it will zap a few of our test samples with enough current for long enough to satisfy the UL and other requirements.

jbartos also linked to a site with welding transformers which also seem right in the range of what we need, which includes lots of overkill: we need 24 kA today, so the least we should actually buy or build realistically is, like, 60 kA with up to 100 kA short-circuit, and even that might not be enough someday.

DougMSOE also mentioned furnace transformers, which indeed one of our foundrymen runs with his own generator and are a possible solution.

I like this page, thinking outside a bigger box:

http://proof.vision-marketing.com/neeltran/power.htm

ScottyUK and elecricuwe, my apologies for abusing the term IMPEDANCE when I believe I meant simple resistance, but please don't halt your dialog; it is helpful to me, anyway.

Thanks again, everyone. My bosses are appreciative of the knowledge and help you all have shared so generously, and my first requisition that I'm sure they will approve for this project will be a donation to this forum.

As ever, any other ideas or questions are appreciated. I'll report what we learn and decide.

--dave
Voltswagen

 

[Voltswagen]

Update 12/5/03~

Last week we went to Harvard where their power distribution team (they bought and maintain an older steam & electric generating station and serve many of their buildings with them) generously demoed the biggest Megger circuit breaker test set, rated at 60 kA with 100 kA short-circuit capabilities, on some of our typical test samples.

It was an impressive device-- it fuzed open a 350 MCM stranded copper wire about a foot long, with one of our grounding connectors attached at one end-- in 3 or 4 seconds at 30 kA, and melted an outer layer of strands on 1000 MCM 61-strand aluminum conductors on each end of a 3-phase connector, in 1.8 seconds at 58 kA! I'll email a few jpeg photos of the tests to anyone who sends me an address. <dave@electricmotioncompany.com>

They have been very pleased with their Megger (after a bad series of experiences with a Hipotronics test set that they mothballed), but had to move it to a facility with 460- or 480-volt service because it wasn't quite as capable at the gen-station with its 208-volt limitations; I think they said they had to buckboost it, and they felt that extra stuff like that on either the input side and especially the more obvious problems of excess conductor lengths and sample resistance on the output side place big limitations on what the thing can do.

We're going to ETI in Baltimore next week for a demo of their 60 kA breaker test set. Unlike other units we've studied, the ETI can automatically adjust output as needed to compensate for the rapid increase in sample resistance as it heats up. Like the Megger, it has 9 taps for coarse adjustment, which cannot be changed during a test, so the main question we hope to find a positive answer to next week is whether the ETI has enough range of auto-adjustment of output to handle the amount of resistance increase that accumulates over the nine required seconds of current, 14 ka for 500 MCM products and 24 kA for 1000 MCM.

The ETI seems to have a lot of extra features that promise to be handy for documenting tests, including the total rms power output through the sample during the test, rather than just a peak max current or an average.

Even at this 60-kA level, these robust devices do not appear to have a lot of redundant or extra capability to perform our most demanding tests. Our typical sample resistances of roughly 200 to 250 microohms are much higher than the breakers normally tested by them, which are rumored to run more like 30 or 40 microohms, and which resistance, we are informed, is best checked at 100 amps rather than the 10 d.c. amps all of our lab microohmmeters are limited to. But a Megger DLRO-10 we have reads to a resolution of 0.1 µO, while the others only report to 1 µO so will the Megger will have to do for now. We're open to any and all suggestions of a good 100-amp microohmmeter for a lab environment.

They claim we can run the ETI off of a 200-amp breaker, even though the input current can go up to 2106 amps briefly (one second, with 2 minutes to cool off afterward) in an overload condition (testing at 60 kA output). We'll therefore be looking carefully at our service entrances and pad-mount transformers.

Thanks again to all the diverse views on this challenging project. I feel strongly that we are going to do this right the first time, carefully but steadily, as we must due to both time demands and financial consequences of buying the wrong equipment, and the wisdom and depth of knowledge offered here so generously has been beyond anything we could have bought or placed a value upon in any way. If we can someday return the favors with the (generally NIST-traceable) capabilities we end up with, I would be glad to attempt do so to the extent practical.

(--dave)

 

[Laplacian]

You could easily get that kind of current from batteries, however controlling that current would be a different matter.

For example, 10 parallel strings of 3 each Enersys Powersafe GU45 can output ~30KA for one minute to a discharge of 1.75vpc.(

http://www.enersysstationary.com/documents/51_45_GU.pdf)

These are stationary batteries and are designed for float service. If you did many 9 second bursts on them in a day, they wouldn't last long. If your testing requirement is such that one 9 second burst per day then they may last 10 years. Turn the chargers off before testing though.

The batteries would hold pretty steady at 6.5vpc @ 25KA for 9 seconds, but you would need fast variable resistance to hold current constant as your test subject changed resistance. Adding more KA's would mean paralleling more strings. You could size chargers reasonable for your utility supply if your testing isn't too frequent.

The only other feasible option without purchasing your own co-generation facility would be to parallel several high power, low voltage rectifiers and control them via PLC.

It must be an interesting position you are in to come up with new testing facilities. I'm at a loss to come up with a way to economically control so much current.

 

[jbartos]

Comment on the previous posting: The lead acid battery internal resistance or impedance is nonlinear. It causes the battery short circuit current peak or overshoot relatively steady short circuit decreasing discharge current by about 20% (in about first 100 milliseconds range). This can be captured on a fast oscilloscope screen. The 20% inaccuracy may be subject to evaluation or reconsideration for the unit under the test.