Arc Flash Reduction Methods On Secondary of Sub-unit 4

[majesus]

I've seen this application time and time again, but I would appreciate any advice and comments on my findings for the following project I’m working on.

My project is a new 3MVA dry type sub-unit installation (step down 12.47kV to 600V) for a Grain Processing Facility.

Approximating the Impedance Data for the XFMR based on "Cutler-Hammer Reference Data" (p.1.5-11 of the 2006 Consulting Application Guide). The dry-type XFMR (80C rise) is modeled as having:
Z=5.75%
X/R=10.22

Note: The following is based on the secondary side.
The short circuit current is about 58 kVA because the Grain Processing load (on the secondary) is about 95% induction motors. (Largest motor is 400hp.) The load current is: 2700 A.



Modeling in Etap, the secondary of the XFMR is assumed to be a switchgear type equipment with an air gap between conductors of 32mm.) My simulations calculate an Arc Flash energy of 330cal/cm2 which (exceeds the Cat 4 level.) This is what I expect.

Now my design goal is to install a secondary protection device to bring the arc flash hazard down to a Cat 2. GOOD LUCK, because this is where I am having my difficulties:

Option 1) Secondary Protection Using Power CB
Power CBs (with ratings greater than 3000A) have typically an instantaneous 4 cycle clearing time. Provided that I coordinate the CB so that the arching current will trip in the instantaneous region, this brings the arc flash to a Category 3 (about 9.66 cal/cm2). (This is the best that a CB can do.) I've been looking at Schneider's Masterpact NW and NT Circuit Breakers with Arc Flash Protection that have faster clearing times, but they are available only up to 2000A. My CB options seem limited.

Option 2) Current Limiting Fuses
In the past, I’ve done research looking at the fastest fuses available for applications that require load currents between 1-5kA. The Bussman KTU fast acting fuse is the fastest available in a Class L format. Using a fuse rated above the XFMR’s secondary current, even with this protection device, the arc flash current flowing through the fuse lies in the clearing time’s tolerance range that using the worst case scenario, the arc flash is not significantly improved. I can always select a smaller current fuse, however, it will conflict with the XFMR’s load capacity.

Option 3) Increasing the Impedance of the XFMR.
I’ve simulated the XFMR as having an impedance of Z=7.5%. The arc flash reduced slightly, but was still in the Cat 3 region.

I’ve looked at the very interesting document “Distribution Equipment Modernization to Reduce Arc Flash Hazards” obtained from here:

http://www.eng-tips.com/viewthread.cfm?qid=203769&page=14

Their scenario is similar to mine. I have simulated their SLD (Figure 2 on page 29) and I’ve obtain similar result as them. However, their SLD has smaller secondary load then mine.


So now I'm stuck as I excercised my options. I would appreciate any comments.
Thanks,
Maj

 

[DanDel]

I expect that you meant your short circuit current is about 58MVA, not 58kVA.

It is typical to have high AF values at transformer secondaries, due to the impedance of the transformer. Increasing this impedance will usually make the AF worse, since it will limit the through-current to a point higher on the upstream device's TC curve, causing a longer clearing time.

Though I believe that Square D/Schneiders make NW/NT C/Bs up to the 4000A range, and the addition of this device will improve the AF result downstream from this C/B, you will still have the same AF on the line side of it.

Fuses are always the quickest-operating, and will provide the best possible AF result (typically), but as you said, it may not be the best device for coordination, and you still have the same problem on the line side.

Time to look at you options and decide what is best for the customer and the system. I personally don't like to sacrifice system coordination for lower AF values. I would instead insist that this particular equipment not be worked on when energized. After all, that's the safest alternative.

 

[dpc]

I think your analysis is probably pretty accurate. Of course using an upstream fuse or breaker to achieve this will result in a loss of coordination with the downstream feeders. And you will have just pushed the 330 cal/cm2 energy level back up to the new breaker or fused switch you are putting in.

Current limiting fuses sized for the max transformer rating probably will not help much.

Getting down to 8 cal or less may not be achievable with that large a transformer.

If you have a local primary breaker on the transformer, another option is to install an overcurrent relay on the 480 V side of the transformer and trip the primary breaker.

So there is no existing main breaker?

 

[majesus]

Yes DanDel, my mistake 58kA, not 58kVA, just automatic hands typing away.

I came to the exact same conclusion as you, but I wanted to test the waters to see what alternatives are out there.

A few things to note (just discussion, not arguing)
"Increasing this impedance will usually make the AF worse, since it will limit the through-current to a point higher on the upstream device's TC curve, causing a longer clearing time."
True, if you were using a fuse or (CB and going from Instantaneous to Short Delay), but if you were in the CB's instantaneous range for both situations like I purposely was, then the later decreased the AF (just a note on what I was doing.)

"Though I believe that Square D/Schneiders make NW/NT C/Bs up to the 4000A range, and the addition of this device will improve the AF result downstream from this C/B, you will still have the same AF on the line side of it."
YUP, you will only have protection upto the protective equipment. One thing, their AF CBs seem to be only rated to 2000A.

http://ecatalog.squared.com/techlib/displaydocument.cfm?id=0613CT0001&action=view

Page 155

Maj

 

[Zogzog]

You can retrofit your breakers with Qucik Trip, allows for .05s clearing with out coordintaion issues.

http://www.utilityrelay.com/QUICK-TRIP_Page.html

 

[majesus]

DPC no secondary main breaker.
Primary is GE Mutlin 575 at the service feed.

The story about this is that the existing XFMR is 3 single phase 1000kVA XFMRs built in 1923. They are massive. 10 ft high, liquid filled, self cooled. The incoming and secondary feeds have very interesting exposed bus work. I really like these XFMR because they are such antiques and been around well 3.5 times longer then myself.

The XFMRs still work, and have no PCB in their liquid. The client wants to remove them because these is no seconadary CB and they are old. Honestly, I'd leave them. They have been around for so long, why bother? If they aren't broken, why fix them?

But that's me.


Here is a photo of them.

http://i48.photobucket.com/albums/f223/majesus/DSC00686.jpg

If I stand next to them, my head reaches a little higher than past half.

 

[majesus]

You can retrofit your breakers with Qucik Trip, allows for .05s clearing with out coordintaion issues.

http://www.utilityrelay.com/QUICK-TRIP_Page.html

Pls correct me if I'm wrong but according to their manual:
THE AC-PRO is a state of art micro-controller based trip unit for use on 3 phase AC breakers."

http://www.utilityrelay.com/PDFs/I-AC-PRO-QT-1.1.pdf

Basically it is a relay that offers different coordination setting for the CB without having to open the case.

If the ARC-flash is already in the the CB's instantaneous region, then this relay doesn't help. The quick-trip website shows a total clearing time of 50mS in their brochure. That is misleading, as those Curves aren't dependant on the QUICK-TRIPT product, but on the CB that is in use.

If you know of any Power CBs that have a clearing time less of 50ms or less (rated for 3000A or greater) I would love to know.

I heard ABB makes some and I shall investigate them after lunch.

 

[slavag]

WOW, Thanks a lot Maj for the photo.
Best Regards.
Slava

 

[WDeanN]

330 cal!! What’s the boundary on this? Somewhere in the next county?

Do you have the option to split the load between transformers? If you can use 3 1000kVA Xfmrs, you will be much better off.

I'd suggest the ABB REA flash detector (

http://www.abb.com/product/seitp328/adfc30c6b75a80bcc1256e4500329ee2.aspx

), but I don't think you'll beat 4 cycles.

Comb your calculations and "sharpen your pencil" as my supervisor says. Look for anything in your data that may be off, or that you can adjust, even with a caveat. Look at scenarios, motor contributions, working distances, etc closely. Although 300 cal sounds about right for 3MVA. Good luck getting that down below 8 cal!

 

[ghostbuster7]

Great Info

I have a very similar situation with approx.160 cal at main switchboard.Curves are very tight.This is a main data processing centre,that can't afford intermittent primary breaker tripouts to satisfy arc flash req'ts.

I am seriously looking at working on this switchboard in dead mode only.The site has lots of back-up diesel generation(not fed through this board) .

 

[majesus]

330 cal!! What’s the boundary on this? Somewhere in the next county?

Well like the article Distribution Equipment Modernization to Reduce Arc Flash Hazards that I mentioned in my first post says:
"At the mill's 470V Bus D we calculated energy values of hundreds of calories per square centimeter and flash hazard boundaries that put personnel completely out of the building!"



Comb your calculations and "sharpen your pencil" as my supervisor says. I LOVE THAT!

 

[unclebob]

I noticed a Cemco/Federal Pioneer CB on your picture (50H , 75H may be ?). Are they using these type of breakers on the 480 V side?

Like DPC, I'd go with a relay on the 480 V side of the transformer and trip the primary breaker.

 

[majesus]

Unclebob you can see that! All the way in the corner of the photo, so tiny and you made it out!!! Holly smokes, I am amazed... truely AMAZED! Yes it is a Cemco/Federal Pioneer CB made in Vancouver a long time ago.

Yes they are still using them all over the plant on the 600V side. That one was just sitting there for god knows how long.

 

[DanDel]

This looks a lot like some network systems in my area. (The buswork coming from the transformers is connected to a fused switch and then a network protector.) No one would expect the AF levels to be low.

What is the urgency of this particular system connection point to be <8cal? Is it just for oil sampling while the transformers are energized?

 

[Zogzog]

Doubt it, this is a dry type transformer

 

[DanDel]

Say what, ZogZog?

 

[DanDel]

Oh, I was talking about the originals in the picture. Sorry.

 

[majesus]

No urgency, < 8 cal/cm2 is the design criteria to avoid wearing combersome PPE. If do-able good, if not then the the risk is addressed and procedures to deal with it are noted.

 

[majesus]

An update:

It turns out that the Arcflash Reduction Maintenance Switch from Cutler Hammer is what is needed. The odd thing I initially looked at the ARMS and I quickly assumed that the ARMS simply turns the instantaneous setting down to a lower value. Where I then went to the Magnum CB’s TC curves and look at their instantaneous total clearing times. Yup, it was approximately 60 mS.

However directly from the Arcflash Reduction Maintenance Switch brochure:

http://www.eaton.com/ecm/idcplg?IdcService=GET_FILE&dID=158535

Myth: All it does is turn the instantaneous setting down
to a lower value."
FALSE: Unlike other products the Arcflash Reduction
Maintenance System involves a separate circuit that acts
faster than the standard "instantaneous" protection.
• Arcflash Reduction Maintenance System is an analog circuit
which provides faster interruption times versus the
digital "instantaneous" protection (~20ms faster).
• Faster interruptions = less fault current let through = less
arc flash energy.



The total clearing times is approx 40mS as shown here in the TC curves (Document 70C1446.PDF):

http://www.eaton.com/ecm/idcplg?Idc...sed&Rendition=Primary&&dDocName=1138850418129

PERFECT! THAT brang my system down to a Cat 2

Which leads to a question (if anyone can answer it): why isn’t this analog method not incorporated into the standard protection scheme to give a faster instantaneous clearing times during normal operation? I’m guessing there is a compromise ie, lack of multiple settings that is possible with the current digital instantaneous system?

 

[alehman]

Thanks majesus. I was under the same impression. In fact I think my salesman presented it that way.