800 Amp Circuit Breaker Failure 1

[jktwn]

Two days ago, we experienced a failure of an 800 amp Square-D moulded case circuit breaker that vaporized the A phase stab connector and bus bar connection. Three large production lines were down as a result. Three IR inspections in the last year (most recent 10/01) failed to identify a hotspot on this breaker leading me to question the failure mechanism in this case. The breaker is in a power panel that houses a 1400A main breaker.

The 480V, 3PH loads on the 800A breaker consist of six 60HP fans, six 1HP fans, three 2HP fans. Automatic sequencing is not provided in the startup of the equipment relying instead on the good sense of the operator to start things correctly. The equipment was installed about 23 years ago.

Has anyone seen a similar failure in a large moulded case CB? I have pictures if they will help. By the way, I plan on sending the breaker back to the mfg for failure analysis.

Thanks

 

[SphincterBoy]

We detected (IR scanner) of a potential molded case circuit
breaker, and were able to change out the breaker.

Since your IR scan did not show any hot spots, the insulation may have simply worn out. You probably
should subscribe to the Doble Engineering (Watertown, MA), and begin performing annual power factor tests for the insulation for equipment older than 15 years. If you
are not familiar with Doble, then Doble can perform
extensive tests of all your equipment for you.

Aging insulation is a huge problem, and is not detectible
with IR scans.

 

[electricpete]

Some circuit breakers have a tulip-shaped female connector with a spring that holds the petals of the tulip tight on the stab. If the spring breaks, it will pop off and failure will occur shortly. I was involved in investigating a number of such failures among a very large population *US military). All of the failures occurred among the same vintage, even though breakers of all ages were in service. It was apparetnly a bad batch of springs. Metallurgical analysis of recovered failed springs (they popped off and lay intact at bottom of cubicle, in spite of devistation in the area of the stab... helping to prove that they came off at the start of the failure) showed fatigue failure and possibly inferior material

IEEEC37.10-1995 is IEEE guide for circuit breaker failure evaluation

 

[mpparent]

I myself have been involved w/ failure of the "rosette" springs that electricpete is talking about. I would wager that's what happened in your case.

 

[jbartos]

Suggestion: Visit

http://www.clockwork-group.com/enricosum.htm

See Square D reliability Main Pro's

http://www.squared.com/us/services_...inPro Brochure.pdf/$file/MainPro Brochure.pdf

(The referenced web may need to be copied into the browser)

 

[jktwn]

Thanks for the responses.
Our infrared thermography inspection program has been effective with only an occassional miss and never one causing this level of downtime. Thus the opening of an RCFA investigation.

belectricpete says: Some circuit breakers have a tulip-shaped female connector with a spring that holds the petals of the tulip tight on the stab.

This thermal-magnetic breaker uses what Square D calls "blow-on" type connectors. They claim that under short-circuit conditions the increased magnetic flux causes the plug on connectors to grasp the panelboard bus bars even tighter. I doubt that this feature adds any meaningful additional connecting force during normal high current events such as motor starting. The connectors are a 5-pronged in-line design made to slide onto a panel bus bar.

Additionally, we have had a couple of these same style breakers in the same panel that would not trip off with the handle. Now we are adding PMs to exercise the breakers on 12 month intervals. Even if the breaker won't turn off with the handle, it should trip under overcurrent conditions, should it not?

 

[electricpete]

If it was a failure totally confined to one phase with no evidence of arc strike to adjacent phase or to ground (is that the case?), then it seems to me the only alternative is a high resistance connection. Based on your description of the failure location, it seems we rule out high resistance main (interrupting) contacts, and focus on the stab contacts (unless there were also some bolted bus contacts within the area destroyed by the fault).

I don't understand exactly how your connector is held tight onto the bus bar, but you should consider if there is anyway that it might loosen or become contaminated. In the case I described they were held together by spring and a breaking of the spring caused a SUDDEN loss of contact pressure.

 

[jktwn]

Clearly, the damage was limited to the A phase stab connection on the circuit breaker. There was no evidence of a phase-to-ground or phase-to-phase short circuit. I agree with you that a high resistance connection at the bus bar to breaker interface was probably the cause of the failure.

The upstream main panel breaker rated at 1400A and the substation breaker rated at 1600A have tripped many times in recent years but never this 800A breaker leading me to suspect that the breaker in question was defective. I suspect that this connection was stressed over many years of operation (these motors often are restarted several times per day) and thermal cycling of the connections eventually did it in.

Hopefully, as this forum develops, a means of embedding images will be developed. In cases such as this, where the connection is difficult to visualize, a picture really is worth a thousand words.

 

[JBD]

What you are describing is a Square D MA frame (800A max trip)I-Line construction breaker and associated I-Line bus bar system. I believe this construction was introduced in the mid 60's. You are correct in that the blow-on design does nothing for non-short circuit currents.

This is from my personal experience only. I have been involved with Square D breakers since entering the field in 1978. I have not seen, or heard, of enough unexplained "stab" failures to question the integrity of the design. Most failures have occured after a breaker was removed and then reinstalled in a panel and no new Plug-on Joint Compound was used, per the manufacturer's instructions. The second common cause of failures has been severe cylical (shock) loading similar to electric welding applications.

All switching mechanisms should be exercised at least yearly. While I have not experienced any OCPD which has not "tripped" due to age, I have seen many that could not be reset or operated manually.

 

[jktwn]

They are on the way. Thanks!

 

[electricpete]

Here is a link to the pictures:

http://geocities.com/pschimpf/temporary_stuff/bkrpics.htm

 

[hendo]

wxashcra,
Could you please email to me the pictures that you have taken of the damage to the mccb.
We have had a similar problem with damage to an mccb while clearing fault levels below the design parameters of the mccb
Regards
Hendo
powerex@tpg.com.au

 

[electricpete]

hendo - Please let me know if you had any problems in accessing the pictures that I posted at the link above... it works fine for me.

 

[peebee]

Wow, wild pics. Blammo.

My guess is the plug just got old and a little loose, maybe started walking off the bus from the daily thermal cycling after 23 years.

Maybe this goes without saying, but I'd recommend bolt-on next time. . . .

 

[jbartos]

Suggestion: Have you performed any root cause analysis?
Similar malfunctions are attributed to:
1. High connection/contact resistance/impedance
2. A combined effect of the overvoltage and lowered surface resistance/impedance.
3. Some crawling creatures could cause this heavy arcing-like damage.