Open Circuited CT Failure Mode 3

[MxGxK]

I need some help. I work for a large industrial plant and have seen a large motor trip that is being attributed to an Open Circuited CT. I would post pics but the company prohibits it.

The CT was located in the connection box and is not used for protection. It was connected to a current transducer and then to a DCS system.

The relay indicated a ground fault using a 3phase core balance CT. This, I do not question.

When the connection box was opened, we found a CT (purpose described above) laying in the bottom of the connection box. The phase cables were hanging and half of the lugs were burnt off. Copper droplets were found on the arrester below. The lugs were connected to the neutral bus bar (ungrounded) and located ~6" from the CT. The micarta (sp?) board around the CT was burnt very bad, more damage on the lug side. The side of the connection box was burnt between the lug and CT, not on the opposite side of the CT. Six inches of the insulation was burnt back from the lugs. The DCS loss of current coincided with the relay trip time.

Question: Did the lug get hot or did we have an open circuited CT?

There was no evidence of arcing or burns on a Weidmuller block were a loose wire (CT Sec) was said to have been found.

Question: How would the CT fail? Could it cause fire at open circuit or does core get hot and melt?

Any thought would be appreaciated.

 

[scottf]

Tough one to answer without more details....

What are the ratings of the suspect CT, i.e. ratio, accuracy/burden rating, etc...

What was the steady state phase current prior to the fault? What was the fault current?

 

[MxGxK]

The suspect CT is 800/5 with ~600Amps primary current. I don't know the rating but I imagine C200. Saturation shouldn't be a concern because of function (my assumption). Phase currents appeared to never exceed FLA. The motor tripped on ground fault when 3ph current sum exceeded 20Amps.

 

[scottf]

Tough to answer for sure, but it doesn't sound like an OC CT.

First, the fact that the DCS lost current at about the same time as the fault means the chance of an OC CT causing the fault is remote. OC CTs can generate a good bit of heat, but not that much and not that fast. The heat generated comes from the CT being in deep saturation, i.e. very high core losses.

I'm having a tough time visualizing the set-up, but based on what you've given, I would bet the the neutral/lugs heated up and that may have caused the CT to fail. I doubt a CT failure caused what you describe...but impossible to say for sure.

 

[MxGxK]

Further explanation of setup.

MV leads pass through Diff CT's (directly above subject CT) on each phase. Neutral end of leads pass back through diff CT's from motor and make turn down to phase CT. After passing through phase CT, the lead terminates on neutral bus bar. This is all in connection box.

The phase CT hangs from top bolts only. When it melted, the core fell out of the case. I can't seem to accept that the lugs would have been burnt off (leaving the 2-hole NEMA pads and one of two crimps bolted to neutral bus)if the CT had open circuited. Assuming a fire results from the OC CT, the primary cable insulation would have been burnt off and possibly resulted in ground fault to the bolts securing CT (nearest ground to phase conductor in CT).

Maintenance personnel are ready to say 'CT Open Circuited' and move on but I'm not sure.

What have you guys/gals seen with open circuited CT's? Arcing at terminals? Winding to winding failure - OC or SC?

My main opposition tried to demonstrate test for me us a test set for testing motor overload relay (current generator of some sort). First, he wanted to demonstrate that 3.5Amps would not cause arcing. So he shorted the leads and generated 3.5Amps. When lead was lifted, there was no arc. I witnessed this but don't believe it to be accurate test. Wouldn't CT secondary current be inductive and be less willing to be extinguished than the test set current? Next, he attempted to drive primary current through a sister CT. As soon as the secondary was opened, the test set current would fall off to nothing. He nor I could explain this. Therefore, nothing was proven. Any thoughts?

 

[waross]

I had an instructor with a lot of industrial experience. He spoke of several "CT opened under load" events. He said that in each case the last thing the electrician saw for several hours was a "BIG ORANGE BASKETBALL".
The last set of CTs I saw tested saturated at around 700 volts on the 5 amp winding. Some CTs will generate much higher voltages if the secondary is opened under load.
I would guess that the test set does not develop enough voltage to saturate an open circuited CT. You were both very fortunate that your test failed.
But, your problem may not have been related to a loose wire on a CT. Possibly a loose cable connection on the neutral bus. Loose connections generate a lot of heat. Copper is an efficient conductor of heat. I have seen cables fail to ground because of heat softening of the insulation, quite a distance from the loose connection that was causing the heat.

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

 

[electricpete]

First, he wanted to demonstrate that 3.5Amps would not cause arcing. So he shorted the leads and generated 3.5Amps. When lead was lifted, there was no arc

When you say he generated 3.5 Amps, do you mean by applying say 40*3.5 Amps to the primary ? (if 40:1 CT). That would be a good test. Just applying a voltage source to the secondary will not recreate the high voltage upon open-circuiting. Yes, there may be some high voltage from open-circuiting the inductive secondary, but that is a whole different animal. For one thing, the inductive kick from open-circuiting an inductance dies out when the secondary current goes away. The high voltage from an open circuit CT secondary won't go away as long as the primary current remains. For another thing, the inductive kick from open circuiting depends on how quickly you open the circuit, the high voltage from open-circuited CT secondary does not.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[scottf]

The voltage present on the secondary of an open circuit CT is a function of the inductance of the winding (why I asked you for ratio and accuracy rating) and the primary current level.

A rough guess would be that for an 800:5A CT with a C200 rating and 600A in the primary would develop about 10-15 kVp. This could cause arcing, but normally the arcing doesn't have enough power behind it to cause the damage you mention. Reason is the RMS voltage is not that high...just the peak is very high, as it's nowhere near a sinusoidal waveform, but rather highly distorted.

Could you please confirm again that the CT you suspect as being open circuited fed the DCS system and that system lost current input at the same time as the ground fault? To me, if this is true, it all be rules out the OC CT as being the root cause.

 

[oldfieldguy]

CT's are pretty simple things to look at but their actual application and operation can get pretty interesting when things go wrong.

One of the more common failures of CT's causing open circuits is the failure of connections. If a poor connection develops high resistance, the CT still tries to "push" current through it, producing heat, which tends to make the connection worse, which produces MORE heat, etc., until the circuit opens, then you may have arcing and combustion byproducts. In the confines of a motor terminal box, this may lead to big problems with flashover, and may have resulted in your incident.

The guy who "demonstrated" the "3.5 amps won't cause arcing" misses the point. Opening the secondary of a CT while primary current is flowing turns the CT into a step-up transformer with a one-turn primary and an XXX turn secondary. Secondary voltages rise to a high value immediately and the weakest insulation in the system fails.

If you want a real demonstration of the open circuited CT, connect a current lead through the window of the CT, pass a considerable current through the window with a clip-lead on the secondary. Wearing voltage-rated gloves and faceshield, yank the cliplead off the secondary terminals. That should give you the arc.

I recommend strongly against Weidmuller or other spring-clip style terminals in CT circuits because too many technicians do not install the conductors in them correctly. Call me old-fashioned, but with CT's I prefer screw-type terminal strips with visible access to the connection, and properly crimped ring lugs.

Further, I have seen a few CT's which open-circuited due to external damage, i.e., the dropped crowbar syndrome.

Not to point fingers, but I have seen more than one CT installation messed up by the instrumentation and control bunch because they do not understand the way CT circuits operate.

old field guy

 

[waross]

If you want a real demonstration of the open circuited CT, connect a current lead through the window of the CT, pass a considerable current through the window with a clip-lead on the secondary. Wearing voltage-rated gloves and faceshield, yank the cliplead off the secondary terminals. That should give you the arc.

You may want to do this from 10 or 15 feet away, with a cord to pull the connection off.
Alternately you may wish to wear a welding helmet with the dark glass in place.
The first time anyway, until you see how much of an arc your particular CT develops. Warning, this may be hazardous to your eyes.

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

 

[electricpete]

I agree with oldfieldguy (and waross). Testing on the secondary of a CT with no current through the primary misses the point. As I tried to explain above, it is NOT a simple interruption of current through an inductance.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[MxGxK]

I apologize. I must have given a poor description of the test. The test lead was passed through the window of the CT with the CT secondary shorted. Approximately 50Amps primary current produced by the test equipment fell to nothing as soon as the secondary was open circuited (using much caution). Something about the CT (flux, magnetic field??) was limiting the amount of current that could pass through the primary with the secondary OC'd. Please explain?

I did not mention before because I didn't want to confuse (it may be too late), but a second failure occurred in same connection box 19 days later. This time phase to phase faults occurred and CT was again melted (even the diff CT's were melted).

I could almost prove that it wasn't an OC CT in either case because the CT was producing current until within seconds of the 1st trip. However, the CT produced current for an additional 40seconds after the second trip. Since this has yet to be explained, my opposition would simply like to discard this piece of info/evidence. The only stored energy device in the box is a surge capacitor. Assuming the arc (which burnt CT and cable insulation) allowed contact of the phase cable and exposed CT secondary, could the capacitor discharge through the CT secondary and resemble FLA? Otherwise, I can't explain this. Any ideas?

Some have assumed that the CT was the problem all along because the new CT was moved from A to B phase and the problem followed the CT. However, I try to explain that B phase is directly in front of the motor leads as they enter/exit the motor. The phase to phase arc fault in this area simply damaged the equipment directly in front of it. Not to mention that the motor leads were damaged in the 1st event. They were then taped and put back into service. I tried to explain that damage or lack of the outer semiconductor layer would not allow the insulation to function very long. These cables have potential differences of 7960V/13.8kV and pass very close to each when entering/exiting the motor.

In either case, would the OC'd CT melt in one minute or less? Am I being hard headed or is there reason to believe it may not have been an OC'd CT?

Thanks to all for your help.

 

[scottf]

50 A primary through an 800:5A C200 CT with a weak source certainly shouldn't produce an arc. Once the CT secondary is open circuited, that impedance is reflected to the primary. Basically, your source couldn't drive any current/enough current through the primary to produce a voltage on the secondary.

 

[rovineye]

I don't know what your test set-up was, the current thru the window (primary) should have nothing to do with the secondary. That current should be present even after the CT melts down. Can you explain your test set-up?

 

[roydm]

It sounds like your instructer put 50 Amps thru the primary using a low voltage source e.g. 6 Volt transformer and variac so you probably only had 1/2 Volt or so.
When you opened the secondary the primary current dropped off because of the increase in impedance.
For example 40:1 CT
.5 x 40 = 20 Volts But just immagine if you had 600 Volt on the primary, you would still have 50 Amps and a lot more volts on the secondary.
We need someone to do a video to settle the issue.
Roy

 

[stevenal]

rovineye,

A transformer is a transformer, changing voltage by the turns ratio and current by the inverse. An open circuit zero current secondary translates to zero current in the primary. While most sources will have voltage enough to saturate the CT and drive current anyway, the relay test set provides only enough to drive a relay.

 

[waross]

Use an AC welding transformer to supply the primary current for the CT under test. Stand well back when you pull the connection off. You may well generate a lot of plasma.

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

 

[danbarber]

We use Neilsen Kuljian CTs. They have been pretty reliable.

http://www.clrwtr.com/NK-Current-Transformers.htm

 

[burnt2x]

Sans a picture of the damage, we can only speculate.I wonder how your repair guys did it out there though!
The fact that the indication was a "ground fault" should have me thinking that was it! Your cable terminations failed and sure enough the lugs should go with it, including CT's and others near the grounded cable termination.
Improperly terminated cables can heat up and blow when the insulation fails. You are right when you said cables should be properly terminated,i.e. semicon tape first, pvc tape next and increase the diameter to reduce the stress,and then semicon tape last before the shielding tape.
I could be wrong. Just my take on the topic.

 

[MxGxK]

I believe the evidence suggests it was not an open circuited CT. The last real piece of evidence is the consultation of ABB about their CT.

Thanks to all for your help.