is there a possible ct fault that gives false high current 4

[electricpete]

Out of curiosity...

Is there any credible fault in a CT or associated wiring that could make the secondary current seen by one of the loads increase above actual?


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

 

[davidbeach]

If you were using a low tap on a multi tap (multi-ratio) CT and somehow one of the higher taps were shorted to the CT secondary circuit. Or, something could have been programmed with the wrong CT ratio.

 

[ScottyUK]

Thinking aloud, how would a shorted turn or number of turns on a CT core affect the apparent ratio? Aside from potentially huge current in the shorted turn(s), off the top of my head I think the CT ratio would also change as seen by the external circuit. No doubt someone has analysed this - I'll maybe have a go later when I get five minutes.



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

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

 

[electricpete]

Thanks David.
Programmed wrong - would be a relay setup issue. I agree with that and other possible malfunctions of relays and indicators.

Let's assume the circuit used to indicate correctly and now we had a relay trip and we are trying to think of all the possibilities related to the ct.

I was thinking about the tap shorting issue.
Fewer turns = higher secondary current tap.
Since everything was working before, I am only going to allow one fault (either an open at one point or a short at one point).

To my way of thinking if I had only one fault which was an added short between turns or between output and a higher-current/lower-turns secondary tap, I would have increased current internally, but decreased current to external devices. To increase the external current I would have to not only add the short but add an open circuit at the original tap... not something I'm considering credible.

Do you guys agree? Or is there some single shorting or opening fault that would cause increased current external to the CT.

(btw, I know someone out there is itching to ask about the trip or give me suggestions for testing the ct... that is not my interest at this point... just want to focus on understanding possible failures of CT).

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

 

[electricpete]

Scotty

Yes, that is a reasonable first thought: Apparent CT ratio goes down and therefore secondary current goes up.

BUT drawing it out a little more, by shorting turns we are providing a loop path which can contribute amp-turns to balance primary without even flowing in the external circuit. I'd be interested to hear your further thoughts when you get 5 minutes.

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

 

[scottf]

There is no credible failure mode that I can think of that would result in a higher than expected current in the secondary wiring, i.e. relay circuit.

As has been said above, an internal wiring short or open-circuit will result in a lower current and not a higher current in the relay loop.

 

[ScottyUK]

ePete,

Have sat and cogitated with a pencil and paper and a pot of tea for a while I agree with you and ScottF that an internal short provides another balancing amp-turn loop other than that through the external burden. The external burden will see a lower current.

No doubt I will post this and then something further will occur to me. C'est la vie.


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

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

 

[BJC]

Could one of the CTs be connected backwards? I will draw the diagrams when I get time, I am working on distant memory of a long ago problem.

 

[DanDel]

Hello electricpete; a possible failure mode might be a loose connection or intermittent shorted turns, which in turn would cause intermittent secondary current which might be seen by some devices as something other than acceptable (spiking, arcing, rate-of-rise, etc). You just have to tell us what device may be connected to the CT.

 

[scottf]

ScottyUK

No offense taken...the winding of the core/coil isn't rocket science for sure. Now...passing 765 kV through a core/coil with a 400 mm diameter grounded shield around it is another story completely

 

[PWR]

I believe DanDel has asked the right question: What type of relay is connected? Was it a phase overcurrent relay operating in the overload region, or instantaneously, or?

 

[electricpete]

Thanks for all those responses. As far as I can tell, there was unanimous agreement, the ct can't increase external current. Except Dandel made a good point about intermittent open may have some unexpected results.

First and foremost, I was interested in the general question. My reason - at our plant for certain critical problems, our problem solving process is called a fault tree. You get together in a conference room with 10 other people and you list every possible cause and then you rule them out by logic if possible and then you develop a plan to troubleshoot the remaining (not ruled out) items in a prioritized manner based on probability of being the cause and based on ease of tests to rule a given item out.

Over the years, we have had various motor trips on a variety of occasions with a variety of relays - electromechanical (GE) and electronic (ITE Circuit Shield) relays providing functions of instantaneous overcurrent , time overcurrent, and high-dropout overcurrent. I would like to have justificaiton to remove the CT from consideration in these scenario's without having to test the CT.

There is one issue that we were fighting yesterday. It is now a closed issue (unless the problem recurs).

It is a long convoluted story. Since you guys asked even after I warned you ....(electricpete /Electrical / 1 Jul 05 8:00 wrote: "btw, I know someone out there is itching to ask about the trip or give me suggestions for testing the ct... that is not my interest at this point... just want to focus on understanding possible failures of CT")...., since you asked I'll give some details.

I'm pretty sure there is not much help you can provide on this long convoluted ill-posed question, but if you are a glutton for punishment and want to read the whole thing, be my guest. I welcome any thoughts on the question at the end. Once again it was not my choice to describe our scenario but you guys asked for it

This is an ITE electronic Circuit Shield 50/51 relay. It protects an 4kv 800hp horizontal sleeve bearing induction motor driving a between-bearings double-suction centrifugal pump through a Thomas Shim Pack coupling. We have two of these relays - one for trip and higher setpoint and one for alarm at lower setpoint. We got a pump trip 13 hours after pump start. At the time of pump trip we happened to be reducing flow in the system. Pump curve shows for this pump that equates to reducing the current in the motor. Flow was at a level where we expect current somwhere in the neighborhood of 70% FLA. Alarm relay is 1.2*FLA and Trip relay is 1.3*FLA. We got the time overcurrent flag for on the trip relay but not the alarm relay (and no instantaneous on either relay). This in itself was a very strong indicator of relay problem since alarm relay should have flagged before the trip rleay. But not exactly able to confirm any problem on the trip relay or the alarm relay. Alarm relay which did not trip is in series on the same secondary CT circuit which would seem to rule out CT altogether in this case, except maybe for DanDel's scenario.... but at this point in time that doesn't seem too probable (although interested to hear comments).

Cable was meggered sat, motor meggered and bridged sat. Our ace relay technician did report some kind of a trip during testing of the alarm relay that was not reproducible, but the story of the circumstances surrounding that trip are completely a mystery to me considering it is a complex test procedure on a relay I am not familiar with, and the story of the trip during testing was relayed from nightshift through several intermediate people. The trip did not occur when they were testing the trip function (all trip tests were sat). I think maybe the trip occured when they were removing testjacks.

It seems pretty close to a "case-closed" scenario that we have a faulty trip relay based on:
1 - the alarm relay didn't trip (lower setpoint) but the trip relay did trip (higher setpoit)
2 - nothing found wrong in the motor or cable.
3 - some anomaly identified during testing of the relay.
4 - Winding and bearing temperatures were normal and rock steady up to the time of the trip.
btw - we don't have any record of current. Wish we had them SEL relays!

The one thing that is bugging me a little bit is that the trip occurred exactly at the time we were manually reducing flow in the system (a very infrequent occurence).... 30 seconds after they began reducing flow. Seems too much a coincidence to be unrelated. I generally don't believe in coincidences. But we don't see any possibility for emi/rfi conducted or radiated interference between the motor operated valve used to control flow and the relay dc power (the only power provided is dc voltage... same dc voltage that is used to sense the contact position). Double checked that no operators were in the switchgear room where handheld radio's could have affected the circuit.

For all useful portions of this pump curve, reducing flow causes reduced BHP (and therefore reduced current). At the extreme right of the curve (very high flows), the bhp vs flow curve starts to turn in the opposite direction, but I am very confident that we were not at that region of the curve (don't want to talk about that either... take my word for it). I am confident we were in the region of the curve where reducing flow reduces BHP.

I have a theory on how reducing flow might possibly cause a trip, but most people would say it is way way way way way out there. I happen to think it is halfway credible (if the trip recurs , it will be my #1 theory). It is based on some additional info which I haven't given you, which most people would consider irrelevant. I'm not going to give that additional info to you because #1 - I feel a little embarassed about the theory, #2 - I don't want to predjudice you. Maybe I'll give you some additional info or give you my theory after you've had awhile to think about it. For now, can you guys think of any possible way that reducing flow can cause an overcurrent trip (again, my rules are that you have to take my word on the flow and shape of the bhp vs current curve - reducing flow reduces bhp).

Sorry to be so convoluted in my discussion. I did warn you that there is probably no useful input you can provide. We will attempt to recreate the flow conditions and take some current measurements on all three phases... will let you know any more info.

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

 

[davidbeach]

electripete,

I think you are probably right to suspect the relays.

<Shameless plug> Basler has a 50/51 relay that is a drop in replacement for the Circuit Shield 50/51 relay, but it isn't a numeric relay. </Shameless plug>

You don't necessarily need an SEL relay, anybody's numeric relay would provide the data you wish you had.

My first suggestion would be to pull the trip relay and test the set curve at several current points and see if anything shows up (you have, I trust, verified that it is actually set as it should be). If that one tests out fine, check the other one.

If the two relays are actually set as you implied, on the same curve, with the same time dial, and only the pickup differs, (are they on the same or different CT circuits?) then one of them failed to operate in the proper manner.

If the two relays are on different CT circuits, what else is on each circuit. Maybe (don't have a clue why the motor current would have gone up) the trip relay was right and something in the alarm relay CT circuit caused the CTs to saturate, so it didn't get enough current, but the trip relay accurately responded.

Trying to recreate the conditions, with lots of good instrumentation, is certainly a good way of finding out what happened. The only problem is that per some unnamed corollary to Murphy's law, the likelihood of a problem repeating itself is inversely proportional to the additional instrumentation added to detect the problem.

 

[DanDel]

'pete, by "manually reducing flow", do you mean by use of a valve or some kind of baffle, or what?

 

[itsmoked]

I'm going where I think DanDel is... if you are blocking the flow to reduce it why not a weak water hammer possibly from turbulance etc? Would that not cause a very brief current spike?

 

[electricpete]

Manually reducing flow in this case means that the operator operates a switch from the control room which drives a motor-operated throttle valve further closed. The flow is monitored by the plant computer and recorded during the whole process.

Touching again on the possible emi interference issue - The motor operated valve and controls are all 480vac (no dc). The relay power is dc. Completely separate power system. Of course there is the ac ct input to the relay but those circuits are located completely within the 4kv switchgear which is nowhere close to the 480vac power to the mov.

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

 

[electricpete]

I agree throttling flow will cause turbulence and I am intrigued by the connection. What would be the mechanisms that turbulence can cause a current spike sufficient to trip our time overcurrent at 130% FLA when the steady load is somewhere in the range 70% FLA.

The time characteristics is relay trips at somewhere in the neighborhood 5x trip setpoit (5x130%) for 5 seconds.

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

 

[itsmoked]

Water hammer I would presume would be a calculatable value (pressure difference) on the supply side of the restriction that is function of the media speed and massflow. If you can get multi-hundred psi pressure spikes from this I can see the pump, for some brief moment, seeing a large differential. But alas 5x130% for 5... That would seem to be long enough for those kinds of transients to wither away.

What is running through this pump? H2O? You are in the powerplant realm aren't you?

 

[electricpete]

Yes, H20 near room temperature in a powerplant.

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

 

[Skogsgurra]

This is exactly the kind of problems I usually am hired to solve.

To avoid discussions about components involved I usually hook up a current clamp and a memory recorder or a computer with a Dacq Card and wait for things to happen. Any other relevant signals are also recorded and the recorder is either trigged by the relay trip or stopped by an informed operator when the trip occurs. By using a reasonably priced recording device I can let the whole thing sit there for months without being too costly. The biggest problem usually is to have someone you can trust look after the equipment 24/24 7/7 so that it hasn't been stopped for some reason.

When the next trip occurs, we know from the recordings what happened. More often than not we then find that the first fault history was wrong or colored by someone's theories or prejudice about what happened.

What if it never happens again? Well then, case solved, I would say. Even if your curiosity isn't satisfied. Sometimes it is an operator or someone working with the equipment that caused a trip without telling anyone about it. Physically impossible events never take place.

Gunnar Englund

http://www.gke.org