Failure of 132 kV Current Transformers 2

[RAgrawal]

Dear Friends,

We are observing failure of 132 kV CTs. The frequency is about 1 no/ year but recently 2 CTs failed in span of 2 months. These CTs are about 10 - 12 years old and are of ABB make.

I want to know:
1. Whats the failure rate of CTs in other location
2. Is there any test which can indicate the health of CTs thereby wanring us before premature failure.

We are annually conducting low voltage Tan delta.

 

[Skogsgurra]

I assume outdoor switchyard. What is the climate? What is the failure mechanism? Are there large temperature swings? Rigid busbars? Vibrations?

How many CTs are there? One failure/year in a population of ten is way too much while it might be acceptable if you have hundreds of them.

CTs are known to do their job for decades without problems. Ten or twelve years is definitely too short a life-time. BTW, a rated voltage tan(delta) would be more conclusive. You do not get the tell-tale PD with LV.

Gunnar Englund

http://www.gke.org

 

[RAgrawal]

I am sorry, i forgot to mention regarding climate condition, population etc.

CTs are outdoor type connected to flexible bus bar and designed to take care of vibration etc. Climate is normally hot. It varies from 50 degree C (in summer) to about 20 dgree C (in winter). We have got 20 bays of CTs i.e. 25 x 3 ~ 75 nos. of CTs and CTs are failing for the last 5 years. During failure CT blasts and catches fire.

As a routine test we carry out low voltage Tan delta, and if low voltag tan delta is high then only high voltage tan delta is carried out.

 

[scottf]

RAgrawal-

Failure within 10-12 yrs is definately premature.

Is the CT a head-type (i.e. cores at the top of the unit) or hari-pin (i.e. cores at the bottom of the unit) design?

What voltage are you meaning when you say DF testing is done at 'low-voltage'?

As a minimum, I would take oil samples of a number of units (all if possible) and do a DGA to look for gassing. Also, DF should be done at at least 10 kV in the field.

You may want to consider taking 1 or 2 of the units currently in-service to a HV laboratory for partial-discharge testing.

 

[ausphil]

We would be very worried with one failure per year, because of the explosive nature of failure of CTs (and VTs for that matter), and the chance of large pieces of porcelain flying around the yard.

Scottf has the right idea, DGA sampling and testing is crucial, becauase you'll probably find that the failure is based on partial discharge causing either tracking along the insulation or significant burning through the insulation, eventually flashing over, and the power arc causing the explosion. DGA will detect very low levels of gases produced by partial discharge in a paper/oil insulation system, and can be tracked over time to give you confidence that you can keep units in service. The more sets of samples you have (in time on the same units, as well as across a number of units) the more confidence you can build up.

Be aware that DDF testing will not pick up damage caused by PD until it is probably too late. Also, ensure that the procedure of sampling and refilling the CT (if you are going to do DGA sampling) ensures that there is no contamination introduced into the CT, because you can cause more problems than you will fix by incorrect sampling techniques.

We have recently found major burning of insulation from partial discharge on a batch of VTs via our DGA routine testing - prior to electrical failure. It really is invaluable.

The other thing you can do whilst you have an oil sample is moisture content, which can tell you if bellows or seals have been compromised.

Finally, with the remaining units of the sets that that contain the ones that have failed, remove them from service and see if they can give you any idea as to why their partner failed. Take them to a test laboratory, perform full PD, DDF and DGA tests on the units, then if there is something wrong, dismantle and find the problem. If there is a PD problem, you may have to unwind the whole of the insulation to find evidence of burning or x-wax (a by-product of discharge that is the precursor to actual burning of the insulation - can be seen under a UV light source on the paper surface). this sort of dismantling can be done with or without the manufacturer present, but the results should certainly be fed back to the manufacturer to enable them top respond and hopefully it may feed back into their design process, so that in the future, premature failure due to that particular failure mode may be avoided.

 

[RAgrawal]

CTs are hari-pin (i.e. cores at the bottom of the unit) design type.

We do DF testing at 48 Volts only.

Oil sample can not be taken for DGA as CTs are hermitically sealed type, no place to take oil sample therefore can not do DGA of oil.

 

[scottf]

I'm surprised that there is no oil-sample port. Being hermetically sealed does not mean that oil samples cannot be taken.

DF tests at 48 V don't strike me as being terribly meaningful.

As for the hair-pin designs, where are you located? Most utilities that we deal with won't accept a hair-pin CT design, as they have a preference for the newer head-type/post-type designs.

 

[RAgrawal]

Mr Scottf,

I am sorry, i gave wrong fact in the beginning. After talking to my maintenance people, i came to know that these CTs have got provision to take oil sample but we do not take samples due to 2 reasons:

1. Suppliers (ABB india) do not recommend to carry out any thing on the CTs at site.
2. We are afraid that if we keep on taking out Oil samples from CTs, oil level may go down and we may have problem of oil top up at site.

As far as our location is concerned, we are located in eastern part of India.

You said that most of the utilities do not accept hair pin CTs, whats the reason of not accepting the hair pin CTs?

 

[scottf]

Well...in the US, we had big problems with them in yrs past.

The hair-pin design brings the high voltage down in to a lower tank...through the cores....and then back up to the top of the unit. As a result, the insulation grading is more complex and there are issues asscoiated with damaging the insulation system during a system fault, as the conductors running down the insulator tend to pull/push apart due to the high fault currents through the primary of the CT. Also, for CT parameters requiring a wound primary winding, the wound-primary in a hair-pin design is much more difficult and subject to the problems mentioned above.

In post-type/head-type CTs, the primary conductor is very short and only runs through the head of the CT. This keeps the energized portion out of the insulator (reducing the chance for explosion) and keeps the inductance of the primary winding very low.

As for the oil sampling, the manufacturer should be able to tell you how much oil is available for sampling. Now days, very little oil is needed to perform a DGA. The CT should allow at least 0.5 liters for sampling and that is a fair number of samples.

 

[tony6]

tony6 16/11/05
Having previously spent over 40years designing HVCTs of all arrangements, including hairpin , top-core (inverted) and "eye-bolt" types I would like to offer my experience on this problem.
Firstly, this is most unlikely to be related to the particular style of CT. On the contrary the hairpin type is the simplest arrangement to insulate and lends itself to full automation of the insulation application giving consistent structure and excellent stress control.

The top-core style however, where an often large earthed core and winding structure is supported inside a high voltage enclosure, is quite difficult to insulate and ensure that there are no large oil-gaps in the transition region between the head and the bushing part.
Failure to eliminate and control these gaps results in regions of high stress and premature failure.

However it should be said that both type can be successfully manufactured given the necessary experience and quality control.
It has been my experience that rapid failures here are difficult to monitor as the time between significant gas generation and detaction and catastophic failure can be quite short.

I would recommend that a still working unit be removed from the field, tested for partial discharge and DGA, subjected to muliple impulse shots at around 450kV and re-tested for PD and DGA.Then stripped for examination, paying particular attention to grading foil dimensions and correct placement.

These sudden failures are often preceded by system switching operations and the system log could be checked for this.

Finnaly as these failures can be extremely hazardous to any pesonnel working in the proximity I recommend they carry a de-tuned or poor quality AM radio receiver as for some time prior to failure the CT can emit very strong radio interference in the regio of 1MHz.

 

[scottf]

tony6-

I respect your experience but I definately disagree on your assessment of hair-pin type verusus head-type.

My company manufactures head-type CTs and we used to (many years ago) manufacture hair-pin type.

There is a reason that hair-pin type are not accepted in most of North America and Europe, as I stated above.

Regardless, I think we've given RAGrawal the advise he needs on looking after the failures.