Transformer 4

[anandmoh]

As per my understanding tandelta(power factor) tests for Transformer windings are not included in the IEC or BS. Should these be included for acceptance of a new transformer?
What values of tandelta for CH,CHL,CL can be acceptable?
In one of our 200MVA single phase 21/400 kV Generator step up transformers, the tandelta of CH exceeds 2%, but the tandelta for CHL and CL are below 0.5%. What could be the reason for this? Vacuum drying at site has not changed the above values.How can we improve the tandelta values.What significance does this carry?
Would you recommend rejecting this transformer?
Should single phase banks be tested for tandelta separately or together?
Kindly educate me on this subject.

Thanking you,
Anandmoh

 

[DanDel]

Are all primary and secondary bus/cables disconnected from the transformer for the tests? Often, leaving even some secondary buswork connected will throw off the primary results.

 

[oldfieldguy]

anandmoh--

That 2% number is bothersome. The expected value for a new transformer this size is less than 0.5%. What does the manufacturer say? Has he reported similar results on other transformers of similar construction?

Single phase banks should be tested separately, with each transformer in the bank evaluated as a stand alone unit. If you're testing them together, you often have external insulation networks involved in the testing, as DanDel mentions.

DanDel points out a very good point. This is often a case in high CH and CL readings. NO EXTERNAL bus or conductors can be connected during the tests. You want ONLY the transformer bushings and internal insulation involved.

I would be having a serious conversation with the manufacturer at this point. If they cannot point to me a valid reason from an engineering and design standpoint for the high CH, assuming that we've done a good job on our tests, then I'd want them to sign off on some liability and for retesting.

That said, if I was stuck with this thing, I'd want an accelerated program of retesting for a couple of years, maybe every 3-6 months, to see if this is a stable or a deteriorating condition.

I recently had an "issue" with a third-party testing company presenting me with a 0.72% power factor on my new transformer. I expected <0.5% and I questioned them on their test procedure. They'd failed to disconnect an "inconvenient" secondary bus even after I'd instructed them that I wanted this done for the tests. They performed a retest and the follow-on result was 0.26%, in keeping with the tests on other windings.

old field guy

 

[anandmoh]

Thank you dandel and oldfieldguy for your valuable comments. Let me assure you that the tests were repeated a number of times considering the suggestions you have pointed out.However do you think that a strip down of the transformer at manufacturers premises could reveal problems with manufacturing.Or could be possibly due to usage of inferior grade of paper or pressboard.If the tandelta of transformers is such an important issue I wonder why this has been kept away from the purview of IEC/BS/IEEE standards.
Yes,I agree that in the absence of any concrete answer to this problem, we may insist for an extended guarantee under the latent defect clause.And we can periodically monitor the deterioration in tandelta values.thaks for all the information provided.
regards,
anandmoh

 

[DanDel]

A strip down of the transformer is not practical, and will not solve the problem, unless you intend to rewind it.

If the transformer(s) have never been energized, you might look into using heat or energization at a low voltage to try to dry them out.

If a transformer is in use (hot) and is shut down for testing and the results are as you say, you may have an insulation problem. Were insulation resistance measurements taken? What were those results?

 

[prc]

There is an IEEE on the limits of tan delta. Clause 6.1.6.2 of IEEE 62-1995(R2006) Guide for Diagnostic Field Testing -Part1 Power Transformers where it is said" Pf more than 1% at 20 C should be investigated.American market always demand for a pf less than 0.5 %for new transformers.

Reported high value may not be due to moisture as other values are OK. This may be due to error in measurement,connections or some materials used in HV winding.Original OEM will be the best judge and go with his recommendations.

Self once had a problem of high tan delta (0.8%) for a trf for America from India.After trying all alternatives when a modern good instrument was used, values came down below 0.4 %.

But in your case if values persist check moisture content by alternative methods and keep a close watch on the unit.

 

[JBinCA]

Admittedly, I'm new to the pf testing world, but I'll add this to the discussion:

Doble brings the value of having tabulated statistical data on a large number of XFMRs, Gens, Bushings, and the like.

I don't know if I'd use the results of the tan delta test and try to interpret it based upon Doble statistics, but you might consider a retest with a Doble unit.

Good Luck.

JB

 

[wfowfo]

What kind of results did you get on the bushings?

 

[waross]

Search a little on Eng-Tips. There was a recent discussion on the subtle differences between Power factor testing and tan-delta testing. Basically the same test but the results are developed with a slightly different formula.
respectfully

 

[anandmoh]

The bushings are the OIP type and were tested at 10kV.the tandelta values were less than 0.5.I am not unaware about the differences between tandelta and power factor as i understand that at low angle delta, tan delta equals cosine(90-delta)and 90-delta is the angle phi.Hence cosine phi approximates tan delta at low delta.Iam however aware that two types of measurements exists-one thro a shearing bridge and the other thro direct watt loss measurement. In this case both type of measurements have yielded similar results.The transformer is a new one and was not tested for tandelta at the factory.

 

[prc]

One more input.Even if the bushing Pf is very high, it will not affect the winding pf.Doble set measures watt loss and hence give the power factor.Bridges measure tan delta.

 

[anandmoh]

Dear friends,
The answer to my problem was found. The problem of high tandelta lay in the bad quality of insulating material used in the divertor switch of the OLTC.This was detected at the manufacturers works.the manufacturer has agreed to replace all fibreglass supporting and insulating material used in the divertor switch.Thank you all for the help and support rendered.
regards,
anandmoh