Fault current higher than trfr rating 1

[veritas]

I have come a across a design where a 65MVA, 66/22kV trfr was specified by a certain party to as follows:

rated short time withstand current @22kV = 12kA
rated peak withstand current @22kV = 30.6kA

I was tasked to do the IEC 60909 fault studies at the trfr 22kV bus and came up with the following:
Ik" = 11.22kA @22kV
Ip = 31.8kA @22kV

I pointed out to the engineer that specified the trfr that the rated peak current is less than the calculated value and that there is <10% margin between rated short time withstand current and Ik. He proceeded to contact the trfr manufacturer who said it was ok. But I have grave misgivings as the manufacturer representative is from an Asian background, has poor English and he actually got some of the parameters mixed up. I am thus not confident that the manufacturer clearly understood the question and implications.

The engineer shrugged his shoulders, said the manufacturer said it was okay and walked away. Client knows nothing about this.

I feel there is no margin of safety to allow for future increase in fault levels and if the trfr is on maximum tap when the 22kV fault occurs the fault current is even higher.

I was asked by the client to assist with the project doing the protection study but feel that it is my duty to inform the client that a trfr will be put into service soon that is not properly rated for the potential fault current duties. The client is not electrical and relies on his consultants.

Any comments? What are considered acceptable design margins for power trfrs fault ratings and potential fault current duties?

Thanks.

 

[prc]

This size of a transformer will have impedance of 10-15 % and the manufacturer will design transformer for a withstand strength for a secondary fault current of 11.3 to 17.1 kA of rms current.Frankly for this size of transformer, peak current is critical from the dynamic mechanical forces and the design margins will be generally as given in IEC60076-5. You have taken the bus fault current,but we have to consider the winding current when the transformer with stand strength is checked. I believe the slight difference is not an issue. It is not a definite cut-off value.

 

[veritas]

I think the situation is actually worse - first of all there is no margin of safety, no margin to allow for future growth in fault level or calculation/modelling error. Surely with such an expensive item of plant the worst case fault scenario should not be right at the limit or exceeding the trfr peak and short circuit withstand capability? I would have like to see at least a 10% safety margin.

Secondly, the trfr has been specified as 11% @50MVA on nominal tap (tap5). It has 17 taps with 1.25% tap increment. Thus at max tap the no-load open-circuit voltage at the 22kV side is 25.88kV which is 17.65% higher than nominal. In addition the Z% is 10.45% at max tap. This means the driving voltage for the fault can theoretically be up to 17.65% above nominal assuming HV voltage is 66kV. With an impedance that actually decreases with increasing tap position does it not point to even higher fault currents than those calculated at nominal tap?

I would probably specify the trfr short circuit ratings based on an infinite bus or very close to it. It can then be deployed anywhere in the network.

BTW this is the only trfr feeding the bus thus bus current is winding current.

Thanks.

 

[prc]

The impedance of system is negligibly small compared to the transformer impedance and usually transformer manufacturers calculate transformer withstand fault current considering infinite primary fault level. Safety margins are taken by individual manufacturers based on their experience and there is no need for additional safety margins on current.

Large transformers rarely see the full effect of nominal peak fault current in service. The peak value depends on the point on voltage wave where the fault initiates and the fault may not be at the terminals but away adding further secondary impedance to fault point. When I started my career in transformer industry some 48 years back, some of the utilities had a standard clause in GTP- Confirm that the transformer will withstand full short circuit current for 5 seconds with maximum asymmetry and infinite fault level on primary without voltage drop in incoming supply. Manufacturers used to readily say YES, fully knowing that there is no way for both parties to check the claim and in reality no system will face such situations. To day engineers understand that the real failure of large transformers occur during the first one or two cycles from mechanical dynamic forces and thermal effects are not a concern. So today short circuit withstand capability is checked as a design test in high power laboratories by subjecting the transformer to desired peak fault current for 0. 25 sec at max,rated and min taps with 3 applications at those taps. Unfortunately short circuit withstand level is not a cut off point and it depends not only on design but the manufacturing,processing and quality issues. Based on the short circuit test experience on similar transformers, manufacturers arrive at withstand capability for his transformers.

 

[Parchie]

@veritas,
Did you consider the impedances of the lines from the high voltage tap as well as the trafo terminals to the low-side bus? Your trafo X/R matters also.

 

[stevenal]

ANSI transformers of this class (IV) are built to customer supplied system impedance specs. For system growth and flexibility, consider specifying infinite bus as suggested above.

 

[veritas]

dpc - most insightful and very helpful feedback. Thanks. The trfr manufacturer in this instance based his rating on the fault currents supplied by the client in the client's datasheet. I agree that it is highly unlikely that the trfr will see the worst case peak calculated by PTW. But I believe that we still need to cater for it. I guess, coming from a protection background the question I always ask is what is the worst that can go wrong - and then cater for it within certain constraints of course. I do not know if the figures you quoted of 0.25s at max, rated and min taps apply to the ANSI world? The IEC standard specifies a 2s rating for the thermal withstand.

Parchie - Yes, it was all considered as I used PTW software to model and do the fault calcs.

It appears my belief that the fault current increases with increasing tap number does not appear to be correct. PTW actually shows a small decreases in current with increase in tap position with the Z% being kept constant. Will need to get to bottom of this as it will have a significant bearing on this discussion.

Any thoughts on this?

Thanks.

 

[dpc]

dpc - most insightful and very helpful feedback.

Yes, but it was from prc, not me.

Actually, I have a question for prc - if the mechanical damage is the primary concern, why is this excluded for transformers if the "non-frequent" ANSI damage curve is used. In the old days, we used to plot the "two second point" and that was it as far as transformer damage was concerned. Now we have the ANSI "Z" curves, but you seem to be saying that the thermal damage is not a major issue.

 

[veritas]

Oops, sorry it was meant for prc.

 

[prc]

DPC, can you give me the ANSI standard number where these Z curves etc are given? These are applicable only for small distribution trfs and I think only given in IEEE color books based on the thermal limits where a cut off point can be easily calculated.(IEC calls for 2 sec for such calculation)

When I was getting a power transformer short circuit tested first time in India(early 1980's- a 12.5 MVA 132 kV Trackside supply transformer),the only reference document available was an IEEE standard on short circuit testing of transformers which was later withdrawn. The duration of test was then 1 second. Later IEEE changed it and today test duration is 0.25 sec, with 6 applications per phase - 4 symmetrical + 2 asymetrical (against IEC 3 asymetrical) but as per IEEE, one symmetrical application shall be for 0.5 sec for category3( 5-30 MVA 3 phase) 1sec for category 2 ( 0.5-5 MVA 3 phase)transformers.As per IEC test duration is 0.25 sec except for transformers of less than 2.5 MVA 3 phase rating ( for them it is 0.5 sec)IEEE C57.12.00-2006 & C57.12.90-2010,IEC 60076-5-2006.

 

[stevenal]

C37.91. Curves are given for all classes of transformers. ANSI also specifies 2s capability in C57.12.00 for most classes, so the curves stop at this point. C57.12.90 is the test code.

 

[prc]

Through fault current duration curves were originally given by C57.109 -1985,1993(R2008) 'Guide for liquid immersed transformer through fault current duration'. In C37.91-2008 Guide for protecting Power transformers, these are elaborated as appendix A ,"Transformer through fault current duration guide. But I could not locate the Z curves.

Except in case of small transformers, the failure is generally due to dynamic mechanical forces which will be max during the first cycle of asymmetrical peak. Thermal with stand capability calculations are done assuming the entire resistance loss during the symmetrical fault current for 2 seconds is stored in copper and then the hot spot temperature in winding shall not exceed 250C with copper and 200 C with aluminum conductor.(clause 7.4 of IEEE C57.12.00-2010)

 

[Parchie]

@veritas,
I did a quick look on your 60909 study results and noted the factor of 2.834 (31.8/11.22)! That would mean you arrived at a K factor of roughly equal to 2 [2.834/sqrt(2)]. This would mean a very big X/R, if I'm not mistaken. [k = 1+𝑒^(−2𝜋𝜏/(𝑋/𝑅))]

My own estimate using X/R= 32 gave me K =1.91 for the first half-cycle(𝜏=1/2). That would make a peak asymmetrical short-circuit current of 11.22 X 1.1414 X 1.91 = 30.3kA!.

 

[veritas]

Parchie

Rechecked everything, got latest data from utility, actual trfr test info from manufacturer and redid the IEC 60909 fault studies. It is a 60MVA, Z% = 12.84% @tap5 and 12.17% @tap17 (max tap). Obtained following results:

Trfr on tap5: Ik" = 10.7kA, Ip = 28.4kA

Trfr on tap17: Ik" = 12.3kA, Ip = 32.8kA

Manufacturer stated that rated short-time withstand current = 12kA and rated peak withstand current = 30.6kA.

I note and agree with prc that it is highly improbably that the trfr will see maximum peak current but it is possible. The study results show that as the trfr taps up the situation becomes more precarious. There is no buszone protection either with the result that any bus fault will be cleared in ~1.3s. Not good either. I believe that we do need to cater for the worst case scenario within limits of course and so this situation is not good. I would have preferred the trfr short circuit ratings to be specified based on infinite bus calcs.

Will get back to you regarding the k factor.

Regards.

 

[veritas]

Regarding the k factor,

X/R = 31 for the trfr and 24.76 at the 22kV bus once the source impedance is factored in as well.

Trfr on tap5: Ik" = 10.7kA, Ip = 28.4kA, k*sqrt(2) = 28.4/10.7 = 2.65 or k = 1.88

Trfr on tap17: Ik" = 12.3kA, Ip = 32.8kA, k*sqrt(2) = 28.4/10.7 = 2.66 or k = 1.89

 

[dpc]

prc,

Pretty sure the requirements are covered in C57.109. There are equations given and the "Z curves" can be calculated. I don't think they are called Z curves in the standard.

This may help as well:

http://ewh.ieee.org/r1/boston/pes/E...rseResources/IEEEPESTransformerProtection.pdf