Parallelling Substation Transformers 5

[hmchi]

I have heard from a Japanese customer of mine that their utilities would run their distribution substations with two or more step-down transformers in parallell [put it another way, with the bus tie breakers normall closed]. He did not see any reason for concern, and yet could not name any advantage of such practice, other than the fact that if one transformer fails, and taken out by the differential relays, the loads would not see an immediate loss of power.

But, after all, transformers do not fail very often --- why such 'theatrics' for such a low risk ?

It seems to me that the disadvantages are :

[1] Increased short circuit current level, either putting system at risk, or necessitate more expensive and more capable fault interrupters.

[2] The risk of circulating currents running from one transformer to another, due to the slight differences in the secondary windings, even if the primary source is identical.

[3] Difficult and exacting differential protection required --- not only expensive, but one false move, you're dead ...

I hope to hear from utility engineers using this practice to enlighten me --- there must be something I missed in our imperfect info exchange --- one of us not communicating in his mother tongue.

I would like specifically to hear about the circulating current issue --- most utilities would not run closed loops at distribution except from the same transformer --- due to the concern of circulating currents. Wouldn't parallelling substation transformers create the same concern ? when the transformers are not perfectly matched ? Can someone cite IEC or IEEE standards on this issue ?

Thanks

 

[jghrist]

The interrupting capability of typical US distribution cutouts is 10kA. UG loadbreak elbows also have a 10kA fault limit. Limiting the available fault to 10kA on a 12.5 kV system with 9% impedance power transformers requires limiting the base rating to 20 MVA. Paralleling transformers greater than 10 MVA therefore will result in too high a fault current for the typical US distribution system.

 

[hmchi]

The overhead distribution lines reduces the fault current quite rapidly though. A 40MVA PXF with 13.8kv secondary, 12% impedance, source side X/R ratio of 17, would have 14KA fault duty at the substation, but 8.5KA 1 kilometer away from the sub, and 6KA at 2 KM from the sub.

 

[ohhn]

Mr hmchi:
We have in paralell operation 2X150 MVA, 3X 100 MVA 230/138 KV transformers with high voltage OLTC and we haven't had any troubles, we also have 2X25 MVA, 2X50 MVA 138/13.8 KV with OLTC on the secondary sides and we we haven't had any troubles either...
As for the short circuit current, all our CB's are rated 25 KA no matter the voltage level and we haven't reached this fault current on distribution or transmission substations, so we're OK with any combination of paralell transformers...

 

[kith]

In a distribution network, zero downtime is of paramount importance with bulk customers.Paralleling two 40MVA Txers would imply that on a through fault, you would lose both Txers and completely shut down the customer. Running the two Txers on a split bar would optimise availability.
In Zimbabwe, where posssible, Txers are operated on a split bar arrangement.
As far as paralleling theory is concerned, the following conditions must be satisfied:
THe % impedance must be within 10% of each other.
The MVA rating must be roughly the same(not obligatory).
The ratio must be identical.
The vector group must be the same.

 

[jOmega]

rbulsara

"Spot Networking is expensive and requires careful engineering and hence not popular in commercial applications."


F Y I ....

I see this done quite often in commercial buildings (offices, hospitals, hotels & resorts, etc.) and in the Water/wastewater industry; i.e., 4 x 2500 kVA, 13.2kv/480v.

jO

 

[asymptote]

In the UK it is normal practice to operate BSP (bulk Supply Points) with 132/33kVYd 90MVA and Primary SS having 33/11kVYy or Dy 38MVA transformers in parallel. The protection is not very complex in either case and the circulating current is used via negative sequence compounding to keep each TX within one tap on a random control system. The system fault levels are 1000MVA and 250MVA respectivly. The system has one salient advantage with continuity of supply to the customer under single fault conditions.

 

[hmchi]

asymptote, Thanks for your real-world experiences. I am trying to understand the intent of the design :

[1] I assume that no more than 2 transformers are parallelled in the substations you mentioned above.

[2] Are the sources to the 2 transformers 'reasonably independent', meaning not very likely for both to suffer outage simultaneously ? like being fed from different substations, different routes ...

[3] If one transformer fails [internally, taken out by the transformer differential relays], the other takes over the load instantly and automatically, which is the principal advantage --- any statistics of how often this happens ? or if the likelihood of a underground cable fault is X, would you say that the likelihood of transformer internal fault being X/4 or even less likely ?

Thank you in advance for your response.

 

[Raenz]

From a NZ experience, which commonly follows UK practice due to shared engineering heritage, most CBD type zone substations (2 transformers) are run in parallel - often with an inter-trip to take both the HV and LV CB's out at once. Most common reason is maintaining continuity of service to commercial and industrial customers with the loss of a HV supply circuit. Second most common reason - lack of space to install substations in CBD areas, so get as much load as possible out of one site almost implies that paralleling of transformers to ensure load sharing. Often the LV is also paralleled for heavy load areas.

Transformers are usually purchased as ONAN with over-load ability or dual rated (ONAN/ONAF or ONAN/OFAF) to handle the increased load. My present utility runs all of our 33kV and the associated transformers in parallel (with suitable protection) and we do not have any issues - all transformers are fitted with paralleling controllers based on minimising circulating currents.

PLease note that NZ runs on a "Multiple Earthed (Grounded) Neutral" system - which does add some pros/cons that you might not have.

 

[engr9999]

This is one of the most inetresting posts for me. I have come across many transformer trippings due to spurious operation of protective devices like water ingress etc. If transformers are operated in parllel a power outage at down stream could be avoided (so less financial impact for some dist.companies). Do we need to look for any thing other than 0.5 cycle fault current within the interrupting capacity of breakers? please give the details or suggest any website.
I would like to ask a few more questions to see what is the requirement of protection scheme(when paralleled)to have decriminated tripping during actual "down stream fault". The case is two identical transformers with upstream buscoupler closed, LV downstream bus having two incomers and one bus coupler. Follwoing are options.
1. To have one descrimination stage(say 300 msec) between the LV incomer and buscoupler so that incase of fault the buscoupler will trip before the unfaulted side incomer trips. Transformer HV and LV breaker will have same operating time. i.e.no descrimination between them.
2. To have the CTs of LV incomer and buscoupler paralleled(differential kind of wiring) and connect to a inverse relay and use this relay to trip buscoupler and incomer simultanoeusly.
Please suggest which of the above two options is better and why?
Thank you all in advance and hope this will help hmchi also.

 

[Bung]

We have recently been pushed by regulatory requirements for reliability etc to runnning two transformers in parallel instead of one on / on standby. As noted above, this results in increased fault levels, which is bad news for distribution dropout fuses (EDOs)(cutouts in the other half of the English speaking world!). This means that we are limited in how close to the sub we can use them. At the typical overhead conductor sizes we use, EDOs can only be used at distance of 1km or more. If there is a cable before going onto overhead, then the cable length is 3.5km to limit the fault level to 8kA (at 50Hz, EDOs are good for only 8kA, not the 10kA you get at 60Hz). [NOTE: THIS IS ALL BASED ON THE CHARACTERISTICS OF OUR NETWORK - YOURS IS QUITE PROBABLY DIFFERENT!]

But an advantage is the reduced losses, especially near full load of one transformer. This applies equally to I^2X and I^2R loss - on a 25MVA transformer, times about one hundred substations, it adds up. Halving the current quarters the loss. But the primary advantage is supply availability - it was this that was the primary driver for us doing it.

The major headache for protection is the reduced backup cover of your distribution feeders. Each transformer sees only half of the fault current, so you get reduced sensitivity. A 2000A setting on the transformer secondary (25MVA transformer, 11kV) means that anything less than a 4000A feeder fault is "invisible" and so not backed up by the transformer protection. Our typical feeder overcurrent setting is 400A pickup.

Beware of directional protection on the HV side of Dy transformers - an earth fault back-fed by the transformer will not be seen - the delta is not a zero sequence source, so you get no current flow, and hence no protection operation. But this is easily taken care of by appropriate design and the installation of a HV busbar.

Bung
Life is non-linear...

 

[hmchi]

Statistically substation transformer should not see frequent faults, unlike overhead lines or even underground lines. If it does, it seems to me that appropriate steps should be carried out to straighten out this abnormal problem, rather than relying on parallelling transformers to address it.

As for lower losses, etc., one could argue that it is a lot less costly to buy one 50MVA transformer than 2- 25MVA transformer and try to operate them in parallel.

I agree that this is a matter of balancing the priorities - so that there may be disagreements no matter what. What I was trying to get at from the beginning was statistical information from the practitioning utilities such as studies regarding the costs/benefits justifications, etc.

 

[Bung]

All transformers see all faults downstream, whether the fault is on a line or underground cable.

Bung
Life is non-linear...

 

[hmchi]

Bung,

"All transformers see all faults downstream, whether the fault is on a line or underground cable"

Yes, but line faults are supposed to be cleared by the outgoing feeder breakers whereas transformer or bus faults are to be cleared by differential protection, transformer or bus differential.

A line fault should not have caused any of the transformer protection to operate.