Transformers

[Engcan]

Hi Everybody
Me and my sr. engineer during a discussion did not agree on a fundamental law, and I beleive he is correct. But
I need a final word from you guys. Plz analyze the situation below.
A 13.8 KV line is feeding two transformers in a building.
I am not sure of the %Z of the two transformers but I know they are 13.8Kv/600V. Now they are connected on to one switchgear which has a split bus arrangement that means
bus A and bus B. I beleive each T/F is feeding one bus.
One T/f is for lighting load and IT server rooms and the other one is for HVAC load. We want to run two risers on
10 th floor. I suggested that as we have one transfer switch, we can connect both risers to a single bus and use it as a normal supply. But he said no, I am not sure why.
I think this is paralleling of transformers what my concern is if we do it what is the problem?. As both trms are fed from same local utility, we can use this as a normal supply bus(this new connected bus) and another load bus, on which both HVAC and lighting loads will be connected and one transfer switch will be enough otherwise we need two transfer switches. I spoke to some body in local utility he said they use network transformers for this kind of arrangement and this setting is not usual in buildings. Please advise, Is this not possible, if yes what are the implications.
Thanks

 

[waross]

Do I understand that you propose to tie the Lighting/IT bus to the HVAC bus?
A couple of things come to mind.
If the transformers do not have equal percent impedance ratings, you may overload one of them.
The available fault current of both transformers in parallel may exceed the ratings of your panels and breakers.
Switching surges on the HVAC bus may contaminate the IT bus.
A fault on one system will probably take out the primary or secondary protection on both transformers (or everything).

I am confused about your reference to one transfer switch. Is it connected on the Lighting and IT load, or on the HVAC load, or on a dedicated emergency panel? Do you have a standby generator or is the transfer switch for some other purpose?
yours

 

[alehman]

As I understand the switchboard has two separate buses: bus A fed from transformer A, and bus B fed from transformer B. You are proposing to connect a transfer switch with inputs from bus A and bus B, with the output serving the feeder. Is this correct?

If the transfer switch is of the open-transition (break before make) type it will not parallel the two buses.

Perhaps your engineer wants to be sure certain loads are on certain transformers. With a transfer switch, the load could be on either transformer.

 

[mc5w]

The original reason for 2 buses and 2 transformers is to isolate 2 different power quality problem:

1. HVAC equipment creates voltage sags when starting. Reciprocating compressors can create cyclic voltage dips and swells that are very hard on computers. The original reason for using separate single phase lighting and three phase power services was so that the starting current of design A motors would not dim the lights.

2. 5th and 7th harmonics from computer power supplies and electric discharge lighting tend to overheat motors.

Stick with the original game plan and run 1 rise for lights and computers off of the correct bus and additional HVAC equipment off of the other bus.

 

[Engcan]

Thanks.
Yes it is a emergency genset and one transfer switch will be used. Now this project has been set in motion, so we cannot have two ATS, price is available for one. I
have some knowledge about AT switches:
"If the transfer switch is of the open-transition (break before make) type it will not parallel the two buses.
"
When two riser fed from two different breakers/transformers connected to same bus eventually come in parallel by default as our situation is, so how this statement fits in.how an ATS can change the parallel operation.
Suppose we sequence the HVAC loads as gen size is 1.2 MW
so I beleive sags will not be an issue if HVAC loads are started in sequencing.
One more point here:
To start HAVC loads connected to VFD's after a time gap after the power restores through Emerg Gen set, Is there option in VFD's for timer or we need to speak to manufacturer to advise, on how to create a time gap after the power is restored.
Anyways, back to my question,now I am clear this sequencing will mitigate the sags effects on IT equipment, but my question still stands: what will happen if both buses are connected togather. Suppose both transformers have equal mimpoedance and KVA rating, Now what you would say. What can possibly go wrong: (1)Circulating Currents
or what comes through experience.
Again thanks for your time.
Regard's

 

[waross]

The system may have been designed with a tie breaker so that limited power could be supplied to the building with either transformer in the event of a transformer failure. Maybe not.
If you run the transformers in parallel;
1> Fault current levels may exceed the equipment design. This may or may not be an issue. You may want to check the rating on your switchgear. It may possibly be adequate for a parallel feed.
2> More available current in the event of a fault and the potential for more serious damage.
3> Co-ordination of interupting devises. This may be an issue, unless parallel feeding was considered in the original design.
4> Service reliability. If there is a fault that is not cleared by downstream breakers both transformers will go down.
If the system is designed for parallel operation, it may work well, and any faults may be cleared by downstream breakers, leaving your transformers on-line.
5> Reread mc5w's post. If you can live with these issues, and quite possibly you can, it will probably work. The original design engineer will probably have some choice comments however.
With one transfer switch and two systems, Is the transfer switch a high voltage unit on the primaries of the transformers?
Is the transfer switch a special design with 6 poles to accomodate two independent loads?
Does the generator supply only lighting and IT, loads?
yours

 

[alehman]

Apparently I did not understand your system.

If the transformers are in parallel, there will be circulating current - the amount depends on how well the transformers are matched. They can be different KVA if the impedances (resistance and reactance) are well matched. Parallel transformers require special protection (sometimes called network protectors). Fault current will increase when transformers are paralleled.

I agree with mc - don't change the load separation that has been in use unless there is a good reason to do so.

Most VFD's have programmable start delays and ramp times. Check with the mfg or read the manual.

 

[Engcan]

Thanks for all the inputs from you guys. I beleive, we should probably make space for having supply from one transformer instead of getting in to complex analysis.
Just a quick question here,I was reading an application manual from the cummins power(Gen. suplier), I beleive it is called
T-30 application manual. It says here,
"An unbalanced L-N short circuit results in a current of about 7.5 times rated current.At that current level, assuming an initial temperature of about 155 oC, it can take less than 5 sec for the winding to reach 300 oC- the approximate temp. at which immediate , permanent damage to the winding occurs. An unbalanced L-L short takes a few seconds longer to cause the winding to reach 300 oC, and a balanced three phase short takes a little more time."
I am a jr. engineer, all my years I have considered
3 phase bolted fault as the sever one: how come this is true. Am I missing some thing here?
Please advise.
Thanks

 

[alehman]

You have come upon a fundamental difference between generators and other sources. Generators have low zero sequence impedance, which means line-neutral faults normally are higher current than line-line or 3-phase faults. This is why medium voltage generators are always grounded through an external impedance. Normally low-voltage generators can be adequately protected when solidly grounded.

You need to find a good power systems text and read the chapter on symmecrical components if you don't understand sequence networks.