2 UPS with one STS 2

[HawkBrazil]

Dear Friends

I have a problem today in Brazil, related with a 2 UPS's and we need to make a transfer of load between that equipments in case of fail of one.

In a first moment most of us don't can see any problem, but when you discover of that 2 UPS's are double conversion - not true on line - the problem appears.

Anyone here know any solution for this ? The sinus are not at the same position at the entrance of STS, then if the transfer happens, we loose the load.

TKS for all.

Wagner Freitas
WR Freitas Consulting Engineers
Rio de Janeiro - Brazil

 

[ScottyUK]

Are you saying that you require to transfer your load from one UPS to the other if the first UPS fails, without interrupting the load?

If so, you could feed the output of one UPS into the static bypass switch of the downstream UPS, so the UPS's are cascaded, and feed the load from the output of the downstrema UPS. If it fails, it should transfer to bypass and the upstream UPS will feed the load.

It isn't ideal from a redundancy point of view, but it is a recognised way of providing a degree of redundancy in the system without dropping the load.


Hope this helps,

 

[HawkBrazil]

Dear Scotty,

Tks about you attention !!!!

In this case, I don't need to use a STS. Perfect. The only problem, that I don't have mentioned before is the Upstream UPS don't have a static by-pass ( !!!! ), but the configuration that you have mentioned is possible, due to a downstream UPS have that feature.

In a case of fail of the Downstream UPS, the by pass turn on, and the Upstream feeds the load. But if the Upstream shows any problem...bye...bye load !!!! The Upstream UPS protects itself then turns OFF.

But I agree, with your sugestion. The reliability of the system is increased if we use that way. Then I must do that.

Tks once more for your atention !!!!





Wagner Freitas
WR Freitas Consulting Engineers
Rio de Janeiro - Brazil

 

[ScottyUK]

Pleased I could help.

 

[wattsbruin]

I think that you are both missing a serious flaw in this configuration; that is: if the Primary UPS has a catastrophic failure (one that disables both the main and static bypass output) then the load will be lost until the damaged components can be repaired. This could be a many hours if it is a decent sized UPS. This configuration was considered to be an optimum solution many years ago before technolgy advanced sufficently to support a controlable method of paralleling rectifiers.

You might want to stay with your original UPS-STS-UPS configuration and investigate the possibility of using a synchronization unit between the two UPSs; check with the UPS vendor to see if this type of unit is available.
Also, you might want to check the settings on the STS. There are several parameters that could be adjusted to settings other than the factory defaults. These adjustments might compensate for site-specific variations and allow your system to perform as you require it.

 

[DaveMarz]

How old are your UPS units and do they have any communication between them??? Most UPS units produced now can easily be set up to run redundant.

 

[jbartos]

Suggestion: A fast static transfer switch (about 4 ms for the transfer) between two UPSs for the load is required.
The outputs of UPSs should be synchronized.

 

[ScottyUK]

Hi wattsbruin,

I absolutely do recognise that there is a potential problem if the UPS with the STS fails, taking the STS out with it. This would indeed cause loss of supply to the load. My earlier post did note that [blue]"It isn't ideal from a redundancy point of view..."[/blue], which obviously wasn't clear enough. Sorry for not spelling it out more clearly. This configuration is not so common these days for a new installation, but there are a lot of the older configurations out there already. There are a few manufacturers offering redundant systems which allow units to operate in parallel, but if this was an option I'm sure the original poster would have done this already. My understanding was that he was trying to make the best of an existing installation, not replace his UPS with a new one. If only we had that option open to us in every instance!

The UPS without a bypass facility may not be synchronised to the utility line. As a double-conversion type without any ability to transfer to a bypass supply, there would be no reason why the manufacturer would have to synchronise to the utility line, and would quite probably rely on a crystal clock to provide a stable output. This would make it impractical to use an STS between the two UPS outputs, as the two UPS's would not be in synch other than by chance.

 

[HawkBrazil]

Dear friends,

I woulkd like to thanks a lot about that attention.

I had one idea, and I like to ask if somebody have tried this before, and if this have success, or not.

If I put two rectifiers with regulated output, one for each output of thas ups's that i have mentioned before, then I make one CC busbar with one level of tension to be defined later.

After that, from that CC busbar I put two inverters ( to allow redundancy ) and make another busbar, supplied of CA at the level of the loads.

Finally, I put that "thing" inside a very nice box.

The idea is, if I have that two sources in parallel ( with a adequated protection to prevent a propagation of one fail from one to another UPS in any cases )and not a static transfer switch, I can protect the load with a time to transfer "less than zero" (just kidding...), cause I have that two UPS's now in parallel ( things that only a CC can be give to us ).

So,I have three little questions:

1) Can I expect a good results ?, and
2) What name I can give to that thing ?
3) The efficiency of that assembly is too low ?

Thank so much to my friends.

Hawk

Wagner Freitas
WR Freitas Consulting Engineers
Rio de Janeiro - Brazil

 

[JohnHamburger]

It sounds to me that ScottyUK's suggestion will get the job done. Since you are cascading the UPS's, you will need a utility source incase the backup UPS requires service. A panel with two breakers and a mechanical interlock will suffice, since a break in power would be allowed on the backup system. Good luck.

 

[ScottyUK]

Hawk,

Tell us, is this an existing installation that you are working with, or a design for a new installation? If you are working on a new installation, there are several UPS models which will give you a better level of redundancy than the system I proposed in my first reply. I had the impression that it was an existing installation.

I'm out of the office this week, but next week I could probably e-mail you a single-line drawing of a dual utility / redundant UPS / generator system we're using at our plant. It might give you some ideas to take forward.

 

[HawkBrazil]

Scotty,

Yeah, this a old installation where I have a mission of improve the security of the loads at one CPD.

Some complementar data:

- We have one UPS - 30 kVA - without a by-pass feature.1 ond half Y.O.
- We have another UPS - 15 kVA - with a by-pass feature - 6 months old
- The manufacturers is different
- Both UPS´s are double conversion and not true on line.
- Both manufacturers cannot provide any board or adapter to allow any solution for this case.
-

As we can see, we have different vectors at the outlet of both UPS´s. Due to that nobody in Brazil want´s to make a STS, estabilishing a condition of changes in one oh that equipments ( electronic ) to make the sinus match, and the transfer in less than 4 ms can be possible without we loose the load.

So, if I put in a cascade configuration, if a fault happens in a 30 kVA ( upstream - no internal by pass ), I can manually turn on the external by-pass after tuns this UPS off if mains is present, with no loss of the load. Otherwise this UPS feeds the 15 kVA ( downstream ) till his batteries goes out, then the downstream begins the discharge of his batteries. Perfect, the positive view is the improvement of time can we opperate in a batteries and the negative way is the impossibility of use of a external by pass of the downstream UPS, in case of fail, due to a different vectors. It´s only can be possible, if in despite of the fail can i turn the internal by pass on and then the external by pass is turned on.

By my last suggestion, I can turn any ups or by pass in despite of sincronism, extract any UPS, put the mais in one line and a UPS in another line, with no zeros. I don´s have any type of alarm in both UPS´s, then that assembly must have a voltmeters ( one for each ups ) and a ( by one internal battery )relays with a NC contact´s to provide a local and remote audio visual alarm´s in case of fault.

My only question is the operation this system can be perfect, by a my coleagues experiences, or what disturbs can be happen if I use this solution.

Thanks to everybody for the attention.


Wagner Freitas
WR Freitas Consulting Engineers
Rio de Janeiro - Brazil

 

[DaveMarz]

Wagner,

Considering what you are trying to do and the time and money it would cost you might be better off buying two new UPS units and going from there.

Just about every UPS out there now is set up for communications with your network, either through an RS 232 port or an SNMP card. The UPS units will be able to talk directly with each other and share the load that way. This also has the advantage that the UPS computers will constantly check each other and make sure that there are no errors.

The manufacturer becomes responsible for correctly designing the system. This will save you plenty of time and prevent you for second guessing yourself and any downtime you may encounter testing your configuration.

I sell UPS systems for a living and I would be happy to help configure a solution for you.

Dave
davidmarz@yahoo.com

 

[jbartos]

Suggestion to HawkBrazil (Electrical) Feb 2, 2004 marked ///\\If I put two rectifiers with regulated output, one for each output of thas ups's that i have mentioned before, then I make one CC busbar with one level of tension to be defined later.
///What about a single failure criterion on the CC busbar?\\After that, from that CC busbar I put two inverters ( to allow redundancy ) and make another busbar, supplied of CA at the level of the loads.
///One fault on the CC busbar will cause both inverters to be down.\\Finally, I put that "thing" inside a very nice box.
///Would be even worse from a single failure criterion and common mode failure.\\The idea is, if I have that two sources in parallel ( with a adequated protection to prevent a propagation of one fail from one to another UPS in any cases )and not a static transfer switch, I can protect the load with a time to transfer "less than zero" (just kidding...), cause I have that two UPS's now in parallel ( things that only a CC can be give to us ).
///Paralleling of two source double the short circuit current level, causes various detrimental circulating currents, causes a selective coordination of protective devices more complex, etc. Therefore, the static transfer switch between two UPS outputs is normally implemented.\\So,I have three little questions:

1) Can I expect a good results ?,///Reasonably good, but not the best ones.\\ and
2) What name I can give to that thing ?
///Hot paralleled UPSs.\\3) The efficiency of that assembly is too low ?
///Low.\\

 

[HawkBrazil]

Dear JBartos,

Tks once more about your atention,

If I eliminate the CC busbar and maintain only the AC busbar, the SPF only changes your position. Recently I was read one very interesting article at MIL, where they shows the impossibility of elimination of SPF's at all. The only good notice is if we can develop solutions for control of SPF's.

I agree totally with you about dificulties to coordinate and increase of a short circuit level, but by a adequated protection, can we live with this with no suffering. If that solution shows the expected results my company can implement that in many clients in Brazil, where some solution involving 2 or more sources can be useful.

I cannot put a STS, due to both UPS's are double conversion models, and not a True On Line, then the vectors are in diferent position at the outputs ( not sinchronized ).

Tks once more.



Wagner Freitas
WR Freitas Consulting Engineers
Rio de Janeiro - Brazil

 

[ScottyUK]

Hi Wagner,

Check the rating of the transfer switch in the 15kVA UPS. It will probably not be able to carry the full output of the 30kVA unit, although it is often a simple case of changing the thyristor modules used in the STS to a type with a higher current rating when dealing with a relatively small STS. If we were dealing with larger thyristors, it would be necessary to have a close examination of the gate drive circuitry to ensure it could adequately drive the gate of the larger device, but on such a small unit I doubt this will be a problem.

If your 30kVA unit is being under-run at 15kVA or less, the above will not apply.

 

[jbartos]

Comment on HawkBrazil (Electrical) Feb 4, 2004 marked ///\\Dear JBartos,
Tks once more about your atention,
///You are welcome.\\
I cannot put a STS, due to both UPS's are double conversion models, and not a True On Line, then the vectors are in diferent position at the outputs ( not sinchronized ).
///The STS would have to perform the transfer at the minimum voltage difference within 10ms at 50Hz or 8.33ms at 60Hz.
The UPSs output worse case phase shift is 180°, which corresponds to 10ms at 50Hz or 8.33ms at 60Hz. If the STS performs switching within 4ms then there is required either 3.3 ms delay at 60Hz or 6ms delay at 50Hz in the worst delay or phase mismatch situation. This could be an option provided by the STS manufacturer. The best transfer would be with 0 Volts between already disconnected faulty UPS output residual voltage in the rotating loads and the stand-by UPS output voltage.\\

 

[jbartos]

Suggestion: Visit

http://www.energyusernews.com/CDA/A...ures/BNP__Features__Item/0,2584,14487,00.html

for:
The time-honored concern about transferring an HVAC load between unsynchronized power sources is negated with today's ATS and industrial control systems, which include microprocessors that employ sophisticated control and switching strategies.

http://www.caterpillar.com/products/engines_n_power_systems/spec_sheet_library/EPG/pdf/LEXX0524.pdf

http://www.dynagen.com/switches/ge/pdf/ztsBroch.pdf

etc. for more info