Critical UPS systems coordination issues

[Zcon]

Discovered through actual problem and damage that UPS systems are not protected for ground faults properly. Reason most UPS units go into overload on a ground fault at 1/4 of a cycle and the branch circuits they serve are protected by molded case circuit breakers mounted in UPS panels. These breakers trip at 1/2 cycle so a small ground fault on a branch circuit can put all connected loads in danger off dropping out. The fix is special UPS fuses or panels fuses are put in series with the connected loads on the branch circuits. I suspect that many UPS panels are installed this way the panels were installed in my case by contractors in 2006 then about 2010 because of no anti short bushings in a underground pull box a ground fault occurred on a 30 amp double pole sub feed breaker. The ground fault caused the whole new UPS to trip out losing all UPS panel loads at the same time the way it looks to a layman being called out in the middle of the night is everything down with no trips on the molded case breakers and a overload alarm on the UPS serving them all. The UPS was checked out for damage after ohm meter checks were done on the branch circuits the panels that were clear were reenergized the faulty circuit was locked out for safety. In this case severe damage was done to equipment and people came close to being hurt the new control system that the new UPS was serving went completely dead everything quit including redundant power supplies and controllers with no way to control anything and no critical indication of the plant equipment status available all fail mode logic that was programmed in the valve slaves did not work. There is some talk of UPS coordination requirement now on the web but very little in 2010 and I thought the industry would have put out a mandate to repair all systems installed this way but it has not happened these systems are time bombs waiting to happen but nothing is being done. So I went to a code update class and asked if anybody knew about this and nobody did the word is not getting where it needs to be.

Zcon

 

[petersonra]

UPS's in general are not all that reliable. I generally try to avoid them since as often as not they just add unnecessary complications and failure points. If I have to have a UPS I try to make it a DC UPS if possible. They seem a lot more reliable than converting AC to DC and then back to AC.

Coordination between various OCPDs is something that is often overlooked. I just did a look see at a system where 7A feeder breakers on an instrument system were tripping before 2A fuses that fed individual loads. The trip curves showed why. I guess when the system was designed no one looked at the trip curves. I have suggested replacing the 7 A CBs with FNQR fuses. The FNQR trip curves do not overlap with the trip curves of the 2A fuses already there so they should not blow like the 7A CBs.

The 2017 code says this about coordination.

240.12 Electrical System Coordination. Where an orderly
shutdown is required to minimize the hazard(s) to personnel
and equipment, a system of coordination based on the following
two conditions shall be permitted:
(1) Coordinated short-circuit protection
(2) Overload indication based on monitoring systems or
devices

It would thus seem like the code already requires the coordination you are asking for.

Incidentally, this was in the 2008 NEC as well, so it would appear that whoever designed the UPS system did not abide by the code requirements in existence at the time the UPS was supplied.

As an aside, it would appear like someone did not design the system as a whole to be safe if the failure of a single item caused "severe damage ... to equipment" and "people came close to being hurt". That is just a horrible design.

 

[synchro]

. The fix is special UPS fuses or panels fuses are put in series with the connected loads on the branch circuits.

Zcon

I assume that these fuses on the branch circuits would be of the "current limiting" type to limit the magnitude and duration of the fault current?

 

[Zcon]

UPS's in general are not all that reliable. I generally try to avoid them since as often as not they just add unnecessary complications and failure points. If I have to have a UPS I try to make it a DC UPS if possible. They seem a lot more reliable than converting AC to DC and then back to AC.

Coordination between various OCPDs is something that is often overlooked. I just did a look see at a system where 7A feeder breakers on an instrument system were tripping before 2A fuses that fed individual loads. The trip curves showed why. I guess when the system was designed no one looked at the trip curves. I have suggested replacing the 7 A CBs with FNQR fuses. The FNQR trip curves do not overlap with the trip curves of the 2A fuses already there so they should not blow like the 7A CBs.

The 2017 code says this about coordination.



It would thus seem like the code already requires the coordination you are asking for.

Incidentally, this was in the 2008 NEC as well, so it would appear that whoever designed the UPS system did not abide by the code requirements in existence at the time the UPS was supplied.

As an aside, it would appear like someone did not design the system as a whole to be safe if the failure of a single item caused "severe damage ... to equipment" and "people came close to being hurt". That is just a horrible design.

 

[Zcon]

I guess if I had to evaluate 240.12 it is a general statement and can be applied to many situations I just wish there was more about UPS coordination only and the reason it is needed. Yes these are current limiting fuses and myself and a good friend engineer that I got involved looked at the time curves then we did a direct short simulation to prove the problem was fixed with different size FNQR fuses and the UPS never tripped all the circuits got these fuses and I added led lights on the load side of the fuses to provide blown fuse indication. I worry that people could get hurt and damage of gear could occur and want to head off some of it if possible. Just not sure how to get the word out the electrical contractor that installed this system can't be the only one who overlooked the coordination.

Zcon

 

[ron]

I would almost bet that 240.12 ("system coordination") does not apply here in the way you are thinking. It requires short circuit protection and overload indication. Sounds like you are thinking it is similar to selective coordination, which it is not.
I would suggest that you check the calculated available short circuit current at the UPS input and see if the UPS is rated at least that value.
Depending on the UPS design style, it may be a separately derived system, which would require GFP on the output, which may have helped your situation depending on how large the UPS was compared to the 30A branch circuit with the problem.
The fact that you mentioned that an anti short bushing was the problem, if they are using AC cable instead of MC cable or flexible metal conduit for branch circuits, implies that it may be a low end project to the lowest bidder.