Short Circuit Analysis of Closed Transition ATS 2

[PaultheBigFinn]

My company designed synchronizing switchgear to bring four 1 mW generators on line to a single bus to supply power to a large hospital's emergency and stand-by systems. We calculated the available fault current from the generators and specified the short circuit withstand rating well above this at 65 kAIC. We did not design the distribution switchgear which coincidentally was also rated 65 kAIC.

A third engineer was hired by the hospital to perform a short circuit analysis of the hospital’s large and fairly complicated power distribution system. They insist that because the ATSs are closed transition, the equipment must be rated for the utility's short circuit contribution as well as the generator's and as such the rating of the switchgear can not be less than 85 kAIC. This in spite of the fact that the time both systems are exposed to each other is less than 100ms for each transition.

Interestingly the SKM modeling software does not have separate model for closed transition ATS, only ATSs in general. The engineer who did the study constructed his own using two switches connected to a common bus.

I found no exception in the code for such a short time, but then there is no discussion of the impact of ATSs either. Is there a precedent for this?

BTW there are ~31 million seconds in a year and if the system was tested and or used once a week the entire exposure time would be about 10 seconds.

Any thoughts would be appreciated.

 

[alehman]

This has been discussed in a few other threads:

thread238-249427
thread238-239590
thread238-238075

Alan
----
"It’s always fun to do the impossible." - Walt Disney

 

[PaultheBigFinn]

Thanks for the links. I'm glad to see that the argument is well reasoned. As I see it;

Several engineers have questioned the need to increase the ratings of all of the distribution equipment in a facility to protect against a very rare event occurring during the 0.00078% of the time the ATS is in transition. They have submitted proposal to the get an exception in the Code for short term switching events.

The technical committee has rejected these proposals as being too specific and defer to the AHJ.

The AHJ, in my case, is the Fire Marshall, who is not an electrical engineer and who, when given a choice, always picks the most restrictive interpretation, feeling that the expert Code writing engineers should address it.

Let’s all just be glad we don’t get all of the bureaucracy we pay for.

 

[ScottyUK]

Paul,

Consider how much greater the possibility of an incident occurring is when the equipment is being operated compared to when it is being left alone. The most likely way to initiate a problem is to disturb whatever state switchgear is in, and opening or closing a breaker does that quite effectively. Switchgear is very reliable, but you really don't want to be standing near a breaker breaking a fault significantly greater than its rated breaking capacity.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[davidbeach]

I have to agree with ScottyUK; your 0.00078% is only relevant if faults are truly random events. Faults aren't truly random though; switching greatly increases the probability of other things going wrong.

 

[PaultheBigFinn]

One needs to think about the sequence of events when closed transition ATS switching occurs. On a power failure there is no utility to contribute, so going to generator power is not an issue.

When the utility returns the ATS waits until it is synchronized with the generators and the ATS switches back to utility power. The only device changing state is the ATS itself, the circuit breakers in the switchgear and through out the system remain in their original positions.

In my specific case the ATS switch is an actual mechanical device driven from one position to the next. In my 30 years in the industry, I have never seen a closed transition ATS fail "in transition." Once the switch starts moving it continues through the transition point to the alternate position. ATSs seldom fail, but the failures I have seen occurred with the switch frozen in one position or the other. While I agree that the chance of a failure may be a little higher, I think it is largely random. In my case, transition is specified as less than 100ms. That is against ~30 million second in a year.

In the hospital we are working in there are ~80 ATS distributed throughout the million square foot facility. It is only when a failure occurs in one of the four ATSs located in the Penthouse between Substation #2 and the syncronizing switchgeear that the fault levels of ~85 kAIC occur (the rating of the switchgear is 65 kAIC). The longer feeder runs to the other ATSs burn off the fault current before it can reach the ATS and return to the switchboard. Right now I am looking at ways to reduce the fault contribution from the utility to these four ATSs.

As always cost and space are the problem. Anyone have a suggestion?

 

[waross]

I have seen series reactors installed on generator outputs. I would be looking at reactors on the generators. Then I would investigate the possibility of connecting the AVRs downstream of the reactors to compensate for the voltage drops across the reactors.
First though, I would see if the issue may be resolved by tripping the generator breakers ahead of the main breakers.
What is the operating time of the clearing switchgear in relation to the closed transition time?

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[alehman]

Is there any way to increase the circuit lengths (relocate these ATS's)? It might not take a lot to get from 85 to 65k.

Current limiting reactors seem like a possible option.

Alan
----
"It’s always fun to do the impossible." - Walt Disney

 

[waross]

Is there any way to increase the circuit lengths (relocate these ATS's)? It might not take a lot to get from 85 to 65k.

I like this suggestion. I worked on the construction of a world class mine mill where this technique was used to limit the fault current to the unit subs. The specs called for a minimum of 100 feet of cable to each unit sub, the excess if any to be doubled back in the cable trays. Most of the subs were more than 100 feet away, but 4 or 5 where fed with overlength cables to limit the current. Saved a lot of money by avoiding the purchase of higher rated unit subs. As I remember this was on the 13kV system.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[ScottyUK]

In my 30 years in the industry, I have never seen a closed transition ATS fail "in transition."

Until 8 years ago I had never seen a 50MVA transformer blow up either. Then one day it happened, and everything changed for the people who were there. No offence but not having seen something bad happen is a poor argument. At least you are trying to do something about it.

This must be an ACB-based transfer switch. Is the problem on the breaker ratings or on the bus bracing? Have you asked the ACB vendors if their higher rated ACB is physically interchangeable? If it's a bus bracing problem then that's harder to deal with, especially if the manufacturer has no type-test data for the higher fault level.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[rbulsara]

Probability of a fault is not the basis of determining short circuit rating of the equipment.

Rafiq Bulsara

http://www.srengineersct.com

 

[slavag]

Hi
Im fully agree with :
"Probability of a fault is not the basis of determining short circuit rating of the equipment. "

Never say never .

From my point of view, Bill's recommendation , reactors, is good option.

Best Regards.
Slava

 

[xxjohnh]

Standard Industrial practice is to ignore fast closed transition in fault calculations. It is called an engineered solution. But in some cases the authority having durisdiction insists on it, you can protest, but be prepared to lose.

 

[PaultheBigFinn]

Anyone wanting to know more about this should follow the links in the first response by alehman. They were very help full. One of those links led me to Mike Holt’s forum and a post by rbalex, who signs as “Bob, Robert B. Alexander, P.E.”

http://forums.mikeholt.com/showthread.php?t=110459

Apparently there have been proposals to add an exception for short duration dual feeds since 1996, but the committee wants to leave it up to the AHJ.

I’d like to note for those who noted that the rareness of the even shouldn’t be related to safety that it is a triplely rare event (if triplely is a word?). First the system would have to have to experience a bolted fault. Second the system would have to deliver the maximum theoretical short circuit current AND the fault would have to occur in the 3 cycles the ATS was in transition. Let’s not forget the first sentence in the Code: “90.1 Purpose, (A) Practical Safeguarding. The purpose of this Code is the practical safeguarding of persons and properties from hazards arising from the use of electricity.”

Is guarding against an event that is 5 sigma practical?

That said, I am still looking for something to limit the current to my trouble making 480V 1,200A ATSs to guard against the triplely rare event. I don’t have any experience with current limiting reactors and the ones I found are for use with higher voltages, but none for 480V systems. Any suggestions?

 

[waross]

Use longer connection cables.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[ScottyUK]

I guess it comes down to how the legal defence in your country works. In the UK there is the premise of 'reasonably practicable' and in our legal system I think that defending a design which knowingly overstresses equipment by intent would leave the designer with some very difficult questions if a problem occured. If the technology simply did not exist to overcome the problem, or the cost of doing so was hugely disproportionate to the benefit obtained, then there would possibly be a reasonable defence, but the technology does exist and the premium paid for higher rated switchgear is not all that large. That is pretty much indefensible.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[PaultheBigFinn]

We are moving off the topic a little, but there are always things one can do to improve the quality and durability of a design. Where does it end?

Your phrase "a design which knowingly overstresses equipment by intent" doesn't describe the situation in my mind. The odds that the system would experience a bolted failure with the maximum short current possible during the 3 cycles the ATS was in transition are so small that they are for practical purposes none existent.

I feel any cost spent is exorbitant because there is no discernable benefit. Division by almost zero creates a very big number. That said, I have conceded that my client and the local Fire Marshall think there is so perhaps I am wrong. I am in the process of finding a solution to the problem for the lowest possible cost.

We are also looking into the possibility of switching the ATS furthest from the generators first which would allow us to take one or two generators off line and removing their fault contribution.

 

[davidbeach]

First off, for a protection engineer the fault is a given, probability of 1. All the rest of it is just figuring out how bad the fault can be and when is the worst time for the fault to happen to the system.

So, for this case we have the fault as a given, the worst case fault current is bolted, and the worst case time is during switching. None of those are a stretch at all.

The only probability involved is 1, there is no very small number. If you choose to believe that it is a very small number, I can assure you that opposing council will find many witnesses that can convince the jury otherwise.

 

[PaultheBigFinn]

I'm sorry Dave, but you are wrong. If this was correct: Every exterior piece of equipment in the should designed able to surive a direct lighting strike. Every car should be designed to survive a tire blow out at 55 mph without damage. Every airplane should be able to fly through a flock of geese at take off without damage. All buildings should be designed to survive a force 7 earthquake.

The of any of the above is far greater than than the likelyhood that the system would experience a bolted failure with the maximum short current possible during the 3 cycles the ATS was in transition.

The spector of appearing before a jury is ruining the country, IMHO, and I choose not to live my life in fear of lawyers.

This of course is not the problem. The problem is that my client does live in fear of lawyers and now I need a solution to a failure that I believe will not occur.

 

[rbulsara]

Paulthebigfinn:
Davdibeach is not wrong. You are trying to justify something that has no technical basis.

You can take all risk you want, but that does not make it right. Nor are you in a position to claim undue hardship or cost. 100KAIC equipment are routine.



Rafiq Bulsara

http://www.srengineersct.com