Sizing Equipment for Closed Transition from One Source to Another

[billyzee]

Mr. Benjamin Medich had a recent article published in CSE magazine (Sep. 2012). The subject of the article was sizing equipment for closed transition operation. In the article he presented a rationale supporting the notion that fault contributions of all parallel sources must be considered only for systems designed for continuous parallel operation. The conclusion was that there is no need to consider both sources of fault current when sizing equipment for closed-transition transfer schemes. The summary of the article is presented below:

"Based on a review of applicable codes and standards, there is no need to consider both sources of fault current when sizing equipment for closed-transition transfer schemes. However, when making this determination, it is important that the proper electrical interlocking or other supervision techniques are used to ensure that the system cannot be inadvertently placed into a maintained parallel state, which would require that the equipment be sized for the combined parallel sources. This will assure a safe, cost-effective installation that complies with NEC requirements, and is consistent with IEEE guidelines."

I am interested in hearing contrary and supporting opinions.

 

[rbalex]

These are CMP1's comments on the subject. You can draw your own conclusions.

 

[billyzee]

Robert,
Thats great stuff! It isn't hard to figure out your opinion on the matter. You are very persistent. I agree with the position that you took in your proposal. I too have been proposing a similar design philosophy but am being rejected by more powerful entities.

 

[iceworm]

These are CMP1's comments on the subject. You can draw your own conclusions.

I have. The code panel is stuck. In six pages, they only made two statements:

1. 90.4 will handle it.
2. No technical substantiation.

First 90.4 is useless. The idea that one would do design work based on what an AHJ rep might think an any random morning is beyond ludicrous.

As for no substantion - well, only in their mind.

And one person disagreed - sort of. that was even a bit convoluted.

Let's look at the physics:
Following discussion centers on double-ended sub, MTM, operated closed transition switching:
There is this faulty premise that the tie CB, possibily the xfm secondary mains, and the bus bracing have to be rated for the combined SCC of the two transformers. Now, what part of switchgear could be subject to the additive fault current? If the fault is on the bus on either side of the tie, then the tie is only subject to the fault current of one xfm. If the fault is on either side of a secondary main, then that CB is only subject to the fault current of one xfm. What about a fault on the bus its self? Okay, each side leading to the fault is still under the stress of the fault current of one transformer.

So, the only places that are subject to the double fault current is the point of the fault on the bus, or if the fault occurs inside of the of a tie or secondary main. So tell me will the blast inside of a failed 65ka CB be any worse than inside of 25Ka cb. The key point being the CB FAILED INTERNALLY. The interupt rating made no difference.

Now how about the bus failing. Well, does the bus bracing make any difference? No. The fault location is blown into molten slag. And the bus that not destroyed is still in place. Although likely un-useable regardless of bracing.

So, what is gained by making the mains, tie, or bracing equal to double the fault current?

Switchgear mfg get more money. Fused swg mfg get a larger percentage. Safety improved? i don't see it.

There is an issue with feeder CBs from MTM switchgear. Next discussion. One of tyou could pick it up. i'll give you a hink. Most of the data is in IEEE std 493.

Presented with all due respect to code panel members. I sure would not do it. I won't even turn in a proposal. I looked through a few hundred pages of ROPs in the grounding section. Most of these didn't even appear to be written with the idea they would even be considered - like there is a competion for she who gets the most useless proposals.

No wonder the code panel appears to buy REJECT stamps by the car load.

So sayeth the worm

 

[Phil Corso]

Iceworm...

I agree wiih your comment as to the location of maximum current, but the arc will be driven away from the source! Secondly, arcing contaminents can quickly engulf the entire bus!

After having experienced a phase-to-ground fault in the tie-breaker compartment, that quickly became a 3-phase fault (vendor relied on wheels to act as grounding-conductor) a "Probabilistic Risk Assessment" was carried out.

A key improvement was encapsulation of the bus (not just insulation) and to ensere the compartments were properly compartmentalized, up to and including the bus stubs! Another important consideration was the protection/interlocking scheme employed. It had to ensure that at least two of the three breakers involved could sense and operate to clear the fault!

Regards, Phil Corso

 

[rbalex]

Iceworm...

I agree wiih your comment as to the location of maximum current, but the arc will be driven away from the source! Secondly, arcing contaminents can quickly engulf the entire bus!

After having experienced a phase-to-ground fault in the tie-breaker compartment, that quickly became a 3-phase fault (vendor relied on wheels to act as grounding-conductor) a "Probabilistic Risk Assessment" was carried out.

A key improvement was encapsulation of the bus (not just insulation) and to ensere the compartments were properly compartmentalized, up to and including the bus stubs! Another important consideration was the protection/interlocking scheme employed. It had to ensure that at least two of the three breakers involved could sense and operate to clear the fault!

Regards, Phil Corso

I don?t believe anyone has suggested that there are no considerations necessary for properly designing a closed transition transfer scheme; however, nothing included in your "Probabilistic Risk Assessment" indicated there was any need to increase (essentially double) the interrupting duty beyond what the devices actually need to interrupt.

 

[Phil Corso]

Rbalex...

Sorry, I omitted the conclusion! There was no doubling of the the interrupting-duty or the 1st-cyle peak withstand capability!

Phil

 

[billyzee]

I have found that the gold book (IEEE 493) does not address anything specific related to sizing equipment for closed transition. It does not present pros, cons or recommended practices directly related to the original line of inquiry.

I do see that IEEE 493 is useful if I want to present an argument with the premise that sizing the gear for one source only is sufficient because of the improbability of a fault coincident with the short time interval when the two sources in parallel.

I played with the failure rates in the appendices of IEEE 493 and decided that if the system was required to make an average of one transition a week and that the transition resulted in parallel sources for less than 2 seconds then it can be demonstrated that a coincident fault was unlikely (10E-7 to 10E-8). The trouble with using IEEE 493 is that I do not see that it addresses consequences. Even if I could present a flawless calculation that shows a 10E-10 probability of occurrence if the consequences is a blowed-up substation with weeks of repair work then I lose the argument.

So what about the physics of the problem that Worm addressed? What if there is a fault on a downstream feeder simultaneous with the transition. Both sources feed the fault at a value of 1X. The bus is braced for 1X. The circuit breakers are rated to interrupt 1X. At some point in the gear both components of the fault current merge and from there on out the bus and breaker are underrated. So I don?t really see how to use this to support the argument.

I welcome and thank all those that wish to comment or chastise me for my ignorance

 

[rbalex]

IEEE 493 (Gold Book) recommends a failure modes and effects analysis (FMEA). After fighting CMP1 for three Code cycles and while doing a four month study (each) on, what were at the time, the two largest refineries in the world we came to the conclusion that automatic double-ended, secondary selective distribution wasn’t all it was cracked up to be. There are both safer and more reliable distribution systems available. Actually, safety and reliability pretty much go hand in hand.

What is often overlooked is that reliability isn’t as much dependent on multiple sources of power; rather it is more dependent on redundant loads.

I have attached a sketch of several fairly common industrial distribution systems. A formal FMEA will find figure “g” with automatic starts of redundant loads is the most reliable. The tie-breaker introduces addition failure modes.

It is also the safest. I can de-energize one side for maintenance and operate on the other. I don’t have to worry about synchronization. Relaying/coordination is simple. There is no need to increase bracing or interrupting duties.

The above is also the most cost effective.

However, an FMEA using transfer switches will find the load at the mercy of the transfer switches’ reliability.

 

[mayanees]

This topic certainly gets my undivided attention.
I happen to agree with the CMP and think that the systems need to be fully rated for the paralleled condition. My thoughts are that if it were to fault during that 8-16 cycle period, there is the chance, albeit miniscule, that the system could explode. It seems like the rationale is all expense-related, since equipment is available to handle the fault current.
The only time I've ever approved a closed-transition multi-source transfer was for a Utility, and I made it conditional on the equipment being out-of-harms-way such that personnel safety wouldn't be compromised.
But I'm also pragmatic to the extent that I want to see a minimum-essential system implementation. So I would support the idea perhaps if the affected equipment were somehow isolated from human access, and the owner signed off on the potential risk. But I don't see how those types of restraints could be practically implemented.
John M

 

[iceworm]

I do see that IEEE 493 is useful if I want to present an argument with the premise that sizing the gear for one source only is sufficient because of the improbability of a fault coincident with the short time interval when the two sources in parallel.

Yes

I played with the failure rates in the appendices of IEEE 493 and decided that if the system was required to make an average of one transition a week and that the transition resulted in parallel sources for less than 2 seconds then it can be demonstrated that a coincident fault was unlikely (10E-7 to 10E-8).

Stastistics is not my strong suit. This is why I was hoping someone else would pick this up. So what does this mean: Any one M-T-M could expect an incident once ever 100 million years? Or once every 100 million switches - that would be once ever 2 million years?

The trouble with using IEEE 493 is that I do not see that it addresses consequences. Even if I could present a flawless calculation that shows a 10E-10 probability of occurrence if the consequences is a blowed-up substation with weeks of repair work then I lose the argument.

I don't know about losing your argument. Look at the probability of an airliner going down. Or if you really want to get dangerous, look at the probility of getting hurt driving to work. NTSB stats are available. I'm guessing you are safer switching double ended subs.

I happen to agree with the CMP and think that the systems need to be fully rated for the paralleled condition. My thoughts are that if it were to fault during that 8-16 cycle period, there is the chance, albeit miniscule, that the system could explode. It seems like the rationale is all expense-related, since equipment is available to handle the fault current.

Yes, it always comes down to money. How much money is available to mittigate a life threatening fault that has a probability of 1 in 1 million years? Perhaps one should consider leaving the risk as is and mittigating the consequences.

I'm not saying one should or one shouldn't. I'm saying it is a design desision. The code needs to stick with what they do best: 90.1.A, B, C - which does not include industrial design.

Just my thoughts

ice

 

[rbalex]

Yes


Stastistics is not my strong suit. This is why I was hoping someone else would pick this up. So what does this mean: Any one M-T-M could expect an incident once ever 100 million years? Or once every 100 million switches - that would be once ever 2 million years?


I don't know about losing your argument. Look at the probability of an airliner going down. Or if you really want to get dangerous, look at the probility of getting hurt driving to work. NTSB stats are available. I'm guessing you are safer switching double ended subs.


Yes, it always comes down to money. How much money is available to mittigate a life threatening fault that has a probability of 1 in 1 million years? Perhaps one should consider leaving the risk as is and mittigating the consequences.

I'm not saying one should or one shouldn't. I'm saying it is a design desision. The code needs to stick with what they do best: 90.1.A, B, C - which does not include industrial design.

Just my thoughts

ice

Section 90.1(C) does apply here. Despite the common attempts to do otherwise, “This Code is not intended as a design specification or an instruction manual for untrained persons.”


Since you brought up statistics, I thought you might defend the position. I was rummaging through my API-SOEE minutes (I have three years and about eight meetings worth for the time the subject was under API discussion) to document the probabilities that the IEEE 666 Tech committee used. I then got sidetracked on a hazardous locations issue and forgot about it. I still haven’t found the reference; but, I do clearly recall the probabilities against a concurrent downstream fault and an undelayed automatic closed transition (the only real risk) were truly astronomical. Technically, the accepted age of the universe isn’t long enough for such event to have happened once.


Phil Corso acknowledged a fault in the tie breaker still would not require an increased interrupting rating. The only risk is with feeder breakers when there is a concurrent downstream fault during the closed transition.


Indeed, economics is a consideration. If we were as risk averse in hazardous locations, we would only consider Division 1 installations – and maybe not even then – we might just not allow Classified location construction at all.