It is currently Sun May 24, 2020 3:55 pm



Post new topic Reply to topic
Author Message
ekstra   ara
 Post subject: Data Centers
PostPosted: Tue Feb 10, 2015 9:51 am 

Joined: Tue Feb 17, 2009 11:04 am
Posts: 37
Colo Data Center Providers
As it pertains to OSHA and NFPA 70e are Colo Data Centers de-energizing or energizing when installing client branch circuits?


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Data Centers
PostPosted: Wed Feb 11, 2015 7:47 am 
Plasma Level
User avatar

Joined: Tue Oct 26, 2010 9:08 am
Posts: 2174
Location: North Carolina
You have to provide more context than that.


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Data Centers
PostPosted: Mon Feb 16, 2015 4:43 pm 

Joined: Tue Feb 17, 2009 11:04 am
Posts: 37
Cola data centers have multiple clients and expect the servers to be 99.9999% reliable in providing the service.


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Data Centers
PostPosted: Mon Feb 16, 2015 9:18 pm 
Plasma Level
User avatar

Joined: Tue Oct 26, 2010 9:08 am
Posts: 2174
Location: North Carolina
What specific task is being performed?

By the way, OSHA is very clear in a letter of interpretation on the subject that loss of production is not a legitimate justification for live work for utilization equipment.


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Data Centers
PostPosted: Wed Feb 25, 2015 6:28 pm 

Joined: Fri Feb 13, 2015 11:54 am
Posts: 6
Location: Alaska
Data centers that I have been involved with have used busways in their server rooms. The bus plugs for the branch circuits are designed with IP2X compliant (finger-safe) bus stabs and with the bus plugs interlocked such that the built-in disconnect switch must be in the open position when being connected or disconnected to/from the busway. Connecting a bus plug to a live busway should be no more hazardous than plugging in a cord into a live receptacle.

NFPA 70E consistently invokes safety requirements only where workers are potentially "exposed" to live energy. It defines the term, "exposed" as "capable of being inadvertently touched or approached nearer than a safe distance by a person. It is applied to electrical conductors or circuit parts that are not suitably guarded, isolated or insulated." It would be difficult to argue that a modern bus plug design such as the Siemens Sentron system is not suitable guarded, isolated and insulated, and hence can be safely installed on a live busway without requiring electrical PPE.


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Data Centers
PostPosted: Fri Feb 27, 2015 6:10 am 
Plasma Level
User avatar

Joined: Tue Oct 26, 2010 9:08 am
Posts: 2174
Location: North Carolina
70E has TWO requirements on energized work. The first requirement is whether or not equipment is exposed, and there is no argument there. This captures shock hazards. The second requirement is for arc flash. It is not necessary for equipment to be exposed for a task to cause an arc flash. In this case, the argument to be made is when does it go from say cord-and-plug equipment which is defined under 70E, article 110.4(B), to equipment which uses a "plug" but which does not afford the same level of safety as cord-and-plug equipment. OSHA and NEC generally treat equipment rated 240 VAC single phase or less with up to 15-20 A plugs as distinctly different from say a 480 V, three phase, 4-wire plug, or say stabs in a busway, MCC bucket, panelboard, switchboard, or switchgear. Three major differences come to mind immediately:
1. Even without a switch to de-energize the load side, 15-20 A plugs can certainly arc. There is one recorded case in OSHA investigations in 2007 where an IT technician plugged in a laptop into a defective receptacle where it arced at the plug and burned the IT worker's hands, requiring hospitalization due to 2nd degree burns. 70E is intended to ensure survivability, not injury-free, so 70E would not require any sort of arc flash PPE to avoid this injury. Thus even if an arc flash hazard exists this type of equipment is generally acceptable even without a de-energization requirement without special PPE. Nothwithstanding even this design, GE Limitamp starters for instance are fairly well known where the linkage mechanism fails and has even caused at least one fatality and are pretty well known for "blowing up" on occasion. Here is documentation on the fatality that I know of:
http://work.alberta.ca/documents/WHS-PU ... -12-12.pdf
2. Stepping up to 480 V welding receptacles which are common in an industrial setting, the basic plug design is still a NEMA plug. The stabs can still mechanically misalign or bent into one another and cause an arc. However the arc will not self-extinguish and thus the arc flash can be substantial. These plugs have been known to explode while inserting or removing them while energized. Plugs and receptacles such as Cooper "Ark Tite" and Meltrics line of plugs as examples are mechanically interlocked such that shutters block insertion or removal while the receptacle is energized via a safety disconnect. This eliminates the issue, but the "line" side is de-energized.
3. Stepping up to panelboards, switchboards, MCC buckets, and switchgear buckets, there are sometimes mechanical shutters offering a degree of cover over the receptacles similar to the latest 2014 NEC requirement for residential 15/20 A receptacles, but generally there is not enough clearance afforded for workers to place their hands or fingers anywhere near the stabs when they engage into the bus. Irrespective of this design which would be truly shock hazard free, frequently the stabs still misalign and contact either each other or grounded enclosures and induce an arcing fault. ABB reports that 80% of arcing faults in their switchgear are due to the stab mechanism, and IEEE 493 data shows that drawout switchgear reliability is less than its bolt-in cousins.

That all being said, I personally believe that it is possible to design a plug-in arrangement where both shock and arc flash hazards can be eliminated during insertion or removal. If the stabs are spaced out where they cannot contact each other even if two of them are folded over, and the areas where the mating surfaces are at are all insulated, and keeping up with other designs, shutters are used to eliminate shock hazards, and there is a disconnect mechanism (mechanically interlocked to the shutters) on the line side. The sole remaining issue is that when the disconnect is closed, if not all stabs are electrically good connections, an arcing fault can still occur where a pin does not seat solidly into the socket. It may be possible to even design a mechanical interlock to prevent this but I'm going out on a limb here. If all of these precautions were taken then I can agree that ANY stab mechanism, not just limited to busway, can be made where I would argue that section 130.1 of 70E does not apply and that this is just operating equipment, not energized work,under the condition that if an arcing fault were to occur, it would be great enough to not satisfy condition #1 above (not above 1.2 cal/cm^2 or equivalent). This is the most conservative approach. However once you back away from this I would agree that some equipment can be provably shown to be reliable enough to not require this much. But don't fool yourself into thinking that just because there is no shock hazard that an arc flash hazard doesn't exist. And even if it does, it is not carte blanche enough to mean that it falls under 130.1. The definition of an arc flash hazard in 70E includes terms related to the likelihood of occurrence. 70E does not define this numerically because it is industry specific. But if you know your acceptance criteria then in some cases some of the more reliable equipment could be considered acceptable because the likelihood of an arc flash is no more than other comparable serious accidents and fatalities. I would argue though that in the case of data centers, the incident rate of serious injuries is pretty low, so the incident rate for arc flash should be comparable, and well below national averages.


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Data Centers
PostPosted: Mon Mar 02, 2015 6:27 pm 

Joined: Fri Feb 13, 2015 11:54 am
Posts: 6
Location: Alaska
Paul, I don’t have any disagreements with anything you said with the exception of the discussion of bus plugs, which are a special case and are specifically designed to be safely installed or removed without requiring deenergization of the busway. A typical data center server busway system operates at 480V or less, and is designed to be safely configurable without sacrificing uptime.

It seems to me that the core consideration is whether or not the worker is "exposed" to an electrical hazard. The plug-in openings in modern busway bus system are finger safe (IP2X compliant) exactly to the same standard as for a wall receptacle, so there is no possibility of inadvertent contact by a person, and workers are therefore not exposed to live energy as defined in 70E, Article 100.

As you stated, we also have to ensure that workers who may not be exposed to live parts are not also exposed to a likely possibility of arc flash (70E, 130.2). The modern bus plug assembly is provided with narrow plug-in openings in a keyed configuration and with alignment slots to ensure proper installation. They are also interlocked so that the built-in disconnect is maintained in the open position and cannot be closed until the bus plug is fully seated into the busway. The bus plug has no gates or shutters that may malfunction as in the fatality report you cited (thanks – an interesting read!) This prevents any current flow that could result in an arc until the switch is later manually closed under electrically safe conditions.

A modern data center busway system is designed specifically to prevent exposure to hazardous energy and should be able to be safely tapped without credible risk of shock or arc flash. This assumes that all equipment is in good working order - but that is always the base assumption, is it not?


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Data Centers
PostPosted: Tue Mar 03, 2015 4:09 am 
Plasma Level
User avatar

Joined: Tue Oct 26, 2010 9:08 am
Posts: 2174
Location: North Carolina
70E is for work. It assumes everything is properly maintained, not that issues NEVER occur. For instance fixing a problem in a busway assumes the breaker/fuses protecting it work, so PPE specified while doing the work also can do its job. If nothing fails, 70E live work sections are unnecessary.

When looking at safety, engineers have an inherent hard time with it. We have to think 'backwards'. Instead of thinking how to accomplish some goal, the task is to think of every way it can fail, without falling into the chicken little trap of thinking 'everything' which is not only hopeless but can't be analyzed. These lines of thinking get us E-Stop buttons as a catch-all despite research which shows E-Stops don't work because the behavior of the human using them is the issue.

http://www.busplugsblog.com/tag/arc-fla ... -analysis/
http://ecmweb.com/content/case-obsolete ... cal-system

Lots more where those came from. Bus plugs fail occasionally. Busways frequently have high fault currents, so the result is when the do fail, its catastrophic.


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Data Centers
PostPosted: Thu Mar 12, 2015 2:53 pm 

Joined: Fri Feb 13, 2015 11:54 am
Posts: 6
Location: Alaska
@PaulEngr

Those were interesting links, both good examples of hazards arising from obsolete and defective electrical equipment.

In contrast, a modern NRTL-listed bus plug in good repair is every bit as safe to insert or remove from an energized busway as plugging an appliance into a standard NEMA receptacle - safer in some ways, as an appliance plug can be pulled from its receptacle under load causing a switching arc as contact is broken; and when partially inserted, appliance plug prongs are exposed and energized presenting a potential shock hazard.

A modern bus plug cannot be inserted or removed without first opening its disconnect and thus preventing all possibility of arc flash except during switching of the disconnect. The bus way openings are finger safe to the same IP20 standard as a NEMA 5-15R receptacle, however unlike an appliance plug, the bus plug stabs are not exposed during insertion or removal from the energized plug.

I see no reason why a modern, well designed data center bus way system operating at 480 volts or below would be considered less safe than a conventional panelboard supplied distribution system.


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Data Centers
PostPosted: Mon Mar 23, 2015 7:33 am 
Sparks Level

Joined: Wed Oct 29, 2008 6:06 am
Posts: 136
Location: Michigan
Where a worker could potentially make contact with exposed energized parts is NOT the only situation in which an electrical hazard could exist. Even if suitable guarding exists to protect against electric shock hazards, it does not mean that an arc flash hazard can’t still be present, and the likelihood of an arc flash event increases significantly with worker interaction.

The major difference between a 120V, 20A convenience receptacle and a IP2X complaint 480V, 1500A busway is the amount of energy available. Attached are images of a 480V, 2500A busway <15 yrs. old with a large hole where a bolt used to be; note the scorched area on the ground surrounding the fault location.

You stated that the bus plug disconnects are interlocked and therefore it is impossible for the disconnect on the plug to be closed until the plug is fully seated into the busway… I’ve seen brand new right out the box plugs with manufacturer issues, and often times we aren’t pulling a brand new plug out of a box but rather taking something off an in-house shelf that has been previously used (either simple inspection or refurbishment).

If you looked at the manufacturer’s installation instructions posted on the inside of the bus plug door, it likely says to de-energize busway before installation. If performing this task while energized, I at least hope your guys are wearing PPE…


You do not have the required permissions to view the files attached to this post.


Top
 Profile Send private message  
Reply with quote  
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 10 posts ] 

All times are UTC - 7 hours


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Jump to:  
cron
© 2019 Arcflash Forum / Brainfiller, Inc. | P.O. Box 12024 | Scottsdale, AZ 85267 USA | 800-874-8883