Arc Flash Hazard Analysis

[undervoltage]

While we carry the Arc Flash Analysis using SKM or ETAP, we will get the results for Incident energy, Arc Flash Boundry and PPE required for that particular switchgear. Now the results will depend upon Short Circuit and Coordination Studies.

The question I have is that now a days there are protective relays specially designed for Arc Flash Protection which require only Light & Current or only Current or Only Light input to detect the arc and trip the breaker. These systems are not dependent on Short Circuit or Coordination Studies. Can we model such systems in SKM or ETAP and see the impact. Anybody have such experience?

Thanks

 

[exciting]

The arc detection relays still have a certain pickup and fault clearing time. Take this together with circuit breaker operating time to calculate incident energy.

To model this in Powertools or ETAP just create a custom relay with constant operating and clearing time over a large range (say 1A-100kA).

Careful with analyis of results however. Line side faults of feeding circuit breakers may still need overcurrent protection to trip on arcing faults, depending on how your system is set up.

 

[undervoltage]

In modern arc flash units, we prefer to have a main unit and then some trip multiplier units. So for the feeders which will get trip signal from the Trip Multiplier Unit in case of arc flash, how we will model the trip multiplier units for them. Whether we have to model a main relay for each feeder (even if in real there is a trip multiplier unit) or the trip multiplier units are available in softwares libraries.

 

[rbulsara]

I do not believe light and sensor devices are considered overcurrent protection devices. Recognized method is IEEE 1584 and it only considers regular overcurrent devices and only those should be used while calculating the potential hazard.

Light sensors may additional aids, but not a substitute for OCPD, imho.

Rafiq Bulsara

http://www.srengineersct.com

 

[exciting]

rbulsara, light sensors only is a possibility but not often used because of the practical downside (a flash from a camera can set them off). Most setups I've seen use an instantaneous overcurrent trip unit with a blocking function from the light sensor, ie the unit doesn't trip when no flash is detected. You're right about IEEE 1584 not covering this.

As for trip multipliers, you can only use the results of the arc protection system if you're positive the compartment you're labeling is covered by the system and all sources are tripped. How this tripping is achieved (trip multipliers or not) does not really matter as they appear as instantaneous trips when properly setup.

Because SKM and ETAP don't support modeling systems like this you have to 'trick' them with instantaneous trip units on the incoming feeders and manually filter the results yourself. The danger here is that locations not covered by the system can give incorrect results.

A proper way might be to use a differential protection function in the software, but the issue of careful analysis remains.

 

[rbulsara]

It is not a question of incorrect results or tricking the software. The proposed method is simply not recognized and hence invalid.

Rafiq Bulsara

http://www.srengineersct.com

 

[LionelHutz]

IEEE-1584 broadly says to use the protective device characteristics. There is mention of using the time-current curves, which will be a short fixed time due to the arc-flash relay.

It is acceptable to use the protection relay for the arc-flash calculations. ie, a Multilin or SIPROTEC relay tripping curve is acceptable to use, as opposed to just using a fuse curve. So, why wouldn't a protection relay which integrates light sensing also be acceptable?

 

[rbulsara]

I do not believe light sensing amounts to overcurrent protection. But you do not have to go by my opinion either.

Rafiq Bulsara

http://www.srengineersct.com

 

[undervoltage]

The Switchgear where we are installing arc flash system does not have any Multifunction protection relay. So the overcurrent and ground fault protection is provided by the breakers with LSG protection features.

Now lets suppose I model this switchgear in SKM and carry out arc flash analysis choosing IEEE-1584 medthod. I will have certain results for this analysis. These results will be dominated by my breakers tripping chracteristics.

Now in second step, I will add the Arc Flash Relays to the model. These relays operate in 7msec, so the total time will be 7 msec plus the breaker opening time to clear the arc flash. So after I model these relays in SKM and carry out arc flash analysis again as per IEEE, surely my tripping time will be faster and the resulting incident energy will be small.

So as explained above, can I use my simulation study to show that I have reduced the incident energy and hence the hazard by adding the arc flash relays???

 

[undervoltage]

Now how these arc flash relays will work that is a seperate discussion. The system we have choosed is based on light plus current detection. So even if there is camera flash, since the current will be normal, there will be no nuisance tripping.

 

[LionelHutz]

Rafiq - where in that standard does it specify that the protection/clearing can only be achieved by sensing current?

 

[rbulsara]

LH:
I already indicated my opinion. You can do your own research or design.

Rafiq Bulsara

http://www.srengineersct.com

 

[dpc]

Getting back to the OP - your analysis software should allow a specific clearing time to be manually entered for a particular bus. You should refer to the arc detection relay data for its maximum response time and then add in the breaker clearing time as appropriate.

I don't see an issue with using arc detection relays to reduce incident energy - in principle. I do have concerns about attempting to retrofit the sensing fiber in existing equipment and having 99.99% certainty that the sensor will detect all possible arcing faults. If the arc detection relay is provided by a manufacturer in new equipment, then presumably they have designed and tested this, or at least they are assuming the responsibility for proper operation. For a field installation, who would bear the responsibility?

Using current relaying seems a little more fool-proof since an arcing fault will always have current to be detected. Of course anything can fail, but I'm not sure a "standard of care" has been established for retrofitting arc sensing fiber in existing equipment.

David Castor

http://www.cvoes.com

 

[LionelHutz]

Rafiq - So it's really just a personal opinion then? You claimed the standard "only considers regular overcurrent devices" and that a light sensor is an invalid method of mitigating the arc flash hazard. Yet you won't provide info from the standard which would support this claim.

The section from the standard, "4.6 Step 5: Find the protective device characteristics and the duration of the arcs" does discuss using the fuse curve or breaker trip curve but I don't read anything that rules out using other means to clear the fault. I read it as saying that, ultimately, you need to know the clearing time however it is achieved.

David - Installation is always a concern but the manufacturers do have info on the proper installation of the sensor fiber, based on the proximity and amount of fiber that must be exposed to properly trip. Similar concerns about the actual detection and clearing of an arcing fault could also be voiced for any means of clearing an arc fault, since every detection and clearing method can fail.

 

[rbulsara]

LH:

Let me rephrase my view:

IEEE standard is based on trip time AND short circuit current. There is no way to tying the inception or detection of the "light" to the amount of short circuit current and hence the trip time. It does not fit in the IEEE method, the only recognized one I know of. I do not recommend "tricking" the software to arrive at a "desired" result.

If the mfr provides an arc flash label, based on their analysis/testing/theory and AHJ approves it, I have no issues. I will not be the one signing off on it. To that extent, yes it is a personal opinion as to how I would conduct my business. I perform a lot of short circuit and arc flash analysis on professional basis and only based on recognized software that follow IEEE.

This is not an opinion on effectiveness of light sensing and trip to minimize arc flash hazard.

Hope this clarifies my view point.



Rafiq Bulsara

http://www.srengineersct.com

 

[LionelHutz]

I can understand your viewpoint. I didn't read anything in the standard which specifies that the clearing time can only be dervived using protection relay time-current curves. The standard discusses using the time-current curves but I don't read that as meaning the time current curves must be used.

 

[jghrist]

IEEE-1548 establishes equations to calculate incident energy based on the magnitude of arcing current and duration of the worker's exposure to the arc. The only justification I can see for not using the timing of optical arc sensing relays in the calculation would be if you don't consider the relays to be reliable in sensing an arc and initiating a trip. The current duration does not have to be based on any particular relay. The duration does not have to be related to the current magnitude. You could use a distance relay to trip. You could have no tripping device and assume that the worker will move away from the arc in a certain amount of time.

 

[davidbeach]

What jghirst said. It is true that current magnitude and duration are what determine the energy, but there is nothing in the standard that ties duration to current magnitude. What is the TCC of a differential relay look like? There are many good and useful ways of interrupting fault current that will never plot on a conventional TCC.

 

[undervoltage]

Thanks guys for your valueable input.

The attached image shows that SKM provides the facility to model the Fiber/point sensor in Arc Flash Analysis. It require two parameters to be entered by the engineer carrying out the study, one is the operating time of the relay and the second is the breaker opening time.

Now suppose there is no relay, so arc detection will be based on the magnitude of current and the tripping will be dependent on the breakers tripping characteristics.

Now when I will add an Arc Detection Relay, It will detect the Arc based on the magnitude of current (which I will set above the FLA) and the arc flash. Since these devices will provide instantaneous tripping characteristics, so I can see the reduction of Incident Energy due to the reduced tripping time because of arc flash relays.

@dpc, Two main types of sensors are available as far as I know. Fiber and Point. I prefer point sensors because in case of an arc flash it gives you the exact location of the arc fault while since the fiber sensor runs throught out the switchgear compartments, it takes time to find the fault.

 

[JG2828]

undervoltage:

You should still enter the total trip time of the relay. The time will be dependent on which relay you are using and the style of contact/transistor outputs you are using. I've seen as fast as 2 to 2.5ms.
Be careful when using point sensors. Some relays do not monitor the point sensors for fiber breakage like they do with the loop sensors.