That's a pretty long-winded way of saying that (for IEEE 1584 purposes) the total clearing time is 10 ms + breaker time, worst case

How do you deal with line side arcs in your system? The downside of using overcurrent protection in combination with light detection is that you limit the protective zone to downstream of the CTs; line side arcs may not be detected by it. Even if it's detected the line side arc needs to be switched off upstream, how do you signal the upstream breaker to trip?

jvrielink: That's right, it probably couldn't get much longer. However, I do hope that I also got across that I believe the real detection time to be somewhat shorter than 10 ms, and that the statistical approach used in IEEE1584 tends to skew the result towards lower power, as the build-up time is included in the integration time with full weight.

As for the line side arcs, you are right about the upstream faults. The D1000 is thought as a stand-alone, single zone, simple-to-use protection device. The original idea was to always trip the incoming breaker for an detected arc fault anywhere in the system, given that the occurence is rare and the consequences grave. But you are not the only one asking, and we are considering adding a trip fault signal, which will trip the upstream breaker if the arc fault is not cleared by the local breaker (in case of a breaker fault, or the arc burning on the wrong side of the CTs). However, this adds a delay, which increases the burn time and energy involved. To avoid this, one could add another unit for the line zone. Basically, this must be discussed installation by installation, as the needs are different - which is where most users of this forum seems to earn their keep, exactly because it is not easy.

Arcflash Reduction Maintenance Switch

Benefits of Arcflash Reduction Maintenance Switch
• Increased personnel safety - by limiting the available Arc Flash energy
• Simple to Operate.
• Enabled with the circuit breaker door closed by a door mounted lockable switch (no extra PPE required).
• Enabled only for the time required to perform the work
• Preserves overcurrent coordination under normal conditions
• Reduction in incident energy levels may permit reduced levels of PPE
improving worker comfort and mobility.

http://www.eaton.com/ecm/groups/public/@pub/@electrical/documents/content/ct_247849.pdf

From 1 to 3 phase arcing fault

Just found this intresting discussion. Maybe I can contribute.
During 2007 my team developed an arc eliminator/ mitigator in a project partly financed by the Swedish public treasure. Prof. Vernon Coorey helped me with a study called " On various processes that aid the transformation of asingle phase fault into three phase fault in switchgear"
A very short conclusion is "The breakdown of the other phases take place within 10 ms."
During the open arc tests (15kV 40kA) The arc mitigator took over the current in 5-6ms. The visible light from the arcs was like a photoflash, not more.
Tte coming standard UL 2748 "Safety for Arcing Fault Mitigation Equipment" indicates that the superfast mitigation technique will be accepted. I have now a prototype for a low voltage arc mitigator (63kA) intended to be even faster (3ms will be possible)
Do you think such an equipment will be accepted to reduce the PPE requirement class to 0?

What is the final output or action of your arc eliminator? Is it a trip unit? Does it switch impedence into the circuit? I am having trouble understanding how it can take an action to physically reduce the current in 3mSec. I understand that it might do the sensing and command an action in that amount of time, but if the current is not also reduced the arc could still migrate into a phase to phase fault.

Arc mitigation

The test condtions and a real case are slightly different.
In an openarc test you prepare a 3 phase shortcircuit by means of a thin Cu-wire and the current acts as for a clamp circuit. 3ms is from the moment the laboratory push the buttom until the arc eliminator makes a galvanic 3phase shortcircing and grounding which of course will take over the current from the arcs and after 0.5 or sometimes 1s the feeding breaker opens.
A real fault sometime start in one phase and as soon as it becomes 2 or 3 phase the high current and a sufficient light for the sensors will activate the arc eliminator that will operate and stop the increasing preassure as well as the toxic gases. The maximum pressure otherwise will have a peak after approx 10 ms. After 3-6 ms The switchgear is still OK and can be restarted as soon as the reason for the fault is cleared.
But so far I only have a prototype for LV. For MV we have proved that the system works.

What filtering algorothm is used on the current input to achieve a reliable digital value in 10 ms? The consenus a few years ago was that a 1 cycle cosine filter was the best combination of speed and accuracy for use with instantaneous (50) elements - reference the series of papers from the PCIC conference "CT Saturation Calculations - Are they Applicable in the Modern World?" parts 1 - 4.

For the 50 element used to supervise the light element, the fastest the 50 element can respond is 16-20 ms after the high current starts.

Whether it's 16-20 ms or 1 ms is irrelevant when you're tripping through a 3 cycle or 5 cycle breaker. I've researched the MV breaker standards (IEEE C37.04, C37.06, and C37.10) and there's no legitimate way for a user to 'recertify' a breaker to act any faster than the clearing time the manufacturer publishes in the breaker manual - typically 3 or 5 cycle ratings are available. The industry will be a lot more receptive to shaving a few ms off once breaker clearing times are much faster and that few ms improvement becomes significant in the big picture of total clearing time.

Regarding the CTs used to supervise the light element and how the zone of protection is only downstream of the CTs:
The easiest solution is to use an upstream CT located outside of the compartment where the sensor is located. The load current will be higher at that upstream location, but as long as there is still some separation of load current and fault current your scheme can still be sufficiently sensitive to perform when it matters. The IEC 61850 protocol is a great enabling technology for things like this to get current values from upstream to a downstream device.

Ul2748

To explain how the Arc Eliminator can improve the safety for pesonnel and switchgear I quote "The proposed First Edition of the standard for Safety for Arcing Fault Mitigation Equipment, UL 2748, covers devices intended to mitigate arcing faults by creating a lower impedance current path, located within a controlled compartment, to cause the arcing fault to transfer to the new current path."
This is what it is about. There are two options:
1) The light from an open arc detected by an optic sensor inside the cubicle will trip the arc eliminator and create the new current path in 5 ms.for MV (Less for LV).
2) As 1) but combined with high current in the incoming CT. Time difference from 1) is less tham 1ms.
The CB will get a trip order at the same time as the arc eliminator and it will open in its normal way. It has to break the shortcircuit current but that will happen anyway whenthere is an open arc fault. For the CB the breaking is less harmful if the peak current is over.
You can se how the arc eliminator works in a figure in my comment to a member in this forum.

Jankar