Does bus differential protection reduce the clearing time for an arcing fault and should this be considered in performing the arc flash calculations?

Yes, differential relaying can reduce the clearing time for an arc flash occurring between the current transformers of the protected zone. (depending on the 87 relay setting).

The National Electrical Code added this as well to article 240.87 Non-Instantaneous Trip. This article lists differential protection as one of the methods used for higher speed clearing when no instantaneous is available.

87 relaying effectively detects and clears L-G and L-G-L faults extremely quickly because generally it is set up as an "instantaneous" trip, limited only by the speed of the relay and the breaker itself. With vaccuum breakers, getting down to about 50 ms is not difficult.

However, it does nothing for L-L faults. Arc flash calculations are generally concerned with those faults since on a worst case basis, L-L and L-L-L faults are larger in magnitude. The frequency of occurrence on the other hand for most equipment is decidedly in favor of some sort of fault involving grounds. Thus 87 relaying is a great way to reduce the magnitude of the most likely arc flash but it does not affect the worst case. For example in the highly hazardous exercise of inserting a draw out breaker into a cell, if the contacts on the breaker for some reason were to become misaligned and contact a grounded portion of the cell, the 87 relaying would help. However if the contact arm bent and folded over and created a phase-to-phase fault, it would do nothing. Or if in the case of a recent case in South America (shown on a pretty popular Youtube video of an arc flash while a man is racking in a breaker), the breaker was left outdoors on a pier where it became saturated with salts from sea spray. In that case effectively all 3 phases could be considered shorted or at the voltages involved, would rapidly involve into that case, and there is no chance that differential relaying would have provided any significant benefit since it does not improve the tripping characteristics of a L-L fault.

The current (version 7) version of SKM allows for taking into account arc flash relays, differential relays, and similar high speed detection/tripping mechanisms. I'm not sure about other software.

I disagree with PaulEngr, bus differential relaying does not depend on a ground component. Phase currents are summed individually, and any mismatch outside the allowed limit will cause an operation. Line to line means two elements rather than one will detect the fault. Simple Kirchoff, the sum of all currents into the bus of one phase is equal to zero. If the relay fails to measure zero, it is because the presence of another path not being measured; aka a fault. Fast clearing can be achieved independent of load. Perhaps PaulEngr is thinking all currents of all phases are summed together, but this is not the case in any application I know.

Careful: bus differential systems are limited to the zone where CTs will pick up a differential current, and still need a breaker to interrupt the arc. Line side incident energy is probably much higher, at least for the incoming cubicle if you consider them separately. If you have alternate sources connected downstream the load side could be worse as well if downstream devices are slower than the bus differential clearing time (quite likely).

Alternative sources downstream will be cleared in the same time if the fault is within the bus differential zone.

Jim is Correct. Differential protection of a "bus" is protection and should not care whether the fault is arcing or bolted. Simply stated the differential sensing zone is bounded by the CTs used to measure the currents in and out of the zoneand the protected circuit is that which is disconnected from sources by the controlled circuit breakers. Note that in differential protection all CBs bounding the zone could be opened including those being fed by regenerative sources such as large motors.
Important to understand the response time of differential protection based on the fault currents expected. Consult the rely manufacturer to udnerstand all relevant considerations. Note that in LV systems the range of currents that have to be dealt with is significant and that creates application complexity. But, generally very fast response is possible.
When using 87T (transformer differential) need to make sure that the requirement to get around inrush currents do not negatively impact AF protection on the secondary circuits.
One interesting application for secondary substation protection with one relay is where an 87T function is used to protect the transformer and immediate conductors, but a 50/51 function within the same relay is used to protect the secondary conductors that are downstream of the secondary CT location but may be ahead of a secondary Main CB. Digital relays are available that can do this. May be useful where the secondary switchgear is too far from the transformer to easily include it in the zone of protection or where the CTs necessary for differential protection cannot be located within the LV gear for any of various reasons.
Analytical software, to my knowledge, does not normally model 87 protection via traditional curves, however the clearing time (sensing+relay response + CB responce/clearing) can be entered into the analysis. Note that in LV CBs operated via shunt trip some additional time must be added for the shunt trip versus the internal flux shifter normally used by the integral trip.

This zone is the protected bus.



All posible sources are tripped, so more than one breaker is needed.

We've successfully modelled the SEL-487B functions in both in SKM along with 67 directional elements for MV networked systems. I am model these again in ETAP and Power Analytics for different projects and hope to be able to report these tools work as well. In SKM you need to manually edit the one-line diagrams to show graphically what the logic behind the dialoge boxes provide "under the hood". I saw reductions in incident energy consistent with what I expected for more sensitive fault levels. We designed a "Fast Bus Isolation" logic scheme to open all breakers on the bus, but have enough control power for the combined trip coil currents. We also use the same scheme to isolate a bus when any breaker on the MV bus fails to trip after an appropriate command duration time.

thank you for your post your post helps me alot.