Question: Has anyone employed Ground Fault Neutralizers having Residual Current Compensation for the purpose of mitigating arc flash potential on a medium voltage system? Such systems are often used in Europe, but not for this purpose. If used for arc flash mitigation, a fault would result in instantaneous tripout of whichever outgoing feeder had the fault on it, where in Europe these are more used to permit continued operation for sporadic ground faults usually on overhead lines, preventing customers from being blacked out despite presence of a recurring ground fault.

Martin R

I have a little trouble with the terminology across the pond but typically North American utilities rub either multipoint grounding or totally ungrounded with the latter most common in my experience. Underground and coal mines in the US and all mining in Canada must use high resistanc grounding with ground fault tripping and ground continuity monitors by regulation. Many heavy industrial plants, chemical plants and refineries and surface metal/nonmetal mines run resistance grounding (high or low resistance) for reliability reasons especially at medium voltage and especially as codes mandate ground fault detection and tripping above about 1000 A. Inductive or resonance grounding is much less common and not preferred. With high resistance grounds the idea is to use a fixed resistance low enough to allow a current of no bigger thab 1/3 the system ground impedance (Z0) so that charging currents are conducted away and transient voltages are minimized but ground fault current is otherwise minimized to keep the resistor size relatively small, typically 5-25 A short circuit design. Low resistance is chosen to minimize transients while controlling fault currents to reasonable levels, typically either 100 or 400 A although I have seen it as high as 2000 A. In this case the resistor is typically rated for 1-10 second duty and fast tripping is required whereas with high resistance continuous duty resistors are available and ground fault tripping speed is relatively noncritical to the point where some sites run alarm-only with manual coordination to locate faults. The gains achieved with variable impedances don't seem to be beneficial.

This is becoming the norm outside of medium voltage (2 kV to 10 kV) in both directions but is not allowed by Code below 300 V and becomes problematic to implement except as a low resistance system above 10 kV (transient voltages are bigger).

Once you have mitigated the phase to ground exposure, what do you do about the remaining phase to phase and three phase hazard?

I know of an incident that began as a line to ground fault on a resistive grounded medium voltage system. By the time it tripped out, it had gone line to line.

In a low resistance system this is very true and happens. In a hgh resistance system it just doesn't have enough energy to sustain itself. This is of course from experience and as of yet we don't have a definitive way of defining adequate minimum energy to sustain an arc. One system I have here is effectively like a utility...it's a 23 kV wye system with a 2000 A resistor...it is barely "low resistance" if it wasn't for the fact that the bolted fault current exceeds 20 kA at the substation.

The answer to your question though is that the only way that you can mitigate and eliminate line-to-line faults (to say nothing of 3-phase faults) is if the equipment design guarantees isolation of phases. This is not impossible but it's a very steep requirement. Some current GIS siwtchgear as well as some solidly insulated systems definitely seem to qualify for this condition. But simple insulated bus does not qualify. I've seen this first hand and there are published reports where arc flash travels right through insulated bus. And a lot of metal clad gear that claims to be 100% isolated between phases is anything but totally isolated. So you have to look at this with a pessimistic point of view.

But the real value here is that you reduce the likelihood of an arcing fault by approximately an order of magnitude. This will show up not in your arc flash hazard analysis which remains completely unchanged but in your risk assessment where you estimate the likelihood of an arc flash occurring in the first place. And if you look at the risk compared to other activities, it may mean that the risk is so low that PPE is unnecessary. For instance with racking out drawout breakers the likelihood of an arc flash occurring is right on the cusp of what would be acceptable. By reducing the likelihood of an arc flash by a factor of 10, this may change the task from "PEE required" to "PPE not required". The hazard does not change at all because we're looking at worst case (3 phase arcing fault) but the likelihood changes significantly.

This is similar to if you allow operation of breakers without PPE based on a risk assessment or the new tables in 70E-2015. If it is just routine servicing on equipment fed by a molded case breaker then no PPE would be required if equipment is properly maintained (do a visual inspection and exercise the breaker annually). If however the equipment just faulted then the condition of the breaker is now suspect until it has been inspected and PPE is required. These specific PM requirements are from NEMA AB-4 whch is the maintenance standard referenced by all molded case breaker manufacturers and available for free from NEMA's web site.