I was wondering if anyone can tell me why differential relays are ignored in the arc flash calculations, specifically in the EasyPower software. I'm assuming other software ignore them also. Is it because they are generally used for transformer protection? Does IEEE say that they cannot be considered?
I do know the way to have them modeled within the software, I just want to know the reason they are ignored?

At least SKM does not "ignore" them. Since they normally are set up with instantaneous tripping they are assumed to trip in 1 cycle and give the minimum time to trip. Same thing with arc flash relays and di/dt relays which work similarly as far as giving "instantaneous" trip speeds (in 1 cycle). Usually the net result is that incident energy is extremely low as a result.

That's in the current version of SKM. In the 6.50 and earlier versions it also ignored these types of relays.

Don't forget to include lockout relay time if used.

I am not a user of EasyPower. I suggest contacting their support team or searching the user manual to determine how you can correctly model your application.

I'll share my experience with SKM. The software by default looks at the upstream protective devices to determine clearing times for the arc flash incident energy calculations.

As an example, if I wanted to implement bus differential protection in the software, I need to modify the clearing time for the specific bus in my model.
The new calculation for incident energy now takes into account the differential protection clearing time.

With EasyPower, you can put a user defined time for the bus being faulted. I will take the operate time of the relay, the operate time of any LOR in the circuit and the breaker operate time, add all those together and use that as the time for the bus to be cleared.
This is under the Arc Flash Hazard tab > Trip Times for this Bus > drop down box User-defined times

In SkM yo can model the differential protection.

It is a three step process.

First you have to activate the instantaneous protection in the arc flash sub view of a bus. Second, you run the arc flash study. Third, on the arc flash table results, you manually change the tripping time of the differential relay (and if necessary, the opening time of the breaker associated with the differential relay).

In SKM, a few steps were left out in the previous reply.

Select which buses are in the differential relay zone.
Click on each bus, Arc Flash Instantaneous Protection Box, type in the Differential Relay Name (ID) (See the attached file.)
Click on the in the "Available " box.
After you have done this for all buses, run the Arc Flash module/program.
Manually type in the tripping time which is the sum of differential relay trip time + lock out relay time + breaker contact parting.

As the user above mentions, Easypower has the ability to model Diff Relays, but you have to manually enter in the information. Here is a link to an EP article on the subject:

http://feedback.easypower.com/knowledge ... -easypower

Also, as I'm sure you know, differential relaying only adds protection for the panel you are differentiating around (measuring the in and out fault current). Differential protection will not improve safety of the system downstream of said panel.

Good Luck,
Brian

Don't forget too that any of these "high speed" protection systems (differential relaying, arc flash relays, or even just plain old instantaneous relays for that matter) are not "1 cycle" as they are advertised. The relay itself may be very fast and some even recognize a fault in 1/4 cycle but then it may or may not go through a "lockout relay" (Electroswitch trigger-handle) which adds 1 cycle delay and then finally to the circuit breaker which typically operate in 3-5 cycles for vacuum breakers but I've clocked a few very old (40 year old) Siemens Allis breakers recently that were tripping in about 100-200 ms especially with the additional slowdowns of both a lockout relay and being GE electro-mechanical relays. Fast was simply not something that they could be accused of.

The point of course is that this adds to the opening time so that a practical limit on opening time at least for switchgear equipment is around 4 cycles or around 50-70 ms. This is way better than inverse time relaying but still not all that fast at least compared to say fuses.

I've been toying with the idea of using a low noise full time SCR or IGBT switch which is essentially a soft starter except instead of using it to reduce votlage, using it to do on-off control or current limiting specifically as protection equipment rather than the traditional use as a soft start. This would allow some very respectable currents and could limit fault currents to almost arbitrarily low levels, eliminating the "shocks" that faults cause, and could simply starve out an arcing fault.

Right now it's just an idea. Need to get my hands on a discarded soft starter to prove out the concept.

Another issue with Differential relays is their zone of protection.

We often get requests to 'add differential protection' in industrial unit substations (<5000kVA, 480V secondaries) to reduce the AF incident energy on the transformer secondary switchgear. Most of these installations have a single set of secondary CTs mounted as close to the transformer windings as physically possible. These requesters do not realize they are better served by a Virtual Main instead of transformer differential protection. For the majority of <600V installations, 'bus differential' relaying unheard of.

I agree that the CT locations are key. Multi-restraint relays are available if you have or can place CTs further downstream. Please explain "Virtual Main."

A Virtual Main uses CTs mounted on the secondary side of a transformer to trip a physical protective device located on the primary side. One of the primary advantages they offer is add protective functions to equipment that was built without a single main device. Their instantaneous settings do not need to accommodate transformer inrush, therefore they often reduce the AF incident energy below that provided by just having a primary side device.

The reason that people start looking at 87 (differential) relaying on transformers is because it shows up on all kinds of lists of "recommended" things that you absolutely must have on a transformer. But often the fine print that says (over 10 MVA) or some such gets dropped. And few people know what it really does...just a way of recognizing a fault to ground that may not be detected on the primary side using simple inverse time protection since the protection has to be set to high to avoid magnetizing currents during startup. It's also in the same category as sudden pressure rise (differential pressure) relief valves that are frequently recommended for 10 MVA or larger.

I know because I've either asked about it myself or been prodded about it by an insurance company as a recommendation and then had to look into it.

As far as I know there are no "transformer protection" tutorials so there is a lot of mystery about this aspect of transformers. If you want to prove it to yourself, take a picture of a transformer level gauge some time and a photo of the transformer itself. Ask someone to show you where "full" and "empty" are at in the tank. Most will tell you that empty means that the tank is empty or nearly so and that full means it's almost to the top of the tank, like a fuel gauge. This is not true at all of course but again...we don't have a basic training guide.

Just to clear the record: While it is difficult to imagine a multiple line fault not involving ground inside a transformer, differential relaying will detect those faults as well as line to ground faults.

Differential relaying can have a hard time detecting turn to turn faults, so it is generally applied with overcurrent and restricted earth fault and sudden pressure relaying.

The 10 MVA recommendation is from the old electro-mechanical world. With inexpensive multi-function micro-processor relaying available, we've applied it much lower. The cost is justified when safety is enhanced.

I don't find it so mysterious. Start with Kirchoff's Law. Adjust for transformer turns ratio, delta-wye phase shifts, CT ratio, load tap changing, and allow for some CT and relay error. The instruction manual and setup software generally take you through it step by step. And if the relay shows differential current when load is applied, check again.

Not saying that the relays themselves are a mystery but the "transformer standards" that are out there give a list of recommended instrumentation without a shred of information on what it is or does.

There really isn't a good way to detect turn-to-turn faults at all. Some insurance companies push using a turns ratio test and power factor and the test is almost synonymous with Dobles. For initial testing it makes sense...validating that the transformer ratio is correct. But once in service, it's another matter. With say 1,000 turns a turn-to-turn fault will be a 0.1% or smaller difference in turns ratio but the test has en accuracy typically around 0.5 to 1%. In addition if an actual turn-to-turn fault exists, the heat at the fault will quickly show itself in other ways well before the test could ever be considered predictive in any way.