I work for a university that has its own utilities in terms of heat and power. I have been tasked with performing an initial arc flash study on all campus building secondary systems. I am using SKM and have a question how it models the maximum bolted short circuit current. Instead of using an infinite bus at each building transformer, I have values for momentary, interrupting, and 30 cycle short circuit current at our main campus distribution building, but I do not know which value to use. I want to make the calculation conservative, but considering the difference between momentary and interrupting is on the order of 20 kA, I would like to be sure. Due to difference in clear times on the system OCPDs, some are shown to trip in the arc flash analysis at one cycle, while others are stopped at the simulation default of 2 seconds. I would think the single cycle trip would see the momentary fault, while the 2 second trip would see either an interrupting fault or 30 cycle fault, but there does not seem to be any options on how to segregate the current values. Any help would be appreciated.

First off based on your description I guarantee that your incident energy numbers will be way too low.

When approaching a power study you start with the short circuit study, then progress through a coordination study, and finally the arc flash study because the arc flash study depends on getting the first two correct. The first study estimates the actual short circuit currents which are then compared to the ratings you mentioned (momentary, interrupting, 30 cycle withstand) to ensure that the equipment is sufficiently rated to withstand the actual expected short circuit current. Note that there is an ANSI method which is popular in North America as well as multiple IEC standards popular in other parts of the world. All of the various methods of estimating short circuit current make various assumptions which deliver different results. From a pure short circuit study perspective it is acceptable to overestimate the short circuit current but not to underestimate. However when it comes to arc flash it is critical to be as close as possible. Thus within SKM you can deliver results using various IEC or ANSI models but when it actually does the calculation for an arc flash study, it uses the IEC Comprehensive method which is the most accurate that SKM offers. Note also that for the same reasons modelling conductors is usually considered optional when doing a power system study for short circuit purposes but because this frequently grossly overestimated short circuit ratings, the model needs to be as comprehensive as possible (up to a point). Keep in mind also that from a power system modelling point of view, a transformer is literally nothing more than an impedance and it causes us to go through a transformation to relate impedances, voltages, and currents from one side to the other. SKM handles the transformation automatically.

When you set up your model, the ideal model would start at the generators and model the entire system. However when looking at a utility, grids and networks are involved and it quickly becomes very difficult to model all possible combinations with SKM, and you'd be trying to model something completely outside your system. Instead you have to request the information from the utility to be able to model their system and then you put one or more utility sources into your SKM model.

Once this is done the next step is to do a coordination study. This is where you put in all of your protective devices and plot the curves (TCC's), for 50/51 type overcurrent protective devices and fuses. When it comes to differential relaying, arc flash sensors, and zone protective relaying, you need to model this separately and basically ignore SKM results because it can't handle these systems. The completion of this study will be able to determine the opening time at any given fault current.

Finally we get to the arc flash study. This takes the previous two studies and based on the estimated short circuit (not withstand ratings), determines a corresponding arcing current. The arcing current is applied to the TCC to determine the arcing time and also with the system voltage and equipment design parameters (more data specific to the arc flash study) finally determines an incident energy value.

In times past it was much easier to do a power system analysis. For instance one could simply assume an infinite current source, ignore cable impedance, and calculate a transformer short circuit current. This is easily done with transformer size and impedance values alone as long as it was assumed that the load side did not contain significant reactive sources such as large (>150 HP) generators, motors, or very long cables (typically 1000 feet or more of medium voltage shielded cable). These mild assumptions work for most commercial and many industrial operations. And again if the only issue is insuring that the prospective short circuit does not exceed withstand ratings of equipment, this is really all that is needed. But this simply won't work for arc flash studies and this is why the old rules of thumb and "quick methods" are no longer applicable.

Thanks for the in depth response! I understand most of what you said, and I think I have most of it covered. I have all cable, transformer, and equipment impedances, lengths, etc from the campus distribution building (which is at a medium voltage of 12470 V) out to each campus building step down transformer, along with all of the building systems after the voltage is stepped down from 12470 V to 480 V. So I don't think it is so much how the system is modeled, as I believe that is as accurate as I can make it, but it is how I should model the utility. A model of the power plant and the distribution system out to our campus distribution system has been done, and from the short circuit analysis performed on that model I have the momentary, interrupting, and 30 cycle short circuit currents at the campus distribution building. My main question is that for the campus building arc flash analysis, can I model the system as starting from the campus distribution building with an SKM utility connection connected to the bus with the short circuit current from the previous plant short circuit study as the three phase contribution? And if so, which current should I use, momentary, interrupting, or 30 cycle? I understand that the best way would be to just connect the entire plant model up to the campus distribution model, but I currently do not have enough buses in my SKM model to make the model that large. Thanks in advance!

If I understand correctly you have not included the utility in the model. You need to ask the utility for the fault current and X/R where they tie into your system. Then run scenarios based on various generators on line with the utility.

As far not enough buses, what people do is build and upper level model of say the 12.47kV dist system and find the fault currents at tie points such as where a transformer is tied to the system. A new system is built starting at this transformer with the fault currents found by upper level system.

I do have the utility modeled and have accounted for all scenarios. For each scenario, I have the X/R ratio, momentary fault current, interrupting fault current, and 30 cycle fault current. What I am wondering, is for an arc flash study, which value of current do I use, momentary, interrupting, or 30 cycle?

And I have those values at our main campus distribution building (basically just a building housing all of our medium voltage switchgear). I then have all cable lengths and impedances out to each individual building transformer and then the subsequent building systems at 480 V and 208 V, so it is no problem to march the fault current out to each building and through each building transformer from the campus distribution building. What I do not know is which current to use as my fault current, since from my available values, I have momentary, interrupting, and 30 cycle and from what I can tell, SKM does not distinguish between the three, it only asks for one value of fault current.

After talking with SKM tech support, going back and forth a few times, and doing some experimenting on my models, it appears that the root of my problem is that the original SCC study performed on the power plant was done to ANSI standards, while the arc flash report performed by the SKM software prefers to use results from what it calls a comprehensive study. The difference is that the ANSI test reports currents for a three phase fault, momentary fault, interrupting fault, and 30 cycle fault, with the interrupting fault current changing with time, while the comprehensive test reports initial symmetrical RMS fault (1/2 cycle), and asymmetrical peak interrupting amps (at 1/2, 2, 3, 5, and 8 cycles).

The SKM arc flash test prefers to use the initial symmetrical RMS fault current from the comprehensive test, which just so happens to be very close to the three phase and interrupting fault currents from the ANSI test. The two values are not equivalent, however, as they are calculated differently and as already stated, one changes with time while the other does not.

Long story short, the answer to my question is: use the interrupting current.....but double check it by running the comprehensive SCC test on the original power plant model and make sure the ANSI interrupting current is the same as the comprehensive initial symmetrical RMS current.

Thanks to those who helped!

Unless SKM tells you differently it is my understanding that yes the arc flash model overrides the configuration settings for the short circuit model and opts for the "comprehensive" version which is actually an IEC standard. And the statement that there are different "assumptions" is correct. Essentially that model uses the best known method which I believe the closest way to describe it is in the IEEE Beige book.

But more to the point unless my documentation is wrong, it models the various inductive sources using their respective transient values as a constant and then drops down to the symmetrical fault current. From this point I'm not 100% sure exactly how it is doing the model but the obvious way would be for instance to increment time say every 1/2 cycle and do another pass through the model solving the resulting network. Once the transient time is exceeded then the contribution from the various inductive sources would be removed from the model on subsequent passes. I may also be all wet on this but this is the point where sometimes SKM documentation is lacking to say the least. This is as opposed to say ETAP which does it differently. Instead of modeling each inductive source individually it models the ttransient time for all sources as the same global value and varies their contributions.

The statement that each of the various short circuit standards uses different assumptions is very correct. One of the major reasons though that this exists is historical. At one time believe it or not we did power studies using vellum, draftsmen, and this wondrous device called a slide rule which was later replaced by an incredibly sophisticated electronic machine known as a calculator. Many real world problems have been solved using free paper sources available at local bars which can also be used for wiping your face or soaking up condensation from a glass of your favorite brew. In that environment, we can't really be using a higher end laptop with an SKM key dangling out the side of it anyways. You can effectively "do" the ANSI method with a short note card full of tables of the assumptions and a calculator. I kid you not because for small systems or for considering changes to existing systems often this is a lot faster than making a modification (with no undo!!) and waiting for SKM to grind it out. And yes, I'm quite aware of the "windowed" recalculation, and how eventually it degrades or crashes or simply produces nonsense results.

SKM uses the Comprehensive methods as explained to you by the previous responses.

But I do not agree with the answer that you have to use the ANSI Interrupting duty for modeling your utility. If you have the ANSI RMS Momentary duty available on your previous results, I would use it. The reason is that the ANSI Interrupting duty is valid for several cycles AFTER the arc is initiated. Additionally, the ANSI Momentary RMS is the closest value to the Comprehensive Initial Symmetrical RMS that the Arc Flash Module from SKM uses for arc flash calculations.

With respect to the answer that the Comprehensive method is an IEC standard calculation. I do not believe this is the case. SKM has two additional short circuit calculation models besides ANSI and the Comprehensive. These additional calculations methods are according to IEC standards. If the Comprehensive is a calculation according to IEC standard, why having additional and separate IEC calculation procedures.

The Comprehensive method is just a an OHMIC method. All sources of power, transformers, motors, etc. are converted in impedance equivalents. These impedances are considered as vectors (R + jX) and are operated as such to reduce to a final Thevenin equivalent impedance. There is no consideration on IEC impedances in this method.