As we all know, incident energy levels can be greater at lower fault current levels depending on protective devices. This makes using the infinite bus scenario not a guarantee of providing the "worst case" incident energy. It's also well known that the available fault current provided by the utility is a "snapshot in time" of the system, and the current at any given time can be greater or less than this.

So several questions,
1. have people had luck requesting "max" and "min" available fault currents from utilities?

2. when having to use an infinite bus situation, do people do both a low and high value scenario to compare for worst case?
2a. If so, has there been any papers written on this?
2b. Has anyone here performed some testing on their own in this regard?

We're trying to fine tune how we handle this in our models and I'm wondering what others have found and use as their common practice.

Thanks!

https://ieeexplore.ieee.org/document/8052570/

The interesting outcome is that except for mildly inverse or other very flat curves, if the fault current falls in the inverse time region of the breaker/fuse, the incident energy is at a maximum at the lowest current on that curve...lower always provides more energy as fault current decreases in inverse time.

If however you are in a definite time condition then the opposite, more intuitive result applies. Incident energy increases as fault current increases. For example if you are in the "instantaneous" tripping region or if you are at a point where the "2 second rule" applies.

This explains a lot. First it explains why infinite bus is NOT appropriate in most scenarios particularly with main breakers where for coordination reasons instantaneous settings are either set so high that they will never trip or turned off completely, and also why we tend to see a curious pattern where increases in distance or cable length of some kind result in the nonintuitive result of increasing incident energy.

A third and even more interesting result is the fact that you can largely predict maximum incident energy through a given circuit breaker without actually modelling the system by knowing just the breaker's time current curve and a smattering of minor easily determined data. Due to this property knowledge of the utility fault current is less necessary. It helps to quickly go directly to the worst case settings when fault current is not known rather than trial-and-error values. Simply plugging in the "extremes" of the TCC's is all that is necessary.

The information provided by the utilities varies between companies and even within the same company. It is very frustrating!

In 2010, IEEE published an article “Impact of Available Fault Current Variations on Arc Flash Calculations,� which demonstrated this method does not always provide the most conservative approach when performing an Incident Energy Analysis. It was recommended the Incident Energy Analysis be performed with a second scenario with 50% of the provided available fault current.

We have adapted this approach when the utility gives us the secondary current using infinite primary current. We will double the transformer impedance which reduces the fault current by 50%. We then use the worst case results between the two scenarios.

PaulEngr,

Let me see if I understand the process. Assuming you use the two second cutoff, find where 2 seconds intersects your TCC. Use this current, two seconds, and arc length to calculate the incident energy downstream. Perform another IE calculation using the the time and current at the point on the curve where it transitions to instantaneous. Then use the higher of the two values for selecting the most conservative PPE. Did I get this right?

Almost.

First, you might get different results if you have multiple inverse time curves. If you have just an LI breaker then what you've described is probably going to work but if it's LSG then you should check the intermediate (long to short term transition) point. I have no idea whether or not as a general rule this will give a higher or lower result compared to the 2 second time point.

For the two second result, there is one minor issue to be aware of. We're looking for an arcing current at that value so this is not the same as the available fault current. IEEE 1584 calculates a lower arcing current and if you are doing this in a software setting it's going to take your number and calculate the arcing current when in fact that value you want to use at 2 seconds is the arcing current. It would be easy to do this calculatiion by modifying the IEEE 1584 Excel worksheet (or doing the calculation by hand) but not so easy with commercial software.

The instantaneous region though follows what you expect to happen...higher current results in higher incident energy, without an upper limit other than the point where we exceed the interrupting capacity of the breaker so it will no longer open. So we use the "instantaneous" (opening time) of the breaker but we use the available fault current that we are given to calculate incident energy using the conventional procedure. The infinite bus assumption for this purpose might be preposterous but we've got to use our best estimate for this part. The arcing current calculation of IEEE 1584 of course would apply here so the unmodified calculations can be used. Using the instantaneous current trip point is a lower limit at which this occurs, and there is really no upper bound except maybe the breaker current withstand (if it is not undersized).

In the paper Paul is referring to - when using multiple curves (LT/ST/I) you can actually have differing results, and multiple inflection points. In general, it was noticed that point where the 2 second rule intersected the TCC there was the highest IE, in simple curves. This was mainly because of the "change in direction" of the curve (inflection point). The intent was to find the worst case IE based on time and current when all you have is the TCC. This is mathematically easy to do if you have the curve equations and some math software. While the approach is easy in concept it is difficult in practice because a changes in settings requires a recalculation.

If I had software that could find the maximum IE based on the protection settings I would still have to evaluate if the results are real or if the worst case cannot be obtained (too low of current). The method does provide the worst case for the protective device, sometimes that is all that is needed.

I still tend to use the half current method as mentioned by Robert. It is simple and quick. Calculate the IE with a very high utility contribution, run the fault calculation, reduce the utility contribution until the fault current is half. Run two scenarios and take the worst case for each device.

Many utilities still do not give any data, give only infinite bus and typical transformer impedance.