Attached is a typical pad-mount transformer dimension drawing. Modeling has been an issue for me for several years. I have settled on HOA for both the MV Compartment and the LV compartment. I have used both actual bus bar gap dimensions and default.

One issue I have is that in every case, the MV cabinet terminations use booted connections. No bus bar are exposed under normal conditions. For MV <=15kv IEEE 1584 treats this as a 3ph fault. Above 15kV, software gives the option of 3ph, LL, or LG fault. What are our operators really at risk for; 3ph, LL, or LG?

The LV cabinet is another issue as the terminals are typically exposed. The IE is usually excessive at the cabinet.

This has come up on here before. Below is an excerpt from one such discussion you might find interesting. Actual measured incident energy on the secondary side of a pad mount with a 480V secondary was much lower than software models calculate due to bus gap. Lots of photos and data at the EPRI link below. RE: The primary, yes no exposed live parts but you normally cannot access the primary compartment without first opening the secondary compartment so is it a moot point?

I'm following up on this subject as earlier today we had a presentation from a pad mounted transformer manufacturer. He was sharing information on some of their strategies to reduce incident energy at the transformer secondary. He was showing calculated IE over well over 100 cals and clearing times of around 6 seconds in his SKM model. I asked him if he knew anything about testing that showed in reality much lower IE due to the large bus gap and open space in the secondary cabinet causing arcs to self extinguish very rapidly. He did not.

One other thing to add, a couple years ago our oil testing company started to refuse to take samples on energized pad mounts due to too much IE at the secondary and that's when we decided to go with the remote ports.

In any case, after the discussion I looked up a couple of the things referenced by some in this thread and found the following links those, like me, who were not previously familiar, might want to check out.

Thanks to all who took the time to reply.

EPRI Testing: https://ieeexplore.ieee.org/document/6107577/

NESC T410-1 Presentation: http://industrial.tecgen.com/files/ced5 ... c-2012.pdf

I agree on the low-voltage side. HOA is not a good approximation. Tests show that durations are at maximum only a few cycles. For the medium-voltage compartment of live-front, padmounted transformers, HOA is probably a good approximation. Results from tests found peak incident energies that were up to about 2.5 to 3 times that predicted by IEEE 1584-2002. (There was a lot of variability, depending on the type of unit.) That's probably a decent match to HOA. We need to re-analyze the data to compare to the new models to see which matches best.

Here are some more links:

Medium-voltage results for live-front transformers and switchgear

480-V IAS paper

Update LV paper presented at the IEEE ESW

Fixed link to the presentation linked by bbaumer

What is HOA?

never mind. I found it myself.
HOA – Horizontal electrodes in open air

You won't have a 3-phase fault on the primary side with shielded insulated elbows. I'd use ArcPro with single phase faults or the NESC Table 410-2, which is based on ArcPro.