As some of you know the new ARCPRO 3.0 is out for download. They have added 3 phase to the options.

Has anyone heard or know which is better? More accurate?

They don't seem to agree???

I doubt they will ever agree all that well. ArcPro is basically a physics modelling approach whereas IEEE 1584 empirical method (to distinguish it from the other models in that method) is based on curve fitting to test data. They might be close but never identical because the underlying equations are totally different both in form and nature.

But here's the fundamental importance of ArcPro. The test set for IEEE 1584 only allows the curve fitting to be extended up as high as 15 kV. Above that point I've heard recent rumor that Kinetrics (ArcPro) has tested above 15 kV but I haven't seen any published results. Since ArcPro is physics based theoretically it will work with ANY conditions unlike IEEE 1584 that is far more limited so ArcPro is what we have to model distribution voltages above 15 kV.

I don't like OSHA's argument for using ArcPro. Basically above 15 kV they say that Duke Heat Flux (another physics model) is too low, Lee is too high (no argumen there!) and thus ArcPro is "just right" based purely on the results it produces, not test data. For all we know Duke heat flux model might be closer to reality.

Hi Paul, is it possible to test ArcPro above 15 kV to determine what's really going on above that point?

Yes, testing is possible. But it's a tricky question.Take for instance IEEE 1584-2002 which is based on about 300 tests.The validation of the model is based on using those tests to compare the predicted result to the actual results. If the goal is to validate ArcPro over a wide range, you'd need a wide range of tests which would run into millions of dollars worth in testing. And even if one were to spend all the money privately it limits the model to how far the data gets distributed....if it's not shared publicly, it's about as useful as the existing claim from Kinetrics that the model is based on actual tests conducted at Kinetrics but not released publicly. As you can imagine then this is not really very practical.

The alternative then is to create a mock up of a specific scenario with specific equipment and test that. This wouldn't be very representative of validating ArcPro "overall" because we're limiing ourselves to a single specific situation but at least it validates it at a single point. This sort of test could be done relatively inexpensively. For instance predicting incident energy in vaults at this time is best done with actual equipment testing.