I deal with hydro facilities that have 13.8kV synchronous generators and pump motors with differential protection. At each Unit switchgear there are panels for equipment located (electrically) between the Unit and the Unit breaker. I am not sure of the best way to calculate the worst-case arc flash hazard at this equipment.

If there is a fault in the differential zone, both the Unit breaker and the field breaker will open. Is there an accepted method to account for the Unit fault contribution while the field collapses and the machine spins down? Is this fault contribution significant?

Of course it is unlikely that a qualified worker would open these panels while a Unit was running, but the goal is to label the worst case scenario...

Thank You,

Arcflash71

I don't know of an "officially recommended" method, but what I have managed to do for our large motors and generators is to model and graph the decrement curve. You can then either use this decrement curve directly, and integrate it for the total arc flash energy, or, as I did, approximate it. For the majority of the generators that we have, which are probably significantly smaller then yours, I found that by assuming 100% fault contribution for 10 cycles, then decreasing to 100% of full load current for the remaining time, I achieved a pretty good approximation of the total energy available. And yes, this fault contribution can be significant, especially for large generators and motors.

Thanks

Thanks for the fast reply.

I'm hoping to come up with an easy and conservative way to handle this situation that will work for a range of Unit machine sizes.

The simulation software that I am using is currently configured to take the maximum momentary fault level (from Utility and Unit machine in parallel) and assume that max fault persists for a user defined time period. In other words the software isn't set up to account for the Unit machine fault current decay/decrement. I'm a little bit nervous about configuring the software to do this as there are already so many variables to check when doing QC on our study results. Plus machine characteristics are not my strength and it would take time to get that right.

The simplest solution to my current problem would be to simply add cycles based on a conservative assumption for the size of the machine. You mention using 10 cycles to do this, can you (or anyone else) recommend a reference that correlates machine size with fault decay characteristics?

Regards