Can adjustments to the thermal trip settings on breakers reduce arc flash energy levels at down stream devices? I saw a proposal to turn these down during maintenance activities to reduce the incident energy level while performing troubleshooting.

You mean INST settings?

I did some adjustments for one client after their Arc Flash study reveled a high hazzard on their switchgear on the load side of the breaker (line side of the MCC it was feeding).. They redid the coordination study and turned down INST level and times. In order to do it, we did an isolation from the Utility and turned them down and re-energized. Between the time we found the high levels and when we did the adjustments we left the higher label on the gear, so that the hazzard was known...

Both settings can have an effect. Run the calculation once and look at the arcing fault current. Now look at your coordination curve. Anything which reduces the time to fault at the arcing fault current (which generally also affects the second calculation at the bolted fault current) will reduce the arc flash incident energy by a direct, linear ratio. If you reduce the opening time by 50%, so does the arc flash incident energy in a direct, linear relationship.

With fuses, it works somewhat differently because the current limiting effect also plays a major role. Sometimes this has the effect of reducing the time to trip, sometimes increasing it. Note that strict use of IEEE 1584 does not model this.

However all that being said, consider this. You specifically referred to a thermal magnetic unit. Would you trust that the trip setting is accurate without performing a time current test on the breaker after adjusting it? If it's an electronic one I'd say no problems with making the adjustment but my experience with thermal magnetic trip units is that they have so many problems with drift and accuracy that you are taking a very big risk that it might not be calibrated after adjusting to the new settings, even with a brand new unit. I wouldn't trust doing regular changes to one to control arc flash. Doing it once and then testing is another story though. I wouldn't hesitate to do it with a solid state trip unit (and in fact it is in our procedures for some areas).

When you have reached just about the limit of this approach is when you get down to around 50-60 ms with high current/medium voltage systems (vacuum contactors). With some smaller molded case units 1 cycle (15-20 ms) tripping is getting pretty common.

Many trip units are now coming with multiple sets of settings that can be switch activated...basically supporting a "maintenance" switch. Older units have to be manually adjusted.

The settings that were looked at were on a breaker that just had thermal magnetic trip settings. If these were INST settings would it be a recommended practice?

Its not uncommon to find T&M breakers with the magnetic setting cranked to the max. Most times people simply don't know what to set it for and turn it up. The best way is to perform a load (amp recorder, etc) and short circuit study of the branch circuit in question. Or simply reduce the mag setting a little at a time to a point above where nuisance tripping isn't a problem. I work on the principal to start low and work up.