In the plastics industry we use a lot of resistance heaters. Often they are run in parallel from one controller. Every now and then I open the electrical boxes and disconnect everything so that I can check the resistance of each heater separately. I know from calculations (R=V^2/W) what the resistance should be given the voltage and wattage of the heater in question, but how far from that value is _too_ far? I'd like to be able to predict failure, but I really don't have much data to go on at this point.

Mike,
I have worked on a lot of proprietary equipment that had no manuals. I assume you are having a problem with failure(s) since you are disconnecting them all to check them.I have had trouble with this also. I would number them all , do an individual resistance measurement and an amp draw (if possible) then run to failure. Make some type of chart to see if there is any correlation between the measurements and the failure rate. It would be a start.

For submerged heaters a megger test to ground is a good test condition and predict failure. In my experience most heaters are run to failure. If you could inspect the heater sheath for deteriation I suppose it might be a leading indicator. In some cases a megger might show deterating sheath also.