141123-0924 EST
StephenSDH:
In an earlier post of mine I suggested that you probably don't want to do the vector math to determine the power in each leg of your delta load from the delta line currents. To do this in your PLC is probably a waste of effort.
I later suggested, and partially by implication, that you compare the known correct currents, obviously by having measured the mold when it was known to be good at a known voltage, and during operation comparing the actual currents with the reference currents. Implied was that these current values at the known voltage must be adjusted for the actual voltage.
For a single resistance failure I indicated how you could know the delta leg where the failure was.
To make this work in your PLC you have a normalized table for the current in each phase at 600 V for each different mold. A separate table for each different mold. The table is generated experimentally. When heaters are replaced in a mold, then a new table has to be created for that mold, and updated in the PLC.
We have to assume that all phase voltages have a 1 to 1 relationship, but vary together as line voltage varies. Otherwise the problem gets difficult. If we assume the tables have been normalized to 600 V, then a voltage measurement is needed in addition to the three line currents. Each measured line current is adjusted by Vin/600 and compared with its associated table value. This is a relatively simple operation in a PLC, easily understood, and convient for trouble shooting.
Since the mold has temperature feedback control we can assume that resistance values are constant. It is not clear if there is a single temperature sensor and only one contactor per mold. The current probably should be measured near the turn off time of the contactor. If your measuring resolution is such that accuracy is in the range 0.1% of nominal maximum expected current, then I would expect that a 1% change in resistance could be detected, this is just an off hand judgement.
.