Twelve motors on the case line. Five are on PowerFlex drives, the other seven are line fed off plain starters. Management wants predictive maintenance and gave me a budget that covers a few CTs and one 8 channel analog card for the CompactLogix.
I've seen plenty of belts and bearings go on this line. The motor gets loud, then it gets hot, then it lets go on a Sunday. We have already got current off the five drives over EtherNet/IP for nothing, and runtime hours per motor in the PLC.
What I'm not sure about is whether a CT on the seven line fed motors tells you anything real or whether it's just something pretty for the dashboard. And if it is real, which of them, because the card only has eight channels. Which of the seven would you put the eight channels on, or is the money better spent somewhere else altogether?
Historian side first. Can you log twelve tags at 1 s and keep them a year, and do you have the line state and product code as tags alongside? Current on its own is half a number.
Ignition, plenty of room, we keep two years. Line state and product code are already tags, so is runtime hours per motor.
Current is the cheapest useful number off a motor. A bearing on its way, a belt too tight or slipping, a coupling out of line, they all shift the current at the same load by a few percent, weeks before you hear it. The catch is load: current alone lies, so trend it beside runtime and product, same motor, same product, month on month.
Pick by what stops the line when it dies. Palletizer infeed and the two transfers, forget the reject conveyor. Split core 4-20 mA CTs, one phase per motor, into the analog card. They go round the lead with the starter locked out, not with the bucket live.
The current trend approach is here: https://plctr.com/utilizing-plc-for-condition-monitoring-and-predictive-maintenance-in-manufacturing/
The infeed motor's belt ate a bearing last year. That one first, I reckon.
1 s scan, 30 s average into the historian, and one report a month per motor at the same product code. Small deadband or it fills the disk with noise.
Four CTs on the four motors that matter, the infeed, two transfers and the case erector feed. Drives give me the other five. Trending since.
Two weeks in the infeed sits at 8.2 A on the standard case, transfer 2 at 5.1. Flat so far, which is fine, I guess that's the baseline.
Went with CTs on the line stoppers only, current trended against runtime and product code, and the drives cover the other five for nothing.
The three I didn't fit will still surprise me, and one phase only means an unbalance won't show. Thanks dan, scott.
Unbalance shows up as heat. Put the IR gun on those three when you walk past.