Shipping Sorter OEE calculation

I have a conveyor sorter and would like to calculate OEE.
Cartons are inducted, scanned, and sorted.
There is no planned downtime during a shift.
So I will consider total counts by shift for everything.

There are several lanes the cartons are diverted to, but I am only considering the total right now to fiqure out the sorter’s efficiency.
If the lane is full or jammed, or barcodes cannot be read, the cartons are recirculated to be re inducted.
I have counts for inducted cartons, recirculated cartons, diverted cartons, and sorter runtime.

So far I think I should use:
Availability = runtime/shift time

Performance= (infeed/runtime) / OEE std rate
so, there isn’t a OEE std rate for shipping cartons of which are all various sizes so there really cannot be a standard throughput here.
but, if I can use infeed = diverted + recirculation, and use the maximum value of diverts I have seen in my sample size as std rate, this might be close.

I might as well just disregard this, if I calculate performance like I mentioned.

I would appreciate any feedback, and alternatives.


The system will give you the proper availability as long as the downtime events are recorded correct. If you know sorter runtime then do you know the times that the sorter is not running? You could use that to trigger a downtime event.

In order to calculate OEE our system needs at least a product code that has a standard rate defined at the outset of the run. That is the values we record during the run so we can determine the ratio between what should have been processed and what was processed. Are you saying that you have an idea of what that is now, or that you would determine it after the run? If it is after the run we do not accomodate that currently.

Since you can create an infeed tag and set the values via scripting or expressions then it is up to you to decide what value you want to assign to it while the run is running.

It sounds like you have a line with a different part being processed every time. part 1 may have a standard time to complete of 1 second, where as part 2 might have a standard of 10 seconds.

Rather than having an infeed as ‘number of boxes’, you could have an infeed of ‘number of simple box equivelant’. So when you scan a box, you look up it’s dimensions, and say ‘this is a 3-simple-box-equivelant’.

The cost of implementing the “simple-box-equivelant” measurement will give you the difference between a system that gives you a meaningful OEE or a box-ticking OEE. It is hard to use an OEE figure to recommend changes when the data has noise associated with ‘yes, but we had a lot of large boxes that day’.