Thks for the elaboration.
I did a little searching on yr query out of curiosity. At the risk of boring people, i hv offered a little more background in hope to encourage more responses.
The commingling feature is clearly a major grain / flour / milk / silo / usebin / food flow reactor (traceability) problem.
The difficulty, from a food recall POV, is illustrated in its simplest form by this example / process requirements –
com1 - produce traceability initiative, best practice repacking,commingling.pdf 130.74KB
Various FS incidents have shown the safety / financial risks involved in inadequate traceability provisions.
The industry scale / nature of the problem is detailed here –
com2 - Traceability in Food Systems.pdf 1.26MB
( the IFT final report detailing proposed "solutions" / testing is attached (tr2) at following link. Impressive but not a quick read -
A fundamental difficulty as you imply is the definition of what is a lot / batch whatever.? I presume at the moment you are using an arbitrary classification method,eg time, eg 24 hrs or a straight batch-by-batch system (of mixed input data) where the production order allows. Clearly, if used, the larger the time causes increase in documentation / any recall. Obviously any cleaning activity offers a natural separator.
I expect you are aware of this current US program in baking business –
The program mentions processor responsibilities due the commingling aspect from, I guess, a recall POV. Approach looked analogous to earlier attachment to me. I did not see any suggestions for answering yr specific question.
There are some theoretical solutions to yr problem in print particularly for grain silos although the maths can be bewildering (to me anyway). Some are within actual case studies and are practically followed up but they seem to be specific to the systems discussed. (typical approach is to postulate a layered structure of different inputs in usebin and work from dispersion / flowrate / time formulae. i doubt this is what you are seeking unless you can match such arrangements ?
One (but presumably software) solution appears to involve RFIDs. At a cost of course. i expect you are familiar with this already.
Repeating yr request, interesting to know what other people are doing.
Rgds / Charles.C
PS - added later i subsequently saw this baking report/summary which seems to imply the likely continuation of typical (I anticipate) status quo of mixed lot data. Plus the inevitability of software usage becoming routine -
Baking Traceability in the New Era of Food Safety,2012.pdf 460.1KB