Batch manufacturing in food and beverage and pharmaceutical industries requires knowing the genealogy, or history, of any ingredient materials. When the source container or vessel contains several lots of the ingredient material, the control system engineer must make some assumptions and do some math.
One choice, often referred to as plug flow, is to assume that there is no mixing between the lots—that they are stacked on top of each other in the vessel as if there were an invisible membrane between each consecutive lot. The control system assumes that until the volume or weight of Lot A is dosed out, that lot and only that lot is being dosed. When Lot A is gone, the control system then assumes only Lot B is being dosed. At most, there will be only one product batch that has some of both Lot A and Lot B in it. Risked recall cost is very high when this assumption is used because a significant number of batches/lots of product must be included due to the ridiculous underlying assumption. Generally, a plug flow assumption should not be used.
Dosing genealogy might be determined by modifying the plug flow approach, assuming a band of mixing between consecutive ingredient materials. Factors that affect the thickness of the mixing band include how material is loaded into the vessel, the viscosity of the material, the shape of the vessel, time the lots have resided in the vessel, etc. Recall costs are appropriate because, in the case of a recall on ingredient Lot B, only batches dosed from bands A-B, B and B-C need be recalled. This approach can be appropriate when ingredient-mixing tendencies are well known.
A third approach assumes that the ingredient materials are fully mixed. Until the vessel is drained and cleaned, any dosing genealogy is assumed to include some of all ingredient material lots introduced into the vessel. When there is a long duration between drain-clean cycles, ingredient Lot A is assumed to be in many product batches/lots, risking very high recall costs. On the other hand, maintaining a short duration between drain-clean cycles is also expensive. Producers might be driven to design a system using single-ingredient-lot vessels or containers—this very safe approach is often dictated.
A combination approach to determining dosing genealogy also assumes complete mixing of all ingredient lots in the source vessel. The dosing genealogy is assumed to be the percentage of each ingredient lot remaining in the vessel. Volumes removed for each ingredient lot are accumulated, much as with plug flow. At some small remaining amount, the ingredient lot is assumed to be completely removed from the vessel. This approach offers reasonable recall cost risk without necessitating drain-clean cycles, making it an attractive approach when the extremely safe third approach, above, is not dictated.
With only slightly more math, the control system engineer can determine dosing genealogy with any of these approaches while a new ingredient lot is being added to the vessel. Allowing material to enter and leave the vessel concurrently increases equipment availability and, often, overall equipment effectiveness (OEE).
A risk analysis process is used to choose how dosing genealogy should be determined in a specific situation. Factors mentioned above, and others, must be carefully considered. Producers who determine dosing genealogy appropriately balance risk and cost using engineering analysis to make the right assumptions and do the right math.
Timothy S. Matheny, P.E., is president of ECS Solutions Inc., a certified member of the Control System Integrators Association (CSIA). For more information about ECS Solutions, visit its profile on the Industrial Automation Exchange.
