Modeling knowledge about aseptics from cases of product loss

Summary form only given. Aseptic processing and filling is a food industry technology. No unified applied aseptics knowledge domain exists apart from individual cases of product loss experienced at factories. Case based reasoning technology was selected to develop an information handling system. A prototype system demonstrated the potential advantages. Complex cases of large product loss are relatively rare and often involve conflicting accounts of what actually happened. They proved difficult to interpret retrospectively. The initial focus was on routine cases of reports from production operators, mechanics, quality assurance tests and support material from records such as machine downtime sheets and data loggers on equipment. Models of routine domain knowledge were represented in various, cross-referencing case libraries used both to assist solving and to document complex cases. This initial system can be used to run inductive retrievals of records and reports. Suspect quality assurance samples can be related to evidence of a cause in production reports and records. Detailed trends in problems and faults can be examined and explored