Learning material flow models for manufacturing plants from data traces

摘要: Models describing the material flow of discrete manufacturing systems are important documentation artefacts and basis for a comprehensive understanding underlying processes. The analysis such models allows deriving key performance indicators enabling assessment current system implementation. However, manual modeling as well up-to-date model maintenance is an error-prone costly task. In effort to allow automatic derivation models, this paper introduces concept Material Flow Petri Nets (MFPNs) presents learning algorithm their generation based on recorded PLC I/O data. proposed has been evaluated case study laboratory plant with successful results.