Generating a Production Process Simulation Model using Process Mining

Abstract

Digital Twin, information technology and simulation integration technology to achieve Industry 4.0, is in the spotlight with the start of the age of big data-based manufacturing. Digital twin simulations, however, also often adopt modeling techniques, including process experts, facility requirements, layout, and statistics based on manual data collection methods. This has many drawbacks and obstacles in reinforcing the evolving operating environment. This study, therefore, suggested the successful development of a simulation model of the manufacturing process that can be automatically renewed only with data in a changing environment through process mining, which uses event-log information, efficient control can be used, such as scheduling, inventory, and load management, and resource allocation through complex production processes. The suggested methods of simulation of event logbased manufacturing processes vary from the related current studies that have not been covered the manufacturing process's specific features in the model of process simulation, such as entity segmentation/assembly patterns, queuing disciplines, batch operations, and pausing patterns. In order to check its validity, this approach was applied to the real steel making process that represents the manufacturing industry, and furthermore, this study suggests the usage of the simulation model to produce the recommended production plan.