프로세스 마이닝을 활용한 제강공정 시뮬레이션 모델 생성 및 최적화

Abstract

Digital twin is an advanced simulation environment that can reflect real-world models into the virtual world. By applying digital twins to manufacturing processes, production prediction and process optimization can be achieved to improve process efficiency. Simulation model generation and process optimization technologies are key technologies in digital twin. In this study, three steps were taken to generate process simulation model and optimal schedule for steelmaking process. First, an event log error repair method that reflects the characteristics of the process was presented. After that, a simulation model that reflects the characteristics of the steelmaking process, such as idle patterns, queue rules, and continuous activity patterns, was constructed, and the effectiveness of the model was confirmed by comparing the simulation log with the actual log. Finally, the process optimization problem was defined by simplifying the entire process, and an optimization model was built using the process mining method used to generate simulation model. The performance of the optimization model was verified by comparing the simulation logs based on the optimization schedule and original schedule.