Knowledge model for decision support to manage schedule disturbance in steel making

Abstract

Higher production rates combined with consistent product quality and efficient energy use are the key objectives to be solved by production scheduling in the iron and steel industry. Scheduling systems used to assign activities to resources often assume the generated schedule will remain workable for the foreseeable future. Due to the dynamic nature of the steelmaking process however, it is often difficult to maintain the original short-term schedule. Unplanned events or disturbances can disrupt plans requiring modification actions or even rescheduling. Frequent rescheduling often results in instability and lack of continuity in detailed schedule execution. The schedule disturbance management is a manual process and requires many years of experience. The thesis presents a knowledge model for decision support to manage schedule disturbance in steelmaking. Literature review shows the lack of research in developing a knowledge model approach for decision making in steelmaking. Manufacturing process such as steelmaking is `process centric'. The thesis presents a novel knowledge elicitation approach called XPat, which is suitable for engineering process knowledge capture. XPat is used to identify knowledge intensive tasks in steelmaking scheduling. Managing schedule disturbance is recognised as the most knowledge intensive task within the scheduling process. Problem solving knowledge of different types of disturbance in steelmaking is captured. The thesis presents a novel task template for managing the schedule disturbance. A knowledge model of the disturbance management is developed following CommonKADS methodology. The knowledge model is implemented through a design model. The design model helps in developing a prototype decision support system (DSS) to manage schedule disturbance. The system helps the users to make right decisions and implement consistency in the management process. The XPat methodology, the knowledge model and finally the prototype are validated using a number of techniques such as case studies, workshops, paper-based simulation, and user trials. It is observed that the prototype DSS is capable of providing effective decision support to manage schedule disturbance.