| Abstract: | This paper deals with accurate recipe implementation for abnormal condition management in a batch process using a case‐based reasoning (CBR) approach. A set of new problems can be solved by reusing proven process solutions. The proposed system integrates quantitative and qualitative parameters for adaptation of cases. A novel methodology to generate accurate recipes and to adapt to the processes is introduced during normal and abnormal conditions. In particular, the differences between current conditions and the references (recipes) should be managed to prevent any hazardous conditions arising. The processes are evaluated using their similarity to the past cases. This intelligent approach distinguishes plausible cases, generates accurate recipes, and adapts to new situations. The aim is to use the offline historical process data and safety related information in order to propose changes and adjustments in the processes. |
