| Artificial intelligence techniques were used, in this project, to develop methodologies for identifying potential chemical manufacturing schemes within existing production facilities. Major emphasis was placed toward the development of a chemical process knowledge-based system, a composite object oriented database, and a process/equipment planner.; Process equipment and raw material requirements for a batch chemical process were obtained through batch process modeling. Object oriented classes were developed and used to model chemicals, chemical mixtures, process equipment, and basic production facilities. These object-based models were interconnected through relational data structures to form an efficient composite modeling network. A natural language interface was developed integral to the process simulator and was used for implementing process recipe instructions during a process simulation session.; A composite object schema was used for modeling data and information regarding the existing production facility at different levels of abstraction. The factory data structure was made to consist of production equipment modules which in turn were designed to contain functional equipment groups such as tanks, condensers, and pumps. Equipment groups were further divided into individual equipment items. Other data modeling objects were created for providing import links to equipment specific vendor databases. The complete factory model was constructed, and was used in identifying new equipment requirements through composite modeling by modifying its internal structure as needed to accommodate the equipment requirements of a new chemical process.; A planner was developed, in conjunction with graph search techniques, for generating different equipment allocation plans using an objective function based on new equipment costs. Final search selection criteria was based the number of retrofit solutions identified which met specified new equipment costs. |
