Abstract - Ydstie - 9726115 Distributed Processing Units (DPU's) pick up information as needed to perform control calculations, local optimization, data reconciliation, filtering and/or dynamic simulation. Processing speed and reliability are high due to parallelism and redundancy - a distributed system can continue to function even if a "stable" DPU goes down. Distributed computation makes it easy to update hardware and software. Members of an engineering team can change process equipment and modify control algorithms in sequence or parallel without loosing stability. Control systems in industry are distributed for reasons of computational efficiency, safety and cost. The PI's plan to develop systematic procedures and computational tools for the synthesis of distributed control systems. The research will address the following problems: (1) how to design distributed control systems; (2) how to perform dynamic simulation in a distributed network of computers; and (3) how to address the interaction between design and control. The broader aim is to develop a theoretically consistent and workable theory for chemical process control which takes into account the process physics in an effective manner.
