Process Monitoring and Supervisory Control via On-line Simulations: An Application to a Complex Reactor System

In this paper, a framework for implementing complicated process models on-line with real-time process data is presented. This framework is applied to a detailed process model for a complex reaction process. The model calculations provide estimates of unmeasured, but important, process variables including key product quality indicators. Furthermore, the model provides indications of many process variables which are measured. The estimates of process variables can be compared to their measured counterparts to provide a basis for fault diagnosis when plant and model are not in agreement.