Collaborative Decision Support during Process Operations Using Heterogeneous Software Agents

This paper elucidates the methodology of employing agent-based technique for detecting and diagnosing faults in chemical processes. The proposed agent methodology has been shown to be able to effectively combine various fault diagnosis methodology and exploit the capability of parallel computing technology in the process plants. We also show that the integration and collaboration among heterogeneous methods is essential to overcome the numerous hurdles of effectively monitoring process transitions given its high level of complexness. Additionally, the framework shows promises in effectively managing information from various entities of a process plant, further enhancing the decision making of plant management. The framework is illustrated using a case study of the Tennessee Eastman Process.