Intelligent supervision based on multi-agent systems: application to petroleum industrial processes

This work presents a methodology for the design of an intelligent supervisory system that combines the principles of fuzzy logic, the Internal Model Control (IMC) architecture and the paradigm of Multi-Agent Systems (MAS). The methodology has been conceived to be applied in an intelligent supervisory system, specifically for two kinds of complex petroleum industrial processes: the gas-oil separation process and the oil-heating process. The supervision proposal takes into account the fact of using standard local supervisors schemes connected between themselves and to a global supervisor so that local objectives in each process can be met, thereby letting the global or social objective be obtained through the application of basic mechanism of communication, cooperation and coordination; where these objectives have been previously defined and structured in a hierarchical manner. The paper includes some computational simulations performed under MATLAB / SIMULINK and the results obtained show a good overall system performance.