An architecture of multi-agent system applied to fossil-fuel power unit

This paper proposes an architecture of a multiagent system applied to a Fossil Fuel Power Unit (FFPU). An agent-based approach has been dealt in this paper for the entire control system applied to the plant, and simulations demonstrate the usefulness of this multi-agent approach. The model consists of a Central Agent (CA), Set-point Agent (SA), Feedforward Agent (FeFA), Feedback Agent (FeBA), Limiter & Scaling agent (LSA), and Threat/Emergency Agent (TA). The communication protocols for the agents are defined in finite state machine theory, and Matlab/Simutink/Stateflow is used to simulate these communication protocols. The overall performance of this system is robust and adaptable. The CA is the key agent sending the message requests to different agents and monitoring their status for an efficient operation of the plant at different operating regions.