Performance of gas turbine power plants controlled by multiagent scheme

In latter years the idea of artificial intelligence has been focused around the concept of a rational agent. An agent is a (soft-ware or hardware) entity that can receive signals front the environment and act upon that environment through output Signals. In general an agent always tries to carry out an appropriate task. Seldom agents are considered as stand-alone systems. Their main strength can be found in the interaction with other agents in several different ways in a multiagent system. In the present work, multiagent system approach will be used to manage the control process of a single-shaft heavy-duty gas turbine in Multi Input Multi Output mode. The results will show that the multiagent approach to the control problem effectively counteracts the load reduction (including the load rejection condition) with limited overshoot in the controlled variables (as other control algorithms do) while showing good level adaptivity readiness, precision, robustness and stability.