Estudiar y aplicar un algoritmo de control en procesos de propagación en sistemas hídricos : análisis y simulación de sistemas de propagación representados como sistemas multiagentes.

Abstract

A water system composed of tanks is an example of a physical system in which the interaction between the components is essential for its operation. In this work, we explore the potential of representing this system as a multi-agent system. In this hydro-tank system, the desired water level in the tanks is propagated through the system by the interconnection that exists among them. Modeling a water system as a multi-agent system consists of the following phases: first, establishing flow relationships to identify the water flow through the connected tanks; then, describing the attributes of the system through fluid dynamics; and finally, identifying the patterns that arise from the influence between the tanks and allows any connection configuration to be represented in connected tank systems. Distributed control, which coordinates the actions of each agent to achieve a common system objective, is essential to coordinate multiple tanks in the system. Each tank is considered as an agent that communicates with others by sharing information about its status and actions. The control algorithm is based on the ""leader-follower"" concept, taking advantage of the influence of the system model to propagate the reference signal of the desired level in the tanks. In other words, it uses the water level signal from one tank as a reference to control the water level in another tank. The controller used in each tank is of the Proportional (P) and Proportional-Integral (PI) type.