Ingeniería Electrónica
Diseño e implementación de un controlador basado en el operador de koopman para un sistema multivariable de cuatro tanques
Nonlinear dynamical systems represent one of the greatest challenges in the field of modeling and analyzing complex phenomena due to the diversity of behaviors they can exhibit. Unlike linear systems, their dynamics can include multiple equilibrium states, abrupt changes in behavior, and high sensitivity to initial conditions—characteristics that make it difficult to construct accurate models using traditional approaches based solely on physical principles. When these systems present multiple interrelated variables, operational constraints, or processes that are difficult to measure, classical methodologies may prove insufficient to adequately describe their evolution. In this context, data-driven system identification emerges as an effective alternative for understanding and representing complex dynamics from experimental information. This approach allows for the construction of mathematical models capable of approximating the system's actual behavior without relying exclusively on a detailed physical description. Parametric methods offer the advantage of employing defined structures that facilitate the interpretation and analysis of results, while nonparametric approaches provide greater flexibility for capturing complex relationships present in the data. Consequently, the combination of modeling based on physical knowledge and identification from data constitutes a fundamental strategy for improving the understanding, prediction, and control of nonlinear systems present in multiple areas of engineering and science