Método de control mediante lógica difusa para convertidores DC-DC de cuarto orden.
Loading...
Date
2026
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Universidad de Concepción
Abstract
Los convertidores DC-DC de cuarto orden (SEPIC, Zeta y Cuk) tienen ventajas significativas en aplicaciones de electrónica de potencia, sin embargo, presentan comportamiento no lineal y complejidad dinámica que dificultan el diseño de controladores tradicionales. Este trabajo consiste sobre el diseño, implementación y validación experimental de un método de control basado en lógica difusa para estos convertidores.
Se desarrollaron modelos matemáticos para las tres topologías y se realizó análisis en frecuencia que demostró las limitaciones fundamentales. Se diseño un controlador difuso tipo Mamdani con dos entradas (error y derivada del error), una salida (incremento del ciclo de trabajo), siete funciones de pertenencia gaussianas y 49 reglas de inferencia, aplicable directamente a las tres topologías sin necesidad de volver a sintonizar.
La validación experimental se realizó mediante un prototipo de 10 [W] basado en un convertidor SEPIC, construido con la maquina Voltera V-One e implementado en microcontrolador STM32G474 operando a frecuencia de conmutación 50 [kHz]. También se desarrolla una interfaz gráfica de usuario en Python para monitoreo y control en tiempo real.
Finalmente, los resultados experimentales demuestran mejoras del controlador difuso sobre el controlador PID: reducción del 23-25% en tiempo de subida, 15-24% en tiempo de asentamiento y 47-52% en IAE durante pruebas de rechazo a perturbaciones. Con esto, el controlador difuso mostro desempeño dinámico superior, de esta manera validando su efectividad para convertidores DC-DC de cuarto orden.
Fourth-order DC-DC converters (SEPIC, Zeta, and Cuk) offer significant advantages in power electronics applications. However, they exhibit nonlinear behavior and dynamic complexity that make traditional controller design difficult. This work consists of the design, implementation, and experimental validation of a fuzzy logic-based control method for these converters. Mathematical models were developed for the three topologies and frequency analysis was performed, demonstrating the fundamental limitations. A Mamdani-type fuzzy controller was designed with two inputs (error and error derivative), one output (duty cycle increase), seven Gaussian membership functions, and 49 inference rules, directly applicable to the three topologies without the need for retuning. Experimental validation was performed using a 10 [W] prototype based on a SEPIC converter, built with the Voltera V-One machine and implemented in an STM32G474 microcontroller operating at a switching frequency of 50 [kHz]. A graphical user interface was also developed in Python for real-time monitoring and control. Finally, the experimental results demonstrate improvements of the fuzzy controller over the PID controller: a reduction of 23-25% in rise time, 15-24% in settling time, and 47-52% in IAE during disturbance rejection tests. With this, the fuzzy controller showed superior dynamic performance, thus validating its effectiveness for fourth-order DC-DC converters.
Fourth-order DC-DC converters (SEPIC, Zeta, and Cuk) offer significant advantages in power electronics applications. However, they exhibit nonlinear behavior and dynamic complexity that make traditional controller design difficult. This work consists of the design, implementation, and experimental validation of a fuzzy logic-based control method for these converters. Mathematical models were developed for the three topologies and frequency analysis was performed, demonstrating the fundamental limitations. A Mamdani-type fuzzy controller was designed with two inputs (error and error derivative), one output (duty cycle increase), seven Gaussian membership functions, and 49 inference rules, directly applicable to the three topologies without the need for retuning. Experimental validation was performed using a 10 [W] prototype based on a SEPIC converter, built with the Voltera V-One machine and implemented in an STM32G474 microcontroller operating at a switching frequency of 50 [kHz]. A graphical user interface was also developed in Python for real-time monitoring and control. Finally, the experimental results demonstrate improvements of the fuzzy controller over the PID controller: a reduction of 23-25% in rise time, 15-24% in settling time, and 47-52% in IAE during disturbance rejection tests. With this, the fuzzy controller showed superior dynamic performance, thus validating its effectiveness for fourth-order DC-DC converters.
Description
Tesis presentada para optar al título de Ingeniero/a Civil Electrónico/a.
Keywords
Reguladores eléctricos, Vehículos eléctricos, Almacenamiento de energía