Diseño, implementación y control de posición angular mediante rueda de reacción de un gisóscopo.

Abstract

The following report deals with the design and implementation of an angular position control simulator using reaction wheels. The work was carried out as part of the Final Integrative Project for the Electronic Engineering program at the National University of Río Negro, which is a continuation of the work carried out in a supervised professional internship in conjunction with the Bariloche Atomic Center (CAB). The motivation for this work is related to satellite attitude control using reaction wheels, a system used to maintain a fixed orientation with respect to another element or reference frame outside the satellite's frame of reference. The experimental setup attempts to represent the dynamics of the attitude control problem using a reaction wheel as an actuator. The goal is to achieve a device that can be used academically to design and implement different control algorithms in practice. The first stage of the work consisted of the design of the experimental model along with the sensors, actuators, and modules that make up the complete system, such as controllers, power stage, communication stage, etc. The physical assembly of the system was completed, with operational validation through unit testing of each part. The second stage consisted of the characterization and identification of the various plant parameters, leading to the creation and validation of a mathematical model to represent it. Various experimental identification and data analysis techniques were used, seeking a correlation between the expected results and the measurements obtained. The final stage consisted of the design and implementation of different control schemes, aimed at studying the plant's constraints and proposing a robust and stable control system that meets the given requirements. Based on the measurements and results obtained, a discussion was conducted and conclusions drawn. A functional prototype was created, in which five control algorithms were implemented, designed based on the identified model, achieving results of less than 5 seconds in settling time and 5% steady-state error.