Sistema de inferencia difuso tipo Takagi Sugeno Kang para el diseño de un controlador de temperatura en un horno industrial de tratamiento térmico.

Abstract

The present work describes the design of a conventional PI+D type temperature controller, as well as a diffuse TSK controller that allows the minimization of the operation time of the post welding heat treatment in an industrial furnace. Currently, industrial furnace control systems are based on manual control or conventional PID control techniques. One of the limitations of these controllers is that by increasing the response speed the controller delivers outputs not realizable by the actuator. The investigation begins with the acquisition of the data with the use of a PLC and an HMI. The diffuse model TSK of the plant is identified by the process reaction curve, optimized by recursive least squares. The pole assignment technique allows us to show the PI+D controller and obtain the tuning parameters for monitoring the heat treatment curve. The diffuse TSK controller is initialized from the conventional controller and optimized by a downward gradient. To verify the temperature monitoring of the designed controllers, heat treatment curves of different steel thicknesses were simulated in the Matlab software. The results showed that the manual control of the process takes from 2 to 4 hours longer than the established time applying the conventional controller, this time decreases from 30 to 45 minutes, and with the diffuse controller a minimization of 15 minutes is achieved concerning the time of the conventional controller. It is recommended to initialize a fuzzy controller from a previously tuned conventional controller to avoid damage and unwanted responses in the implementation within the industry.