Design of a Fuzzy Controller Applied to Level Control in Industrial Processes

The present work was based on research around artificial intelligence, specifically on intelligent control methods and techniques. The objective is to use one of these control strategies to design and implement an intelligent controller to overcome the PID controller. Once the scientific articles’ study on the theory of intelligent control is concluded, the methodology to design the controller is defined. This will be based on fuzzy logic using a Mamdani-type model. The fuzzy rules are obtained from the operator's experience and the membership functions. Initially, the designed controller is evaluated based on the results obtained in the simulations, for which a classic PID controller is also designed to perform a comparative analysis between both; all of this is developed in the MATLAB/Simulink environment. The response of a controller to a real plant is very different from a test carried out in simulation software, and this leads us to implement a training kit with a level control system to evaluate the controllers. Once again, these controllers are subjected to qualitative and quantitative analysis to demonstrate that the fuzzy controller presents better results than the classic ones. Furthermore, this fuzzy controller does not need dynamic equations of the system to be controlled; by performing a small qualitative analysis of the system's behavior, a linguistic relationship is established between the state variables of the system and the control action. Not depending on the mathematical model for its design makes it a notable feature in this type of intelligent controller.