DefinePK

Abstract

In this article, the control of a pilot pressure plant was developed by means of artificial neural networks, using the training and learning possibilities that these provide. The same control algorithms used in industrial plants can be applied in this pilot plant. With the implementation of this pilot pressure plant, a comparison was made between conventional control techniques and intelligent control in terms of efficiency and usefulness. This comparison was carried out by means of experimental data considering the results obtained with the conventional controller and the controller proposed by neural networks. A series of perturbations were performed once the system was in its steady state to obtain the response times of both control methods and determine the efficiency and advantages of intelligent control. The control was performed on the Arduino Mega board in serial communication with Matlab®(R2021b) to visualize the variables and thus observe the system behavior in real time.