Implementation of a PID-neural network with PLC

PID (Proportional-Integral-Derivative) controllers are widely used in industry, since they are robust and easy to design. To make a system behave, as it is desired, it is enough to determine the appropriate PID coefficients. But, sometimes it would be really difficult to obtain mathematical model of system. Recently, as an alternative, PID like neural networks for abovementioned situations, are used for control purposes. In this study, a neural network of which input-output functions are chosen by the help of classical PID structure has been realized with the programmable logic controller (PLC). The applied training method in realization deals with major disadvantages of neural networks long learning and convergence times.