This article presents a hybrid system for tuning PID controllers, combining the practical methods used in the industry (Ziegler-Nichols) with the adaptive methods based on Fuzzy Rules (RF) and Structured Artificial Neural Networks (RNA-E), applied in a plant for an industrial friction welding process. The parameters of the controller gain vector with Proportional, Integral and Derivative (PID) actions, are determined by the second method of Ziegler-Nichols (ZN) and by an RNA-E with online adjustments using a fuzzy system. The performance of PID controllers is evaluated through computational experiments based on mathematical models described by transfer functions.
