A RBF Neural Network-Based Parameters Tuning for an ADRC Regulator of Electrode Wire Feed Mechanism: Arc Welding Applications

Abstract

Abstract The electrode wire feeding mechanism (EWFM) is a closed-loop system that is commonly utilized in power controlled arc welding machines to achieve better performance for different electrode wire diameters. This study presents parameters self-tuning method based on RBF neural network for active disturbance rejection controller (ADRC) of a welding EWFM, and establishes a real-time testing system based on the dSPACE platform. First, an ADRC control strategy is developed to enhance the tracking performance and robustness of a welding EWFM in a multi-source disturbance environment. Second, an RBF-based parameters tuning method is provided to correctly determine and adjust the gains of the suggested ADRC regulator. Finally, to confirm the considered strategy, the real-time tests are conducted. The findings demonstrate that the suggested ADRC regulator with RBF-based gains tuning algorithm has a considerable disturbance rejection capability, small overshoot, fast response, and high precision which can improve the stability and quality of the arc welding process.