Welding data adquisition based on FPGA

Abstract

This paper presents the use of FPGA in data acquisition and digital preprocessing of the electric current signal of resistance welding stations. This work demonstrates that electric current has enough information to classify this kind of welds in mass production industries. Parameters extracted with the FPGA excite a classifier that accept o reject the welding junction. This system has been developed using a neural classifier and installed in a welding station of General Motors in Cádiz (Spain). Results confirm the validity of this novel approach.