Process monitoring and adaptive quality control for robotic gas metal arc welding
The aim of this research was to develop an adaptive quality control strategy for robotic gas metal arc welding of thin steel sheets. Statistical methods were used to monitor and control the quality of welds produced. The quality of welds cannot be directly measured during welding. It can however be estimated by correlating weld quality parameters to relevant process variables. It was found sufficient to do this using welding current and voltage transient signals only. The strategy developed was problem solving oriented with emphasis on quality assurance, defect detection and prevention. It was based on simple algorithms developed using multiple regression models, fuzzy regression models and subjective rules derived from experimental trials. The resulting algorithms were used to control weld bead geometry; prevent inadequate penetration; detect and control metal transfer; assess welding arc stability; optimise welding procedure; prevent undercut; detect joint geometry variations. Modelling was an integral part of this work, and as a feasibility study, some of the models developed for process control were remodelled using 'Backpropagation' Artificial Neural Networks. The neural network models were found to offer no significant improvement over regression models when used for estimating weld quality from welding parameters and predicting optimum welding parameter. As a result of the work a multilevel quality control strategy involving preweld parameter optimisation, on line control and post weld analysis was developed and demonstrated in a production environment. The main emphasis of the work carried out was on developing control models and means of monitoring the process on-line; the implementation of robotic control was outside the scope of this work. The control strategy proposed was however validated by using post weld analysis and simulation in software.