Abstract
A new method is developed to monitor joint quality based on the information collection and process in spot welding. First, twelve parameters related to weld quality are mined from electrode displacement signal on the basis of different phases of nugget formation marked by simultaneous dynamic resistance signal. Second, through correlation analysis of the parameters and taking tensile-shear strength of the spot-welded joint as evaluation target, different characteristic parameters are reasonably selected. At the same time, linear regression, nonlinear regression and radial basis function (RBF) neural network models are set up to evaluate weld quality between the selected parameters and tensile-shear strength. Finally, the validity of the proposed models is certified. Results show that all of the models can be used to monitor joint quality. For the RBF neural network model, which is more effective for monitoring weld quality than the others, the average error validated is 2.88% and the maximal error validated is under 10%.
京公网安备 11010502051620号
