Resistance spot welding is a crucial assembly process for vehicle body manufacturing. The quality of the weld joint significantly influences the rigidity and durability of the vehicle; therefore, it is necessary to inspect the weld quality. The indentation depth of the resistance spot welding joint is widely used as an indicator to evaluate the quality of welds. However, because indentation tests on resistance spot welds are typically performed by field workers, additional time and process are required for the tests. Moreover, several attempts to apply automatic methods have shown limitations in high efficiency and prediction accuracy. To address this problem, we measured electrode displacement using a linear variable differential transformer during resistance spot welding in this study. In addition, we established an estimated regression model using the measured electrode displacement data to predict the indentation depth. Multiple regression models were estimated through stepwise regression analysis, and the significance of the model was analyzed through analysis of variance and residual analysis. Indentation depth prediction was performed after the resistance spot welding process using the proposed regression model, and prediction accuracy higher than 93% was achieved. The coefficient of determination obtained for this model was 94.72%.

No abstract

No abstract