The Study On Quality Evaluation Method Of Sedan Body Joints In Resistance Spot Welding

Resistance spot welding is the most important method in the manufacture and assembly of sedan body because of its concentration energy, highly efficient etc.However, the welding process during which thermal-mechanical coupling effect happens is an extremely complex and sensitive process. Any random failure factors will threaten the appearance quality of sedan body joints, and also cause the instability of joints strength. At present, in order to ensure the quality of solder joints, most of the sedan bodies'manufacturers usually check the solder joints quality of strength and appearance by some methods, such as the designed redundant joints, the artificial damage method of sampling and visual observation after welding and so on. Therefore, an online, non-destructive and reliable method of spot welds quality evaluation has a major significance to the actual production of sedan body. Based on two sources of information, surface digital images of spot welds and the electrode displacement signals, in order to judge the quality of spot welds' appearance and strength, some research work were carried out as follows:1)Based on the platform of image acquisition, surface images of some kinds of solder joints such as normal joints, expulsion joints and sticking-electrode joints were acquired and preprocessed. Thereinto,through the analysis of gray images'and its histogram of solder joints, the iteration method and OSTU method were used to automatically extract the threshold for image segmentation. After comparative analysis of the experimental, iteration method was chosen as the final image threshold segmentation of solder joint;2) Through the analysis of the shape and geometric characteristics of joints binary images, the perimeter L, area A,elongation E and density C were selected as the characteristics parameters, and the relationship between the four parameters and welding parameters was revealed based on a lot of experiments. At last,the three parameters of L, A and E were extracted as the characteristic parameters to identify appearance defects of joints. On the basis, an evaluation model was established for the appearance defects of joints based on SVM (support vector machine);3)The electrode displacement signals were obtained and analyzed on normal and fault conditions such as electrode axial dislocation, part warpage, voltage fluctuation etc, which were also possible failure factors in actual production. The results showed that the electrode displacement signals under fault condition showed different degrees of singularity. Then, characteristic parameters representing the singularity were extracted from the electrode displacement curve. After that, a first-class SVM recognition model was established for singular solder joints, which can effectively identify the singular of signals.4) Based on the recognition results of the fist-class model, characteristic parameters as the representation of singularity joints' quality were extracted from the electrode displacement signal of fault state welding. Then, a second-class SVM model of solder joints quality evaluation was established based on shear strength and indentation depth of solder joints. Further, the quality of singular solder joint was identified.5) By analyzing the relationship between the features of the joints'surface images and the nugget formation process of welding joints, surface images were divided into three characteristic zones from which areas of three centric circles were extracted as characteristic parameters to characterize joints'nugget area. Then, SVM (support vector machine) model of classification was established and the nugget area of joints' cross-sectional was considered as the evaluation criteria. Experimental results show that the model can predict and classify the nugget area effectively.