Monitoring The Degree Of Electrode Tip Attrition In Resistance Spot Welding

Resistance spot welding with focus on energy, small deformation, highproduction efficiency and easy to implement automation features, making itindispensable in the assembly process of car body connection method. In the spotwelding process, electrode wear and tear is inevitable, when the little electrodeattrition, no significant effect on the quality of welding spot, but when the wearreaches a certain level, it will seriously affect the solder joint quality, and evencause weldment burn. Electrode tip attrition is main indicating the variation ofelectrode tip shape. Indentation of welding spot is a result from joint action of theelectrode, hot, force and electricity on workpiece surface. Therefore, the shapecharacteristics of indentation can be indirectly reflect the variation of electrode tipshape. In this thesis, To judge the degree of electrode tip attrition as aim, based onSVM, using the surface image of welding spot as information sources, in order topredict the degree of electrode tip attrition the following detailed research workwere carried out:1) Established a measurement system centering on laser sensors, dedicatedmeasurement devices and computer systems to obtain the degree of electrode tipattrition. Explored a multi-point many times measurement method of the grids thestate for electrode tip, the measured height value h can be expressed the situationof electrode tip. Through comparing height value between attrition and newvariation of electrode tip shape, the results show that, the more increase ofwelding numbers, the greater fall in the height of electrode tip, namely that thegreater of electrode tip wear. Due to the change of the electrode tip height Δh canreflect the variation of electrode tip shape, therefore, extracting the average spheredecline altitudes of electrode tip ΔH or an average loss of spherical used tocharacterize the degree of electrode tip attrition.2) Building the welding spot image acquisition device, ensured that in thesame conditions for the indentation of welding spot surface image. Through theanalysis of gray images’ and its histogram of solder joints, the iteration methodand OSTU method were used to automatically extract the threshold for the solderjoint digital image segmentation. The result shows that satisfactory binary imagecan be obtained by the iterative method. The characteristic parameters center round area S0, in the center of the circle close to a ring of an area S1, the area ofelectrode tip and workpiece direct contact S, the center circle area and the area ofcontact area rate K₁which reflect the surface image of welding spot, and amongwhich the S₀, S, K₁as characteristic parameters can be used to predict the degreeof electrode tip attrition, because they and the average sphere decline altitudes ofelectrode tip ΔH correlation is relatively large.3) A model of support vector machine （SVM） for predicting the degree ofelectrode tip wear was established between the parameters S₀, S, K₁as the inputvector and ΔH as the target vector. Test result shows, the correlation coefficientbetween model prediction and actual measured values were0.9907, the meansquare error was0.0238. It shown that the characteristic parameters of the solderjoints can monitor the loss of the sphere from electrode tip attrition, theestablishment of SVM forecasting model can realize estimating the degree ofelectrode tip wear.