A Model for Predicting the Wear Degree of Electrode Tip
A new method is put forward to predicting the degree of electrode tip wear based on a laser measurement and digital image of the surface joint indentation. First, in order to monitoring the degree of electrode tip wear, the decline altitudes of sphere ΔH that can indicate variation of electrode tip shape are measured by means of the laser measurement system. Second, through the correlation analysis between the parameters S0, S1, S, K1 reflecting digital image characteristic of joint indentation and the decline altitudes of sphere ΔH, S0, S, K1 are extracted as characteristic parameters of monitoring electrode tip wear. At last, a model of support vector machine (SVM) for predicting the degree of electrode tip wear is established between the parameters S0, S, K1 as the input vector and ΔH as the target vector. Test result shows, the correlation coefficient between model prediction and actual measured values are 0.9907. The prediction model can realize estimating the degree of electrode tip wear.