Online monitoring of precision optics grinding using acoustic emission based on support vector machine

This paper aims to accomplish online monitoring of precision optics grinding with processing condition factors based on theoretical analysis and through grinding experiments. The model for monitoring surface quality of optical elements online (OSQMM) which contains identification model (IM) and interpolation·factor-support vector regression (i?f-SVR) is proposed. IM is applied to analyze and determine which kind of processing condition factors and which kind of its feature parameters are the best one to be used for online monitoring. i?f-SVR which contains the effect factor (fe) and interpolation function (I) to overcome the drawbacks of existing SVR models is applied to predict the monitoring thresholds. The grinding experiments were designed and performed. The influences of technological parameters (e.g., grain size of the grinding wheel, grinding depth, speed of the grinding wheel, speed of the worktable, and materials of workpiece) and processing condition factors (e.g., acoustic emission, grinding force, and vibration) on the surface quality were investigated and analyzed by IM. i?f-SVR was trained and established by the data which were gained through the experiments. After that, the other grinding experiments were performed to apply and verify OSQMM. The results were that the accuracy of alarm for roughness was 85.19 % and the accuracy of alarm for surface shape peak–valley value was 75.93 %. The results show that this method can be effectively applied to monitor the precision optics grinding process online.