Sensor Based Modeling and Monitoring of Surface Quality in a Lapping Process Using a Nonlinear Stochastic Model

We present a sensor-based approach to estimate the quality of surfaces obtained during lapping-a common industrial polishing process-using an array of vibration sensors. We report the derivation and experimental validation of an analytical nonlinear stochastic model of lapping dynamics and the development of an online surface quality estimator for lapping based on combining sensor information with the model. The experiments were conducted on a Lapmaster bench-top lapping machine instrumented with three accelerometers. The results show that the model correctly captures the dynamics underlying the measured vibration signals including the shape of the complex frequency spectrum. The features were extracted from the acquired sensor data based on the model. The features were found to lead to linear dynamic maps that can estimate surface roughness to within 1% error under experimental conditions. Such predictability can help in reducing the operating cost of the lapping operation by minimizing scrap and rework.