Abstract
The creep-in depth beyond the initial wheel/workpiece contact needs to be investigated in precision grinding to achieve the desired tolerance. The response of acoustic emission (AE) in metal cutting is quick and reliable for the in-process monitoring of creep-in depth. The relationship between AE and grinding conditions is derived theoretically and experimentally and shows the primary effect of the creep-in depth. AE is also found to be proportional to the approaching feed rate. On-line experiments are performed at various thresholds of AE, approaching feed rate, and grinding velocity. The experimental results are used to train a back-propagation network and construct a regression model. Under different grinding conditions the monitoring system can successfully recognize the creep-in depth accurately because of the character of the fault tolerance capability of the neural network. One can control the creep-in depth within 1 μm after the initial contact by setting the proper threshold value and the approaching feed rate of the monitoring scheme. The proposed monitoring technique is verified by experiments.