Model Reconstruction of AWJ Separation Speed Driven by Cutting Mechanisms and Cutting Data

Abrasive water jet (AWJ) has demonstrated its significant advantages in cutting some special materials or some difficult-to-machine materials due to its cold machining characteristics and efficient material removal capabilities. However, the complexity of factors affecting the material removal process makes it extremely difficult to predict the cutting process. Currently, using an empirical model to predict the separation speed has reduced trial cuts greatly. However, predicting error of this empirical model makes the tool less efficient. As more cutting data has been collected, it is possible to build a more efficient model for predicting the separation speed. In this study, an empirical model and a separation speed dataset have been integrated to build an integration model. Based on the training, a new separation speed model has been reconstructed. Through experiments, this new model improves the predicting accuracy greatly compared with the empirical model. At the same time, this model is physically interpretable compared with a pure data-driven model. The method explored in this paper could be applied to solve other multiphysics field coupling problems in the future.