A Novel Approach Based on Knowledge Graph and XGBoost to Characterize and Predict Position- and Speed-Dependent Dynamics of Machine Tool Structure

The change of machine tool structure will cause variation in the dynamic characteristics of the tool, which will result in changes to the stability of the tool and poor machining quality. Some researchers have studied that the dynamics of the MTS are expected to change with the slight change of position and speed of MTS. However, there is almost no effective way to estimate modal parameters, as existing methods of estimating the modal parameters of MTS based on computer-aided engineering and experimental modal analysis can not extend to different combinations of positions and speeds. Simultaneously, the vibration signals of different measurement points on the MTS under different combinations of positions and speeds are isolated each other and data access are too inconvenient, leading to an inefficient and time-consuming work. Furthermore, it is difficult to visually characterize high-dimensional position- and speed-dependent dynamics of machine tool structure. To overcome these issues, it is the first time that we propose to construct the dynamics of machine tool structure knowledge graph (DMTSKG) to associate entire data under different combinations of positions and speeds to visually characterize and predict high-dimensional dynamic characteristics. The analysis results show the effectiveness and performance of our proposed method.