Real-Time Temperature Estimation in the Undetectable Region of Motorized Spindle

The motorized spindle is a key component of high-end CNC machining centers. However, motorized spindles possess the characteristics of compact structure, large heat generation, and complex thermal mechanisms, making it difficult to directly measure and estimate the temperature of internal regions. To address this issue, a real-time temperature estimation method is proposed based on the fusion of mechanism and data. Firstly, by analyzing the structure and thermal properties of the motorized spindle, a thermal resistance network model. Then, the model parameters are optimized by combining the temperature field simulation data and the measured temperature from multiple operating conditions. Applications of monitoring the temperature of the undetectable region show that real-time temperature monitoring is crucial to the overheating protection of motorized spindles. Furthermore, by subdividing the grid and fuzzy interpolation calculation, a refined transient temperature field distribution of the motorized spindle is generated in real time. In the verification of the accuracy of indirectly estimated temperature distribution for the internal undetectable region, the average error of the estimated steady-state temperature for the undetectable region is 2.96°C and the maximum error is 8.65°C. In addition, a thermal fault monitoring method was proposed to monitor faults of undetectable heat sources and has higher accuracy and faster response speed than the typical temperature threshold method. This approach integrates finite element simulation technology, thermal resistance network modeling, and data-driven methods to achieve real-time monitoring of temperature fields in the otherwise inaccessible regions of the motorized spindle and has high industrial application potential in the fields of intelligent spindles and digital twins.