主軸溫度感測器網路校正技術之發展與狀態監測 應用

工具機主軸長時間高速運轉時，軸承、馬達與冷卻系統的熱變化會影響加工精度。若溫度感測器 長期使用後產生漂移，也會降低熱補償與狀態判斷的可靠度。本文介紹一種以自校型溫度感測器為基準 的主軸溫度感測器網路校正技術，利用內嵌微型定點囊產生約47℃ 相變化參考溫度，再透過類神經網路 修正多點低成本感測器偏差。實驗結果顯示，校正後多點溫度誤差可控制於 ± 0.5℃ 內，並可結合主軸 跑合平台進行溫度監測、異常診斷與溫升預警，提升主軸長期監測與智慧維護能力。

variation in machine tool spindles affects machining accuracy, especially during longterm high-speed operation. In addition, sensor drift after long-term use may reduce the reliability of thermal compensation and condition assessment. This article introduces a spindle temperature sensor network calibration technique based on a self-calibrating temperature sensor. A miniature fixed-point cell embedded in the sensor provides a phase-change reference temperature of approximately 47℃, and a neural-network-based model is then used to correct the deviations of multiple low-cost temperature sensors. Experimental results show that, after calibration, the temperature errors of multiple sensing points can be controlled within ± 0.5℃. The proposed technique can also be integrated with a spindle run-in platform for temperature monitoring, abnormal condition diagnosis, and temperature-rise warning, thereby enhancing long-term spindle monitoring and intelligent maintenance capability.