運用微感測器於切削溫度之即時監測

Abstract

在切削作業中，刀尖溫度之量測方法仍有諸多困難，因為最高溫度發生位置位於刀具與切屑之接觸面，若使用紅外線熱像儀容易受到切屑或切削液的干擾。若使用熱電偶，亦難以直接到達該溫度面作量測，且必須於刀具做鑽孔加工，增加成本。 為了提高監控系統在經濟上之效益，本研究開發出適用於車削加工之溫度感測器，此感測器採用微熱敏電阻，擁有相當高之熱敏指標B值，使得溫度量測範圍以及分辨率皆得以提升，另一方面實現感測元件微小化以及小範圍內量測多點溫度值的可能。不但能夠大幅降低感測器對溫度之反應時間，更可以實現陣列溫度場之量測。為達成在線監控(Online Monitoring)之目標，本研究使用有限元素法配合精簡模型理論(Model Order Reduction)以提高運算速度，並利用微熱敏電阻量測之溫度與精簡模型求解逆向熱傳問題(Inverse Heat Conduction Problem, IHCP)，並以共軛梯度法推估刀尖平均溫度。 研究結果證實，本論文提出之溫度感測器可對溫度變化做出快速的反應。在非切削實驗中，精簡模型能準確預測刀尖平均溫度變化，由此確認本論文提出之切削溫度即時監測技術具有相當之準確性，非切削實驗中，平均差異為5%。最後並以車削加工為例，說明本研究提出之量測方法之可行性。

In metal cutting operation, there are still many difficulties in measuring cutting temperature, because the location of the maximum temperature occurs at the contact surface between the tool and the chip. Using an infrared thermal imager is susceptible to interference by chips or cutting fluid. If a thermocouple or thermistor is used, it is difficult to directly measure the tool tip temperature. Drilling a hole in the tool must be done, and it will increase costs. At the same time, cutting temperature is a good indicator for tool wear or tool condition monitoring. In order to apply cutting temperature to monitoring in machining process monitoring, this study developed a novel temperature sensor for turning. It is a micro-thermistor made of NiO with a very high value of the thermal index B, so the resolution is very high compared to other types of temperature sensors. Moreover, miniaturized sensors can make multi-point temperature measurement within a small range possible. The tool tip temperature is estimated by the finite element method(FEM) with the measure temperature, which is known as inversed heat conduction problem(IHCP). In this study, the model order reduction(MOR) is used to increase computing speed of the FEM. In solving the IHCP, the conjugate gradient method is adopted to estimate the average temperature of the tip. The results confirmed that the temperature sensor developed in this study can respond quickly to changes of cutting temperatures. In the non-cutting experiments, the average tool tip temperature can be accurately and quickly predicted with the micro-thermistor array and the conjugate gradient method with MOR. Finally, the proposed methodology is applied to estimate the tool tip temperature in turning.