工具機設備出貨妥善率對於工廠產能具有關鍵性影響，現有的機台組裝精度量測與跑合測試僅由人工進行檢驗，僅能達成機台規格誤差量測與補償。此外，輪圈動平衡問題被視為影響運動穩定性的重要因素，當轉子在旋轉時，因質量分佈不均産生離心力，在軸承上會引起振動，産生雜訊和加速軸承磨損，嚴重影響産品性能與壽命。
為解決上述兩問題，本論文開發兩種智能化模組，分別為整機測試模組與輪圈動平衡預測機制，並應用於輪圈加工臥式車床。整機測試模組基於機台跑合資料收集與分析，透過分析機台耗電量與實際運動表現，建構一個可自動監測組裝完成機台各項運轉性能表現的系統。而輪圈動平衡預測機制使用振動原始資料，並運用深度學習架構進行預測分析，針對不平衡重量與角度進行預測。
在結果上，整機測試模組可即時評估機台整體運作效能，提供與比較機台組立(人員差異) 和電控設定之差異性；至於輪圈動平衡預測機制，則其重量之平均絕對誤差小於1.15克，而角度平均絕對誤差則為7.97度。綜上所述，本研究所提出的兩種模組與機制，均可成功應用於臥式車床，優化機台自身的功能性與生產力。

The availability of machine tool is a critical impact for the factory capacity. The existing measurements and tests are done manually. They can only conduct the measurements and compensations of the machine specification error. In addition, wheel dynamic balance is considered to be an important factor affecting the rotatory stability. When the rotor rotates, the centrifugal force is generated due to uneven mass distribution, which causes vibration on the bearing, creates noise, accelerates bearing wear and seriously affects product performance and life.
Therefore, this research develops two types of intelligent modules and applies them to the horizontal lathe machine, which is the Integrated-Test Module and Wheel Dynamic Balance Prediction Mechanism. Integrated-Test Module is designed based on the collection and analysis of machine tool running data. Through analyzing machine power consumption and the actual performance, an automated system can be constructed to monitor various operational performances of an assembled machine tool. Wheel Dynamic Balance Prediction Mechanism uses vibration raw data and the deep-learning method in modeling and prediction analysis for the unbalanced weights and angles.
In conclusion, Integrated-Test Module can integrate two variables, assembly operator and electronic control setting, in real time, to evaluate the overall machine operational efficiency. And for Wheel Dynamic Balance Prediction Mechanism, the mean absolute error of the weight is 1.15 grams, and the mean absolute error of angle is 7.97 degrees. Consequently, the two modules and mechanisms proposed in this research can be successfully applied to the horizontal lathes to optimize the functionality and productivity of the machine tools.

[1] Fan-Tien Cheng, Hao Tieng, Haw-Ching Yang, Min-Hsiung Hung, Yu-Chuan Lin, Chun-Fan Wei, Zih-Yan Shieh “Industry 4.1 for Wheel Machining Automation, ” IEEE Robotics and Automation Letters, vol. 1, no. 1, pp. 332-339, Jan. 2016.
[2] Fei Liu, Jun Xie, Shuang Liu, “A method for predicting the energy consumption of the main driving system of a machine tool in a machining process, ” Journal of Cleaner Production, vol. 105, no .1, pp. 171-177, Oct. 2018.
[3] Gianluca Tristo et al., “Real time power consumption monitoring for energy efficiency analysis in micro EDM milling, ” The International Journal of Advanced Manufacturing Technology, vol. 78, no. 9-11, pp. 1511-1521, Jun. 2015.
[4] S.Kara, W.Li , “Unit process energy consumption models for material removal processes, ” CIRP Annals. - Manuf. Technol., vol. 60, no. 1, pp. 37-40, 2011.
[5] A.Vijayaraghavan, D.Dornfeld, “Automated energy monitoring of machine tools, ” CIRP Annals. - Manuf. Technol., vol. 59, no. 1, pp. 21-24, 2010.
[6] K.Prażnowski, J.Mamala, “Classification of the road surface condition on the basis of vibrations of the sprung mass in a passenger car, ” 2016 IOP Conf. Ser.: Materials Science and Engineering, Sep. 2018, pp. 012-022.
[7] Marco Matteucci, Alessandro Gilocchi, “Wheel balancer with means for determining tyre uniformity, ” U.S. Patent 8739624B2, Oct. 15, 2017.
[8] Paolo SOTGIU, Lillo Gucciardino, Francesco Braghiroli, “Wheel balancer with means for determining tyre uniformity, ” U.S. Patent 8739624B2, Jun. 03, 2014.
[9] Zhang Wang, “A New Deep Learning Model for Fault Diagnosis with Good Anti-Noise and Domain Adaptation Ability on Raw Vibration Signals, ” Sensors., vol. 17, , pp. 77-98, February 2017.
[10] Luyang Jing, Ming Zhao, Pin Li, Xiaoqiang Xu, “A convolutional neural network based feature learning and fault diagnosis method for the condition monitoring of gearbox, ” Measurement., vol. 111, pp. 1-10, Dec. 2017.
[11] Balance Quality of Rotating Rigid Bodies, ISO 1940, Mechanical vibration, 2003.
[12] LeCun Yann et al, “Learning algorithms for classification: A comparison on handwritten digit recognition,” Neural networks: the statistical mechanics perspective pp. 261-276, 1995.