臺灣博碩士論文加值系統

本研究以一種創新的方法與流程，利用智慧刀把蒐集端銑時的動態加工資料
用以檢測主軸健康度的模型，並專注於綜合加工機之主軸鬆弛的狀態評估。
工具機的發明大幅提高了生產效率和精度為產業帶來了許多效益。因應時代
要求和趨於複雜的加工環境與流程，工具機本身之狀態導致的問題卻需要付出高
昂的時間與維護成本，間接影響了本該提升的效率與精度。國內工具機廠屬綜合
加工機佔最多數，但因為加工程序複雜且多樣化，缺乏加工過程中的精準監控，
導致人機配比雖然提高，但加工效率下降的問題。目前開發之檢測系統，大多利
用感測器讀取機台資料並用以預測刀具狀態、讀取加工相關應力等；但工具機加
工時，無法直接量測主軸的實際狀態，主軸若是發生異常難以即時發現，對有精
密加工需求的產業來說帶來的損耗恐怕難以估計。
本方法使用智慧刀把模組能擷取加工時各種參數的優勢，取得正常加工資料
與異常加工資料，並將其利用於模型建立，模型可將實際加工資料與正常加工資
料進行比對，判斷工具機主軸是否處於正常狀態或異常狀態並給予一組評估分數
供以判斷，其中擷取之參數經處理後消除單位與數值之落差，使判斷標準注重於
各切削力特徵之間的關係，藉以適用於不同的加工環境。同時也比較了不同機器
學習演算法在針對主軸鬆弛、偏轉的情況下的泛用性與準確性，且利用不同工具
機驗證模型。

This study introduces an innovative approach and process using a smart tool
holder to collect dynamic machining data during end milling. The developed model
focuses on evaluating the health status of machine tool spindles, specifically
addressing spindle loosening in the domain of machining center.
The proposed method captures various parameters during processing, obtaining
data from both normal and abnormal machining conditions. These data are then
utilized to build a model that compares actual machining data with normal machining
data. The model assesses whether the machine tool spindle is in a normal or abnormal
state, providing an evaluation score. Processed parameters eliminate unit and
numerical discrepancies, emphasizing the correlation between various cutting force
features to make the assessment criteria adaptable to different machining center.
The study also compares different machine learning algorithms for their
generality and accuracy in detecting spindle loosening and deflection. The model is
validated across various machine tools, demonstrating its applicability and
effectiveness in real-world scenarios.

中文摘要 ............ i
Abstract ............ ii
誌謝 ............ iii
目錄 ............ iv
表目錄 ............ vi
圖目錄 ............ vii
符號表 ............ x
第一章 緒論 ............ 1
1.1 前言 ............ 1
1.2 研究動機與目的 ............ 1
1.2.1 研究動機 ............ 1
1.2.2 研究目的與方向 1
1.3 文獻回顧 ............ 2
1.4 論文架構 ............ 3
第二章 系統架構介紹 ............ 4
2.1 研究架構設計理念 ............ 4
2.2 實驗設備的選用及其功能介紹 ............ 6
2.2.1 智慧刀具的選用 ............ 6
2.2.2 綜合加工機的選用 ............ 8
2.2.3 刀具的選用 ............ 11
2.2.4 加工材料選用 ............ 12
第三章 實驗設計與方法討論 ............ 13
3.1 實驗設計動機 ............ 13
3.2 加工實驗設計 ............ 14
3.2.1 實驗設計 ............ 15
3.3 資料處理 ............ 19
3.3.1 處理方法 ............ 20
3.4 模型建立與特徵選取方法 ............ 22
3.4.1 特徵選取 ............ 22
3.4.2 模型選擇 ............ 23
第四章 實驗結果與分析 ............ 26
4.1 初步特徵選擇 ............ 27
4.1.1 箱型圖與相關係數 ............ 28
4.2 二元分類模型 ............ 45
4.2.1 邏輯式回歸與隨機森林 ............ 45
4.2.2 K means ............ 48
4.3 同機台泛用性判斷 ............ 49
4.4 不同機台泛用性與健康度評比 ............ 51
4.4.1 線性回歸的判斷 ............ 52
4.4.2 梯度提升與 SVM 適用性 ............ 57
4.4.3 投票分類器(Voting Classifier) ............60
第五章 結論與未來展望 ............ 66
5.1 結論 ............ 66
5.2 未來展望 ............ 67
參考文獻 ............ 68
Extended Abstract ............ 71
Reference ............ 74

[1]經濟部產業技術司. (2023). 2023/2024產業技術白皮書 . 台經院.
[2]KHOSHDARREGI, M. R., & ALTINTAS, Y. (2016). ACCESSING THE CUTTING FORCES IN MACHINING PROCESSES: AN OVERVIEW. PROCEDIA CIRP, 46, 601-606.
[3]MA, J.-W., WANG, F.-J., JIA, Z.-Y., XU, Q., & YANG, Y.-Y. (2014). STUDY OF MACHINING PARAMETER OPTIMIZATION IN HIGH-SPEED MILLING OF INCONEL 718 CURVED SURFACE BASED ON CUTTING FORCE. INT. J. ADV. MANUF. TECHNOL., 75, 269–277.
[4]TETI, R., JEMIELNIAK, K., O’DONNELL, G., DORNFELD, D. (2010). ADVANCED MONITORING OF MACHINING OPERATIONS. CIRP ANNALS, 59(2), 717-739. ISSN 0007-8506. DOI: 10.1016/J.CIRP.2010.05.010.
[5]LIANG, Q., ZHANG, D., WU, W., & ZOU, K. (2016). METHODS AND RESEARCH FOR MULTI-COMPONENT CUTTING FORCE SENSING DEVICES AND APPROACHES IN MACHINING. SENSORS, 16, 1926.
[6]SARAIE, H., SAKAHIRA, M., IBARAKI, S., MATSUBARA, A., KAKINO, Y., & FUJISHIMA, M. (2003). MONITORING AND ADAPTIVE CONTROL OF CUTTING FORCES BASED ON SPINDLE MOTOR AND SERVO MOTOR CURRENTS IN MACHINING CENTERS.
[7]LIANG, Q., ZHANG, D., WU, W., & ZOU, K. (2016). DESIGN AND ANALYSIS OF A SENSORSYSTEM FOR CUTTING FORCE MEASUREMENT IN MACHINING PROCESSES. SENSORS,16(1), 70.
[8]MATSUBARA, A., & IBARAKI, S. (2009). MONITORING AND CONTROL OF CUTTING FORCES IN MACHINING PROCESSES: A REVIEW. INTERNATIONAL JOURNAL OF AUTOMATION TECHNOLOGY, 3(4), 445-456.
[9]RIZAL, M., GHANI, J. A., NUAWI, M. Z., & CHE HARON, C. H. (2015). DEVELOPMENT AND TESTING OF AN INTEGRATED ROTATING DYNAMOMETER ON TOOL HOLDER FOR MILLING PROCESS. MECHANICAL SYSTEMS AND SIGNAL PROCESSING, 52-53, 559-576.
[10]LI Y, ZHAO Y, FEI J, ZHAO Y, LI X, GAO Y. DEVELOPMENT OF A TRI-AXIAL CUTTING FORCE SENSOR FOR THE MILLING PROCESS. SENSORS. 2016; 16(3):405. HTTPS://DOI.ORG/10.3390/S16030405
[11]SHLENS, J. (2014). TUTORIAL ON PRINCIPAL COMPONENT ANALYSIS.
[12]SWATHI, P., & POTHUGANTI, KARUNAKAR. (2020). OVERVIEW ON PRINCIPAL COMPONENT ANALYSIS ALGORITHM IN MACHINE LEARNING. INTERNATIONAL RESEARCH JOURNAL OF MODERNIZATION IN ENGINEERING TECH-NOLOGY AND SCIENCE, 2.10, 241-246.
[13]GUYON, I., & ELISSEEFF, A. (2003). AN INTRODUCTION TO VARIABLE AND FEATURE SELECTION. JOURNAL OF MACHINE LEARNING RESEARCH, 3, 1157-1182.
[14]BREIMAN, L. (2001). RANDOM FORESTS. MACHINE LEARNING, 45(1), 5-32.
[15]FRIEDMAN, J. H. (1999). GREEDY FUNCTION APPROXIMATION: A GRADIENT BOOSTING MACHINE. ANNALS OF STATISTICS, 29(5), 1189-1232.
[16]LOGISTIC REGRESSION: AN OVERVIEW. LAWRENCE M. HEALY. EASTERN MICHIGAN UNIVERSITY, COLLEGE OF TECHNOLOGY, MARCH 20, 2006.
[17]MACQUEEN, J. (1967). SOME METHODS FOR CLASSIFICATION AND ANALYSIS OF MULTIVARIATE OBSERVATIONS. IN PROCEEDINGS OF THE FIFTH BERKELEY SYMPOSIUM ON MATHEMATICAL STATISTICS AND PROBABILITY, VOLUME 1: STATISTICS (PP. 281-297). UNIVERSITY OF CALIFORNIA PRESS.
[18]ROYAL HOLLOWAY, UNIVERSITY OF LONDON COMPUTER LEARNING LAB. (N.D.). SUPPORT VECTOR MACHINES. RETRIEVED OCTOBER 11, 2023, FROM HTTPS://CML.RHUL.AC.UK/SVM.HTML#
[19]NASSIF, A. B., ABU TALIB, M., NASIR, Q. M. H., & DAKALBAB, F. (2021). MACHINE LEARNING FOR ANOMALY DETECTION: A SYSTEMATIC REVIEW. IEEE ACCESS, PP(99), 1-1.
[20]LI, Y., & LUO, Y. (2020). PERFORMANCE-WEIGHTED-VOTING MODEL: AN ENSEMBLE MACHINE LEARNING METHOD FOR CANCER TYPE CLASSIFICATION USING WHOLE-EXOME SEQUENCING MUTATION. QUANTITATIVE BIOLOGY (BEIJING, CHINA), 8(4), 347–358. HTTPS://DOI.ORG/10.1007/S40484-020-0226-1
[21]OSAMOR, V. C., & OKEZIE, A. F. (2021). ENHANCING THE WEIGHTED VOTING ENSEMBLE ALGORITHM FOR TUBERCULOSIS PREDICTIVE DIAGNOSIS. SCIENTIFIC REPORTS, 11(14806).
[22]SARMENTO, R., & COSTA, V. (2017). INTRODUCTION TO LINEAR REGRESSION. IN DOI: 10.4018/978-1-68318-016-6.CH006
[23]PLATT, J. C. (1999). PROBABILISTIC OUTPUTS FOR SUPPORT VECTOR MACHINES AND COMPARISONS TO REGULARIZED LIKELIHOOD METHODS. IN ADVANCES IN LARGE MARGIN CLASSIFIERS.
[24] SARHAN, A. A. D., MATSUBARA, A., SUGIHARA, M., SARAIE, H., IBARAKI, S., & KAKINO, Y. (2006). MONITORING METHOD OF CUTTING FORCE BY USING ADDITIONAL SPINDLE SENSORS. JSME INTERNATIONAL JOURNAL SERIES C, 49(2).