臺灣博碩士論文加值系統

近年來隨著航太及汽車工業的蓬勃發展，因此製造業競爭愈加激烈，然而目前對於產品的要求不只限於加工精度正確與表面粗糙度良好，對於生產的速度要求更加重視。當製造過程以高速及多樣且不間斷的進行銑削加工時，刀具的磨耗更會直接的影響產品的良率，因此刀具壽命的預測更顯得重要，然而目前國內製造業大多採用加工一定數量的工件後便直接更換刀具，亦或者架設昂貴的影像檢測設備來達到即時監控之目的，但是這樣的方式也增加了生產成本。有鑑於此，本研究開發一套簡易且快速智能化刀具壽命預測系統，能夠提供使用者當前切削刀具之情形。研究主要分為刀具壽命預測與雲端物聯網兩個系統，首先刀具壽命預測系統是利用多分量測力計與擷取Fanuc馬達驅動器電流，再以回歸分析與倒傳遞類神經網路進行主軸切削電流與切削力建模與分析，並比較切削力預測判斷能力何者較為符合系統之需求，經由實驗後預測結果得知多項式回歸分析準確度高達90.99%，倒傳遞類神經準確度高達90.5%，因此能夠透過擷取切削過程中的電流變化，以達到預測實際平均切削合力，後續透過鑑別刀具幾何及鋁合金7075-T6之間切削力係數，找出刀具與材料的切削力關係，由實驗結果得知切削力係數準確度高達93.65%，因此當鑑別出切削力係數後不論任何加工參數便能獲取理論平均切削力，並藉由電流預測出的平均切削合力比較後，透過本研究自行開發的智能化刀具壽命預測系統，利用存取Fanuc控制器Macro的共用變數，讓機台自行對本研究預測系統下達擷取與計算的能力，從中達到判斷當前刀具壽命情況。在雲端物聯網系統中，經由Fanuc Focas二次開發程式對機台進行資料擷取，並透過MQTT上傳至PostgreSQL資料庫進行資料管理及儲存，讓使用者能夠擴增不同刀具與材料之切削關係，最後以Node-red雲端網頁平台呈現，最後希望透過本研究以達到提升國內工具機產業的附加價值。

In recent years, the aerospace and the automobile industry are flourishing, and the competition of the manufacturing industry will be fierced. However, the requirements for products are not limited to correct dimensional accuracy and good surface finish, and more attention is paid for the rate of production. When the manufacturing process is fast diverse uninterrupted milling, the yield of product will be directly affected by the cutting tool wear. Therefore, it is more important to predict the tool life. However, domestic manufacturing rechanges cutting tools that were used to mill the workpieces for a periodic time, or they will buy an expensive image equipment in order to inspect the tool for real-time monitoring purposes, but this approach also increases production costs. In view of this, this study developed a simple and fast prediction system of tool life that can provide users with cutting life state of tool. This study is divided into two systems, tool life prediction and internet of things. First, tool life system uses a dynamometer to measure cutting force, and the system uses Servo Guide to catch spindle current, and then spindle current and cutting force were modeled and analyzed by using regression analysis and back-propagation neural network. Furthermore, we compare the ability of cutting force prediction to meet the accuracy of the system. The fitting results of cutting force showed that the polynomial regression was around 90.99%, and the back-propagation neural network was around 90.50%. Therefore, the system is possible to predict the actual cutting force by capturing the spindle current during the cutting process. Next, this study calculates the cutting force coefficients between the cutting tool and the aluminum alloy 7075-T6 that were helpful for the prediction of the cutting force, and the prediction accuracy is about 93.65%. Therefore, the average cutting force can be obtained regardless of any processing parameters after the cutting force coefficients are identified, and it can be predicted by spindle current. This research developed the intelligent prediction system that uses the capturing and calculation functions of the prediction system to determine the tool life. In the internet of things system, the secondary development system captured the machine's information, and the information of the mcahine will be uploaded to the PostgreSQL database for managing and storing, finally the information was presented on the Node-red web.

摘要....................................i
英文摘要(Abstract)....................................ii
誌謝....................................iii
目錄....................................iv
表目錄....................................vi
圖目錄....................................vii
符號說明....................................ix
第 一 章 緒論....................................1
1.1 研究背景與動機....................................1
1.2 研究目的....................................2
1.3 論文架構....................................2
第 二 章 文獻回顧....................................3
2.1 切削力建模及預測相關研究....................................3
2.2 刀具壽命相關研究....................................9
第 三 章 研究架構與方法....................................15
3.1 研究架構與流程....................................15
3.2 硬體及系統架構....................................16
3.2.1 硬體設備....................................16
3.2.2 實驗機台介紹....................................17
3.2.3 實驗刀具與切削材料....................................17
3.2.4 主軸電流量測軟體....................................18
3.2.5 刀具磨耗量測硬體....................................18
3.3 馬達驅動器系統....................................19
3.3.1 馬達系統模型....................................20
3.3.2 伺服控制系統....................................20
3.3.3 主軸電流擷取....................................22
3.4 回歸分析....................................24
3.4.1 回歸分析基本概論....................................24
3.4.2 線性回歸模型....................................24
3.4.3 多項式回歸分析....................................26
3.4.4 判定係數....................................26
3.5 倒傳遞類神經網路....................................27
3.5.1 倒傳遞類神經網路基本概論....................................27
3.5.2 倒傳遞類神經網路學習過程....................................28
3.5.3 Levenberg Marquardt演算法....................................30
3.5.4 倒傳遞神經網路之轉移函數....................................32
3.6 端銑刀切削力模型解析....................................33
3.6.1 切削力解析....................................33
3.6.2 平均銑削力模型....................................35
3.6.3 基於平均銑削力模型與線性回歸鑑別切削力係數....................................35
3.7 刀具磨耗與壽命....................................36
3.7.1 刀具磨耗定義....................................36
3.7.2 刀具壽命定義....................................38
3.8 雲端平台建置....................................39
3.8.1 MQTT通訊協定介紹....................................39
3.8.2 PostgreSQL資料庫介紹....................................39
3.8.3 雲端系統架構介紹....................................40
第 四 章 實驗結果....................................42
4.1 主軸切削電流模型....................................42
4.1.1 主軸電流與切削力量測實驗....................................42
4.1.2 回歸分析之切削電流模型....................................45
4.1.3 倒傳遞類神經分析之切削電流模型....................................46
4.1.4 主軸切削力預測結果比較....................................47
4.2 刀具壽命切削實驗....................................49
4.2.1 切削力係數鑑別實驗....................................49
4.2.2 平均切削力與回歸分析....................................50
4.2.3 刀具磨耗切削結果....................................52
4.3 智能化刀具壽命預測系統....................................55
4.3.1 機台監控系統介面....................................55
4.3.2 回歸分析與倒傳遞網路預測模型介面....................................56
4.3.3 智能化刀具壽命預測系統....................................57
4.4 雲端平台系統建置....................................59
4.4.1 PostgreSQL 資料庫建立....................................59
4.4.2 MQTT 雲端物聯網管理系統....................................60
4.4.3 雲端網頁建置....................................61
第 五 章 結論與未來展望....................................63
參考文獻....................................64
附錄一....................................67
附錄二....................................68
Extended Abstract....................................70

[1]S. Kurada, C. Bradley, “A review of machine vision sensors for tool condition monitoring”, Computer in Industry, 34 (1997), pp. 55-72.
[2]M. Castejon, E. Alegre, J. Barreiro, L.K. Hernandez, “On-line tool wear monitoring using geometric descriptors from digital images”, International Journal of Machine Tools and Manufacture, 47 (2007), pp. 1847-1853.
[3]Joyson Menezes, Mark A. Rubeo, Kadir Kiran, Andrew Honeycutt, and Tony L. Schmitz,
“Productivity Progression with Tool Wear in Titanium Milling”, University of North Carolina at Charlotte,2016.
[4]Stein, J. L., Colvin, D., Clever, G., and Wang, C. H., “Evaluation of DC Servo Machine Tool Feed Drives as Force Sensor”, ASME Journal of Dyanamical Systems, Measurement, and Control, 108(1986), pp.279-288.
[5]Andreas Albrechta, Simon S. Park, Yusuf Altintas, Günter Pritschowa, “High frequency bandwidth cutting force measurement in milling using capacitance displacement sensors”, International Journal of Machine Tools and Manufacture, 45(2005), pp.993-1008.
[6]Suleyman Yaldiz, Faruk Unsacar, Haci Saglam, “Comparison of experimental results obtained by designed dynamometer to fuzzy model for predicting cutting forces in turning”, Materials & Design,27(2006), pp.1139-1147.
[7]Dohyun Kim, Doyoung Jeon, “Fuzzy-logic control of cutting forces in CNC milling processes using motor currents as indirect force sensors”, Precision Engineering,35(2011),
pp.143-152.
[8]M.Postel, D. Aslan, K. Wegener, Y. Altintas, “Monitoring of vibrations and cutting forces with spindle mounted vibration sensors”, CIRP Annals,68(2019), pp.413-416.
[9]Deniz Aslan, Yusuf Altintas, “Prediction of Cutting Forces in Five-Axis Milling Using Feed Drive Current Measurements”, 23(2018), IEEE, pp.833-844.
[10]Cyril Drouillet, Jaydeep Karandikara, Chandra Natha, Anne-Claire Journeaux, Mohamed El Mansori, Thomas Kurfess, “Tool life predictions in milling using spindle power with the neural network technique”, Journal of Manufacturing Processes, 22(2016), pp.161–168.
[11]Muhammad Rizal, Jaharah A. Ghani, Mohd Zaki Nuawi, Che Hassan Che Haron, “Online tool wear prediction system in the turning process using an adaptive neuro-fuzzy inference system”, Applied Soft Computing, 13(2013), pp.1960-1968.
[12]A. Gouarir, G. Martinez Arellano, G. Terrazas, P. Benardos and S. Ratchev, “In-process Tool Wear Prediction System Based on Machine Learning Teachniques and Force Analysis”, Science Direct, 77(2018), pp.501-504.
[13]Pavel Kovac, Marin Gostimirovic, Dragan Rodic, Borislav Savkovic, “Using the temperature method for the prediction of tool life in sustainable production”, Measurement, 133(2019), pp.320-327.
[14]Yinfei Yang, Yuelong Guo, Zhiping Huang, Ni Chen, Liang Li, Yifan Jiang, Ning He, “Research on the milling tool wear and life prediction by establishing an integrated predictive model”, Measurement, 145(2019), pp.178–189.
[15]葉怡成,類神經網路模式應用與實作,儒林圖書有限公司,1993年1月初版,2000年4月7版.
[16]TaeguTec Member IMC Group：http://tw.taegutec.com
[17]Y. Altintas, ”Manufacturing Automation: Metal Cutting Mechanics, Machine Tool Vibrations, And CNC Design”, page 44, Cambridge University Press, 2012.
[18]周國華,”工具機應用設計”, 翰蘆圖書出版有限公司,2014年11月二版.
[19]陳昭亮,郭庭毓,2016,“工具機伺服進給系統模型建立與調機之研究”,國立中興大學,碩士論文.
[20]方世榮,張文賢, “統計學導論”, 華泰文化, 2014年1月.
[21]梁淑芬,趙景明,2006,“導入LM法之平行倒傳遞演算法”,國立中正大學,博士論文.
[22]林其德,郭春寶,2012,“鎳基合金端銑刀切削溫度與刀具壽命研究”,國立中正大學,博士論文.
[23]范揚志,陳文欽,2006,“應用類神經網路與基因演算法於射出成型製程參數最佳化之研究”,國立中華大學,碩士論文.
[24]林伯尚,張文陽,2018,“基於倒傳遞類神經網路之刀具壽命預測與表面粗糙度”,國立虎尾科技大學,碩士學位論文
[25]林俊良, “智慧型控制分析與設計”, 全華圖書股份有限公司,第三版.
[26]沈耘生,張文陽,2018,“工具機銑削力與主軸熱變形補償分析”,國立虎尾科技大學,碩士學位論文
[27]周國華, “工具機切削技術研究”,翰蘆圖書,2017
[28]陳威辰,王俊志,2015,“端銑刀刃口幾何對系統穩定性影響之研究”,國立成功大學,碩士學位論文
[29]Y. Altintas,” Manufacturing Automation: Metal Cutting Mechanics, Machine Tool Vibrations, And CNC Design”, page 44, Cambridge University Press, 2012.
[30]ISO, “Tool Life Testing with Single-Point Turning Tools”ISO, 5th darft proposal, ISO/TC 29/ WG 22(secretariat,37), 91(1972).
[31]Fanuc, ”Fanuc Servo Guide Operator’s Manual”
[32]Daniel Johansson, Sören Hägglund, Volodymyr Bushlya, Jan-Eric Ståhl, “Assessment of Commonly Used Tool Life Models in Metal Cutting”, Procedia Manufacturing, 11(2017), pp.602 – 609.
[33]Hagan, M. T. and M. B. Menhaj, “Traning Feedforward Network with the Marquardt Alogorithm”, IEEE Transaction on Neural Networks, Vol. 5, No. 6. 10(1994).
[34]陳昭亮,張哲豪,2014,“工具機伺服控制虛擬技術之研究”,國立中興大學,碩士學位論文
[35]G. Rizzoni, “ Principles and Applications of Electrical Engineering”, McGraw-Hill, 2007.
[36]李韋霆,張文陽,2016,“高速鋼端銑刀銑削不銹鋼之刀具磨耗與工件表面粗糙度研究”,國立虎尾科技大學,碩士學位論文
[37]江孟桓,張文陽,2016,“智能化工具機即時加工檢測與伺服系統調機”,國立虎尾科技大學,碩士學位論文
[38]連義宏,張文陽,2016,“基於田口式分析刀刃數切削316不鏽鋼特性”,國立虎尾科技大學,碩士學位論文