(3.229.120.26) 您好!臺灣時間:2021/04/10 22:47
字體大小: 字級放大   字級縮小   預設字形  
回查詢結果

詳目顯示:::

我願授權國圖
: 
twitterline
研究生:黃培鈞
研究生(外文):Huang,Pei-Jyun
論文名稱:人工智慧應用於汽車鈑金焊接品質監測
論文名稱(外文):Quality Monitoring of Vehicle Sheet Metal Welding with Artificial Intelligence
指導教授:吳英正
指導教授(外文):Wu,Ying-Jeng
口試委員:吳尚德嚴家銘
口試委員(外文):Wu,Shang-TehYen,Chiaming
口試日期:2018-06-12
學位類別:碩士
校院名稱:國立雲林科技大學
系所名稱:機械工程系
學門:工程學門
學類:機械工程學類
論文種類:學術論文
論文出版年:2018
畢業學年度:106
語文別:中文
論文頁數:74
中文關鍵詞:電阻點焊阻抗功率人工智慧類神經網路
外文關鍵詞:Resistance spot weldingImpedancePowerArtificial intelligenceNeural network
相關次數:
  • 被引用被引用:0
  • 點閱點閱:195
  • 評分評分:系統版面圖檔系統版面圖檔系統版面圖檔系統版面圖檔系統版面圖檔
  • 下載下載:0
  • 收藏至我的研究室書目清單書目收藏:2
電阻點焊被廣泛的使用在各種工業上,例如:汽車、航空等,所以點焊的好壞非常重要。點焊品質檢驗一般可分為兩大類,破壞性檢驗與非破壞性檢驗。破壞性檢驗會使材料的成本增加且耗時;非破壞性檢驗設備較為昂貴,且需要專業的檢驗人員。為了降低成本與檢測時間,Shih-Fu Ling&Lixue Wan(2000,[1])應用阻抗的概念,擷取焊接過程中適當的阻抗特徵值,監測焊接品質。除了阻抗與阻抗微分曲線的使用外,在焊接過程中會有能量的損失變化,所以本論文在電阻點焊中加入功率曲線的量測與計算。在焊接過程中,焊點有無噴火花兩者之間的曲線有所差異,所以將其分開做處理,並且以柯任修(2017,[5])選擇的五種先端徑條件以及四種鈑材條件,所得到的阻抗及功率特徵值,訓練人工智慧當中的類神經網路,之後以標準鈑材及標準先端徑,但不同材質及厚度之試驗片,檢視網路之成效,最後本論文以柯任修(2017,[5])已完成的設備為基礎,在硬體方面新增觸發器及USB模組,實現了能夠直接使用於實際生產線上點焊製程的監測系統。
Resistance spot welding is widely used in various industries such as vehicles, aviation, etc., so spot welding is very important. In general, spot welding quality inspection can be divided into two categories, destructive testing and non-destructive testing. Destructive inspections increase the cost of materials and time consuming; non-destructive inspection equipments are more expensive and require professional inspectors. In order to reduce cost and monitoring time, Shih-Fu Ling & Lixue Wan (2000, [1]) applies the concept of impedance, extracts the appropriate impedance eigenvalues in the welding process, and identifies the welding quality. In addition to the use of impedance and impedance differential curves, there will be energy changes in the welding process, so this thesis adds measurement and calculation of power curve in resistance spot welding. In the welding process, there is a difference in the curve between the two whether the welding spot has sparks or not, so it is treated separately, and select five kinds of tip diameter conditions and four kinds of metal plate conditions by Jen-Hsiu Ko(2017, [5]). The obtain impedance and power eigenvalues are trained on neural network among artificial intelligence. Afterwards, the effectiveness of the network was examined using standard caskets and standard tip diameters, but with test specimens of different materials and thicknesses. Finally, based on the equipment completed by Jen-Hsiu Ko(2017, [5]), this thesis added triggers and USB modules to the hardware,and completed a monitoring system that can be directly used in spot welding processes on actual production lines.
摘要.....................................i
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 研究目的與動機........................1
1.2 文獻回顧.............................1
1.3 研究方法.............................2
1.4 論文架構.............................2
第二章 電阻點焊阻抗及功率變化之量測.......4
2.1硬體架構..............................4
2.1.1 感應線圈(Rogowski coil,羅氏線圈)...5
2.1.2 積分器.............................5
2.1.3 低通濾波器.........................8
2.1.4 放大器.............................11
2.1.5 觸發器.............................13
2.1.6 類比數位轉換器(ADC)規格.............13
2.1.7 DSK(DSP Starter Kit)簡介...........14
2.1.8 USB模組簡介........................14
2.2 電阻點焊時序..........................15
2.3 電阻點焊等效電路......................16
第三章 數位訊號處理......................18
3.1 數位帶通濾波器........................18
3.1.1 帶通濾波器的設計....................19
3.2 希爾伯特轉換(Hilbert Transform).......21
3.2.1 希爾伯特轉換(Hilbert Transform)的設計....21
3.3 阻抗及功率曲線........................22
第四章 類神經網路.........................24
4.1 網路架構.............................25
4.1.1 人工神經元.........................25
4.1.2 激勵函數(Activation Function)......25
4.1.3 倒傳遞類神經網路(Back-Propagation Neural Network)...26
4.2 訓練演算法...........................27
第五章 訓練樣本實驗規劃...................34
5.1 先端徑種類...........................34
5.2 鈑材條件.............................35
第六章 類神經網路的訓練及測試..............36
6.1 點焊品質監測程序......................37
6.2 焊點火花判斷網路(網路1)...............38
6.2.1 網路1特徵選擇及網路架構.............39
6.2.2 網路1判別結果......................41
6.3 焊點無火花之品質判斷網路(網路2).......41
6.3.1 網路2特徵選擇及網路架構.............41
6.3.2 網路2判別結果......................43
6.4 焊點有火花之品質判斷網路(網路3).......44
6.4.1 網路3特徵選擇及網路架構.............44
6.4.2 網路3判別結果......................46
6.5 試驗片判斷之結果......................47
6.6 線上焊接實驗之成效....................50
6.6.1 線上焊接實驗鈑材組成................50
6.6.2 線上焊接實驗參數及判別結果...........51
6.6.3 線上實驗結果討論....................52
第七章 結論與建議.........................54
7.1 結論.................................54
7.2 建議.................................56
參考文獻.................................58
附錄.....................................59
附錄A 網路1之嘗試.........................59
附錄B 網路2之嘗試.........................61
附錄C 網路3之嘗試.........................62

[1] Shih-Fu Ling & Lixue Wan, 2000, Monitoring a spot welding process via electrical input impedance, in Proceeding of the 19th International Conference on Experimental Mechanics, pp. 348–351.
[2] 謝宗賢,2002,DSP於阻抗監測之應用,雲林科技大學機械所碩士論文.
[3] 鄭守益,2003,類神經網路之點焊品質鑑別,雲林科技大學機械所碩士論文.
[4] 林志偉,2004,焊接品質鑑別之研究,雲林科技大學機械所碩士論文.
[5] 柯任修,2017,結合類神經網路與電阻點焊阻抗變化應用於汽車鈑金焊接品質鑑別,雲林科技大學機械所碩士論文.
[6] B.M.Wilamowski & H. Yu, 2010, "Improved Computation in Levenberg Marquardt Training", IEEE Trans. on Neural Networks, vol. 21, no. 6, pp.930–937.
[7] Alan V. Oppenheim & Ronald W. Schafer, 1999, Discrete-time signal processing, 2nd Edition, Prentice Hall, pp. 792-795.
[8] A.R.Collins (n. d.).MISCELLANY: Miscellaneous technical articles by Dr.A.R. Collins,retrieved April 5, 2018,from http://www.arc.id.au/FilterDesign.html

電子全文 電子全文(網際網路公開日期:20230621)
QRCODE
 
 
 
 
 
                                                                                                                                                                                                                                                                                                                                                                                                               
第一頁 上一頁 下一頁 最後一頁 top
系統版面圖檔 系統版面圖檔