小波理論與類神經網路在超音波檢測之應用 潘永振、葉競榮
E-mail: [email protected]
摘 要
鋼鐵材料在焊接時難免會產生瑕疵，焊接瑕疵可以粗分為平面狀及立體狀兩大類。由於前者在應力作用下其尖端的應力集 中因素很大，因此極易成長及造成設施的破裂，所以在新建工程中不論其大小如何均需予以檢出，因此在以超音波檢測焊 道時如何正確辨識平面狀瑕疵是很重要的課題。 本研究首先準備具有各種不同瑕疵的焊接試片，瑕疵發生於不同部位並有 不同的大小。然後利用超音波儀器檢出瑕疵信號。瑕疵信號經小波分析萃取其特徵後再以類神經網路予以辨識分類。在本 研究中平面狀瑕疵及立體狀瑕疵的平均正確辨識率分別高達94%及90.19%，可見所設計的處理程序確能辨識不同的種類的 瑕疵，未來在進一步改進並經過實用的考驗後，相信一定可以成為一套實用及可靠的智慧型瑕疵辨識系統。
關鍵詞 : 超音波檢測 ; 類神經網路 ; 小波理論 ; 平面狀瑕疵 目錄
目錄 封面內頁 簽名頁 授權書...iii 中文摘要...iv 英文摘要...v 誌 謝...vi 目錄...vii 圖目錄...ix 第一章 緒
論...1 1.1 研究動機與目的...1 1.2 文獻回顧...2 1.3 論文架 構...3 第二章 超音波檢測...4 2.1 非破壞性檢測概述...4 2.2 超音波檢 測...5 2.2.1 超音波產生之原理...5 2.2.2 音波的種類...6 2.2.3 超音波檢測方 法...9 2.3 不同瑕疵之回波特性...10 第三章 小波理論...11 3.1 小波轉換簡 介...11 3.2 時頻分析...11 3.3 離散小波轉換...15 3.4 多重解
析...16 3.5 多重解析度之金字塔架構...21 第四章 類神經網路...28 4.1 類神經網路發 展史...28 4.2 類神經網路基本架構...28 4.3 人工神經元模型...29 4.4 類神經網路之分 類...31 4.5 類神經網路之運作原理及特性...33 4.6 倒傳遞演算法...34 第五章 實驗方法與結 果...37 5.1 實驗流程...37 5.2 實驗結果...42 第六章 結論...49 參 考 文 獻...50
參考文獻
參考文獻 [1] AWS, “AWS D1.1 Structural Weldung Code — Steel 1996” ,American Welding ociety, 1996.
[2] A. Grossmann, J. Morlet, “Decomposition of Hardy function into square integrable wavelet of constant shape,” SIAM J. Math Anal.15(4):pp.736-783,1984.
[3] S. Mallat, “A Theory for Multiresolution Signal Decomposition: The Wavelet Representation,” IEEE Transaction on Pettern Analysis and Machine Intelligence, Vol.11, No.11, pp.674-693, July 1989.
[4] Ingrid Daubechies, “Orthonomal bases of compactly supported wavelets,” Communications on Pure and Applied Mathematics, Vol. XLI, pp.909-996, 1988.
[5] Ingrid Daubechies, “The wavelet transform, time frequency localization and signal analysis,” IEEE Transactions on Information Theory, Vol. 36, No. 5, pp.961-1005, 1990.
[6] W. S. McCulloch, and W. Pitts, “A logical Calculus of the Ideas Immanent in Nervous activity,” Bulletin of Mathematical Biophysics 5:
pp.115-133, 1943.
[7] J.J. Hopfield, “Neural Networks and Physical with Emergent Collective Computational Abilities,” Proc. Natl. Acad. Sci USA,,1982.
[8] D. E. Rumelhard, G.E. Hintion, and R. J. Willians, “Learning Representations by Back-Propagation errors,” Nature, 323, pp.533-536, 1986.
[9] P. Carter, “Experience With the time-of-flight diffraction technique and an accompanying portable and versatile ultrasonic digital recording system”, Brit. J. of NDT, Sept. 1984, pp 354~361.
[10] J.,Verkooijen, “TOFD used to replace radiography”, INSIGHT, Vol. 37(6), pp 433-435, June 1995.
[11] J.J. Wild, “The use of ultrasonic pulses for the measurement of biological tissue and the detection of tissue density changes”, Surgery, Vol.
27 , pp. 183-188, 1950.
[12] J.-S. R. Jang, C.-T. Sun, and E. Mizutani “Neuro-Fuzzy and Soft Computing”, Prentice-Hall International, Inc, 1997.
[13] CHIN-TENG and C.S GEORGE LEE “Neural Fuzzy systems” , Prentice-Hall International, Inc.1999.
[14] K. Harumi, Y. Ogura and M. Uchida, “Ultrasonic Defect Sizing”, Japanese Tip Echo Handbook, Tip Echo Group of 210 and 202 Sub-committee of Japanese Society for Nondestructive Inspection, 1989.
[15] C. Yeh and R. Zoughi, “Sizing Technique for Slots and Surface Cracks in Metals”, Materials Evaluation, Vol. 53, No 4Apr. 1995, pp.496-501.
[16] A.H Spinks “A Portable Digital Ultrasonic Image recording system”, Paper presented at the 6th Asian-Pacific Conference on Non Destructive Testing at Blenheim, New Zealand,5-9 March 1990.
[17] Ingred Daubechies, “The wavelet transform, time-frequency localization and signal analysis”, IEEE Tran Information Theory, Vol. 36, No 5, Sep. 1990.
[18] A. Abbate, J. Koay, J. Frankel, S. C. Schoeder, and P. Das, “Signal detection and noise suppression using a wavelet transform signal processor: Application to ultrasonic flaw detection,” IEEE Trans. Ultrason., Ferrolect., Freq. Contr., Vol. 44, pp 14-27,Jan 1997.
[19] Robi Polikar ,Lalita Udpa, Satish S.Udps, and Tom Taylor, “Frequency Invariant Classification of Ultrasonic Weld Inspection Signals”, IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, Vol.45,NO.3,May 1998.
[20] Virginio Cantoni and Alfredo Petrosino “, Neural Recognition in a Pyramidal Structure,” IEEE Transactions on Neural Networks, Vol. 13 , No .2 , March 2002.
[21] Jun-Youl Lee, Muhammad Afzal, Satish Udpa, Lalita Udpa and Peter Massopust, “Hierarchical Rule Based Classification of MFL Signals Obtained from Natural Gas Pipeline inspection,” IEEE Transactions on Neural Networks, 2000.
[22] 單維彰,“凌波初步”,全華科技圖書,台北,民國88年.
[23] 吳學文、黃啟貞、陳碧冠、葉競榮, “超音波檢測法初級”,中華民國非破壞檢測協會,1988.
[24] 葉競榮、徐鴻發,“超音波檢測法中級”,中華民國非破壞檢測協會,1990.
[25] 張智星,“MATLAB程式設計與應用”,清蔚科技,2000.
