本研究主旨在探討非破壞性檢測技術對鋁板裂紋長度的應用，利用有限元素法模擬不同裂紋長度試片之訊號情況，並透過產波器和示波器收集實際訊號，運用人工智慧的方式準確辨識出鋁板上裂紋的長度。
本研究在實驗方面，先將壓電材料黏貼至鋁板表面，利用產波器激振一片壓電片產生訊號，同時利用示波器連接另一片壓電片接收訊號。此外本研究採用abaqus軟體，以有限元素法模擬平板蘭姆波之暫態波傳，並建立與實際鋁板尺寸大小、厚度相符、壓電材料黏貼位置一致的模型，並設定適合此模型的網格尺寸、時間增量，再依據不同的裂紋長度建立出不同裂紋長度的模型，模擬不同裂紋長度下的蘭姆波訊號，將模擬產生的蘭姆波訊號當作人工智慧資料。
本研究建立兩種不同人工智慧模型分別為以數值數據當作資料、以實驗數據當作資料，建立人工智慧辨識裂紋長度之模型，運用人工智慧深度學習 (Deep Learning) 演算法中的人工神經網路 (Artificial Neural Network, ANN)，進行資料分析，以建立最佳的演算法模型，以數值數據當作人工智慧訓練集資料辨識測試集資料，共建立兩種模型，分別為辨識以mm為單位之整數裂紋長度模型(例如3 mm、5 mm等)、以mm為單位非整數裂紋長度模型(例如3.2 mm、4.2 mm等)，於整數裂紋長度辨識準確度約為 100 %；於非整數裂紋長度辨識準確度約為 80 %，以實驗數據當作人工智慧訓練集資料辨識測試集資料，也建立兩種模型，分別為辨識整數裂紋長度模型、非整數裂紋長度模型，於整數裂紋長度辨識準確度約為 100 % ；於非整數裂紋長度辨識準確度約為 100 %。

The purpose of this study is to explore the application of non-destructive techniques to detect crack lengths in aluminum plates. In the experiments, the piezoelectric transducers were adhered on the surface of the aluminum plates. A function generator was used to apply a voltage to a piezoelectric transducer to generate a signal, while an oscilloscope connected to another piezoelectric transducer received the signal. Additionally, the finite element software ABAQUS was used to simulate the Lamb wave propagations in the plate. The aluminum plate model was established that matches that in the experiments The simulated Lamb wave signals obtained in the models were used as the data for artificial intelligence.
This study establishes two different artificial intelligence models, one using numerical data and the other using experimental data, to create a model for crack length identification. Artificial neural networks (ANNs), a deep learning algorithm, were utilized for data analysis to develop the optimal algorithm models. Using numerical data as the training set for artificial intelligence to identify the test set data, two models were developed: one for identifying integer crack lengths and another for identifying crack lengths with crack length of non-integer in mm. The accuracy of the integer crack length in mm identification model is approximately 100%, while the accuracy for the non-integer crack length in mm model is about 80%. Using experimental data as the training set for artificial intelligence, two models were also established for identifying integer and non-integer crack lengths, with an accuracy of about 100% for both

[1]S. Gholizadeh, “A review of non-destructive testing methods of composite materials,” Procedia Struct. Integr., vol. 1, pp. 50–57, 2016.
[2]H. Lamb, “On waves in an elastic plate,” Proc. R. Soc. Lond. A Math. Phys. Sci., vol. 93, no. 648, pp. 114–128, 1917.
[3]D. Worlton, “Ultrasonic testing with lamb waves,” Office of Scientific and Technical Information (OSTI), Richland, Washington, 1956.
[4]V. Giurgiutiu, J. Bao, and W. Zhao, “Active sensor wave propagation health monitoring of beam and plate structures,” in Smart Structures and Materials 2001: Smart Structures and Integrated Systems, Bellingham, Washington, 2001.
[5]S. H. Díaz Valdés and C. Soutis, “Health monitoring of composites using Lamb waves generated by piezoelectric devices,” Plast. Rubber Compos., vol. 29, no. 9, pp. 475–481, 2000.
[6]A. De Luca, D. Perfetto, A. De Fenza, G. Petrone, and F. Caputo, “Guided wave SHM system for damage detection in complex composite structure,” Theoretical and Applied Fracture Mechanics, vol. 105, p. 102408, 2020.
[7]A. De Luca, D. Perfetto, A. De Fenza, G. Petrone, and F. Caputo, “Guided waves in a composite winglet structure: Numerical and experimental investigations,” Composite Structures, vol. 210, pp. 96–108, 2019.
[8]A. De Luca, D. Perfetto, A. Polverino, A. Aversano, and F. Caputo, “Finite Element Modeling Approaches, Experimentally Assessed, for the Simulation of Guided Wave Propagation in Composites,” Sustainability, vol. 14, no. 11, p. 6924, 2022.
[9] N. Markovic, D. Stojic, R. Cvetkovic, V. Radojicic, and S. Conic, “Numerical modeling of ultrasonic wave propagation - by using of explicit FEM in ABAQUS,” Facta Univ Arch Civ Enge, vol. 16, no. 1, pp. 135–147, 2018.
[10]C. A. C. Leckey, K. R. Wheeler, V. N. Hafiychuk, H. Hafiychuk, and D. A. Timuçin, “Simulation of guided-wave ultrasound propagation in composite laminates: Benchmark comparisons of numerical codes and experiment,” Ultrasonics, vol. 84, pp. 187–200, 2018.
[11]G. F. Gomes, F. A. De Almeida, D. M. Junqueira, S. S. Da Cunha, and A. C. Ancelotti, “Optimized damage identification in CFRP plates by reduced mode shapes and GA-ANN methods,” Engineering Structures, vol. 181, pp. 111–123, 2019.
[12]D. Perfetto, A. De Luca, M. Perfetto, G. Lamanna, and F. Caputo, “Damage Detection in Flat Panels by Guided Waves Based Artificial Neural Network Trained through Finite Element Method,” Materials, vol. 14, no. 24, p. 7602, 2021.
[13]J. Wu, X. Xu, C. Liu, C. Deng, and X. Shao, “Lamb wave-based damage detection of composite structures using deep convolutional neural network and continuous wavelet transform,” Composite Structures, vol. 276, p. 114590, 2021.
[14]G. Azuara, M. Ruiz, and E. Barrera, “Damage Localization in Composite Plates Using Wavelet Transform and 2-D Convolutional Neural Networks,” Sensors, vol. 21, no. 17, p. 5825, 2021.
[15]C. Su, M. Jiang, S. Lv, and S. Lu, “Improved Damage Localization and Quantification of CFRP Using Lamb Waves and Convolution Neural Network,” IEEE Sensors J., vol. 19, no. 14, pp. 5784–5791, 2019.
[16]P. Pandey, A. Rai, and M. Mitra, “Explainable 1-D convolutional neural network for damage detection using Lamb wave,” Mechanical Systems and Signal Processing, vol. 164, p. 108220, 2022.
[17]B. Xu, L. Tong, T. Bi-wan, and H. Xin-jing, “A continuous leakage real-time localization method based on space phase image of elastic wave field with improved CNN,” Measurement, vol. 224, p. 113894, 2024.
[18]V. Vy, Y. Lee, J. Bak, S. Park, S. Park, and H. Yoon, “Damage localization using acoustic emission sensors via convolutional neural network and continuous wavelet transform,” Mechanical Systems and Signal Processing, vol. 204, p. 110831, 2023.
[19]B. Le, T. Le, T. Luu, D. Ho, and T. Huynh, “Fault Assessment in Piezoelectric-Based Smart Strand Using 1D Convolutional Neural Network,” Buildings, vol. 12, no. 11, p. 1916, 2022.
[20]P. Pandey, A. Rai, and M. Mitra, “Explainable 1-D convolutional neural network for damage detection using Lamb wave,” Mechanical Systems and Signal Processing, vol. 164, p. 108220, 2022.
[21]S. Yu, C. Fun, Q. Chen, B. Gao, and J. Guo, “A convolutional neural network method for damage location based on raw guided Lamb wave technique,” IEEE Far East NDT New Technology & Application Forum, pp.63-67,2021
[22]B. Jaffe, R. S. Roth, and S. Marzullo, “Properties of piezoelectric ceramics in the solid-solution series lead titanate-lead zirconate-lead oxide: Tin oxide and lead titanate-lead hafnate,” J. RES. NATL. BUR. STAN., vol. 55, no. 5, p. 239, 1955.
[23]洪嘉志，“ 蜂巢式佈置之蘭姆波壓電感測器陣列應用於大面積平板結構的健康偵測 ” 國立成功大學機械工程學系，2023.