臺灣博碩士論文加值系統

刀具的磨耗量是影響加工件精度的重要因素，以往判斷換刀時機都是以加工後的成品毛邊的多寡，或是以累積的加工數量來決定是否換刀。太早進行換刀將會使刀具成本提升，而太晚換刀則會使工件品質下降，且會提高工件的不良率，導致工件材料成本的提升，綜合上述問題，及時的刀具磨耗檢測就顯得格外重要。本文將以銑削的振動訊號來進行分析與探討，針對加工的刀具進行刀具磨耗的自動化智慧診斷機制（Automatic Intelligent Diagnosis Mechanism，AIDM）的判斷，透過AIDM中兩種不同的訊號處理方法進行特徵擷取，並代入不同的分類器進行辨識。AIDM內共有五種診斷機制，第一種與第二種皆透過有限脈衝響應（Finite Impulse Response Filter，FIR）濾波器搭配近似熵理論（Approximate Entropy，ApEn）來進行特徵擷取，並將特徵擷取後之特徵分別代入倒傳遞類神經網路（Back Propagation Neural Network，BPNN）以及支持向量機（Support Vector Machine）來做辨識，第三種至第五種診斷機制皆透過分數階Chen-Lee混沌系統來進行非線性特徵映射，取其混沌系統動態誤差之重心點以及混沌動態誤差分布圖來做為狀態識別特徵，分別代入倒傳遞類神經網路、支持向量機以及卷積神經網路（Convolutional Neural Network，CNN）來做辨識。透過實驗結果比較，第一組實驗訊號將端銑刀A的銑削訊號進行模型訓練，端銑刀B的銑削訊號進行測試下，在特徵擷取為分數階Chen-Lee混沌系統搭配倒傳遞類神經網路進行辨識有著98.66%最高的辨識率，而第二組實驗訊號將端銑刀A以及端銑刀B的銑削訊號進行模型訓練，端銑刀C的銑削訊號進行測試下，在特徵擷取為分數階Chen-Lee混沌系統搭配卷積神經網路進行辨識89.33%的最高辨識率，並在實驗結果證實AIDM模型對於輸入訊號可以選擇最適合的分類模型來辨識刀具目前壽命狀況。

The amount of tool wear is an important factor that affects the accuracy of the processed parts. In the past, the timing of tool change was determined by the amount of burrs in the finished product, or the cumulative number of processing to determine whether to change the tool. performing a tool change too early will increase the cost of the tool, while changing the tool too late will reduce the quality of the workpiece, and will increase the defect rate to increase the material cost. In view of the above problems, timely tool wear detection appears especially important. This article will analyze and discuss the vibration signals of milling, and determine the Automatic Intelligent Diagnosis Mechanism (AIDM) of tool wear for the processed tools, and use two different signal processing methods in AIDM for feature extraction , And substitute different classifiers for identification. There are five kinds of diagnostic mechanisms in AIDM. The first and the second are through the Finite Impulse Response Filter (FIR) filter with Approximate Entropy (ApEn) theory for feature extraction.The extracted features are respectively substituted into Back Propagation Neural Network (BPNN) and Support Vector Machine (SVM) for identification. The third to fifth diagnosis mechanisms are all through fractional Chen-Lee chaos. The system performs nonlinear feature mapping, and takes the center of gravity of the chaotic system dynamic error and the chaotic dynamic error distribution map as state recognition features, and substitutes them into the Back Propagation Neural Network, Support Vector Machine, and Convolutional Neural Network (CNN). Through the comparison of experimental results, the first set of experimental signals modeled the milling signal of the end mill A and the milling signal of the end mill B was tested, and the feature extraction was a fractional-order Chen-Lee chaotic system with inverted transitive neurons. Network recognition has the highest recognition rate of 98.66%, and the second set of experimental signals modeled the milling signals of end mill A and end mill B, and the milling signals of end mill C were tested under feature extraction The fractional-order Chen-Lee chaotic system is equipped with a convolutional neural network to identify the highest recognition rate of 89.33%, and the experimental results confirm that the AIDM model can select the most suitable classification model for the input signal to identify the current life status of the tool.

摘要 I
Abstract III
致謝 V
目錄 VI
圖目錄 VIII
表目錄 XI
第一章 序論 1
1.1 前言 1
1.2 文獻回顧 2
1.3 論文架構 5
第二章 硬體簡介與實驗架構 6
2.1 實驗硬體簡介 6
2.2 實驗架構 10
第三章 訊號處理方法及理論 15
3.1 Automatic Intelligent Diagnosis Mechanism 15
3.2 有限脈衝響應濾波器 16
3.3 近似熵理論 17
3.4 混沌理論 18
3.5 Chen-Lee混沌系統動態誤差 19
3.6 分數階Chen-Lee混沌系統動態誤差 21
3.7 倒傳遞類神經網路 24
3.8 支持向量機 27
3.9 卷積神經網路 29
3.9.1 卷積運算（Convolutional Calculation） 30
3.9.2 池化（Pooling）31
3.9.3 激活層（Activation Layer）33
第四章 實驗結果與討論 34
4.1 銑削刀具磨耗現象分析及討論 34
4.2 有限脈衝響應濾波器搭配近似熵理論特徵擷取結果 40
4.3 分數階Chen-Lee混沌系統非線性特徵映射結果 42
4.4 倒傳遞類神經網路以及支持向量機其訓練以及辨識結果 53
4.5 卷積神經網路其訓練以及辨識結果 55
4.6 實驗結果與討論 63
第五章 結論 65
參考文獻 67

[1]A. Ancarani, C. Di Mauro, and F. Mascali, "Backshoring strategy and the adoption of Industry 4.0: Evidence from Europe," Journal of World Business, vol. 54, pp. 360-371, Jun 2019.
[2]S. Salloum, J. Z. Huan, and Y. L. He, "Random Sample Partition: A Distributed Data Model for Big Data Analysis," Ieee Transactions on Industrial Informatics, vol. 15, pp. 5846-5854, Nov 2019.
[3]F. Montori, L. Bedogni, and L. Bononi, "A Collaborative Internet of Things Architecture for Smart Cities and Environmental Monitoring," Ieee Internet of Things Journal, vol. 5, pp. 592-605, Apr 2018.
[4]B. Jinsong, G. Yuan, Z. Xiaohu, Z. Jianguo, and J. Xia, "A Data Driven Model for Predicting Tool Health Condition in High Speed Milling of Titanium Plates Using Real-Time SCADA," Procedia CIRP, vol. 61, pp. 317-322, 2017/01/01/ 2017.
[5]Y. Q. Dai and K. P. Zhu, "A machine vision system for micro-milling tool condition monitoring," Precision Engineering-Journal of the International Societies for Precision Engineering and Nanotechnology, vol. 52, pp. 183-191, Apr 2018.
[6]B. S. Prasad and M. P. Babu, "Correlation between vibration amplitude and tool wear in turning: Numerical and experimental analysis," Engineering Science and Technology-an International Journal-Jestech, vol. 20, pp. 197-211, Feb 2017.
[7]J. Downey, S. Bombiński, M. Nejman, and K. Jemielniak, "Automatic Multiple Sensor Data Acquisition System in a Real-time Production Environment," Procedia CIRP, vol. 33, pp. 215-220, 2015/01/01/ 2015.
[8]I. Abu-Mahfouz and A. Banerjee, "Drill Wear Feature Identification under Varying Cutting Conditions Using Vibration and Cutting Force Signals and Data Mining Techniques," Procedia Computer Science, vol. 36, pp. 556-563, 2014/01/01/ 2014.
[9]T. Mohanraj, S. Shankar, R. Rajasekar, N. R. Sakthivel, and A. Pramanik, "Tool condition monitoring techniques in milling process — a review," Journal of Materials Research and Technology, vol. 9, pp. 1032-1042, 2020/01/01/ 2020.
[10]I. Abu-Mahfouz, "Drilling wear detection and classification using vibration signals and artificial neural network," International Journal of Machine Tools & Manufacture, vol. 43, pp. 707-720, May 2003.
[11]C. S. MacInnes, "FFT-based joint bearing estimation," Ieee Journal of Oceanic Engineering, vol. 27, pp. 774-779, Oct 2002.
[12]J. D. Rangel-Magdaleno, R. D. Romero-Troncoso, R. A. Osornio-Rios, E. Cabal-Yepez, and A. Dominguez-Gonzalez, "FPGA-Based Vibration Analyzer for Continuous CNC Machinery Monitoring With Fused FFT-DWT Signal Processing," Ieee Transactions on Instrumentation and Measurement, vol. 59, pp. 3184-3194, Dec 2010.
[13]W. Deng, S. J. Zhang, H. M. Zhao, and X. H. Yang, "A Novel Fault Diagnosis Method Based on Integrating Empirical Wavelet Transform and Fuzzy Entropy for Motor Bearing," Ieee Access, vol. 6, pp. 35042-35056, 2018.
[14]Z. Q. Huo, Y. Zhang, P. Francq, L. Shu, and J. F. Huang, "Incipient Fault Diagnosis of Roller Bearing Using Optimized Wavelet Transform Based Multi-Speed Vibration Signatures," Ieee Access, vol. 5, pp. 19442-19456, 2017.
[15]J. Kopac and S. Sali, "Tool wear monitoring during the turning process," Journal of Materials Processing Technology, vol. 113, pp. 312-316, Jun 15 2001.
[16]M. C. Lu and E. Kannatey-Asibu, "Flank wear and process characteristic effect on system dynamics in turning," Journal of Manufacturing Science and Engineering-Transactions of the Asme, vol. 126, pp. 131-140, Feb 2004.
[17]R. E. Haber, J. E. Jimenez, C. R. Peres, and J. R. Alique, "An investigation of tool-wear monitoring in a high-speed machining process," Sensors and Actuators a-Physical, vol. 116, pp. 539-545, Oct 29 2004.
[18]M. Malekian, S. S. Park, and M. B. G. Jun, "Tool wear monitoring of micro-milling operations," Journal of Materials Processing Technology, vol. 209, pp. 4903-4914, Jun 1 2009.
[19]B. Kaya, C. Oysu, H. M. Ertunc, and H. Ocak, "A support vector machine-based online tool condition monitoring for milling using sensor fusion and a genetic algorithm," Proceedings of the Institution of Mechanical Engineers Part B-Journal of Engineering Manufacture, vol. 226, pp. 1808-1818, Nov 2012.
[20]G. S. Hong, M. Rahman, and Q. Zhou, "Using neural network for tool condition monitoring based on wavelet decomposition," International Journal of Machine Tools & Manufacture, vol. 36, pp. 551-566, May 1996.
[21]Z. W. Huang, J. M. Zhu, J. T. Lei, X. R. Li, and F. Q. Tian, "Tool Wear Predicting Based on Multisensory Raw Signals Fusion by Reshaped Time Series Convolutional Neural Network in Manufacturing," Ieee Access, vol. 7, pp. 178640-178651, 2019.
[22]葉鎧瑋, "AFDM: 可支持分類模型推薦功能的自動化錯誤診斷機制," 碩士, 電機工程學系, 國立交通大學, 新竹市, 2019.
[23]B. N. Mohapatra and R. K. Mohapatra, "Performance Analysis on Frequency Response of Finite Impulse Response Filter," Procedia Computer Science, vol. 79, pp. 729-736, 2016/01/01/ 2016.
[24]Y. A. Grebenko and R. I. Polyak, "Choosing a FIR filter order for the linearization of an analog LPF phase response," in 2018Systems of Signal Synchronization, Generating and Processing in Telecommunications (SYNCHROINFO), 2018, pp. 1-3.
[25]X. Gao, X. Yan, P. Gao, X. Gao, and S. Zhang, "Automatic detection of epileptic seizure based on approximate entropy, recurrence quanti fi cation analysis and convolutional neural networks," Artif Intell Med, vol. 102, p. 101711, Jan 2020.
[26]J. Zhao, C. Yang, Z. Xue, Y. Leng, and Z. Wang, "A method for identifying the fault current of DC traction power supply system based on EMD approximate entropy," in 2017 International Conference on Green Energy and Applications (ICGEA), 2017, pp. 18-21.
[27]W. L. Chu, M. W. Huang, B. L. Jian, and K. S. Cheng, "Analysis of EEG entropy during visual evocation of emotion in schizophrenia," Ann Gen Psychiatry, vol. 16, p. 34, 12/01 2017.
[28]H. T. Yau, C. C. Wang, J. Y. Chang, and X. Y. Su, "A Study on the Application of Synchronized Chaotic Systems of Different Fractional Orders for Cutting Tool Wear Diagnosis and Identification," Ieee Access, vol. 7, pp. 15903-15911, 2019.
[29]T. Tarumi, G. W. Small, R. J. Combs, and R. T. Kroutil, "Infinite impulse response filters for direct analysis of interferogram data from airborne passive Fourier transform infrared spectrometry," Vibrational Spectroscopy, vol. 37, pp. 39-52, Jan 14 2005.
[30]L. Li, W. Chen, X. Shao, and Z. Wang, "Analysis of amplitude-integrated EEG in the newborn based on approximate entropy," IEEE Trans Biomed Eng, vol. 57, pp. 2459-66, Oct 2010.
[31]R. Hornero, D. Alvarez, D. Abasolo, F. del Campo, and C. Zamarron, "Utility of approximate entropy from overnight pulse oximetry data in the diagnosis of the obstructive sleep apnea syndrome," IEEE Trans Biomed Eng, vol. 54, pp. 107-13, Jan 2007.
[32]H. T. Yau, S. Y. Wu, C. L. Chen, and Y. C. Li, "Fractional-Order Chaotic Self-Synchronization-Based Tracking Faults Diagnosis of Ball Bearing Systems," Ieee Transactions on Industrial Electronics, vol. 63, pp. 3824-3833, Jun 2016.
[33]B. L. Jian, C. C. Wang, J. Y. Chang, X. Y. Su, and H. T. Yau, "Machine Tool Chatter Identification Based on Dynamic Errors of Different Self-Synchronized Chaotic Systems of Various Fractional Orders," Ieee Access, vol. 7, pp. 67278-67286, 2019.
[34]Y. C. Kuo, C. T. Hsieh, H. T. Yau, and Y. C. Li, "Research and Development of a Chaotic Signal Synchronization Error Dynamics-Based Ball Bearing Fault Diagnostor," Entropy, vol. 16, pp. 5358-5376, Oct 2014.
[35]S. K. Yang, C. L. Chien, and H. T. Yau, "Control of chaos in Lorenz system," Chaos Solitons & Fractals, vol. 13, pp. 767-780, Mar 2002.
[36]H. T. Yau, C. C. Wang, C. T. Hsieh, and S. Y. Wu, "Fractional order Sprott chaos synchronisation-based real-time extension power quality detection method," Iet Generation Transmission & Distribution, vol. 9, pp. 2775-2781, Dec 3 2015.
[37]C. J. Lin, X. Y. Su, K. T. Yu, B. L. Jian, and H. T. Yau, "Inspection on Ball Bearing Malfunction by Chen-Lee Chaos System," Ieee Access, vol. 8, pp. 28267-28275, 2020.
[38]J.-H. Chen, "Controlling chaos and chaotification in the Chen–Lee system by multiple time delays," Chaos, Solitons & Fractals, vol. 36, pp. 843-852, 2008/05/01/ 2008.
[39]R. Scherer, S. L. Kalla, Y. Tang, and J. Huang, "The Grünwald–Letnikov method for fractional differential equations," Computers & Mathematics with Applications, vol. 62, pp. 902-917, 2011/08/01/ 2011.
[40]G. Jumarie, "Modified Riemann-Liouville derivative and fractional Taylor series of nondifferentiable functions further results," Computers & Mathematics with Applications, vol. 51, pp. 1367-1376, May 2006.
[41]M. Heidari and H. Shamsi, "Analog programmable neuron and case study on VLSI implementation of Multi-Layer Perceptron (MLP)," Microelectronics Journal, vol. 84, pp. 36-47, Feb 2019.
[42]S.-H. Oh, "Error back-propagation algorithm for classification of imbalanced data," Neurocomputing, vol. 74, pp. 1058-1061, 2011/02/15/ 2011.
[43]X. Nie and W. X. Zheng, "Multistability and Instability of Neural Networks With Discontinuous Nonmonotonic Piecewise Linear Activation Functions," IEEE Trans Neural Netw Learn Syst, vol. 26, pp. 2901-13, Nov 2015.
[44]L. Ding, L. Xiao, K. Q. Zhou, Y. H. Lan, and Y. S. Zhang, "A New RNN Model With a Modified Nonlinear Activation Function Applied to Complex-Valued Linear Equations," Ieee Access, vol. 6, pp. 62954-62962, 2018.
[45]Y. Wang, H. W. Li, and X. W. Li, "A Case of On-Chip Memory Subsystem Design for Low-Power CNN Accelerators," Ieee Transactions on Computer-Aided Design of Integrated Circuits and Systems, vol. 37, pp. 1971-1984, Oct 2018.
[46]L. Koskinen, A. Paasio, and K. R. I. Halonen, "Motion estimation computational complexity reduction with CNN shape segmentation," Ieee Transactions on Circuits and Systems for Video Technology, vol. 15, pp. 771-777, Jun 2005.
[47]A. Krizhevsky, I. Sutskever, and G. E. Hinton, "ImageNet classification with deep convolutional neural networks," Communications of the ACM, vol. 60, pp. 84-90, 01/01 2017.
[48]C. Szegedy, W. Liu, Y. Jia, P. Sermanet, S. Reed, D. Anguelov, et al., Going deeper with convolutions, 2015.
[49]K. Simonyan and A. Zisserman, "Very Deep Convolutional Networks for Large-Scale Image Recognition," arXiv 1409.1556, 09/04 2014.
[50]J. C. Wang, J. Lin, and Z. F. Wang, "Efficient Hardware Architectures for Deep Convolutional Neural Network," Ieee Transactions on Circuits and Systems I-Regular Papers, vol. 65, pp. 1941-1953, Jun 2018.
[51]F. Y. Guo, R. F. He, and J. W. Dang, "Implicit Discourse Relation Recognition via a BiLSTM-CNN Architecture With Dynamic Chunk-Based Max Pooling," Ieee Access, vol. 7, pp. 169281-169292, 2019.
[52]S. H. Wang, Y. Y. Jiang, X. X. Hou, H. Cheng, and S. D. Dui, "Cerebral Micro-Bleed Detection Based on the Convolution Neural Network With Rank Based Average Pooling," Ieee Access, vol. 5, pp. 16576-16583, 2017.
[53]G. Wang, G. B. Giannakis, and J. Chen, "Learning ReLU Networks on Linearly Separable Data: Algorithm, Optimality, and Generalization," Ieee Transactions on Signal Processing, vol. 67, pp. 2357-2370, May 1 2019.
[54]L. L. Alhadeff, M. B. Marshall, D. T. Curtis, and T. Slatter, "Protocol for tool wear measurement in micro-milling," Wear, vol. 420, pp. 54-67, Feb 15 2019.