ii Acknowledgements I want to express my most sincere gratitude to my supervisors, Associate Professor Hong Geok Soon and Professor Wong Yoke San. I want to thank them for their motivation, support, and critique about the work. Their depth of knowledge, insight and untiring work ethic has been and will continue to be a source of inspiration to me. Thanks to Mr.Simon Tan, and all the technicians at Advanced Manufacturing Lab of NUS for their kind and quick technical assistance during my experiments. I have also benefitted from discussion with many of seniors and colleagues. In particular Dr.Kommisetti V R S Manyam, Dr. Yu Deping, Ms. Wang Qing, Dr. Wu Jiayun and others in the Control and Mechatronics Lab. I also would like to thank National University of Singapore for offering me research scholarship and research facilities. I benefitted from the abundant professional books and technical Journal collection at NUS library. Finally, I would like to devote the thesis to my family for their love and understanding. National University of Singapore NUS iii Table of Content Acknowledgements ii Table of Content . iii Summary vii List of Table ix List of Figures . xi List of Symbols . xiv Nomenclature . xv Chapter Introduction 1.1 Tool Condition Monitoring 1.2 Motivation and Challenges 1.3 Objectives 1.4 Organization of the Thesis . Chapter 2.1 Literature Review . TCM Review 2.1.1 Background 2.1.2 Overview of Tool Condition Monitoring 12 National University of Singapore NUS iv 2.2 Prognostic Methodologies for General Health Monitoring Review 15 2.2.1 Reliability Analysis . 16 2.2.2 Condition Monitoring Based Prognostic . 18 2.3 Review for TCM Prognostics and its Challenge 23 Chapter Long-term Prognostics . 29 3.1 Introduction 29 3.2 The Feasibility of Using VDHMM to Model the Machining Process . 32 3.2.1 HMM Based Prognostics . 34 3.2.2 Modeling the tool wear as stationary Markov process and Markov process with explicit state duration. 37 3.2.3 VDHMM parameter estimation . 42 3.2.4 Case study results and discussion 53 3.3 3.3.1 Adaptive VDHMM Fundamentals 66 3.3.2 Experimental setup 72 3.3.3 Results and discussion . 73 3.4 states Adaptive VDHMM based prognostics . 65 Adaptive-VDHMM based feature selections for different number of 81 National University of Singapore NUS v 3.4.1 Feature extraction methodology 83 3.4.2 Results and discussion . 87 3.5 Conclusion . 94 Chapter Short-term Prognostics . 99 4.1 Introduction 99 4.2 Theoretical background . 104 4.2.1 Delay Coordinate Embedding . 104 4.2.2 Function Approximation Methodologies . 105 4.2.3 Bayesian MLP for regression 108 4.3 Results and discussion . 109 4.3.1 Force prediction . 109 4.4 Combination of Short term and long term prognostics 118 4.5 Conclusions 121 Chapter Conclusions and Future Work 123 5.1 Conclusions and Contributions 123 5.2 Conclusions 123 5.3 Contributions 124 National University of Singapore NUS vi 5.4 Future work 126 5.4.1 Variable cutting condition study for long-term prognostics 126 5.4.2 Study the signal properties for short-term prognostics 127 References 128 National University of Singapore NUS vii Summary Tool condition monitoring (TCM) plays an important role in modern manufacturing system. At present, the researches in TCM focus mainly on diagnostics of different tool conditions. Such detection approaches may sometime be too late in avoiding damage or quality issues associated with the worn or broken tool. Hence, this thesis proposes a TCM prognostic system which aims to enable a future scheduling decision as well as optimal tool replacement time. The proposed TCM prognostic system consists of two parts: long term prognostics, and short term prognostics. In the long term prognostics, the remaining useful life (RUL) is prognosticated by utilizing variable duration Hidden Markov model (VDHMM). VDHMM overcomes the state duration limitation of the traditional HMM. The relation between the model structure and the tool wear process is studied to understand and address issues regarding the factors that affect the prognostic results. It is found that VDHMM with Gaussian distribution as the state duration offers effective prognosis for the machining conditions studied. Several features have been derived from the force signal captured during the machining process and identified to correlate with tool conditions. As appropriate selection of these features affects the prognostic results, a feature selection method is proposed. The method identifies and selects a sub-set of complementary and supportive features. The proposed method is compared with the feature ranking methods, which rank the features based on their relevance with tool wear progress. National University of Singapore NUS viii The result shows that features considered relevant to tool wear may worsen the prognostic results, while those considered not relevant to tool wear could improve the prognostic results, when the latter features complement and support other features. Arising from variations in material characteristics of the workpiece and tool, there might be unexpected or pre-matured tool wear occurring before the failure time expected from the long-term prognostics. Hence, a short term prognostics capturing the short term dynamics of the tool wear is proposed. This is achieved by adding a cutting force prediction part to a diagnostic system. Different cutting force prediction structures are analyzed. It is found that Sauer’s local linear model can achieve reasonable prediction accuracy and the shortest computation time. National University of Singapore NUS ix List of Table Table 3.1 Experimental conditions and equipment . 53 Table 3.2 The tool wear table 57 Table 3.3 The A-D test for the samples 60 Table 3.4 The residual lives take each state as the initial state of the life tests 61 Table 3.5 The natural mean residual life . 61 Table 3.6 Cutting conditions 73 Table 3.7 average model for the training sets 74 Table 3.8 the transformation matrix for the training sets 74 Table 3.9 The MAPE of MRL taking each state as the initial. 78 Table 3.10 The cross validation . 81 Table 3.11 Summary of features 87 Table 3.12 The feature sets selected by the proposed method 90 Table 3.13 feature selection by LDA . 91 Table 4.1 The smallest embedding dimension for all the test and their average 110 Table 4.2 Compare the ASIE and calculation time for Sauer’s local linear approach, Global linear approach and non-linear (MLP) approach . 112 National University of Singapore NUS x Table 4.3 Selected features 113 Table 4.4 Successful rate for each testing sets. . 115 Table 4.5 Two prognostic scenarios 118 Table_Apx A-1Cutting condition 142 National University of Singapore NUS xi List of Figures Figure 2.1 Tool wear definition . Figure 2.2 Tool wear measurement. 10 Figure 2.3 TCM process 13 Figure 3.1 The on-line prognostic process for a target test at time t. 37 Figure 3.2 a left-to-write Markov process . 39 Figure 3.3 A Markov process with the explicit state duration distribution . 40 Figure 3.4 the feature & and its GMM model for fresh, moderately worn and worn tool. 45 Figure 3.5 feature & 10 and its GMM model for fresh, moderately worn and worn tool . 45 Figure 3.6 The probability of features vectors belongs to fresh, moderately worn and worn tools. 46 Figure 3.7 different combinatiosn of k0 and λ0 and their corresponding loglikelihood. . 52 Figure 3.8 experimental setup 54 Figure 3.9 Geometry of a face milling operation 54 Figure 3.10 Cutting force sample for fresh cutting tool 55 National University of Singapore NUS References 131 [35] P. H. Fu, A. D.; King, G. A., “intelligent tool condition monitoring in milling operation,” Southampton Inst., (England). Systems Engineering Faculty., pp. 10p, 1998. [36] P. H. Fu, A. D.; King, G. A., “On- line tool condition monitoring based on a neurofuzzy intelligent signal feature classification procedure,” in international Conference on Soft Computing and Information/Intelligent System, pp. 183199, 2001. [37] E. Jantunen, “A summary of methods applied to tool condition monitoring in drilling,” International Journal of Machine Tools & Manufacture, vol. 42, no. Copyright 2002, IEE, pp. 997-1010, 2002. [38] D. Rodriguez-Salgado, I. Cambero-Rivero, and F. J. Alonso-Sanchez, “Feature selection for tool condition monitoring in turning processes,” Materials Science Forum, vol. 526, no. Copyright 2007, The Institution of Engineering and Technology, pp. 97-102, 2006. [39] K. Zhu, S. Wong Yoke, and S. Hong Geok, “Wavelet analysis of sensor signals for tool condition monitoring: a review and some new results,” International Journal of Machine Tools & Manufacture, vol. 49, no. Copyright 2009, pp. 537-53, 2009. [40] Y. Altintas, “in-process detection of tool breakage using time series monitoring of cutting forces,” International Journal of Machine Tools and Manufacture, vol. 28, no. Compendex, pp. 157-172, 1988. [41] S. A. Kumar, H. V. Ravindra, and Y. G. Srinivasa, “In-process tool wear monitoring through time series modelling and pattern recognition,” International Journal of Production Research, vol. 35, no. Copyright 1997, IEE, pp. 739-51, 1997. [42] S. Y. Liang, and D. A. Dornfeld, "detection of cutting tool wear using adaptive time series modeling of acoustic emission signal," American Society of Mechanical Engineers, Production Engineering Division (Publication) PED. pp. 27-38. [43] D. Jianfei, K. V. R. Subrahmanyam, W. Yoke San et al., “Bayesian-inferencebased neural networks for tool wear estimation,” International Journal of Advanced Manufacturing technology, vol. 30, no. Copyright 2007, pp. 797807, 2006. [44] N. E. Huang, Z. Shen, S. R. Long et al., The Empirical Mode Decomposition and the Hilbert Spectrum for Nonlinear and Non-stationary Time Series Analysis, Proc. R. Soc. Lond. A March 1998 vol. 454 no. 1971 pp. 903-995 1996. National University of Singapore NUS References 132 [45] G. G. Leisk, N. N. Hsu, and N. E. Huang, "Application of the Hilbert-Huang transform to machine tool condition/health monitoring." AIP Conf. Proc. 615, pp. 1711-1718, 2002. [46] Y. Peng, “Empirical Model Decomposition Based Time-Frequency Analysis for the Effective Detection of Tool Breakage,” Journal of Manufacturing Science and Engineering, vol. 128, no. 1, pp. 154-166, 2006. [47] D. E. Dimla, Jr., P. M. Lister, and N. J. Leighton, "Tool condition monitoring in metal cutting through application of MLP neural networks," IEE Colloquium on Fault Diagnosis in Process Systems (Digest No.1997/174). pp. 9-1, 2002. [48] D. E. Dimla, Jr., "A modular tool wear monitoring system in a metal cutting operation using MLP neural networks and multivariate process parameters," UKACC International Conference on CONTROL '98 (IEE Conf. Publ. No.455). pp. 296-301, 2002. [49] R. Srinidhi, K. T. Veeramanju, M. Sukumar et al., "A multi layer perceptron (MLP) based on back propagation technique for flank wear analysis of single point cutting tool inserts," Proceedings of the 2007 International Conference on Artificial Intelligence. ICAI 2007. pp. 247-52, 2007. [50] N. Fang, P. S. Pai, and N. Edwards, “Prediction of built-up edge formation in machining with round edge and sharp tools using a neural network approach,” International Journal of Computer Integrated Manufacturing, vol. 23, no. Compendex, pp. 1002-1014, 2010. [51] J. Sun, G. S. Hong, M. Rahman et al., "The application of nonstandard support vector machine in tool condition monitoring system," Proceedings, DELTA 2004 - Second IEEE International Workshop on Electronic Design, Test and Applications. pp. 295-300, 2004. [52] C. Scheffer, and P. S. Heyns, "Monitoring of turning tool wear using vibration measurements and neural network classification," Proceedings of the 25th International Conference on Noise and Vibration Engineering, ISMA. pp. 921-928, 2000. [53] K. Jing, K. Ni, F. Chang-jian et al., "Tool wear monitoring and failure prediction based on hybrid SOM-DHMM architecture," CIS Workshops 2007. International Conference on Computational Intelligence and Security Workshops. pp. 886-9, 2007. [54] P. Sartain, A. B. T. Hopkins, K. D. McDonald-Mair et al., "A framework for self-diagnosis and condition monitoring for embedded systems using a SOM- National University of Singapore NUS References 133 based classifier," 2008 NASA/ESA Conference on Adaptive Hardware and Systems Conference. pp. 417-23, 2008. [55] A. K. S. Jardine, D. Lin, and D. Banjevic, “A review on machinery diagnostics and prognostics implementing condition-based maintenance,” Mechanical Systems and Signal Processing, vol. 20, no. 7, pp. 1483-1510, 2006. [56] E. A.Elsayed, Reliability Engineering: Addison Wesley Longman, 1996. [57] W. R. Blischke, and D. N. P. Murthy, Reliability: Modeling, Prediction, and Optimization: John Wiley &Sons, 2000. [58] F. k. J.L.Bogdanoff, Probabilistic Models of Cumulative Damage, USA: John Wiley & Sons, 1985. [59] Y. Wu, G. S. Hong, and Y. S. Wong, "HMM with Explicit State Duration for Prognostics in Face Milling," 2010 IEEE Conference on Robotics, Automation and Mechatronics (RAM 2010). pp. 218-23, 2010. [60] S. Z. Yu, “Hidden semi-Markov models,” Artificial Intelligence, vol. 174, no. 2, pp. 215-243, Feb, 2010. [61] D. Banjevic, and A. K. S. Jardine, “Calculation of reliability function and remaining useful life for a Markov failure time process,” IMA J Management Math, vol. 17, no. 2, pp. 115-130, April 1, 2006, 2006. [62] T. Rosqvist, “Bayesian aggregation of experts' judgements on failure intensity,” Reliability Engineering & System Safety, vol. 70, no. 3, pp. 283289, Dec, 2000. [63] J. M. Vannoortwijk, R. M. Cooke, and M. Kok, “A Bayesian Failure ModelBased on Isotropic Deterioration,” European Journal of Operational Research, vol. 82, no. 2, pp. 270-282, Apr 20, 1995. [64] S. Yong, M. Lin, and J. Mathew, “Prediction of system reliability for single component repair,” Journal of Quality in Maintenance Engineering, vol. 13, no. Copyright 2007, The Institution of Engineering and Technology, pp. 111124, 2007. [65] M. M. K. Modarres, Vasiliy Krivtsov Reliability Engineering and Risk Analysis: NY: Marcel Decker, Inc, 1999. [66] R. E. Barlow, j. B. Fussell, and M. D.Singpurwalla, Reliability and Fault Tree Analysis: Theoretical and Applied Aspects of System Reliability and Safety Assessment. Society for Industrial and Applied Mathematics, 1975. National University of Singapore NUS References 134 [67] K. M. Goh, B. Tjahjono, T. Baines et al., "A review of research in manufacturing prognostics," 2006 4th IEEE International Conference on Industrial Informatics. pp. 417-422, 2006. [68] eger, J.-B., Iung, B., "Ships fleet-wide management and naval mission prognostics: Lessons learned and new issues", Prognostics and Health Management (PHM), 2012 IEEE Conference on, pp: - [69] Y. Li, S. Billington, C. Zhang et al., “Adaptive prognostics for rolling element bearing condition,” Mechanical Systems and Signal Processing, vol. 13, no. 1, pp. 103-113, 1999. [70] S. J. Engel, B. J. Gilmartin, K. Bongort et al., "Prognostics, the real issues involved with predicting life remaining," IEEE Aerospace Conference Proceedings. Vol.6, pp. 457-470, 2000. [71] A. Ray, and S. Tangirala, “Stochastic modeling of fatigue crack dynamics for on-line failure prognostics,” Control Systems Technology, IEEE Transactions on, vol. 4, no. 4, pp. 443-451, 1996. [72] M. J. Roemer, E. O. Nwadiogbu, and G. Bloor, "Development of diagnostic and prognostic technologies for aerospace health management applications." pp. 3139-3147 vol.6, 2002. [73] D. Chelidze, "Multimode damage tracking and failure prognosis in electromechanical systems." Proc. SPIE 4733, Component and Systems Diagnostics, Prognostics, and Health Management II, pp. 1-12, July 15, 2002. [74] G. J. Kacprzynski, M. J. Roemer, G. Modgil et al., "Enhancement of physicsof-failure prognostic models with system level features." Aerospace Conference Proceedings, IEEE, vol.6, pp. 6-2919-6-2925, 2002. [75] C. H. Oppenheimer, and K. A. Loparo, "Physically based diagnosis and prognosis of cracked rotor shafts." Proc. SPIE 4733, Component and Systems Diagnostics, Prognostics, and Health Management II, 122 pp. 122-132, July 15, 2002. [76] J. Qiu, B. B. Seth, S. Y. Liang et al., “DAMAGE MECHANICS APPROACH FOR BEARING LIFETIME PROGNOSTICS,” Mechanical Systems and Signal Processing, vol. 16, no. 5, pp. 817-829, 2002. [77] A. B. J.Luo, K. Pattipati, L. Qiao, M. Kawamoto, S. Chigusa, “An interacting multiple odel approach to model-based prognostics,” System Security and Assurance, vol. 1, pp. 189-194, 2003. National University of Singapore NUS References 135 [78] D. W. Brown, P. W. Kalgren, C. S. Byington et al., “Electronic prognostics A case study using global positioning system (GPS),” Microelectronics Reliability, vol. 47, no. 12, pp. 1874-1881, 2007. [79] S. E. Franklin, and J. A. Dijkman, “The Implementation of Tribological Principles in an Expert-System (Precept) for the Selection of Metallic Materials, Surface Treatments and Coatings in Engineering Design,” Wear, vol. 181, pp. 1-10, Feb, 1995. [80] F. Francois, E. Escande, and J. Bigeon, “A Methodology for Knowledge Representation in Expert-System, in the Field of Electrical-Engineering,” Ieee Transactions on Magnetics, vol. 30, no. 5, pp. 3656-3659, Sep, 1994. [81] A. K. Garga, K. T. McClintic, R. L. Campbell et al., “Hybrid reasoning for prognostic learning in CBM systems,” 2001 IEEE Aerospace Conference Proceedings, Vols 1-7, pp. 2957-2969 , 2001. [82] T. I. Liu, J. H. Singonahalli, and N. R. Iyer, “Detection of roller bearing defects using expert system and fuzzy logic,” Mechanical Systems and Signal Processing, vol. 10, no. 5, pp. 595-614, 1996. [83] C. K. Mechefske, “Objective machinery fault diagnosis using fuzzy logic,” Mechanical Systems and Signal Processing, vol. 12, no. 6, pp. 855-862, 1998. [84] L. P. Heck, and J. H. McClellan, "Mechanical system monitoring using hidden Markov models." Acoustics, Speech, and Signal Processing, 1991. ICASSP91., 1991 International Conference pp. 1697-1700 vol.3, 1991. [85] C. J. Leggetter, and P. C. Woodland, “Maximum likelihood linear regression for speaker adaptation of continuous density hidden Markov models,” Computer Speech and Language, vol. 9, no. Copyright 1995, IEE, pp. 171185, 1995. [86] H. M. Ertunc, K. A. Loparo, and H. Ocak, “Tool wear condition monitoring in drilling operations using hidden Markov models (HMMs),” International Journal of Machine Tools and Manufacture, vol. 41, no. 9, pp. 1363-1384, 2001. [87] C. Kwan, X. Zhang, R. Xu et al., "A novel approach to fault diagnostics and prognostics.", Robotics and Automation, 2003. Proceedings. ICRA '03. IEEE International Conference, vol.1, pp. 604-609, 2003. [88] C. Scheffer, H. Engelbrecht, and P. Heyns, “A comparative evaluation of neural networks and hidden Markov models for monitoring turning tool wear,” Neural Computing & Applications, vol. 14, no. 4, pp. 325-336, 2005. National University of Singapore NUS References 136 [89] Z. Xiaodong, R. Xu, K. Chiman et al., "An integrated approach to bearing fault diagnostics and prognostics." , American Control Conference, 2005. Proceedings of the 2005, vol. 4, pp. 2750-2755, 2005. [90] Z. Kunpeng, “multi-categories tool wear classification In micro-milling,” PhD Thesis, Department Of Mechanical Engineering, Nus, Singapore, 2007. [91] J. H. Yan, M. Koc, and J. Lee, “A prognostic algorithm for machine performance assessment and its application,” Production Planning & Control, vol. 15, no. 8, pp. 796-801, Dec, 2004. [92] V. Leiva, M. Riquelme, N. Balakrishnan et al., “Lifetime. analysis based on the generalized Birnbaum-Saunders distribution,” Computational Statistics & Data Analysis, vol. 52, no. 4, pp. 2079-2097, Jan 10, 2008. [93] L. L. Ku, and T. C. Huang, “Sequential monitoring of manufacturing processes: an application of grey forecasting models,” International Journal of Advanced Manufacturing technology, vol. 27, no. 5-6, pp. 543-546, Jan, 2006. [94] T. Sauer, “Time series prediction by using delay coordinate embedding (data set A),” Time Series Prediction: Forecasting the Future and Understanding the Past, volume XV of SFI Studies in the Sciences of Complexity, pp. 175-175, 1994. [95] R. C. M. Yam, P. W. Tse, L. Li et al., “Intelligent predictive decision support system for condition-based maintenance,” International Journal of Advanced Manufacturing technology, vol. 17, no. Copyright 2001, IEE, pp. 383-391, 2001. [96] W. Q. Wang, M. F. Golnaraghi, and F. Ismail, “Prognosis of machine health condition using neuro-fuzzy systems,” Mechanical Systems and Signal Processing, vol. 18, no. Compendex, pp. 813-831, 2004. [97] C. Mian-Yun, "Uncertainty analysis and grey modeling" Proceedings. First International Symposium on Uncertainty Modeling and Analysis (Cat. No.90TH0334-3). pp. 469-473, 1990. [98] T. H, and T.Murata, “Basic investigations of tool wear,” Trans. ASME J Eng.Ind., no. 85, 1963. [99] A. Molinari, and M. Nouari, “Modeling of tool wear by diffusion in metal cutting,” Wear, vol. 252, no. 1-2, pp. 135-149, Jan, 2002. [100] E. Usui, T. Shirakashi, and T. Kitagawa, “analytical prediction of three dimensional cutting process,” American Society of Mechanical Engineers (Paper), no. Compendex, Vol.100, Issue.2, 1977. National University of Singapore NUS References 137 [101] T. Kitagawa, K. Maekawa, T. Shirakashi et al., “Analytical prediction of flank wear of carbide tools in turning plain carbon steels (part 1) - characteristics Equ. of flank wear,” Bulletin of the Japan Society of Precision Engineering, vol. 22, no. Compendex, pp. 263-269, 1988. [102] K. Maekawa, T. Kitagawa, T. Shirakashi et al., “Analytical prediction of flank wear of carbide tools in turning plain carbon steel. Part 2. Prediction of flank wear,” Bulletin of the Japan Society of Precision Engineering, vol. 23, no. Compendex, pp. 126-133, 1989. [103] H. G. Lee, and D. G. Lee, “Tool life model for abrasive wet micro-drilling of ceramic green bodies,” International Journal of Machine Tools and Manufacture, vol. 44, no. 7-8, pp. 839-846, 2004. [104] X. Luo, K. Cheng, R. Holt et al., “Modeling flank wear of carbide tool insert in metal cutting,” Wear, vol. 259, no. Copyright 2006, IEE, pp. 1235-40, 2005. [105] S. E. Oraby, and D. R. Hayhurst, “Tool life determination based on the measurement of wear and tool force ratio variation,” International Journal of Machine Tools & Manufacture, vol. 44, no. 12-13, pp. 1261-1269, Oct, 2004. [106] P. G. Maropoulos, and B. Alamin, “Integrated tool life prediction and management for an intelligent tool selection system,” Journal of Materials Processing Technology, vol. 61, no. Compendex, pp. 225-230, 1996. [107] C. Ping Yi, and H. Yeong Dong, “An improved neural network model for the prediction of cutting tool life,” Journal of Intelligent Manufacturing, vol. 8, no. Copyright 1997, IEE, pp. 107-15, 1997. [108] H. Wiklund, “Bayesian and regression approaches to on-line prediction of residual tool life,” Quality and Reliability Engineering International, vol. 14, no. Copyright 1999, IEE, pp. 303-9, 1998. [109] S. E. Oraby, and D. R. Hayhurst, “Tool life determination based on the measurement of wear and tool force ratio variation,” International Journal of Machine Tools & Manufacture, vol. 44, no. Copyright 2004, IEE, pp. 1261-9, 2004. [110] Y. Ao, and B. Wang, “Monitoring of drill process and prediction for remaining useful life of drill tool,” Jixie Gongcheng Xuebao/Journal of Mechanical Engineering, vol. 47, no. Compendex, pp. 177-181, 2011. [111] D. Songlin, R. R. A. Izamshah, J. Mo et al., “Online Tool Life Prediction in the Machining of Titanium Alloys,” Key Engineering Materials, vol. 458, no. Copyright 2011, pp. 355-61, 2011. National University of Singapore NUS References 138 [112] A. G. Rehorn, J. Jiang, and P. E. Orban, “State-of-the-art methods and results in tool condition monitoring: a review,” The International Journal of Advanced Manufacturing Technology, vol. 26, no. 7, pp. 693-710, 2005. [113] L. R. Rabiner, “A tutorial on hidden Markov models and selected applications in speech recognition,” Proceedings of the IEEE, vol. 77, no. 2, pp. 257-286, 1989. [114] A. Ljolje, and S. E. Levinson, “Development of an acoustic-phonetic hidden Markov model for continuous speech recognition,” Signal Processing, IEEE Transactions on, vol. 39, no. 1, pp. 29-39, 1991. [115] C. Mou-Yen, K. Amlan, and Z. Jian, “Off-line handwritten word recognition using a hidden Markov model type stochastic network,” Pattern Analysis and Machine Intelligence, IEEE Transactions on, vol. 16, no. 5, pp. 481-496, 1994. [116] I. T.Nabney, Netlab Algrithms for Pattern Recognition-(Advances in pattern recognition): Springer, 2003. [117] S. E. Levinson, “Continuously variable duration hidden Markov models for automatic speech recognition,” Computer Speech and Language, vol. 1, no. Copyright 1987, IEE, pp. 29-45, 1986. [118] Stephens ， M.A., “ EDF Statistics for Goodness of Fit and some Comparisons,” Journal of the American Statistical Association, vol. 69, pp. 730-737, 1974. [119] J. Yamagishi, K. Ogata, Y. Nakano et al., "HSMM-based model adaptation algorithms for average-voice-based speech synthesis," 2006 IEEE International Conference on Acoustics, Speech, and Signal Processing (IEEE Cat. No. 06CH37812C). Vol.1, pp. 77-80, 2006. [120] A. A. Houshmand, and E. Kannateyasibu, “Statistical Process-Control of Acoustic-Emission for Cutting-Tool Monitoring,” Mechanical Systems and Signal Processing, vol. 3, no. 4, pp. 405-424, Oct, 1989. [121] J. Sun, G. S. Hong, M. Rahman et al., “Identification of feature set for effective tool condition monitoring by acoustic emission sensing,” International Journal of Production Research, vol. 42, no. 5, pp. 901-918, Mar 1, 2004. [122] Y. S. Tarng, “Study of Milling Cutting Force Pulsation Applied to the Detection of Tool Breakage,” International Journal of Machine Tools & Manufacture, vol. 30, no. 4, pp. 651-660, 1990. National University of Singapore NUS References 139 [123] Y. S. Tarng, Y. W. Hseih, and S. T. Hwang, “Sensing Tool Breakage in Face Milling with a Neural-Network,” International Journal of Machine Tools & Manufacture, vol. 34, no. 3, pp. 341-350, Apr, 1994. [124] Elbestawi M. A, M. J, and P. T. A., “Process monitoring in milling by pattern recognition,” Mechanical system and signal processing, vol. 3, no. 3, pp. 305315, 1989. [125] D. Lin, and V. Makis, “Recursive filters for a partially observable system subject to random failure,” Advances in Applied Probability, vol. 35, no. Copyright 2003, IEE, pp. 207-27, 2003. [126] N. H. Packard, J. P. Crutchfield, J. D. Farmer et al., “Geometry from a time series,” Physical Review Letters, vol. 45, no. Copyright 1980, IEE, pp. 712-16, 1980. [127] J. Zhang, H. Shu-Hung Chung, and L. Wai-Lun, “Chaotic time series prediction using a neuro-fuzzy system with time-delay coordinates,” IEEE Transactions on Knowledge and Data Engineering, vol. 20, no. Copyright 2008, The Institution of Engineering and Technology, pp. 956-64, 2008. [128] M. B. Kennel, R. Brown, and H. D. I. Abarbanel, “Determining embedding dimension for phase-space reconstruction using a geometrical construction,” Physical Review A (Statistical Physics, Plasmas, Fluids, and Related Interdisciplinary Topics), vol. 45, no. Copyright 1992, IEE, pp. 3403-11, 1992. [129] M. Casdagli, and A. S. Weigend, “Exploring the continuum between deterministic and stochastic modeling (data sets A,B,D),” Time Series Prediction: Forecasting the Future and Understanding the Past, volume XV of SFI Studies in the Sciences of Complexity, pp. 347-347, 1994. [130] Y. Shao, and K. Nezu, “Prognosis of remaining bearing life using neural networks,” Proceedings of the Institution of Mechanical Engineers Part IJournal of Systems and Control Engineering, vol. 214, no. 13, pp. 217-230, 2000. [131] G. V. P. Wang, “Fault prognostics using dynamic wavelet neural networks,” AI EDAM-Atificial Intelligence for Engineering Design Analysis and Manufacturing, vol. 15, pp. 383-391, 2001. [132] P. Wang, and G. Vachtsevanos, “Fault prognostics using dynamic wavelet neural networks,” Ai Edam-Artificial Intelligence for Engineering Design Analysis and Manufacturing, vol. 15, no. 4, pp. 349-365, Sep, 2001. [133] P. B. R.B. Chinnam, “A neuro-fuzzy approach for estimating mean residual life in condition-based maintenance systems,” International Journal of Materials and Product Technology, vol. 20, pp. 166-179, 2004. National University of Singapore NUS References 140 [134] K. Jemielniak, M. Szafarczyk, and J. Zawistowski, “Difficulties in Tool Life Predicting when Turning with Variable Cutting Parameters,” CIRP Annals Manufacturing Technology, vol. 34, no. 1, pp. 113-116, 1985. National University of Singapore NUS Appendix A Appendix A. 141 Ball nose milling experimental set up Six cutters are used in the ball nose milling experiment: 07BX1, 31PN4, 33PN6, 34PT1, 18SC3, 09BX3 The cutters are different from one another by its cutter geometry and coating but they are all 6mm Alignment-Tool carbide ball-nose end with three flutes. For all the cutting processes, Inconel 718, which is used in Jet engines, is used as the work-piece material. It is a precipitation-hardenable nickel-chromium alloy and is considered as a hard-to-cut material. Its chemical composition is 53% Ni, 19% Cr, 18% Fe, 5% Nb, 3% Mo, 0.9% Ti, 0.5% Al and C balance. Figure_Apx A-1 shows the machining process. As shown in Figure_Apx A-1, during the cutting process, the upper face of the material is cut with horizontal lines from the top edge to the bottom edge. After 320 cutting times, the cutter will start again at the top edge of the material for another similar 320 cuts. The cutting face is 112.5 mm wide and 40 mm high. So one cutter in the whole process will travel (0.1125m x 640 cuts =) 72m-long distance. National University of Singapore NUS Appendix A 142 Figure_Apx A-1The machining process. The running conditions throughout the experiment with six cutters are set up according to Table_Apx A-1: the spindle speed of 10360 rounds per minute (RPM) – which is considered as high speed. The cuts generated are 0.125mm wide and 0.25mm deep. So the time taken for one cut is around seconds and the whole process of 640 cuts for each cutter takes around 2560 seconds or near 43 minutes. Table_Apx A-1Cutting condition Spindle speed 10360RPM Feed rate 1.555m/min Depth of cut 0.25mm Three sensors are used to collect different kinds of data during the cutting process: accelerometer for vibration signal, acoustic emissions (AE) sensor for AE signal, and Dynamometer for force signal. National University of Singapore NUS Appendix A 143 The data acquisition is collected by a PCI 1200 board with the sampling rate of 50 KHz per channel, which means 50,000 data points can be recorded each second from each sensor. The raw cutting force data is illustrated in Figure_Apx A- and Figure_Apx A-. fresh tool worn tool 120 12 100 cutting force /N cutting force /N 10 80 60 40 20 0 -2 1.01 1.02 1.03 1.04 1.05 1.06 1281.1 1281.11 time /s 1281.12 1281.13 1281.14 1281.15 time /s Figure_Apx A-2 Cutting force sample for fresh cutting tool Figure_Apx A-3 Cutting force sample for worn cutting tool Figure_Apx A-4 shows the setup of the experiments mentioned above with all the components and their positions. Figure_Apx A-4 Experiment setup National University of Singapore NUS Appendix A 144 After a period of cutting, the cutter begins to wear out and its condition becomes worse than before as in Figure_Apx A-. a) Location of measurement in original cutter b) Original After 64 cuts After 192 cuts After 320 cuts Figure_Apx A-5 Tool wear of a cutter after a period of cutting Therefore, the tool wear of each cutter is measured after a period of cutting for a number of times using Olympic microscope. National University of Singapore NUS Appendix B 145 Appendix B. The cross validation results in Section 3.3 Testing sets T1 T2 T4 T10 T12 mean standard diviation 28.0% 25.1% 24.4% 40.1% 24.4% 28.4% 6.7% T1 T2 T4 T7 T12 28.0% 25.1% 24.4% 1.5% 24.4% 20.7% 10.8% T1 T4 T7 T10 T12 28.0% 37.0% 1.5% 40.1% 24.4% 26.2% 15.2% T1 T2 T3 T4 T12 28.0% 25.1% 27.5% 24.4% 24.4% 25.9% 1.7% T1 T3 T4 T10 T12 28.0% 27.5% 24.4% 40.1% 24.4% 28.9% 6.5% T1 T3 T4 T7 T12 28.0% 27.5% 24.4% 1.5% 24.4% 21.2% 11.1% T1 T2 T4 T6 T12 28.0% 25.1% 24.4% 10.1% 24.4% 22.4% 7.0% T1 T4 T6 T10 T12 Training sets T5 T5 T5 T5 T5 T5 T5 T8 T8 T8 T8 T8 T8 T8 T9 T9 T9 T9 T9 T9 T9 T11 T11 T11 T11 T11 T11 T11 T6 T6 T6 T6 T6 T6 T3 T3 T3 T3 T7 T7 T10 T7 National University of Singapore T7 T10 T2 T10 T2 T2 T10 NUS Appendix B T5 T5 T5 T5 T5 T5 T5 T5 T8 T8 T8 T8 T8 T8 T8 T8 146 T9 T9 T9 T9 T9 T9 T9 T9 T11 T11 T11 T6 T6 T6 T6 T3 T3 T3 T7 T3 T3 T3 T7 T7 T7 T10 T10 T7 T7 T10 T10 T10 National University of Singapore T2 T2 T2 T10 T2 T2 T2 T2 28.0% 24.4% 10.1% 40.1% 24.4% T1 T4 T6 T7 T12 28.0% 37.0% 29.0% 9.0% 44.0% T1 T3 T4 T6 T12 28.0% 27.5% 24.4% 10.1% 24.4% T1 T2 T4 T11 T12 73.5% 22.1% 47.5% 35.5% 27.1% T1 T4 T10 T11 T12 73.5% 47.5% 21.3% 35.5% 27.1% T1 T4 T7 T11 T12 73.5% 47.5% 7.4% 35.5% 27.1% T1 T3 T4 T11 T12 73.5% 74.8% 47.5% 35.5% 27.1% T1 T4 T6 T11 T12 73.5% 47.5% 59.6% 35.5% 27.1% NUS 25.4% 10.7% 29.4% 13.1% 22.9% 7.4% 41.1% 20.5% 41.0% 20.7% 38.2% 24.5% 51.7% 21.8% 48.6% 18.5% [...]... technology, tool condition monitoring systems based on suitable intelligent sensors can be developed, such as those for cutting force or torque monitoring, vibration monitoring, spindle/drive power or current monitoring, acoustic emission monitoring, and temperature monitoring Most of the existing on- line TCM systems tend to focus on the acquisition of relevant tool wear degradation information and the... long- term and short -term The long- term prognostics determine the remaining useful tool life as a random variable, and could contribute to the scheduling of future machining requirements The short -term prognostics target at tool failure occurring before the expected tool life obtained by long- term prognostics, with emphasis on the damage avoidance and optimal tool replacement time Therefore, the objectives... the cutting edge region This weakening causes plastic deformation under normally applied load Once this deformation occurs, the blunt edge causes additional heating due to rubbing Hence, weaken the tool material further and eventually leading to tool failure 2. 1 .2 Overview of Tool Condition Monitoring Tool condition monitoring (TCM) aims at identifying and predicting the cutting tool state, by apply... considered as long- term prognostics, as the prediction horizon is from the present time until the tool failure The prognostic of POF has a shorter prediction horizon, and so can be considered as short -term prognostics The challenge for short -term prognostics lies in National University of Singapore NUS Chapter 1 Introduction 5 the computation time The tool life is relative short compared with the machine... reliability analysis is not accurate enough for a special aging process 2. 2 .2 Condition Monitoring Based Prognostic As condition monitoring (CM) techniques were developed, prognostics began to involve CM information In the literature, a CM -based prognostic system is dependent on the following critical factors[67]: Current health situation Historical health situation Past maintenance history Expected usage... analysis, and condition monitoring based prognostics Finally, the prognostic methodologies for TCM applications are discussed Chapter 3 presents the establishment of a long- term prognostic system, which aims at prognosticating the RUL of the cutting tool Firstly, the modeling of the tool wear process as a non-stationary Markov Chain is discussed A VDHMM is established for the prognostic of the RUL The prognostic. .. prognostic information In this section, both reliability National University of Singapore NUS Chapter 2 Literature Review 16 analysis and condition monitoring based prognostic methodologies for general health monitoring are reviewed 2. 2.1 Reliability Analysis The reliability analysis for the machining process involves modeling the failure times using different probability distribution functions and considering... Therefore, if the prognostic algorithm is quite complex, the computation time might exceed the prediction horizon In this case, the prognostic is not achieved, regardless of the accuracy of the result 1.3 Objectives The purpose of a tool condition monitoring system (TCM) is primarily to provide tool condition information for making decision on tool replacement It can also enable optimal decision to... Figure 3 .21 Prognostic results for T4 77 Figure 3 .22 Prognostic results for T6 77 Figure 3 .23 Prognostic results for T7 78 Figure 3 .24 prognostic results for T 12 78 National University of Singapore NUS xiii Figure 3 .25 select the training-testing sets for cross validation 80 Figure 3 .26 The proposed feature selection process 89 Figure 3 .27 The MAPE for different... Establish a prognostic system capable of long- term and short -term prognostics National University of Singapore NUS Chapter 1 Introduction  Verify the prognostic results by case studies  6 Analyze the factors which affect the prognostic results The long- term prognostics are achieved by a variable duration Hidden Markov model (VDHMM) and adaptive-VDHMM This VDHMM is an update of conventional HMM by . Reliability Analysis 16 2. 2 .2 Condition Monitoring Based Prognostic 18 2. 3 Review for TCM Prognostics and its Challenge 23 Chapter 3 Long- term Prognostics 29 3.1 Introduction 29 3 .2 The Feasibility. regression 108 4.3 Results and discussion 109 4.3.1 Force prediction 109 4.4 Combination of Short term and long term prognostics 118 4.5 Conclusions 121 Chapter 5 Conclusions and Future Work 123 . Conclusions and Contributions 123 5 .2 Conclusions 123 5.3 Contributions 124 vi National University of Singapore NUS 5.4 Future work 126 5.4.1 Variable cutting condition study for long- term