Smart Machining Systems: Modelling, Monitoring and Informatics (Springer Series in Advanced Manufacturing)

✍ Scribed by Kunpeng Zhu

• Publisher: Springer
• Tongue: English
• Leaves: 420
• Category: Library

✦ Synopsis

This book provides the tools to enhance the precision, automation and intelligence of modern CNC machining systems. Based on a detailed description of the technical foundations of the machining monitoring system, it develops the general idea of design and implementation of smart machining monitoring systems, focusing on the tool condition monitoring system.

The book is structured in two parts. Part I discusses the fundamentals of machining systems, including modeling of machining processes, mathematical basics of condition monitoring and the framework of TCM from a machine learning perspective. Part II is then focused on the applications of these theories. It explains sensory signal processing and feature extraction, as well as the cyber-physical system of the smart machining system.

Its utilisation of numerous illustrations and diagrams explain the ideas presented in a clear way, making this book a valuable reference for researchers, graduate students and engineers alike.

✦ Table of Contents

Preface
The Motivation of This Book
The Content of This Book
Acknowledgement
Contents
1 Introduction to the Smart Machining System
1.1 The Development of Modern Manufacturing System
1.2 Modern Machining Technology
1.2.1 High Precision Machining
1.2.2 High Speed Machining
1.2.3 Green Machining
1.2.4 Smart Machining
1.3 The Smart Machining System
1.3.1 Intelligent Process Planning
1.3.2 The Process Simulation and Optimization
1.3.3 The Machining Process Monitoring
1.3.4 The Intelligent Control
1.3.5 The Database and Big Data Analytics
1.3.6 Smart Machine Tool
1.4 The Trends of Smart Machining System
References
2 Modeling of the Machining Process
2.1 The Machining Process Modeling Methods
2.1.1 Modeling Based on Cutting Mechanics
2.1.2 Modeling Based on Machine Tool Vibration
2.1.3 Modeling Based on Numerical Simulation
2.1.4 Modeling Based on Measurement Information
2.1.5 Modeling Based on Artificial Intelligence (AI)
2.1.6 Modeling Method Combining Data and Cutting Mechanics
2.2 Principles of Chip Formation
2.2.1 Chip Formation
2.2.2 Mechanical Model of Chip Formation
2.2.3 Divisions of Deformation Zones
2.3 Cutting Forces
2.3.1 Sources of Cutting Forces
2.3.2 Joint and Component Cutting Forces and Cutting Powers
2.3.3 Empirical Models of Cutting Forces
2.3.4 Affecting Factors of Cutting Forces
2.4 Cutting Heat and Temperatures
2.4.1 Generation and Transfer of Cutting Heat
2.4.2 Cutting Temperatures and Their Distributions
2.4.3 Modeling of Temperature Fields
2.5 Milling Process Modeling and Control
2.5.1 Types of Milling Cutters
2.5.2 Milling Types
2.5.3 Milling Parameters and Cutting Layer Parameters
2.5.4 Milling Forces
2.5.5 The Milling System Dynamics
2.6 High-Speed Machining
2.6.1 Introduction to High-Speed Machining
2.6.2 Advantages of High-Speed Machining
2.6.3 Modeling of the Three-Dimensional Instantaneous Milling Force
2.7 Control of Machining Process
References
3 Tool Wear and Modeling
3.1 Types of Tool Wear
3.1.1 Crater Wear
3.1.2 Flank Wear
3.1.3 Boundary Wear
3.1.4 Tool Wear Criteria
3.2 The Formation of Tool Wear
3.2.1 Mechanical Wear
3.2.2 Adhesive Wear
3.2.3 Diffusion Wear
3.2.4 Chemical Wear
3.2.5 Thermoelectric Wear
3.3 Tool Usability and Its Relationship with Cutting Parameters
3.3.1 Tool Life
3.3.2 Tool Life Equation
3.3.3 Tool Breakage
3.4 Modeling of Tool Wear
3.4.1 Abrasive Wear Rate Model
3.4.2 Adhesive Wear Rate Model
3.4.3 Diffusion Wear Rate Model
3.4.4 Comprehensive Wear Rate Model
3.4.5 Intelligent Tool Wear Model
3.5 Tool Wear Modeling in High-Speed Milling
3.5.1 Tool Flank Wear Conditions
3.5.2 Modeling of Tool Flank Wear
3.5.3 Generalization of the Tool Wear Model
3.5.4 Analysis of Tool Wear Model
References
4 Mathematical Foundations of Machining System Monitoring
4.1 Machining System Monitoring
4.1.1 The Content of Machining System Monitoring
4.1.2 The System of Machining Process Monitoring
4.2 The Content of the Machining Process Monitoring System
4.2.1 Signal Detection
4.2.2 Feature Extraction
4.2.3 State Recognition
4.2.4 Decision-Making and Control
4.3 The Methods of Machining Process Monitoring
4.3.1 Introduction
4.3.2 Stochastic Process Based Methods
4.4 Parameter Estimation Methods
4.4.1 Least Square Estimation
4.4.2 Yule-Walker Estimation
4.4.3 Maximum Likelihood Estimate
4.5 Time Series Analysis in Condition Monitoring
4.5.1 The Auto-Regression Model AR(N)
4.5.2 The Auto Regression Moving Average Model ARMA(n, m)
4.6 The Machining State Description
4.6.1 Typical Anomaly State of the Machining Process
4.6.2 Process Model Based State Feature Extraction
4.7 Identification of Machining Process
4.7.1 Overview of Process Modeling
4.7.2 Model of Machining Process and Identification Method
4.7.3 The Time Series Identification of the Machining State
4.7.4 Identification of the Cutting Force
4.7.5 Neural Network Identification of Machining Process
4.8 The Common Measurement Methods and Characteristics
References
5 The Smart Machining System Monitoring from Machine Learning View
5.1 The Condition Monitoring Methods
5.1.1 Empirical Analysis
5.1.2 Statistical Method
5.1.3 Intelligent Method
5.2 Smart Machining System Monitoring (MSM) as a Machine Learning Problem
5.2.1 Feature
5.2.2 State
5.2.3 Classifier
5.3 The MSM System Content
5.3.1 Signal Preprocessing
5.3.2 Feature Extraction and Selection
5.3.3 State Classification
5.4 Feature Selection Method
5.4.1 Effective Criteria for Monitoring Features
5.4.2 Optimal Monitoring Feature Group Selection
5.4.3 The Bidirectional Search Algorithm for Feature Selection
5.5 Machine Learning Method
5.5.1 Bayesian Classifier
5.5.2 Fisher Linear Discriminant
5.5.3 Principal Components Analysis
5.5.4 Kernel Principal Components Analysis
5.5.5 Support Vector Machines
5.5.6 Artificial Neural Network (ANN)
5.5.7 K-Nearest Neighbor (KNN)
5.5.8 Case Study: MSM with Self-Organizing Map (SOM)
5.6 Deep Learning
5.6.1 Introduction to Deep Learning
5.6.2 Sparse Autoencoder (AE)
5.6.3 Deep Belief Neural Network (DBN)
5.6.4 Convolution Neural Network (CNN)
5.6.5 Recurrent Neural Network (RNN)
5.6.6 Challenges of Deep Learning Approaches in MSM Process Monitoring
References
6 Signal Processing for Machining Process Modeling and Condition Monitoring
6.1 Signal Processing in Condition Monitoring
6.1.1 Overview of Condition Monitoring
6.1.2 Signal Processing Issues in Condition Monitoring
6.2 Signal Space, Linear System, and Fourier Transform
6.2.1 Signal Spaces and Inner Product
6.2.2 Fourier Transform
6.2.3 Linear System, Sampling Theorem, and Convolution
6.3 Spectrum Analysis of Machining Signals
6.3.1 The Spectrum of Machining Signals
6.3.2 Spectrum Characteristics of Stochastic Signals
6.4 Correlation Analysis
6.4.1 Autocorrelation Function
6.4.2 Cross-Correlation Function
6.5 Common Signal Features in Time and Frequency Domain
6.5.1 Feature Parameters in the Time Domain
6.5.2 Feature Parameters in the Frequency Domain
6.6 Wavelet Analysis
6.6.1 Limitation of Fourier Methods
6.6.2 Continuous Wavelet Analysis (CWT) and Its Time–Frequency Properties
6.6.3 Discrete Wavelet Transform and Its Implementation
6.6.4 Wavelet Basis Function
6.6.5 Wavelet Packets Decomposition
6.6.6 Some Remarks on Wavelet Transform
6.7 Sparse Decomposition of Signals
6.7.1 Compressive Sensing
6.7.2 Sparse Decomposition Over Pre-defined Dictionaries
6.7.3 Greedy Algorithms
6.7.4 Dictionary Learning for Redundant Representation
References
7 Tool Condition Monitoring with Sparse Decomposition
7.1 Introduction
7.2 Sparse Coding for Denoising (Heavy Non-Gaussian Noise Separation)
7.2.1 Introduction
7.2.2 Noise Properties in Micro-milling
7.2.3 Sparse Representation in the Time–Frequency Domain
7.2.4 Sparse Representation as a Convex Optimization Problem
7.2.5 Case Studies
7.3 Sparse Representation for Tool State Estimation
7.3.1 Sparse Coding of Wavelet Packet Decomposition Coefficients
7.3.2 The Discriminant Dictionary Learning
7.3.3 Fast Tool State Estimation Without Signal Reconstruction
7.3.4 Experimental Validation
7.3.5 Results and Discussions
References
8 Machine Vision Based Smart Machining System Monitoring
8.1 Machine Vision System for Machining Process Monitoring
8.1.1 Introduction
8.1.2 The State-of-the-Art
8.2 Digital Image Acquisition and Representation
8.2.1 Image Acquisition of the Monitored Objects
8.2.2 CCD Sensor
8.2.3 CMOS Sensor
8.2.4 Representation of Digital Images
8.2.5 Digital Image Processing
8.3 Machine Vision System for Micro Milling Tool Condition Monitoring
8.3.1 The Micro Milling Tool Condition Monitoring
8.3.2 Tool Wear Inspection System
8.3.3 Tool Wear Inspection Method
8.3.4 Experimental Verification
8.3.5 Conclusions
References
9 Tool Wear Monitoring with Hidden Markov Models
9.1 Introduction
9.2 HMM Based Methods
9.2.1 Hidden Markov Models
9.2.2 Three Problems of Hidden Markov Models
9.3 Hidden Markov Models Based Tool Condition Monitoring
9.3.1 HMM Description of Tool Wear Process and Monitoring
9.3.2 The Framework of HMMs for TCM
9.3.3 Hidden Markov Model Selection: Continuous Left–Right HMMs
9.3.4 Selection of the Number of Gaussian Mixture Components
9.3.5 On the Number of Hidden States in Each HMM
9.3.6 Estimation of the HMM Parameters for Tool Wear Classification
9.3.7 Tool State Estimation with HMMs
9.4 Experimental Verifications
9.4.1 Experiment Setup
9.4.2 HMM Training for TCM
9.4.3 HMM for Tool Wear State Estimation
9.4.4 Moving Average for Tool Wear State Estimation Smoothing
9.4.5 On the Generalization of the HMM-Based Algorithm for TCM
9.5 Diagnosis and Prognosis of Tool Life with Hidden Semi-Markov Model
9.5.1 Hidden Semi-Markov Model for Degradation Process Modeling
9.5.2 On-Line Health Monitoring via HSMM
9.6 Experimental Validation
9.6.1 Case Study
9.6.2 Feature Extraction and Quantization
9.6.3 Training of HSMM for Tool Wear Monitoring
9.6.4 Diagnosis and Prognosis Results
References
10 Sensor Fusion in Machining System Monitoring
10.1 Multi-sensor Information Fusion Principle
10.2 Multi-sensor Information Fusion with Neural Networks
10.3 Sensor Fusion with Deep Learning
10.3.1 Problem Formulation
10.3.2 The Unit of Pyramid LSTM Auto-encoder
10.3.3 The Structure of the Pyramid LSTM Auto-encoder
10.3.4 The Training Method
10.3.5 Computational Efficiency
10.3.6 Experimental Validation
10.3.7 Conclusion
References
11 Big Data Oriented Smart Tool Condition Monitoring System
11.1 The Big Data Issues in Manufacturing
11.2 The Big Data Analytics in Smart Machining System
11.2.1 The Big Data Challenges and Motivation
11.2.2 Related Works
11.3 The Framework of Big Data Oriented Smart Machining Monitoring System
11.3.1 The Monitoring System Architecture
11.3.2 The Big Data-Oriented Formulation of TCM
11.4 The Functional Modules and Case Study
11.4.1 Sparse Coding Based Data Pre-processing
11.4.2 In-process Workpiece Integrity Monitoring
11.4.3 Heterogeneous Data Fusion and Deep Learning
11.4.4 Intelligent Tool Monitoring and Wear Compensation
11.5 Case Study
11.6 Summary
References
12 The Cyber-Physical Production System of Smart Machining System
12.1 Introduction
12.2 The Cyber-Physical System in Manufacturing
12.2.1 The Definition
12.2.2 The CPS Features
12.3 The CPS of Machine Tool and Machining Process
12.3.1 The State-of-the-Art
12.3.2 The CPS of Machine Tool
12.3.3 The CPS of Machining Process
12.4 A CPPS Framework of Smart Machining Monitoring System
12.4.1 Induction
12.4.2 The Smart CNC Machining Monitoring CPPS Structure
12.4.3 Case Studies
12.5 Summary
References