Smart Monitoring of Rotating Machinery for Industry 4.0 (Applied Condition Monitoring, 19)

This book offers an overview of current methods for the intelligent monitoring of rotating machines. It describes the foundations of smart monitoring, guiding readers to develop appropriate machine learning and statistical models for answering important challenges, such as the management and analysis of a large volume of data. It also discusses real-world case studies, highlighting some practical issues and proposing solutions to them. The book offers extensive information on research trends, and innovative strategies to solve emerging, practical issues. It addresses both academics and professionals dealing with condition monitoring, and mechanical and production engineering issues, in the era of industry 4.0. Contents 6 Vulnerabilities and Fruits of Smart Monitoring 8 1 Introduction 8 1.1 The Ultimate System 8 1.2 What Is Smart Monitoring? 10 1.3 Smart Systems Versus Smart Staff 10 2 Evolution of Condition Monitoring Systems 11 2.1 Early Days 11 2.2 Expansion of Stationary Distributed Systems 11 2.3 Industrial Internet-of-Things 12 3 CMS Interaction with Human 12 3.1 Selection 12 3.2 Configuration 13 3.3 Operation 14 3.4 Maintenance Planning 15 4 Recommendations for Selection of Suitable System 15 5 Summary 16 References 16 A Tutorial on Canonical Variate Analysis for Diagnosis and Prognosis 17 1 Introduction 17 2 Canonical Variate Analysis for Diagnosis 20 2.1 The Basic Framework of CVA 20 2.2 Determination of the Number of Retained States 23 2.3 Determination of Fault Threshold 24 2.4 Extensions of CVA—Canonical Variate Dissimilarity Analysis 25 2.5 Industrial Case Study—Canonical Variate Analysis 27 3 Canonical Variate Analysis for Prognosis 28 3.1 CVA-Based State Space Models 28 3.2 Determining the Number of Retained States 29 3.3 Example of Using CVA State Space Model for Prognosis 30 3.4 CVA-Based Data Driven Models 30 4 Conclusion 35 References 35 A Structured Approach to Machine Learning Condition Monitoring 38 1 Introduction 39 2 Machine Learning 40 2.1 Deep Learning 41 2.2 Advantages and Drawbacks of the Machine Learning Supervised and Unsupervised Techniques in CBM 42 3 Development of Classifiers with Machine Learning Algorithms 45 4 Model Development Workflow 50 5 Conclusions 56 References 57 A Structured Approach to Machine Learning for Condition Monitoring: A Case Study 60 1 Introduction 61 2 Random Forest 62 3 Deep Learning/Autoencoder 63 4 Problem Description 64 4.1 Preliminary Test on Rotary Test Rig 65 4.2 XTS Test Rig 69 4.3 Autoencoder for Anomaly Detection 73 5 Conclusions 76 References 80 Dynamic Reliability Assessment of Structures and Machines Using the Probability Density Evolution Method 82 1 Introduction 83 2 The Probability Density Evolution Method 86 2.1 The PDEM Equation 86 2.2 Physical Interpretation of the PDEM 87 2.3 Dynamic Reliability Assessment Using PDEM 88 3 Dynamic Reliability Assessment of Structures 90 3.1 Offline PDEM-Based Reliability Assessment Method 90 3.2 Online PDEM-Based Reliability Assessment Method 91 3.3 Case Study: Cantilevered Beam 92 4 Dynamic Reliability Assessment of Machines 97 4.1 Extra Considerations for Dynamic Reliability Assessment of Machines 97 4.2 Case Study: Bearing 98 5 Discussion and Future Research Directions 102 5.1 Future Research Directions 103 References 104 Rotating Machinery Condition Monitoring Methods for Applications with Different Kinds of Available Prior Knowledge 107 1 Introduction 107 2 Prior Knowledge in Condition Monitoring 108 2.1 Engineering Knowledge 110 2.2 Knowledge Extracted from Machine Learning Algorithms 111 3 Case Study 114 3.1 Data Availability: Level 0 115 3.2 Data Availability: Level 1 116 3.3 Data Availability: Level 2 117 4 Conclusions and Recommendations 118 References 118 Model Based Fault Diagnosis in Bevel Gearbox 120 1 Introduction 120 2 Dynamic Modelling of One Stage Straight Bevel Gearbox 122 3 Modelling of Mesh Stiffness Function 124 3.1 Mesh Stiffness Model of a Healthy Bevel Gear 124 3.2 Mesh Stiffness Model of Bevel Gear with a Missing Tooth Fault 126 4 Simulation and Results 127 4.1 Dynamic Response of a Healthy Bevel Gear System 127 4.2 Dynamic Response of a Bevel Gear System with Missing Tooth Fault 128 5 Experimental Validation 130 6 Conclusion 135 References 135 Investigating the Electro-mechanical Interaction Between Helicoidal Gears and an Asynchronous Geared Motor 137 1 Introduction 137 2 Experimental Set Up 138 3 Results 139 4 Conclusion 146 References 146 Algebraic Estimator of Damping Failure for Automotive Shock Absorber 148 1 Introduction 149 2 Vehicle Model 150 3 Proposed Algebraic Estimator 151 4 Results of Simulation 152 5 Conclusion 154 References 155 On the Use of Jerk for Condition Monitoring of Gearboxes in Non-stationary Operations 157 1 Introduction 157 2 Dynamic Model 159 3 Numerical Simulations 160 3.1 Stationary Operating Conditions 160 3.2 Non-stationary Operating Conditions 162 3.3 Influence of Noise 165 4 Conclusion 165 References 166 Dynamic Remaining Useful Life Estimation for a Shaft Bearings System 168 1 Introduction 168 2 Methodology 170 3 Validation of the Proposed Approach 171 3.1 Experimental Setup 171 3.2 Results and Discussion 172 4 Conclusion 176 References 176