Residual Life Prediction and Optimal Maintenance Decision for a Piece of Equipment

This book addresses remaining life prediction and predictive maintenance of equipment. It systematically summarizes the key research findings made by the author and his team and focuses on how to create equipment performance degradation and residual life prediction models based on the performance monitoring data produced by currently used and historical equipment. Some of the theoretical results covered here have been used to make remaining life predictions and maintenance-related decisions for aerospace products such as gyros and platforms. Given its scope, the book offers a valuable reference guide for those pursuing theoretical or applied research in the areas of fault diagnosis and fault-tolerant control, remaining life prediction, and maintenance decision-making. Foreword Preface Summary Contents 1 Introduction 1.1 Background 1.2 Equipment Life Prediction 1.2.1 Fundamental Concept of Life Prediction 1.2.2 Literature Review on Life Prediction 1.3 Maintenance Decision of Equipment 1.3.1 Fundamental Concept and Classification of Maintenance 1.3.2 Literature Review on Maintenance Decision of Single-Component System 1.3.3 Literature Review on Maintenance Decision of Multi-component System References 2 Residual Life Prediction Based on Wiener Process with Nonlinear Drift 2.1 Introduction 2.2 Definition of Wiener Process 2.3 Degradation Modeling Based on Wiener Process with Nonlinear Drift 2.4 Probability Density Function of the Residual Life 2.5 Parameter Estimation 2.5.1 Off-Line Estimation of Common Parameters and Hyper-Parameters 2.5.2 Real-Time Updating of Random Parameter 2.6 Case Study 2.6.1 Problem Description 2.6.2 Results and Discussions 2.7 Summary of This Chapter References 3 Residual Life Prediction Based on Wiener Process with Abrupt Changepoint 3.1 Introduction 3.2 Degradation Model with Abrupt Changepoint Based on Wiener Process 3.2.1 Wiener-Process-Based Degradation Model 3.2.2 Changepoint Detection in Performance Degradation Process 3.3 Conjugate Distribution of Prior Distribution of Exponential Family 3.4 Bayesian Online Changepoint Detection Algorithm 3.5 Empirical Bayesian Method for Determining Prior Distribution 3.5.1 Improved Forward–Backward Algorithm 3.5.2 Joint Distribution of Changepoint Markers at Adjacent Detection Moments 3.5.3 EM Algorithm 3.6 Residual Life Prediction Based on Bayesian Online Changepoint Detection 3.7 Case Study 3.8 Summary of This Chapter References 4 Gamma Process-Based Degradation Modeling and Residual Life Prediction 4.1 Introduction 4.2 Definition of Gamma Process 4.3 Parameter Estimation for Gamma Process 4.3.1 Method of Moments 4.3.2 Maximum Likelihood Estimate 4.4 Residual Life Prediction Based on Gamma Process 4.4.1 Life Distribution 4.4.2 Residual Life Distribution 4.4.3 Reliability Function 4.4.4 Example Verification 4.5 Degradation Modeling Based on Gamma Process with Environmental Impact 4.5.1 Problem Description 4.5.2 Residual Life Distribution 4.5.3 Maintenance Decision 4.6 Summary of This Chapter References 5 Inverse Gaussian Process-Based Degradation Modeling and Residual Life Prediction 5.1 Introduction 5.2 Definition of Inverse Gaussian Process 5.3 ER-Based Parameter Estimation 5.3.1 Parameter Estimation Based on Single Specific Equipment’s Degradation Data 5.3.2 Fusion of Fixed Parameters Based on ER 5.4 Derivation of Residual Life Distribution 5.5 Experimental Verification 5.6 Summary of This Chapter References 6 Degradation Modeling and Residual Life Prediction Based on Support Vector Machine 6.1 Introduction 6.2 SVR Principle 6.2.1 Primal and Dual Problems 6.2.2 Sparsity of SVR 6.2.3 Kernel Function 6.3 Residual Life Prediction Method Based on GA-Optimized SVR 6.3.1 Problem Description 6.3.2 Basic Ideas 6.3.3 Specific Steps of the Method 6.3.4 Case Study 6.4 Residual Life Prediction Method Based on SVR and FCM Clustering 6.4.1 Problem Description 6.4.2 Basic Ideas and Specific Steps 6.4.3 Case Study 6.5 Summary of This Chapter References 7 Degradation Modeling and Residual Life Prediction Based on Fuzzy Model of Relevance Vector Machine 7.1 Introduction 7.2 Fuzzy Model Based on Relevance Vector Machine 7.2.1 Mathematical Description of Fuzzy Model 7.2.2 Fuzzy Model Based on Relevance Vector Machine 7.2.3 Uniform Approximation of Fuzzy Model Based on Relevance Vector Machine 7.3 Fuzzy Model Identification Based on Relevance Vector Machine 7.3.1 Structure Identification 7.3.2 Parameter Identification 7.3.3 Fuzzy Model Identification Algorithm Based on RVM and Gradient Descent Method 7.4 Degradation Modeling and Residual Life Prediction 7.5 Case Study 7.5.1 Description of Simulation System for Continuous Stirred Tank Reactor 7.5.2 Simulation Experiment and Results 7.5.3 Result Analysis 7.6 Summary of This Chapter References 8 Degradation Modeling and Reliability Prediction Based on Evidence Reasoning 8.1 Introduction 8.2 Degradation Modeling Based on Evidence Reasoning 8.2.1 Structure and Expression Form of Prediction Model 8.2.2 Degradation Modeling and Prediction Under the ER Framework 8.2.3 Utility Based Construction of the Numerical Outputs 8.3 Recursive Algorithms for Updating the ER-Based Prediction Model 8.3.1 Recursive Parameter Estimation Algorithm Based on Judgment Output 8.3.2 Recursive Parameter Estimation Algorithm Based on Numerical Output 8.4 Case Study 8.4.1 Problem Description 8.4.2 Reference Points of Teliability Data 8.4.3 Degradation Modeling and Prediction Model 8.4.4 Simulation Results Based on Judgment Output 8.4.5 Simulation Results Based on Numerical Output 8.5 Summary of This Chapter References 9 Weight Optimization-Based Particle Filter Algorithm for Degradation Modeling and Residual Life Prediction 9.1 Introduction 9.2 Particle Filter Algorithm Based on Weight Optimization 9.2.1 Particle Filter Algorithm and Characteristic Analysis 9.2.2 Particle Filter Algorithm Based on Weight Optimization 9.3 Degradation Modeling with Weight Optimization-Based Particle Filter 9.3.1 Description of Degradation Process 9.3.2 Parameter Estimation 9.4 Residual Life Prediction 9.5 Numerical Simulation 9.6 Summary of This Chapter References 10 Degradation Modeling and Residual Life Prediction Based on Grey Predcition Model 10.1 Introduction 10.2 Grey Predcition Model 10.2.1 Classical Grey GM (1, 1) Model [12] 10.2.2 Improved Grey Predcition Model 10.3 Residual Life Prediction Based on Improved Grey Predcition Model 10.4 Case Study 10.5 Summary of This Chapter References 11 Optimal Inspection Policy for Deteriorated Equipment Based on Life Prediction Information 11.1 Introduction 11.2 Inspection Strategy and the Optimization Objective Function 11.3 Optimal Inspection Policy Based on Residual Life Prediction 11.3.1 Optimal Inspection Period of Equipment When Unknown G(X) 11.3.2 Optimal Inspection Period of Equipment with G(X) Known 11.4 Optimal Inspection Policy for Inertial Platform 11.5 Summary of This Chapter References 12 Cooperative Predictive Maintenance of Two-Component System with Limited Resources 12.1 Introduction 12.2 Cooperative Predictive Maintenance Model 12.3 Maintenance Decision Modeling and Optimization 12.3.1 Estimation of Expected Failure Times 12.3.2 Cost Rate Model 12.3.3 Maintenance Optimization 12.4 Numerical Simulation 12.5 Summary of This Chapter References