This book presents systematic overviews and bright insights into big data-driven intelligent fault diagnosis and prognosis for mechanical systems. The recent research results on deep transfer learning-based fault diagnosis, data-model fusion remaining useful life (RUL) prediction, etc., are focused on in the book. The contents are valuable and interesting to attract academic researchers, practitioners, and students in the field of prognostics and health management (PHM). Essential guidelines are provided for readers to understand, explore, and implement the presented methodologies, which promote further development of PHM in the big data era.
Features:
- Addresses the critical challenges in the field of PHM at present
- Presents both fundamental and cutting-edge research theories on intelligent fault diagnosis and prognosis
- Provides abundant experimental validations and engineering cases of the presented methodologies
Author(s): Yaguo Lei, Naipeng Li, Xiang Li
Publisher: Springer
Year: 2022
Language: English
Pages: 291
City: Singapore
Preface
Contents
About the Authors
1 Introduction and Background
1.1 Introduction
1.1.1 AI Technologies for Data Processing
1.1.2 Big Data-Driven Intelligent Predictive Maintenance
1.1.3 Big Data Analytics Platform Practices
1.2 Overview of Big Data-Driven PHM
1.2.1 Data Acquisition
1.2.2 Data Processing
1.2.3 Diagnosis
1.2.4 Prognosis
1.2.5 Maintenance
1.3 Preface to Book Chapters
References
2 Conventional Intelligent Fault Diagnosis
2.1 Introduction
2.2 Typical Neural Network-Based Methods
2.2.1 Introduction to Neural Networks
2.2.2 Intelligent Diagnosis Using Radial Basis Function Network
2.2.3 Intelligent Diagnosis Using Wavelet Neural Network
2.2.4 Epilog
2.3 Statistical Learning-Based Methods
2.3.1 Introduction to Statistical Learning
2.3.2 Intelligent Diagnosis Using Support Vector Machine
2.3.3 Intelligent Diagnosis Using Relevant Vector Machine
2.3.4 Epilog
2.4 Conclusions
References
3 Hybrid Intelligent Fault Diagnosis
3.1 Introduction
3.2 Multiple WKNN Fault Diagnosis
3.2.1 Motivation
3.2.2 Diagnosis Model Based on Combination of Multiple WKNN
3.2.3 Intelligent Diagnosis Case Study of Rolling Element Bearings
3.2.4 Epilog
3.3 Multiple ANFIS Hybrid Intelligent Fault Diagnosis
3.3.1 Motivation
3.3.2 Multiple ANFIS Combination with GA
3.3.3 Fault Diagnosis Method Based on Multiple ANFIS Combination
3.3.4 Intelligent Diagnosis Case of Rolling Element Bearings
3.3.5 Epilog
3.4 A Multidimensional Hybrid Intelligent Method
3.4.1 Motivation
3.4.2 Multiple Classifier Combination
3.4.3 Diagnosis Method Based on Multiple Classifier Combination
3.4.4 Intelligent Diagnosis Case of Gearboxes
3.4.5 Epilog
3.5 Conclusions
References
4 Deep Transfer Learning-Based Intelligent Fault Diagnosis
4.1 Introduction
4.2 Deep Belief Network for Few-Shot Fault Diagnosis
4.2.1 Motivation
4.2.2 Deep Belief Network-Based Diagnosis Model with Continual Learning
4.2.3 Few-Shot Fault Diagnosis Case of Industrial Robots
4.2.4 Epilog
4.3 Multi-Layer Adaptation Network for Fault Diagnosis with Unlabeled Data
4.3.1 Motivation
4.3.2 Multi-Layer Adaptation Network-Based Diagnosis Model
4.3.3 Fault Diagnosis Case of Locomotive Bearings with Unlabeled Data
4.3.4 Epilog
4.4 Deep Partial Adaptation Network for Domain-Asymmetric Fault Diagnosis
4.4.1 Motivation
4.4.2 Deep Partial Transfer Learning Net-Based Diagnosis Model
4.4.3 Partial Transfer Diagnosis of Gearboxes with Domain Asymmetry
4.4.4 Epilog
4.5 Instance-Level Weighted Adversarial Learning for Open-Set Fault Diagnosis
4.5.1 Motivation
4.5.2 Instance-Level Weighted Adversarial Learning-Based Diagnosis Model
4.5.3 Fault Diagnosis Case of Rolling Bearing Datasets
4.5.4 Epilog
4.6 Conclusions
References
5 Data-Driven RUL Prediction
5.1 Introduction
5.2 Deep Separable Convolutional Neural Network-Based RUL Prediction
5.2.1 Motivation
5.2.2 Deep Separable Convolutional Network
5.2.3 Architecture of DSCN
5.2.4 RUL Prediction Case of Accelerated Degradation Experiments of Rolling Element Bearings
5.2.5 Epilog
5.3 Recurrent Convolutional Neural Network-Based RUL Prediction
5.3.1 Motivation
5.3.2 Recurrent Convolutional Neural Network
5.3.3 Architecture of RCNN
5.3.4 RUL Prediction Case Study of FEMTO-ST Accelerated Degradation Tests of Rolling Element Bearings
5.3.5 Epilog
5.4 Multi-scale Convolutional Attention Network-Based RUL Prediction
5.4.1 Motivation
5.4.2 Multi-scale Convolutional Attention Network
5.4.3 Architecture of MSCAN
5.4.4 RUL Prediction Case of a Life Testing of Milling Cutters
5.4.5 Epilog
5.5 Conclusions
References
6 Data-Model Fusion RUL Prediction
6.1 Introduction
6.2 RUL Prediction with Random Fluctuation Variability
6.2.1 Motivation
6.2.2 RUL Prediction Considering Random Fluctuation Variability
6.2.3 RUL Prediction Case of FEMTO-ST Accelerated Degradation Tests of Rolling Element Bearings
6.2.4 Epilog
6.3 RUL Prediction with Unit-to-Unit Variability
6.3.1 Motivation
6.3.2 RUL Prediction Model Considering Unit-to-Unit Variability
6.3.3 RUL Prediction Case of Turbofan Engine Degradation Dataset
6.3.4 Epilog
6.4 RUL Prediction with Time-Varying Operational Conditions
6.4.1 Motivation
6.4.2 RUL Prediction Model Considering Time-Varying Operational Conditions
6.4.3 RUL Prediction Case of Accelerated Degradation Experiments of Thrusting Bearings
6.4.4 Epilog
6.5 RUL Prediction with Dependent Competing Failure Processes
6.5.1 Motivation
6.5.2 RUL Prediction Model Considering Dependent Competing Failure Processes
6.5.3 RUL Prediction Case of Accelerated Degradation Experiments of Rolling Element Bearings
6.5.4 Epilog
6.6 Conclusions
References
Glossary
