Deep learning for prognostics and health management: State of the art, challenges, and opportunities. (15th October 2020)

• Record Type: Journal Article
• Authors: Rezaeianjouybari, Behnoush; Shang, Yi
• Is Part Of: Measurement. Volume 163(2020)
• Publication Date: 2020-10-15
• Subjects: Prognostics and health management -- Deep learning -- Fault diagnosis -- Anomaly detection -- Domain adaptation; AAE Adversarial autoencoders -- ACGAN auxiliary classifier generative adversarial network -- AdaBN Adaptive batch normalization -- AE Autoencoder -- AFSA Artificial fish swarm algorithm -- AHKL Auto-balanced high-order Kullback-Leibler -- AI Artificial Intelligence -- BLSTM Bi-directional Long Short Term Memory -- CAE Contractive Autoencoder -- CBLSTM CNN- Bi-directional LSTM -- CD Contrastive divergence -- CDBN Convolutional Deep Belief Network -- CLSTM CNN-LSTM -- CNN Convolutional Neural Network -- CORAL Correlation alignment -- CPS Cyber-physical-systems -- CPU Central Processing Unit -- CUDA Compute Unified Device Architecture -- CVAE Conditional variational autoencoder -- DA Domain Adaptation -- DAD Deep Anomaly Detection -- DAE Denoising Autoencoder -- DBM Deep Boltzmann Machine -- DBN Deep Belief Network -- DL Deep Learning -- DNN Deep Neural Network -- DQN Deep Q-Network -- EMA Exponential moving average -- FFT Fast Fourier Transform -- GAN Generative Adversarial Network -- GDA Generalized discriminant analysis -- GPU Graphics Processing Unit -- GDBM Gaussian Bernoulli DBM -- GRU Gated Recurrent Unit -- GRU-ED GRU Encoder-Decoder -- HHT Hilbert-Huang transform -- HI Health Indicator -- IaaS Infrastructure as a Service -- IIoT Industrial Internet of Things -- JSD Jensen-Shannon Divergence -- KL Kullback-Leibler -- KNN k-nearest neighbors -- LSTM Long Short Term Memory -- LSTM-ED LSTM Encoder-Decoder -- MCMC Markov chain Monte Carlo -- MLP Multi-Layer Perceptron -- MMD Maximum mean discrepancy -- MSCNN Multi-scale convolutional neural network -- NAS Neural Architecture Search -- PaaS Platform as a Service -- PHM Prognostics and Health Management -- PSO Particle Swarm Optimization -- PSR Phase Space Representation -- RBF Radial basis function -- RBM Restricted Boltzmann Machine -- RKH Reproducing kernel Hilbert -- RL Reinforcement Learning -- RNN Recurrent Neural Network -- SaaS Software as a Service -- SAE Stacked Autoencoder -- SCDA Smooth conditional distribution alignment -- SDAE Sparse Denoising Autoencoder -- SDAE-NCL Stacked denoising autoencoder network with negative correlation learning -- SGD Stochastic gradient descent -- SML Stochastic maximum likelihood -- SNR Signal to Noise Ratio -- SPEV Spectrum-principal-energy vector -- SSAE Sparse Stacked Autoencoder -- SSDAE Stacked sparse denoising autoencoder -- STPN Spatiotemporal pattern network -- SVM Support Vector Machine -- TDConvLSTM Time-distributed Convolutional LSTM -- TL Transfer Learning -- TPU Tensor Processing Unit -- VAE Variational Autoencoder -- WGAN Wasserstein generative adversarial network -- WJDA Weighted joint distribution alignment -- WPI Wavelet Packet Image -- WPT Wavelet Packet Transform; Weights and measures -- Periodicals; Measurement -- Periodicals; Measurement; Weights and measures; Periodicals; 530.8
• DOI: 10.1016/j.measurement.2020.107929
• Languages: English
• ISSNs: 0263-2241

Abstract

Highlights: The state-of-the-art deep models in PHM applications have been overviewed. The models are classified into generative, discriminative and hybrid categories. The transfer learning and domain adaptation in the context of PHM are discussed. Important challenges and future research directions have been provided. Abstract: Improving the reliability of engineered systems is a crucial problem in many applications in various engineering fields, such as aerospace, nuclear energy, and water declination industries. This requires efficient and effective system health monitoring methods, including processing and analyzing massive machinery data to detect anomalies and performing diagnosis and prognosis. In recent years, deep learning has been a fast-growing field and has shown promising results for Prognostics and Health Management (PHM) in interpreting condition monitoring signals such as vibration, acoustic emission, and pressure due to its capacity to mine complex representations from raw data. This paper provides a systematic review of state-of-the-art deep learning-based PHM frameworks. It emphasizes on the most recent trends within the field and presents the benefits and potentials of state-of-the-art deep neural networks for system health management. In addition, limitations and challenges of the existing technologies are discussed, which leads to opportunities for future research.