A multimodal and hybrid deep neural network model for Remaining Useful Life estimation

Al-Dulaimi, Ali; Zabihi, Soheil; Asif, Amir; Mohammadi, Arash

doi:10.1016/j.compind.2019.02.004

Cited by 213 publications

(78 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In a method resembling the principle of random forest, Zhang et al constructed a multiple DBN ensemble to maximize two conflicting objectives: accuracy and diversity. Composite models using LSTM with RBM [190] and 1D CNN [176], [182] were also investigated recently, and quite competitive performance was reported.…”

Section: Prognosismentioning

confidence: 99%

A Review on Deep Learning Applications in Prognostics and Health Management

et al. 2019

View full text Add to dashboard Cite

Deep learning has attracted intense interest in Prognostics and Health Management (PHM), because of its enormous representing power, automated feature learning capability and best-in-class performance in solving complex problems. This paper surveys recent advancements in PHM methodologies using deep learning with the aim of identifying research gaps and suggesting further improvements. After a brief introduction to several deep learning models, we review and analyze applications of fault detection, diagnosis and prognosis using deep learning. The survey validates the universal applicability of deep learning to various types of input in PHM, including vibration, imagery, time-series and structured data. It also reveals that deep learning provides a one-fits-all framework for the primary PHM subfields: fault detection uses either reconstruction error or stacks a binary classifier on top of the network to detect anomalies; fault diagnosis typically adds a soft-max layer to perform multi-class classification; prognosis adds a continuous regression layer to predict remaining useful life. The general framework suggests the possibility of transfer learning across PHM applications. The survey reveals some common properties and identifies the research gaps in each PHM subfield. It concludes by summarizing some major challenges and potential opportunities in the domain.

show abstract

Section: Prognosismentioning

confidence: 99%

A Review on Deep Learning Applications in Prognostics and Health Management

et al. 2019

View full text Add to dashboard Cite

show abstract

“…Training samples Training time [s] ARIMA SVM [6] 39.6843 -20631 -DCNN [1] 18.4480 1286.7 20631 -LSTM [5] 16.17 338 20631 714.53 DCNN [11] 12.61 273.7 17731 -WELM [5] 13.78 267.31 20631 5.04 HDNN [9] 13.017 245 20631 - The proposed approach performances are compared with a set of other recent approaches in the literature. The results from Table 2 indicate that the proposed algorithm has the ability to achieve a low score value depending on less training samples and less training time.…”

Section: Rmse Scorementioning

confidence: 99%

“…In [8], a new data-driven approach is introduced by Ben Ali et al by training a Simplified Fuzzy Adaptive Resonance Theory Map (SFAM) neural network with Weibull Distribution (WD) to avoid time domain fluctuation during RUL prediction. Al-Dulaimi et al [9] developed a mini batch hybrid deep NN that takes two paths for RUL estimation; the multidimensional feature extraction based on Long Short Term Memory (LSTM) and convolutional NN and the prediction path via a fusion algorithm. Wen et al [10] constructed feature mapping and training schemes based on ensemble residual CNN and validated their training model with the K-fold cross-validation method after constructing several learners to predict RUL accurately.…”

Section: Introductionmentioning

confidence: 99%

“…The results from Table 2 indicate that the proposed algorithm has the ability to achieve a low score value depending on less training samples and less training time. [11] 12.61 273.7 17731 -WELM [5] 13.78 267.31 20631 5.04 HDNN [9] 13.017 245 20631 - The training model was tested using new data (unknown to the model) from a test set of FD001. Additionally, the example of curve fitting of the RUL target function in Figure 7 proves that the proposed approach has an acceptable performance.…”

mentioning

confidence: 99%

See 1 more Smart Citation

Aircraft Engines Remaining Useful Life Prediction with an Improved Online Sequential Extreme Learning Machine

et al. 2020

View full text Add to dashboard Cite

The efficient data investigation for fast and accurate remaining useful life prediction of aircraft engines can be considered as a very important task for maintenance operations. In this context, the key issue is how an appropriate investigation can be conducted for the extraction of important information from data-driven sequences in high dimensional space in order to guarantee a reliable conclusion. In this paper, a new data-driven learning scheme based on an online sequential extreme learning machine algorithm is proposed for remaining useful life prediction. Firstly, a new feature mapping technique based on stacked autoencoders is proposed to enhance features representations through an accurate reconstruction. In addition, to attempt into addressing dynamic programming based on environmental feedback, a new dynamic forgetting function based on the temporal difference of recursive learning is introduced to enhance dynamic tracking ability of newly coming data. Moreover, a new updated selection strategy was developed in order to discard the unwanted data sequences and to ensure the convergence of the training model parameters to their appropriate values. The proposed approach is validated on the C-MAPSS dataset where experimental results confirm that it yields satisfactory accuracy and efficiency of the prediction model compared to other existing methods.

show abstract

“…In engineering practice, real-time mastering and managing the health status of equipment is an efficient measure to effectively reduce the probability of failure and improve safety [1][2][3]. Prognostics and health management (PHM) technology is a crucial way to grasp the real-time status of equipment and carry out effective management [4,5], which is mainly divided into two parts.…”

Section: Introductionmentioning

confidence: 99%

Remaining Useful Life Prediction Under Imperfect Prior Degradation Information

et al. 2020

View full text Add to dashboard Cite

The remaining useful life (RUL) prediction is the core of equipment maintenance and decisionmaking. Accurate RUL prediction can make effective maintenance before the failure occurs to reduce the probability of equipment failure. However, in industrial practice, we often face the situation that prior information is insufficient or inaccurate, which could influence our prediction of RUL and reduce the prediction accuracy severely. To solve this problem, we study the issue of RUL prediction under imperfect prior degradation data by reasonably fusing the prior information and the field degradation data. Firstly, based on the linear Wiener degradation process, we prove a conclusion that the random parameter estimation results by the joint Bayesian algorithm and the expectation-maximization (EM) algorithm are the same as that by the maximum likelihood estimation (MLE) method. It shows that the joint estimation method completely overcomes the influence of prior information, and the more iterations, the smaller proportion of prior information. Secondly, we also prove that the random parameter joint updating method for the nonlinear Wiener degradation process has the same characteristics. Then, a heuristic algorithm that reasonably fuses the prior information and the field degradation data is proposed, which controls the number of iterations of the joint updating algorithm based on the credibility of the prior information. It is also applied to the nonlinear degradation process by not iterating the nonlinear coefficients. Finally, the correctness of the conclusion of this paper is verified by the simulated degradation data, and the effectiveness of the proposed heuristic updating method is verified by the actual lithium battery degradation data.

show abstract

A multimodal and hybrid deep neural network model for Remaining Useful Life estimation

Cited by 213 publications

References 30 publications

A Review on Deep Learning Applications in Prognostics and Health Management

A Review on Deep Learning Applications in Prognostics and Health Management

Aircraft Engines Remaining Useful Life Prediction with an Improved Online Sequential Extreme Learning Machine

Remaining Useful Life Prediction Under Imperfect Prior Degradation Information

Contact Info

Product

Resources

About