2022
DOI: 10.1007/s10489-021-03004-y
|View full text |Cite
|
Sign up to set email alerts
|

Deep learning models for predictive maintenance: a survey, comparison, challenges and prospects

Abstract: Given the growing amount of industrial data in the 4th industrial revolution, deep learning solutions have become popular for predictive maintenance (PdM) tasks, which involve monitoring assets to anticipate their requirements and optimise maintenance tasks. However, given the large variety of such tasks in the literature, choosing the most suitable architecture for each use case is difficult. This work aims to facilitate this task

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1
1

Citation Types

0
36
0
1

Year Published

2022
2022
2024
2024

Publication Types

Select...
3
3
2

Relationship

0
8

Authors

Journals

citations
Cited by 104 publications
(37 citation statements)
references
References 131 publications
0
36
0
1
Order By: Relevance
“…The second group is referred to as Knowledge-based methods, and focuses on reducing the need for complex physical models by supplementing them with heuristics associated with each machine. Finally, the Data-driven group consists in those methods that derive their predictions from extracting statistical patterns from the data, primarily through Machine Learning (ML) techniques [170,239,186]. Given the rapid growth of the latter and that our objective is to study continual learning as applied to this context, we will focus on Data-driven solutions in this survey.…”
Section: Predictive Maintenancementioning
confidence: 99%
See 2 more Smart Citations
“…The second group is referred to as Knowledge-based methods, and focuses on reducing the need for complex physical models by supplementing them with heuristics associated with each machine. Finally, the Data-driven group consists in those methods that derive their predictions from extracting statistical patterns from the data, primarily through Machine Learning (ML) techniques [170,239,186]. Given the rapid growth of the latter and that our objective is to study continual learning as applied to this context, we will focus on Data-driven solutions in this survey.…”
Section: Predictive Maintenancementioning
confidence: 99%
“…Then, [239] updates the literature review and introduces a new taxonomy to classify the different methods of this research area. Finally, [186] focus their survey on Deep Learning (DL) methods, comparing different architectures on a known dataset.…”
Section: Predictive Maintenancementioning
confidence: 99%
See 1 more Smart Citation
“…In the process of GM, molds secrete various extracellular enzymes to decompose the starch, lipids, cellulose and other substances in the grain, destroying the shell of the grain and consuming the dry matter in the grain; molds in the decomposition of amino acid products, such as thiols, causing the grain to brown and lose its original luster, while the molds multiply to form colored colony forms, accelerating the discoloration of the grain and reducing the market value of the grain. Some molds also produce toxic substances that endanger the liver, kidneys, digestive tract, and reproductive capacity of humans and animals, causing serious safety hazards [8][9].…”
Section: The Danger Of Gm and Mildewmentioning
confidence: 99%
“…Finally, we conduct a critical analysis of prior work to justify the research issues addressed in this thesis. Interested readers may refer to [87][88][89] for further information and in-depth discussions on the reasons why certain methods are superior to others.…”
Section: Literature Reviewmentioning
confidence: 99%