Application of Deep Learning Algorithms to the Study of the Relationship between Acoustic Emission Signals and Grinding Force Parameters

The article considers prediction of cutting force components based on analysis of acoustic emission (AE) signals using deep learning algorithms. Based on pre-processing and synchronization of experimental data obtained during grinding of a heat-resistant nickel alloy, a training sample based on spec...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Russian engineering research Ročník 45; číslo 6; s. 765 - 770
Hlavní autoři: Mitrofanov, A. P., Rastegaev, I. A., Novikov, A. V.
Médium: Journal Article
Jazyk:angličtina
Vydáno: Moscow Pleiades Publishing 01.06.2025
Springer Nature B.V
Témata:
Accuracy
Acoustic emission
Acoustics
Algorithms
Cutting force
Data compression
Deep learning
Engineering
Engineering Design
Experiments
Grinding
Lubricants & lubrication
Machine learning
Neural networks
Prediction models
Process controls
Sensors
Software
Spectrograms
Synchronism
United States > US
Switzerland
deep learning
spectrograms
acoustic emission
cutting force
grinding
ResNet
ISSN:1068-798X, 1934-8088
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:The article considers prediction of cutting force components based on analysis of acoustic emission (AE) signals using deep learning algorithms. Based on pre-processing and synchronization of experimental data obtained during grinding of a heat-resistant nickel alloy, a training sample based on spectrograms of AE signals is compiled. Using a trained and specially modified ResNet-34 network, a highly accurate (coefficient of determination R 2 = 0.903) predictive model is created.
Bibliografie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:1068-798X
1934-8088
DOI:10.3103/S1068798X25701242