.png){kind=icon}
.png){kind=icon}
Return
Numerical analysis and experimental verification of neodymium butadiene rubber based on constitutive artificial neural networks
.png){kind=icon}
PRE
AI
T
C
J
K
DOI：10.1007/s12206-026-0236-7.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Rubber-like materials are widely used in the automotive industry because of their excellent properties, such as energy absorption and heat resistance. Mechanically, the physical deformation of rubber is quantified by strain energy density functions. This study used a neural network with prior knowledge of rubber mechanics to model the hyperelastic behavior of rubber. Experimental data from uniaxial, equibiaxial, and pure shear tests were used for training; the results were compared with those of conventional Neo-Hookean, Mooney-Rivlin, and 3rd Ogden models. The neural-network-based model demonstrated better fitting results. The trained model was then implemented in commercial finite-element code to perform stress analysis to investigate the load-displacement relationship of an automotive bushing. The simulations showed good agreement with experiments in several loading directions. Therefore, this neural-network-based model can be used to investigate the responses of bushings and to optimize structure designs.
Keywords:
Automotive bushing
Constitutive artificial neural networks
Data-driven modeling
Deep learning
Finite element analysis
Rubber materials
Journal
IF:
1.7
Papers:
481
Citations:
1.2W
Organization
Cited Papers