Hybrid neural network and finite element modeling of sub-base layer material properties in flexible pavements

2007 ◽  
Vol 28 (5) ◽  
pp. 1725-1730 ◽  
Author(s):  
Mehmet Saltan ◽  
Hüseyin Sezgin
Keyword(s):  
Neural Network ◽  
Finite Element ◽  
Finite Element Modeling ◽  
Material Properties ◽  
Flexible Pavements ◽  
Base Layer ◽  
Hybrid Neural Network ◽  
Element Modeling ◽  
Layer Material

scholarly journals Characterization of Maple and Ash Material Properties for the Finite Element Modeling of Wood Baseball Bats

2018 ◽  
Vol 8 (11) ◽  
pp. 2256 ◽  
Author(s):  
Joshua Fortin-Smith ◽  
James Sherwood ◽  
Patrick Drane ◽  
David Kretschmann
Keyword(s):  
Finite Element ◽  
Finite Element Modeling ◽  
Material Properties ◽  
Charpy Impact ◽  
Impact Testing ◽  
Charpy Test ◽  
Ball Impact ◽  
Element Modeling ◽  
The Impact ◽  
Baseball Bats

To assist in developing a database of wood material properties for the finite element modeling of wood baseball bats, Charpy impact testing at strain rates comparable to those that a wood bat experiences during a bat/ball collision is completed to characterize the failure energy and strain-to-failure as a function of density and slope-of-grain (SoG) for northern white ash (Fraxinus americana) and sugar maple (Acer saccharum). Un-notched Charpy test specimens made from billets of ash and maple that span the range of densities and SoGs that are approved for making professional baseball bats are impacted on either the edge grain or face grain. High-speed video is used to capture each test event and image analysis techniques are used to determine the strain-to-failure for each test. Strain-to-failure as a function of density relations are derived and these relations are used to calculate inputs to the *MAT_WOOD (Material Model 143) and *MAT_EROSION material options in LS-DYNA for the subsequent finite element modeling of the ash and maple Charpy Impact tests and for a maple bat/ball impact. The Charpy test data show that the strain-to-failure increases with increasing density for maple but the strain-to-failure remains essentially constant over the range of densities considered in this study for ash. The flat response of the ash data suggests that ash-bat durability is less sensitive to wood density than maple-bat durability. The available SoG results suggest that density has a greater effect on the impact failure properties of the wood than SoG. However, once the wood begins to fracture, SoG plays a large role in the direction of crack propagation of the wood, thereby determining if the shape of the pieces breaking away from the bat are fairly blunt or spear-like. The finite element modeling results for the Charpy and bat/ball impacts show good correlation with the experimental data.

Finite element modeling to estimate the apparent material properties of trabecular bone

2013 ◽  
Vol 14 (8) ◽  
pp. 1479-1485 ◽  
Author(s):  
Sangbaek Park ◽  
Soo-Won Chae ◽  
Jungsoo Park ◽  
Seung-Ho Han ◽  
Junghwa Hong ◽  
...  
Keyword(s):  
Finite Element ◽  
Finite Element Modeling ◽  
Trabecular Bone ◽  
Material Properties ◽  
Element Modeling

A Predictive Analysis of Wall Stress in Abdominal Aortic Aneurysms Using a Neural Network Model

2021 ◽  
Author(s):  
Balaji Rengarajan ◽  
Sourav Patnaik ◽  
Ender A. Finol
Keyword(s):  
Neural Network ◽  
Finite Element ◽  
Finite Element Modeling ◽  
Wall Stress ◽  
Aortic Aneurysms ◽  
Multivariate Adaptive Regression Splines ◽  
Patient Specific ◽  
Clinical Images ◽  
Element Modeling ◽  
Abdominal Aortic

Abstract In the present work, we investigated the use of geometric indices to predict patient-specific abdominal aortic aneurysm (AAA) wall stress by means of a novel neural network (NN) modeling approach. We conducted a retrospective review of existing clinical images of two patient groups: 98 asymptomatic and 50 symptomatic AAA. The images were subject to a protocol consisting of image segmentation, processing, volume meshing, finite element modeling, and geometry quantification, from which 53 geometric indices and the spatially averaged wall stress (SAWS) were calculated. We developed feed-forward NN models composed of an input layer, two dense layers, and an output layer using Keras, a deep learning library in Python. The NN models were trained, tested, and validated independently for both AAA groups using all geometric indices, as well as a reduced set of indices resulting from a variable reduction procedure. We compared the performance of the NN models with two standard machine learning algorithms (MARS: multivariate adaptive regression splines and GAM: generalized additive model) and a linear regression model (GLM: generalized linear model). The NN-based approach exhibited the highest overall mean goodness-of-fit and lowest overall relative error compared to MARS, GAM, and GLM, when using the reduced sets of indices to predict SAWS for both AAA groups. The use of NN modeling represents a promising alternative methodology for the estimation of AAA wall stress using geometric indices as surrogates, in lieu of finite element modeling.

Sign in / Sign up

Export Citation Format

Share Document