Modelling Inhomogeneity of Veneer Laminates with a Finite Element Mapping Method Based on Arbitrary Grayscale Images

Abstract

<jats:p>Failure and deformation behavior of veneer laminates of ring porous wood species vary with the individual arrangement of early- and latewood zones over a veneer sheet. Therefore, a method is presented, where local failure and damage modes are considered for finite element models with respect to forming simulations, during the development process of automotive interior trim parts. Within the mapping tool Envyo, a routine has been realized for the discretization of early- and latewood zones from ash wood veneer surfaces to finite element meshes. The routine performs the following steps: reading a grayscale image of known size and generation of a point cloud based on the number of pixels; transformation and scaling of the generated point cloud to align with a target finite element mesh; nearest neighbor search and transfer of grayscale values to the target mesh element centroids; assigning part and therefore material properties to the target elements based on the mapped grayscale value and user-defined grayscale ranges. Due to the absence of measurement data for early- and latewood, optimization was used to identify locally varying material constants. A set of material input parameters for early- and latewood was created, calibrating the force-displacement response of tensile test simulations to corresponding experimental curves. The numerical results gave a very good agreement to the failure behavior of tensile tests in the loading directions longitudinal and transverse to the fiber orientation. Furthermore, in a stochastic analysis the characteristic distribution of tensile strength and ultimate strain could be verified for the suggested procedure. The introduced modelling approach can be applied for the discrete implementation of inhomogeneity to numerical simulations.</jats:p>

Segmentation for Human Motion Injury Ultrasound Medical Images Using Deep Feature Fusion

“…where R(a) is feature fusion output value of medical ultrasound image, S zj is image pixel subset, J(a) represents the gray image [18,19], t 0 is the image initial structure similarity, and (t(a), t 0 ) denotes the similarity degree.…”