Image-Based Mesh Generation from 3D Data for CAD and CAE€¦ · Image-Based Mesh Generation from...

Image-Based Mesh Generation from 3D Data for CAD and CAE

D. Harman, P. Young, W. Smigaj, G. JamesSimpleware Ltd., Exeter, UK – www.simpleware.com

Introduction

Bene�ts of Image-Based Meshing

» Techniques can be extended to calculating effective material properties using finite-element-based homogenisation

» Complex heterogeneous material approximated with a homogeneous material whose response to external loading resembles as closely as possible that of the original material

» Approach simplifies analysis of complex systems» Built-in FE solver in Simpleware software calculates response of a

cuboidal sample of a material to a sequence of boundary conditions (BCs) associated with certain physics to obtain effective properties

» Homogenisation workflow supports calculation of elasticity (+SOLID module), thermal and electrical conductivity (+LAPLACE module) and absolute permeability (+FLOW)

» Use of smoothed FE-based meshes and solver in Simpleware software offers advantages over grid or voxel-based approaches with stepped surfaces – FE mesh surfaces are more accurate and converge with increasing resolution to actual surface areaFig 4. FE-based homogenisation of a composite in (a) +SLOID; (b) +FLOW; (c) +LAPLACE

» 3D image data acquired from sources such as industrial CT and micro-CTscan be used to visualise and inspect parts, and to characterise material samples

» Processed data can be exported as computational models for FiniteElement Analysis (FEA) and Computational Fluid Dynamics (CFD), as files for Additive Manufacturing (AM), and as NURBS

» Models created from 3D image data enable comprehensive qualitative and quantitative analysis, including non-destructive evaluation of defects and simulation of material properties.

» Image-based meshing techniques developed by Simpleware Ltd. enable rapid conversion of scan data to models for design and simulation applications

Fig 1. Example engine block: (a) import of CT scan data and segmentation of parts; (b) measurement of features

FE-based Homogenisation

» Traditional meshing approaches require a CAD geometry to be generated before meshing. By comparison, image-based meshing can generate models directly from 3D image data

» Meshes can be generated for topologies of arbitrary complexity with any number of constituent materials (multi-part modelling)

» Accuracy of meshes created directly from voxels only limited by the quality of image acquisition and segmentation of materials�

» Image-based meshing technique developed by Simpleware based on an adaptation of the ‘marching cubes’ algorithm to support multiple segmented domains

» Techniques also developed for multi-part surface remeshing: this enables voxel-based meshes to be effectively decimated according to the size/complexity of local features

Conclusions and Applications

(b)(a)

» Use of a single integrated software environment enables complete workflow from scan to design and simulation

» Tools for calculating effective material properties enable comprehensive data to be retrieved from scanned samples

» Techniques suitable for applications such as non-destructive evaluation of scanned parts, characterisation of material samples, analysis of transport properties of porous media

» Image-based techniques also enable lattice structures to be easily added to CAD and image data to reduce weight of parts

» Example in Fig 3. demonstrates a workflow from micro-CT image acquisition to visualisation and segmentation in Simpleware software, image-based meshing to create a mixed hexahedral and tetrahedral mesh, and adaptively remeshed

» Remeshing is completely faithful to the original EVoMaC surfaces, with conforming interfaces and shared nodes

» Mesh inspection ensures decimated mesh is completely representative to the original surfaces�

» Exported models can be used as watertight STL files for Additive Manufacturing, and as volume meshes for FEA/CFD applicationsFig 2. Examples of multi-part CT models: (a) composite; (b) battery; (c) asphalt

(a) (c)

Fig 3. Workflow from scan to mesh (example micro-CT scan)

1. Scan 2. Stack of 2D slices 3. Voxel grid 4. Segmentation

5. Image-based mesh6. Adaptive remeshingShared nodes

Fig 5. Examples: (a) bottle cap (photo and model); (b) 3D printed part with lattice structure

(a) (c)

(b)

(a) (b)

(b)

Digital Industrial Radiology and Computed Tomography (DIR 2015) 22-25 June 2015, Belgium, Ghent - www.ndt.net/app.DIR2015