SAMPE Journal - November/December 2019 - 10

FEATURE / ADDITIVELY MANUFACTURED AUTOCLAVE TOOL

Sample Direction

Sample Location

CTE
mm/ (m * K)

x direction

14.4

x direction

7.28

y direction

41.45

y direction

66.9

z direction

CTE Average/Sample Direction
mm/(m * K)

74.7

10.84

54.18
74.7

Table 1. CTE in x, y, and z directions.

in the x direction have the lowest values due to the
highly oriented fibers in the deposition direction.
The largest CTE value was in the z direction which
is 700% larger than of x directional CTE. Sample
location in the bead highly affect the CTE results. It
is noticed that for the x direction, CTE from sample
1 is twice of CTE from sample 2. This is attrbutided
by the change of the fiber oriention within the
deposited bead[11] where the periphery has more
aligned fibers than the center of the printed bead.
For the simulation framework develpoed in this
work we have used the average value of the two
tested locations (periphery and center) for the
x-direction CTE.
SIMULATION
The expansion of an in-autoclave mold can be
predicted via computational simulation using a
finite element method (FEM). The simulation of
the mold design is challenging for several reasons.
First, the mold geometry to be printed is different
from the original CAD design. The original CAD
design is completely solid throughout the volume,
whereas the printed design incorporates a fill
pattern that is generated from Slicer. The slicergenerated design can vary depending on the
Slicer settings as shown in Figure 5. To perform
an accurate computational simulation, the slicergenerated design must be imported to FEM

software. However, since Slicer does not directly
communicate with the FEM software, the geometry
had to be reconstructed based on a specified
toolpath. The toolpath is the deposition path
along which the extruder moves as determined
by Slicer. The second challenge is a non-uniform
orientation of the anisotropic properties. The fibers
in the printed bead were assumed to be generally
aligned along the toolpath, which dictates the
orientation of the anisotropic thermal/mechanical
properties of the printed. Because the printed part
consists of tortuous toolpath, the material property
orientation in the simulation model must change
as the toolpath direction changes.
The approach presented here for predicting
the thermal expansion of an in-autoclave mold
consists of the following steps. An internally solid
model was created using CAD software (Step
(a)). The STL formatted design file was imported
into Slicer and converted into Gcode describing
deposition toolpaths (Step (b)). The design file was
also sent to FEM software for meshing (Step (c)),
meaning that the entire geometry was spatially
discretized into a large number of finite elements
(FE). During the 3D-printing process simulation
(Step (d)), the specific toolpath was mapped onto
the FE mesh that was based on the original CAD file.
In this step, only the discretized elements that were
identified as lying on the toolpath were activated

Figure 5. Slicer generates a toolpath with multiple options for a print
shape that can be different from the original CAD geometry. The outmost surface geometry is the same, but the internal structure can vary
depending on the Slicer settings.

SAMPE JOURNAL

N OV E M B E R / D E C E M B E R 2 0 1 9

w w w. s a m p e . o r g

http://www.sampe.org

SAMPE Journal - November/December 2019

Table of Contents for the Digital Edition of SAMPE Journal - November/December 2019