Improved Fiber Orientation Predictions for Injection Molded Composites

1
Improved Fiber Orientation Predictions for
Injection Molded Composites

• Charles L. Tucker III and Jin Wang
• Department of Mechanical and Industrial
  Engineering
• University of Illinois at Urbana-Champaign
• John F. OGara and Gabriel DeBarr
• Delphi Research Labs
• NSF/DOE/APC Workshop
• The Future of Modeling in Composites Molding
  Processes
• June 9-10, 2004

2
Fiber Orientation Prediction State of the Art

• Describe orientation using A ?pp?
• Orientation evolves according to Jefferys eqn.
  interaction term
• Hele-Shaw or 3-D mold filling simulation gives
  velocity distribution, D, W

p
3
The Problem Orientation at Short Flow Lengths

• Edge-gated strips, PBT 30 glass fiber
• Measure elastic modulus in flow (E11) and
  crossflow (E22) directions
• Predict modulus using measured or predicted fiber
  orientation

4
Orientation Structure End-Gated Plaque
top
2h
shell flow-aligned
shell
shell
midplane
core
core random or cross-flow (depending on inlet)
bottom
bottom
midplane
top
5
Predicted vs. Measured Orientation, Standard
Model

• In short plaques, predicted core is too narrow
• Leads to over-prediction of E11,under-prediction
  of E22
• 2 mm, slow fill

6
Progress Toward a Better Model

• Hypothesis fibers experience local strain that
  is lower than average
• resin-rich slip layers absorb most of the
  strain
• fibers follow Jeffery-type motion based on local
  strain rate
• Strain Reduction Factor (SRF) (fiber strain
  rate / total strain rate)

7
Experiment vs. SRF Theory

• New theory with SRF20 does a good job for short
  plaquesover a range of thicknesses and filling
  speeds
• 80?90?2 mm
• slow fill speed

8
Further Steps for an Improved Model

• Our simple SRF theory is not objective
• does not behave sensibly in rigid-body
  rotationneed an objective version of the model
• Can only get SRF value by fitting experimental
  data
• need a micromechanics theory to predict SRF

if SRF 20, fiber rotates at 5 RPM!
flow rotates at 100 RPM
9
Long-Fiber/Thermoplastic Composites Predicted
vs. Measured Orientation
Data from Reinhard Hafellner, Advanced Polymer
Engineering Leoben, Austria
specimen plate
10
Gaps Processing of Injection Molded Composites

• Fiber orientation modeling
• capture the transient behavior at short flow
  lengths
• models for long-fiber thermoplastics (include
  migration, fiber breakage, bundle dispersion)
• predict orientation model parameters (CI, SRF)
  as function of fiber volume fraction, l/d, . . .
• models tested in a variety of flow geometries
• Fiber orientation measurement methods
• non-destructive
• capture full 3-D orientation
• use online
• . . .

11
Gaps, continued

• Models for warp/shrink/residual stress
• state of the art predict warp within factor of 2
• reach quantitative accuracy (fiber orientation,
  matrix PVT, . . . )
• incorporate flow-induced crystallization of
  matrix