M. Valliappan et al., TEMPERATURE AND CURE IN PULTRUDED COMPOSITES USING MULTISTEP REACTIONMODEL FOR RESIN, Journal of reinforced plastics and composites, 15(3), 1996, pp. 295-321
A numerical model has been developed to analyze the temperature and de
gree of cure profiles in pultruded composites. In pultruded composites
the resin plays an important role in holding the fibers together in a
structural unit and also transferring and distributing the applied lo
ad to the fibers. The degree of cure of the pultruded composite is an
important phenomenon in the manufacturing process since it can be rela
ted to the mechanical properties. Therefore, the degree of cure of the
resin plays a crucial role in the pultrusion process. The chemical re
action of the resin determines the degree of cure as well as the exoth
ermic energy released by the resin. A one-step reaction model has been
employed for the resin system in most of the previous research. For s
ome resin systems, a one-step model may not produce good results; ther
efore, it was decided to pursue a multiple-independent-step reaction m
odel that more closely follows the actual behavior of the resin. In th
is research, the effect of an approximate multiple-independent-step re
action model for the thermoset epoxy resin SHELL EPON 862/W was studie
d. The numerical model utilized a fixed control volume based finite di
fference approach [1]. This technique was used to solve the coupled, n
on-linear, three-dimensional steady-state energy and species equations
for a cylindrical geometry. The species equation(s) utilized both a o
ne-step Arrhenius reaction rate model as well as a multiple-independen
t-step Arrhenius reaction rate model for the resin. The kinetics param
eters of the resin for a one-step reaction model were obtained from th
e differential scanning calorimeter (DSC) scans and for the multistep
model a regression fit was made to obtain the kinetic parameters from
the DSC scans. The numerical model was used to predict the temperature
and degree of cure for the pultruded composite both inside the die an
d in the post-die region. These numerical results were compared with e
xperimental measurements. The processing variables examined in this st
udy were die-wall temperature setting, pull speed and fiber volume fra
ction.