CCD: Cure Cycle Design
Simulations
Index (at Interactive Learning Tools subsite)
Cure is a process that converts a liquid resin into a solid by chemical
reactions induced by heat. This program simulates the temperature
distributions and degree of cure throughout the thickness of
a composite plate for various mold-heating scenarios.
|
This work was conducted
in a joint effort between the University of Delaware and Michigan State University
under the aegis of the National Science Foundation project
"Manufacturing Education and Training Program in Composite
Materials for the DoD and Durable Goods Industries" as part
of the DARPA Supplementary Education Awards.
|
Background
A major challenge encountered in curing thick-section composites
is control of the heat generated by the chemical reaction in
order to achieve uniform cure and minimum residual stresses.
This problem is aggravated by the poor thermal conductivity of
the material and heat transfer characteristics.
Linking thermodynamic behavior, reaction kinetics, and heat
transfer characteristic provides cure process models, as illustrated
in the algorithms shown above.
The computation is implemented through a finite-difference technique
that gives temperature and degree-of-cure values
at various locations for different times and mold temperatures.
For this instructional treatment, attention is restricted to a flat-plate
geometry. Even the simplified analytical model requires considerable input,
and the response time to obtain output graphs of temperature and degree of
cure distributions is slow. To improve the response time, the analytical
program was used to train a neural network for rapid response. However,
this feature is obtained at the expense of flexibility and the loss of
some accuracy.
Features
The introductory page describes the program and presents the user with four
options: Tutorial (which describes the issues involved in designing cure cycles),
Theoretical Basis (which presents the basis for the computation), and
two computation modes-the more flexible Analytical Program or the fast response
Neural Net Simulation. An example of the Neural Net Simulation page is shown below
Two slide switches, T1 and T2, control the inputs for the mold temperatures;
the dwell time is controlled by a separate slide switch. The mold is heated
up to temperature T1 and held there for the duration of the dwell time. The
temperature is then increased to T2 and maintained at T2 for the remainder
of the cycle. These temperature-time settings are displayed by a grey profile
in the temperature versus time plot.
The red profile displays the temperature and degree of cure at the center plane, the
blue profile at the top and bottom of the plate; and the green profile at the one-quarter
and three-quarter plane. The cure crossover intersection at which the center plane is
curing faster than the surface is displayed as a numerical value.
Provisions are available to vary the activation energy and fiber volume fraction for
sensitivity analyses.
Applications
The neural network simulation can be used to obtain preliminary estimates of the
impacts of various mold heating schedules and system parameters on the design of
cure cycles. Once a satisfactory cure cycle has been designed, the analytical
model should be used to refine the design.
An important feature of both programs is the ability to implement sensitivity
analyses. After an acceptable cure cycle has been obtained, the user may change
various parameters to explore how each parameter affects the cure and identify
the critical parameters.
The general program used to generate these examples is capable of treating much
more complex geometries as well as estimating residual stress development
resulting from the mold-heating program. Further information concerning the
general program may be obtained at the address given below.
Prof. John W. Gillespie Jr.
phone: 302-831-8702
fax: 302-831-8525
e-mail: gillespie@ccm.udel.edu
Center for Composite Materials
201 Composites Mfg. Science Lab
University of Delaware
Newark, DE 19716-3144
|
|