98
there was a 33% increase in MOE for 40% VTC compared to 40% control (see
Figure
4.7). At 40% VTC, there was a continuous path of VTC strands on the surface, therefore,
they had a larger increase in MOE. There was an effect of interface on the MOE and the
effect was larger in the 40% VTC than in the 20% VTC (e.g. the decrease from 100% to
1% glue was larger for 40% VTC).
Table 4.1 has a summary of simulated moduli results of Figure 4.6 and Figure 4.7
along with the experimental data from Rathi (2009). The
absolute values between
simulations and experiments do not agree. For comparison, we looked instead at relative
moduli between VTC specimens and controls. From experimental results, for OSB panels
with 20% VTC, there are was a 6.9% increase in MOE; for OSB panel with 40% VTC
addition, there was a 25% increase in MOE (Rathi 2009). The increase in MOE from
simulations varied with different values of D
t
. From the results of Table 4.1 there was
13.6% increase in MOE for 1% adhesive coverage, a 16.1% increase in MOE for 25%
adhesive
coverage levels, and a 26.1% increase in MOE for 100% glue. For the case of
40% VTC, there was a 30% increase in MOE for 1% adhesive coverage, a 32% increase
in MOE for 25% adhesive coverage and a 37% increase in MOE for 100% adhesive
coverage. The simulations show the potential for increase in MOE is higher when the
glue is better. Thus optimal VTC panels should optimize the glue to get the most
benefit
from the VTC strands.
In average there was 18.6% increase in MOE for 20% VTC and a 32% increase
for 40% VTC which are larger than experimental values of Rathi (2009) but trend in the
same direction. The simulation of 40% VTC gave closer results to experiment than the
20% VTC. This may be because the 20% VTC had only one layer of VTC strands on
each surface so the panel may be more influenced by the initial configuration. However,
in the 40% VTC addition, there were more VTC strands on
the surface, and less
dependence on randomly selected structures.
In Figure 4.6 and 4.7 the error bars represent one standard deviation in the mean
of at least 5 different runs. In Figure 4.6 the error bars for the 20% control and 20% VTC
overlapped. This may be because of the lack of continuous layers of VTC for the 20%
VTC. However, for 40% VTC, there was a continuous layer of VTC strands on the
99
surface layers. At shown in Figure 4.7, there is no overlap of error bars. The error bar
(one standard deviation) for the simulated results are listed in Table 4.1 as well. There
were larger standard deviations for the case of 20% VTC than for 40% VTC addition.
The error bars for 20% and 40% control were also much smaller; suggesting variability in
VTC strand has a larger influence on the deviations.
3000.0
3500.0
4000.0
4500.0
5000.0
5500.0
6000.0
6500.0
0.00
0.01
0.02
0.03
0.04
0.05
Do'stlaringiz bilan baham: