Environmental Engineering Reference
In-Depth Information
ef ective impact on
WAT
,
WA B
and
WAC
can be observed by Al
2
O
3
, SiO
2
and CaO content. High inl uence of K
2
O content is observable for
WAC
and
WAT
.
FS
is most af ected by Al
2
O
3
, SiO
2
, and CaO content. h e inl u-
ence of MgO, K
2
O, and Na
2
O is also notable.
WLF
s are most af ected by
SiO
2
and Al
2
O
3
. CaO is also very inl uential. h e inl uences of MgO, Na
2
O,
and K
2
O were also notable.
VMC
is most af ected by Al
2
O
3
and SiO
2
con-
tent. CaO is also very inl uential. h e inl uences of MgO, Na
2
O, and K
2
O
were also notable.
Table 4.9 shows the inl uence of the input variables on
CSB
,
CSC
,
WAT
,
WA B
,
WAC
,
FS
,
WLFT
,
WLFB
,
WLFC
, and
VMC
, according to sum of
squares, calculated by comparing model predicted values with and without
white noise signal applied, according to table 4.8.
Prediction of the i nal laboratory product parameters evaluated by sec-
ond order polynomial regression models (SOPs) and optimal artii cial neu-
ral network (ANN) in conjunction with sensitivity analysis showed quite
good agreement with experimental results. SOPs and ANN showed high
r
2
values, as well as prediction accuracy, for the observed outputs, which
proved them to be useful in predicting i nal products' quality. Moreover,
the results of ANOVA (for SOPs) correspond very well to sensitivity analy-
sis (for ANN model).
4.3.6
Fuzzy Synthetic Optimization
Knowing i nal product application and optimal values of
CSB
,
CSC
,
WAT
,
WA B
,
WAC
,
FS
,
WLFT
,
WLFB
,
WLFC
, and
VMC
of it, allow ones to i nd
the optimal chemical composition and i ring temperature [2, 3, 33]. Fuzzy
synthetic optimization of the output variables was accomplished in order
to i nd the content of major oxides (SiO
2
, Al
2
O
3
, Fe
2
O
3
, CaO, MgO, Na
2
O,
K
2
O, MnO, and TiO
2
), and i ring temperature that give optimums of
CSB
,
CSC
,
WAT
,
WA B
,
WAC
,
FS
,
WLFT
,
WLFB
,
WLFC
, and
VMC
. Trapezoidal
membership function was used, according to eq. 4.5, in which
a - b
cov-
ered the complete interval of obtained output values, and
m - n
repre-
sented the optimal values for observed product group (table 4.10). h e
optimal parameters, used for FSE evaluation, were given based on our
experience with i red laboratory heavy clay products, and published else-
where [3]. Heavy clay products are divided to three groups, where: Group
I is suitable for the production of solid bricks; Group II can be used for
hollow bricks and blocks, as well as ceiling elements; and Group III is
appropriate for roof tiles and facade elements. h e samples belonging
to Groups II and III could be also used in a light-weighted bricks, as a
primary raw material [2, 3].
Search WWH ::
Custom Search