Information Technology Reference
In-Depth Information
θ
w
j
ij
v
j
r
Fig. 25.
The input and output of Texaco
gasifier
Fig. 26.
Structure of the 3 BP neural
networks
f
=×
g
(250000
−
120000)
+
120000
Fg
f
=×−+
=×−+
o
(45
35)
3
(5)
CO
f
H
(45
35)
35
H
2
2
The values of weights and thresholds of each of the 3 BP neural networks computing
results are listed in Table 1. After finishing the former preparing work, we can begin
the optimization task for the real-time operational parameters. The optimal model is
presented as follows.
Objective function:
Single objective function
The single objective function is effective gas yield defined as equation (6). [17] The
optimal algorithm is 3LM-CDE.
*
*
max{
pf
=× +
[
(
f
f
)] /(Fc
××
Cc
ρ
)}
(6)
Fg
CO
H
c
2
in which
p
, Fc*, Cc*, and
ρ
c
are the effective gas yield, current inlet flowrate of the
coal slurry, current inlet concentration of the coal slurry, and the density of the coal
slurry (=1.086t/m
3
), respectively.
Multi-objective function
The multi-objective function is defined as equation (7). The 4 MO-3LM-CDE
algorithms are used to optimize the results.
maximize
Subject to
[
f
,
f
,
f
]
Fg
CO
H
(7)
2
15000
≤≤
x
35000, 3000
≤ ≤
x
6000, 250
≤≤
x
400.
Fo
Foc
Fw
in which
x
,
Fo
x
, and
F
x
are the 3 optimized control parameters of inlet Total
Oxygen flowrate, inlet Central Oxygen flowrate and inlet flowrate of quench water.
Fo
Search WWH ::
Custom Search