5.5 Further Issues
379
Table 5.7 The dataset and the network structures for view factor approximation
N
Data size
Reduced input
dimension
Method/Neural
network
Hidden layers Neurons
0
315 × 6
1
Regular
interpolation
–
–
1
40,523 × 9
3
Feed-forward 3
10, 10, 10
2
55,185 × 12
6
Feed-forward 3
15, 15, 15
3
89,367 × 15
9
Feed-forward 3
20, 20, 20
4
142,002 × 18 12
Cascadeforward
3
25, 25, 25
5
143,002 × 21 15
Cascadeforward
3
27, 27, 27
6
144,002 × 24 18
Cascadeforward
3
30, 30, 30
7
145,002 × 27 21
Cascadeforward
3
33, 33, 33
8
146,002 × 30 24
Cascadeforward
3
35, 35, 35
9
167,002 × 33 27
Cascadeforward
3
37, 37, 37
10
168,002 × 36 30
Cascadeforward
3
40, 40, 40
5.5.3.4 Validation and Analysis
Structured Packing of Spheres
The commonly packed beds of structured packing are utilized for validation and
analysis. The beds with the lattice of Simple Cubic (SC), Body-Centered Cubic
(BCC), and Face-Centered Cubic (FCC) are given in Fig. 5.118.
Fig. 5.118 The structured packing of Simple Cubic (SC) (a), Body-Centered Cubic (BCC) (b),
and Face-Centered Cubic (FCC) (c)