atmosphere2/PINO.puml at main · arnold-pdev/atmosphere2 · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
@startuml

title
Class diagrammed od PINO software
end title

package Data #DDDDDD {

class X {
x: float
y: float
getPoint()
}

class Points{
<latex>\{ \vec{x_1}, \vec{x_2}, ..\vec{x_N} \} \in \mathbb{R}^2  </latex> : Set
}
Points::x_1 <.. X


class U {
{field}<latex>\vec{v}(\vec{x})</latex> : Vector
{field}<latex>\Theta(\vec{x}) \in \mathbb{R}</latex> : float
{field}<latex>p(\vec{x}) \in \mathbb{R}</latex> : float
getState()
}

class Motion{
 <latex>\mathcal{M}</latex>: BoundaryConditions
{field} u(t) : TimeSeries
}
Motion::u <.. U
Motion::U <..Points

U::p <.. X
U::Theta <.. X
U::v <.. X

class BoundaryConditions{
- thermophysical properties: Scalar
- source terms: ¿?
- <latex>\nu</latex>
- <latex>\rho</latex>
}
Motion::Mu<..BoundaryConditions

}
package "Physics-Informed Neural Operator" #DDDDDD {


class FourierTransform{
frequencies : list
<latex>\mathcal{F}</latex>(u(t))
}
FourierTransform::F <..Motion::u

class NeuralNetwork{
frequencies : list
train()
<latex>L(\hat{y},y)</latex> + physics equations()
}

NeuralNetwork::train <..  FourierTransform::frequencies

class  InverseFourierTransform{
{field}<latex>\hat{u}(t)</latex>: TimeSeries
<latex>\mathcal{F}^{-1}</latex>(frequencies)
}
InverseFourierTransform::F<..NeuralNetwork::frequencies
}

@enduml