WPGNN/utils.py at main · NatLabRockies/WPGNN · 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
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
import copy
import h5py
import numpy as np
np.seterr(invalid='ignore')

def load_data(data_path, normalize=False, scale_factors=None):
    x_dict_list, f_dict_list = [], []

    with h5py.File(data_path, 'r') as f:
        for idx in [idx for idx in f]:
            f_idx = f[idx]

            x_graph = {'globals': f_idx['x/globals'][()],
                         'nodes': f_idx['x/nodes'][()],
                         'edges': f_idx['x/edges'][()],
                       'senders': f_idx['x/senders'][()],
                     'receivers': f_idx['x/receivers'][()]}
            x_dict_list.append(x_graph)

            f_graph = {'globals': f_idx['f/globals'][()],
                         'nodes': f_idx['f/nodes'][()],
                         'edges': f_idx['f/edges'][()],
                       'senders': f_idx['f/senders'][()],
                     'receivers': f_idx['f/receivers'][()]}
            f_dict_list.append(f_graph)

    if normalize:
        x_dict_list, f_dict_list, _ = norm_data(xx=x_dict_list, ff=f_dict_list, scale_factors=scale_factors)

    return (x_dict_list, f_dict_list)

def norm_data(xx=None, ff=None, scale_factors=None):
    x, f = copy.deepcopy(xx), copy.deepcopy(ff)

    N_x = len(x) if (x is not None) else 0
    N_f = len(f) if (f is not None) else 0

    assert scale_factors is not None, 'Scale factors must be provided'

    for i in range(N_x):
        x[i]['edges'] = (x[i]['edges'] - scale_factors['x_edges'][:, 0])/scale_factors['x_edges'][:, 1]
        x[i]['nodes'] = (x[i]['nodes'] - scale_factors['x_nodes'][:, 0])/scale_factors['x_nodes'][:, 1]

        x[i]['globals'] = (x[i]['globals'] - scale_factors['x_globals'][:, 0])/scale_factors['x_globals'][:, 1]

    for i in range(N_f):
        f[i]['nodes'] = (f[i]['nodes'] - scale_factors['f_nodes'][:, 0])/scale_factors['f_nodes'][:, 1]
        f[i]['globals'] = (f[i]['globals'] - scale_factors['f_globals'][:, 0])/scale_factors['f_globals'][:, 1]

    if (N_x > 0) and (N_f > 0):
        return x, f, scale_factors
    elif (N_x > 0):
        return x, scale_factors
    elif (N_f > 0):
        return f, scale_factors

def unnorm_data(xx=None, ff=None, scale_factors=None):
    x, f = copy.deepcopy(xx), copy.deepcopy(ff)

    N_x = len(x) if (x is not None) else 0
    N_f = len(f) if (f is not None) else 0

    assert scale_factors is not None, 'Scale factors must be provided'

    for i in range(N_x):
        x[i]['edges'] = scale_factors['x_edges'][:, 1]*x[i]['edges'] + scale_factors['x_edges'][:, 0]
        x[i]['nodes'] = scale_factors['x_nodes'][:, 1]*x[i]['nodes'] + scale_factors['x_nodes'][:, 0]

        x[i]['globals'] = scale_factors['x_globals'][:, 1]*x[i]['globals'] + scale_factors['x_globals'][:, 0]

    for i in range(N_f):
        f[i]['nodes'] = scale_factors['f_nodes'][:, 1]*f[i]['nodes'] + scale_factors['f_nodes'][:, 0]
        f[i]['globals'] = scale_factors['f_globals'][:, 1]*f[i]['globals'] + scale_factors['f_globals'][:, 0]

    if (N_x > 0) and (N_f > 0):
        return x, f
    elif (N_x > 0):
        return x
    elif (N_f > 0):
        return f

def speed_to_velocity(xx):
    x = np.atleast_2d(copy.deepcopy(xx))

    ws, wd = x[:, 0], -(x[:, 1]+90)*(np.pi/180.)
    u, v = -ws*np.cos(wd), -ws*np.sin(wd)

    if x.shape[0] == 1:
        x = np.concatenate((u, v), axis=0)
    else:
        x = np.concatenate((np.atleast_2d(u), np.atleast_2d(v)), axis=0).T

    return x

def velocity_to_speed(xx):
    x = np.atleast_2d(copy.deepcopy(xx))

    u, v = x[:, 0], x[:, 1]
    ws = np.sqrt(u**2 + v**2)

    wd = 90-np.arctan(v/u)*(180./np.pi)
    wd[u<0] += 180

    # If ws = 0, then no way to recover the direction
    wd[np.isnan(wd)] = 0.

    if x.shape[0] == 1:
        x = np.concatenate((ws, wd), axis=0)
    else:
        x = np.concatenate((np.atleast_2d(ws), np.atleast_2d(wd)), axis=0).T

    return x

def identify_edges(x_loc, wind_dir, cone_deg=15):
    # Identify edges where wake interactions may play a role in power generation
    N_turbs = x_loc.shape[0]

    u, v = speed_to_velocity([10., wind_dir])
    theta = np.arctan(v/u)
    if u < 0:
        theta += np.pi
    R = np.array([[np.cos(theta), -np.sin(theta)],
                  [np.sin(theta),  np.cos(theta)]])
    x_loc = x_loc@R

    x_rel = x_loc.reshape((1, N_turbs, 2)) - x_loc.reshape((N_turbs, 1, 2))

    alpha = np.arctan(x_rel[:, :, 1]/x_rel[:, :, 0])*(180./np.pi)
    alpha[np.isnan(alpha)] = 90.

    directed_edge_indices = ((abs(alpha) < cone_deg) & (x_rel[:, :, 0] <= 0)).nonzero()

    senders, receivers = directed_edge_indices[0], directed_edge_indices[1]

    edges = x_rel[senders, receivers, :]

    return edges, senders, receivers