-
Notifications
You must be signed in to change notification settings - Fork 10
/
relative_L2.py
123 lines (110 loc) · 5.74 KB
/
relative_L2.py
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
import numpy as np
import scipy
import torch
import os
from pinn_model import *
import pandas as pd
device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
def load_data_feature(filename):
data_mat = scipy.io.loadmat(filename)
stack = data_mat['stack'] # N*4 (x,y,u,v)
x = stack[:, 0].reshape(-1, 1)
y = stack[:, 1].reshape(-1, 1)
t = stack[:, 2].reshape(-1, 1)
temp = np.concatenate((x, y, t), 1)
data_mean = np.mean(temp, axis=0).reshape(1, -1)
data_std = np.std(temp, axis=0).reshape(1, -1)
return data_mean, data_std
def load_full_points(filename):
# for each process, load data points within the extended boundary
# 每一个进程读取相应区域延拓边界内的数据点
# 读取原始数据
data_mat = scipy.io.loadmat(filename)
stack = data_mat['stack'] # N*4 (x,y,u,v)
x = stack[:, 0].reshape(-1, 1)
y = stack[:, 1].reshape(-1, 1)
t = stack[:, 2].reshape(-1, 1)
u = stack[:, 3].reshape(-1, 1)
v = stack[:, 4].reshape(-1, 1)
p = stack[:, 5].reshape(-1, 1)
low_bound = np.array([np.min(x), np.min(y), np.min(t)]).reshape(1, -1)
up_bound = np.array([np.max(x), np.max(y), np.max(t)]).reshape(1, -1)
temp = np.concatenate((x, y, t, u, v, p), 1)
x_ts = torch.tensor(x, dtype=torch.float32)
y_ts = torch.tensor(y, dtype=torch.float32)
t_ts = torch.tensor(t, dtype=torch.float32)
u_ts = torch.tensor(u, dtype=torch.float32)
v_ts = torch.tensor(v, dtype=torch.float32)
p_ts = torch.tensor(p, dtype=torch.float32)
return x_ts, y_ts, t_ts, u_ts, v_ts, p_ts, low_bound, up_bound
def compute_L2_norm(filename_train_data, filename_raw_data, filename_predict, norm_status="no_norm"):
# 预测值
x_raw, y_raw, t_raw, u_raw, v_raw, p_raw, low_bound, up_bound = load_full_points(filename_raw_data)
x_pre = x_raw.clone().detach().requires_grad_(True).to(device)
y_pre = y_raw.clone().detach().requires_grad_(True).to(device)
t_pre = t_raw.clone().detach().requires_grad_(True).to(device)
data_mean, data_std = load_data_feature(filename_train_data)
pinn_net = PINN_Net(layer_mat, data_mean, data_std, device)
pinn_net = pinn_net.to(device)
filename_load_model = filename_predict + '/NS_model_train.pt'
pinn_net.load_state_dict(torch.load(filename_load_model, map_location=device))
if norm_status == "no_norm":
u_pre, v_pre, p_pre = pinn_net.predict(x_pre, y_pre, t_pre)
else:
u_pre, v_pre, p_pre = pinn_net.predict_inner_norm(x_pre, y_pre, t_pre)
u_raw_mat = u_raw.numpy()
v_raw_mat = v_raw.numpy()
p_raw_mat = p_raw.numpy()
u_pre_mat = u_pre.cpu().detach().numpy()
v_pre_mat = v_pre.cpu().detach().numpy()
p_pre_mat = p_pre.cpu().detach().numpy()
# 处理数据
L2_u = (np.linalg.norm(u_pre_mat - u_raw_mat) / np.linalg.norm(u_raw_mat)).reshape(-1, 1)
L2_v = (np.linalg.norm(v_pre_mat - v_raw_mat) / np.linalg.norm(v_raw_mat)).reshape(-1, 1)
L2_p = (np.linalg.norm(p_pre_mat - p_raw_mat) / np.linalg.norm(p_raw_mat)).reshape(-1, 1)
L2_uvp_at_moment = np.concatenate((L2_u, L2_v, L2_p), axis=1)
return L2_uvp_at_moment
def L2_norm_at_moment(u_predicted, v_predicted, p_predicted, u_selected, v_selected, p_selected):
L2_u = (np.linalg.norm(u_predicted - u_selected) / np.linalg.norm(u_selected)).reshape(-1, 1)
L2_v = (np.linalg.norm(v_predicted - v_selected) / np.linalg.norm(v_selected)).reshape(-1, 1)
L2_p = (np.linalg.norm(p_predicted - p_selected) / np.linalg.norm(p_selected)).reshape(-1, 1)
L2_uvp_at_moment = np.concatenate((L2_u, L2_v, L2_p), axis=1)
return L2_uvp_at_moment
def rearrange_folders(matching_folders_list):
def compare_by_number(string):
# 提取字符串中的数字部分，并将其转换为整数进行比较
number = int(string[4:])
return number
rearranged_list = sorted(matching_folders_list, key=compare_by_number)
return rearranged_list
if __name__ == "__main__":
layer_mat_1 = [3, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 3] # 网络结构 neural network
layer_mat_2 = [3, 80, 80, 80, 80, 80, 3] # 网络结构 neural network
layer_mat_3 = [3, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 3] # 网络结构 neural network
layer_mat = layer_mat_1
filename_raw_data = './cylinder_Re3900_ke_all_100snaps.mat'
filename_train_data = './cylinder_Re3900_36points_100snaps.mat'
name_of_csv_set = ['exp_3900ke_1', 'exp_3900ke_2', 'exp_3900ke_3', 'exp_3900ke_4', 'exp_3900ke_5']
for name in name_of_csv_set:
folder_path = './data/' + name + '/write/'# 文件夹路径
prefix = 'step' # 文件夹名前缀
# 获取文件夹下所有文件夹名
folder_list = [folder for folder in os.listdir(folder_path) if os.path.isdir(os.path.join(folder_path, folder))]
# 筛选以指定前缀开头的文件夹名
matching_folders = [folder for folder in folder_list if folder.startswith(prefix)]
matching_folders = rearrange_folders(matching_folders)
print(matching_folders)
norm_status = "no_norm"
# norm_status = "inner_norm"
L2_set = np.empty((0, 3))
for dir_name in matching_folders:
file_predict = folder_path + dir_name
L2_norm = compute_L2_norm(filename_train_data, filename_raw_data, file_predict, norm_status)
L2_set = np.append(L2_set, L2_norm, axis=0)
# 存储L2范数信息
df = pd.DataFrame(L2_set, columns=["U", "V", "P"])
if not os.path.exists('./data_csv'):
os.makedirs('./data_csv')
filename_csv = './data_csv/non_dimensional_' + name + '.csv'
df.to_csv(filename_csv, index=False)
print('ok')