markilue
/
self_example
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

model更新

This commit is contained in:
markilue 2022-11-04 10:17:20 +08:00
parent 3a2eadc687
commit 1cf7ba2aa3
12 changed files with 579 additions and 443 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

272
TensorFlow_eaxmple/Model_train_test/condition_monitoring/plot/test_plot.py
View File

@ -25,9 +25,9 @@ from mpl_toolkits.axes_grid1.inset_locator import inset_axes
result_file_name = "E:\self_example\TensorFlow_eaxmple\Model_train_test\condition_monitoring\self_try\compare\mse\RNet_C\RNet_C_timestamp120_feature10_result.csv"
mse_file_name = "E:\self_example\TensorFlow_eaxmple\Model_train_test\condition_monitoring\self_try\compare\mse\RNet_D\RNet_D_timestamp120_feature10_mse_predict1.csv"
mse_file_name = "E:\self_example\TensorFlow_eaxmple\Model_train_test\condition_monitoring\self_try\compare\mse\RNet_D/banda\RNet_D_banda_mse_predict1.csv"
max_file_name = "E:\self_example\TensorFlow_eaxmple\Model_train_test\condition_monitoring\self_try\compare\mse\RNet_D\RNet_D_timestamp120_feature10_max_predict1.csv"
max_file_name = "E:\self_example\TensorFlow_eaxmple\Model_train_test\condition_monitoring\self_try\compare\mse\RNet_D/banda\RNet_D_banda_max_predict1.csv"
source_path = "G:\data\SCADA数据\jb4q_8_delete_total_zero.csv"
@ -36,10 +36,11 @@ model_name1 = "../hard_model/two_weight/banda_joint_epoch14_9872_9863/weight"
file_name = "E:\跑模型\论文写作/DCAU.txt"
# list = [64.16, 65.26, 65.11, 66.6, 67.16, 66.28, 73.86, 75.24, 73.98, 76.7, 98.86, 99.45, 99.97]
def plot_result(result_data):
def plot_result_banda(result_data):
parameters = {
'figure.dpi': 600,
'figure.figsize': (2.8, 2),
@ -51,7 +52,7 @@ def plot_result(result_data):
'legend.fontsize': 5,
}
plt.rcParams.update(parameters)
fig,ax=plt.subplots(1, 1)
fig, ax = plt.subplots(1, 1)
plt.rc('font', family='Times New Roman') # 全局字体样式
font1 = {'family': 'Times New Roman', 'weight': 'normal', 'size': 5} # 设置坐标标签的字体大小，字体
plt.scatter(list(range(result_data.shape[0])), result_data, c='black', s=0.01, label="predict")
@ -64,7 +65,7 @@ def plot_result(result_data):
# plt.text(0, 0.55, "Threshold", fontsize=5, color='red',
# verticalalignment='top')
plt.axvline(result_data.shape[0] * 2 / 3, c='blue', ls='-.', lw=0.5, label='real fault')
plt.axvline(result_data.shape[0] * 2 / 3-50, c='blue', ls='-.', lw=0.5, label='real fault')
# plt.axvline(415548, c='blue', ls='-.', lw=0.5, label='real fault')
# plt.xticks(range(6), ('06/09/17', '12/09/17', '18/09/17', '24/09/17', '29/09/17')) # 设置x轴的标尺
plt.text(result_data.shape[0] * 5 / 6, 0.4, "Fault", fontsize=5, color='black', verticalalignment='top',
@ -78,7 +79,7 @@ def plot_result(result_data):
'pad': 1.5, 'linewidth': 0.1}, fontdict=font1)
indices = [result_data.shape[0] * i / 4 for i in range(5)]
classes = ['01/09/17', '08/09/17', '15/09/17', '22/09/17', '29/09/17']
classes = ['02/25', '03/01', '03/06', '03/12', '03/18']
indices1 = [i / 4 for i in range(5)]
classes1 = ['0', '0.25', 'Threshold', '0.75', '1']
@ -92,22 +93,21 @@ def plot_result(result_data):
plt.xlabel('Time', fontsize=5)
plt.tight_layout()
# plt.legend(loc='best', edgecolor='black', fontsize=4)
plt.legend(loc='best', frameon=False, fontsize=4)
plt.legend(loc='upper right', frameon=False, fontsize=4.5)
# plt.grid()
#局部方法图
# 局部方法图
axins = inset_axes(ax, width="40%", height="30%", loc='lower left',
bbox_to_anchor=(0.1, 0.1, 1, 1),
bbox_transform=ax.transAxes)
axins.scatter(list(range(result_data.shape[0])), result_data, c='black', s=0.001, label="predict")
axins.axvline(result_data.shape[0] * 2 / 3, c='blue', ls='-.', lw=0.5, label='real fault')
axins.axvline(result_data.shape[0] * 2 / 3-50, c='blue', ls='-.', lw=0.5, label='real fault')
plt.axhline(0.5, c='red', label='Failure threshold', lw=0.5)
#设置放大区间
# 设置放大区间
zone_left = int(result_data.shape[0]*2/3-100)
zone_right = int(result_data.shape[0]*2/3)+100
x=list(range(result_data.shape[0]))
# 设置放大区间
zone_left = int(result_data.shape[0] * 2 / 3 -160)
zone_right = int(result_data.shape[0] * 2 / 3) + 40
x = list(range(result_data.shape[0]))
# 坐标轴的扩展比例（根据实际数据调整）
x_ratio = 0.5 # x轴显示范围的扩展比例
@ -130,9 +130,98 @@ def plot_result(result_data):
axins.set_xlim(xlim0, xlim1)
# axins.set_ylim(ylim0, ylim1)
plt.show()
pass
def plot_result(result_data):
parameters = {
'figure.dpi': 600,
'figure.figsize': (2.8, 2),
'savefig.dpi': 600,
'xtick.direction': 'in',
'ytick.direction': 'in',
'xtick.labelsize': 5,
'ytick.labelsize': 5,
'legend.fontsize': 5,
}
plt.rcParams.update(parameters)
fig, ax = plt.subplots(1, 1)
plt.rc('font', family='Times New Roman') # 全局字体样式
font1 = {'family': 'Times New Roman', 'weight': 'normal', 'size': 5} # 设置坐标标签的字体大小，字体
plt.scatter(list(range(result_data.shape[0])), result_data, c='black', s=0.01, label="predict")
# 画出 y=1 这条水平线
plt.axhline(0.5, c='red', label='Failure threshold', lw=1)
# 箭头指向上面的水平线
# plt.arrow(result_data.shape[0]*2/3, 0.55, 2000, 0.085, width=0.00001, ec='red',length_includes_head=True)
# plt.text(result_data.shape[0] * 2 / 3 + 1000, 0.7, "real fault", fontsize=5, color='red',
# verticalalignment='top')
# plt.text(0, 0.55, "Threshold", fontsize=5, color='red',
# verticalalignment='top')
plt.axvline(result_data.shape[0] * 2 / 3-50, c='blue', ls='-.', lw=0.5, label='real fault')
# plt.axvline(415548, c='blue', ls='-.', lw=0.5, label='real fault')
# plt.xticks(range(6), ('06/09/17', '12/09/17', '18/09/17', '24/09/17', '29/09/17')) # 设置x轴的标尺
plt.text(result_data.shape[0] * 5 / 6, 0.4, "Fault", fontsize=5, color='black', verticalalignment='top',
horizontalalignment='center',
bbox={'facecolor': 'grey',
'pad': 1.5,
'linewidth': 0.1}, fontdict=font1)
plt.text(result_data.shape[0] * 1 / 3, 0.6, "Norm", fontsize=5, color='black', verticalalignment='top',
horizontalalignment='center',
bbox={'facecolor': 'grey',
'pad': 1.5, 'linewidth': 0.1}, fontdict=font1)
indices = [result_data.shape[0] * i / 4 for i in range(5)]
classes = ['09/01', '09/08', '09/15', '09/22', '09/29']
indices1 = [i / 4 for i in range(5)]
classes1 = ['0', '0.25', 'Threshold', '0.75', '1']
# 第一个是迭代对象，表示坐标的显示顺序，第二个参数是坐标轴显示列表
plt.xticks([index + 0.5 for index in indices], classes, rotation=25) # 设置横坐标方向，rotation=45为45度倾斜
# pad调整label与坐标轴之间的距离
plt.tick_params(bottom=True, top=False, left=True, right=False, direction='inout', length=2, width=0.5, pad=1)
# plt.yticks([index for index in indices1], classes1)
plt.ylabel('Confidence', fontsize=5)
plt.xlabel('Time', fontsize=5)
plt.tight_layout()
# plt.legend(loc='best', edgecolor='black', fontsize=4)
plt.legend(loc='best', frameon=False, fontsize=4.5)
# plt.grid()
# 局部方法图
axins = inset_axes(ax, width="40%", height="30%", loc='lower left',
bbox_to_anchor=(0.1, 0.1, 1, 1),
bbox_transform=ax.transAxes)
axins.scatter(list(range(result_data.shape[0])), result_data, c='black', s=0.001, label="predict")
axins.axvline(result_data.shape[0] * 2 / 3-50, c='blue', ls='-.', lw=0.5, label='real fault')
plt.axhline(0.5, c='red', label='Failure threshold', lw=0.5)
# 设置放大区间
# 设置放大区间
zone_left = int(result_data.shape[0] * 2 / 3 -100)
zone_right = int(result_data.shape[0] * 2 / 3) + 100
x = list(range(result_data.shape[0]))
# 坐标轴的扩展比例（根据实际数据调整）
x_ratio = 0.5 # x轴显示范围的扩展比例
y_ratio = 0.5 # y轴显示范围的扩展比例
mark_inset(ax, axins, loc1=4, loc2=2, fc="none", ec='k', lw=0.5)
# X轴的显示范围
xlim0 = x[zone_left] - (x[zone_right] - x[zone_left]) * x_ratio
xlim1 = x[zone_right] + (x[zone_right] - x[zone_left]) * x_ratio
# axins.tick_params(bottom=False)
axins.xaxis.set_visible(False)
# Y轴的显示范围
# y = np.hstack((y_1[zone_left:zone_right], y_2[zone_left:zone_right], y_3[zone_left:zone_right]))
# ylim0 = np.min(y) - (np.max(y) - np.min(y)) * y_ratio
# ylim1 = np.max(y) + (np.max(y) - np.min(y)) * y_ratio
# 调整子坐标系的显示范围
axins.set_xlim(xlim0, xlim1)
# axins.set_ylim(ylim0, ylim1)
plt.show()
pass
@ -203,7 +292,7 @@ def plot_Corr(data, size: int = 1):
person = pd_data.corr()
print(person)
# 画热点图heatmap
cmap = sns.heatmap(person, annot=True, cmap='Blues',annot_kws={
cmap = sns.heatmap(person, annot=True, cmap='Blues', annot_kws={
'fontsize': 11
})
classes = ['Gs', 'Gio', 'Gip', 'Gp', 'Gwt', 'En', 'Gft', 'Grt', 'Gwt', 'Et', 'Rs', 'Ap', 'Ws', 'Dw', 'Ges', 'Gt',
@ -234,9 +323,9 @@ def plot_bar(y_data):
'savefig.dpi': 600,
'xtick.direction': 'in',
'ytick.direction': 'in',
'xtick.labelsize': 15,
'ytick.labelsize': 15,
'legend.fontsize': 12,
'xtick.labelsize': 18,
'ytick.labelsize': 20,
'legend.fontsize': 18,
}
plt.rcParams.update(parameters)
plt.figure()
@ -263,12 +352,13 @@ def plot_bar(y_data):
# plt.tick_params(bottom=False, top=False, left=True, right=False, direction='in', pad=1)
plt.xticks([])
plt.ylabel('False Positive Rate(%)', fontsize=18)
# plt.xlabel('Time', fontsize=5)
plt.ylabel('False Positive Rate(%)', fontsize=22)
plt.xlabel('Methods', fontsize=22)
# plt.tight_layout()
num1, num2, num3, num4 = 0.1, 1, 3, 0
plt.legend(bbox_to_anchor=(num1, num2), loc=num3, borderaxespad=num4, ncol=5, frameon=False, markerscale=0.5)
num1, num2, num3, num4 = 0, 1, 3, 0
plt.legend(bbox_to_anchor=(num1, num2), loc=num3, borderaxespad=num4, ncol=5, frameon=False, handlelength=1,
handletextpad=0.45, columnspacing=1)
plt.ylim([-0.01, 5])
plt.show()
@ -282,7 +372,7 @@ def acc(y_data=list):
'ytick.direction': 'in',
'xtick.labelsize': 18,
'ytick.labelsize': 20,
'legend.fontsize': 14,
'legend.fontsize': 18,
}
plt.rcParams.update(parameters)
plt.figure()
@ -308,12 +398,12 @@ def acc(y_data=list):
# plt.tick_params(bottom=False, top=False, left=True, right=False, direction='in', pad=1)
plt.xticks([])
plt.ylabel('Accuracy(%)', fontsize=20)
plt.xlabel('Methods', fontsize=20)
plt.ylabel('Accuracy(%)', fontsize=22)
plt.xlabel('Methods', fontsize=22)
# plt.tight_layout()
num1, num2, num3, num4 = 0, 1, 3, 0
plt.legend(bbox_to_anchor=(num1, num2), loc=num3, borderaxespad=num4, ncol=5, frameon=False, markerscale=0.5)
plt.legend(bbox_to_anchor=(num1, num2), loc=num3, borderaxespad=num4, ncol=5, frameon=False,handlelength=1,handletextpad=0.45,columnspacing=1)
plt.ylim([60, 105])
plt.show()
@ -325,9 +415,9 @@ def plot_FNR1(y_data):
'savefig.dpi': 600,
'xtick.direction': 'in',
'ytick.direction': 'in',
'xtick.labelsize': 15,
'ytick.labelsize': 15,
'legend.fontsize': 12,
'xtick.labelsize': 18,
'ytick.labelsize': 20,
'legend.fontsize': 18,
}
plt.rcParams.update(parameters)
plt.figure()
@ -350,12 +440,16 @@ def plot_FNR1(y_data):
# plt.tick_params(bottom=False, top=False, left=True, right=False, direction='in', pad=1)
plt.xticks([])
plt.ylabel('False Negative Rate(%)', fontsize=18)
# plt.xlabel('Time', fontsize=5)
plt.ylabel('False Negative Rate(%)', fontsize=22)
# plt.tick_params(bottom=False, top=False, left=True, right=False, direction='in', pad=1)
plt.xlabel('Methods', fontsize=22)
# plt.tight_layout()
num1, num2, num3, num4 = 0.16, 1, 3, 0
plt.legend(bbox_to_anchor=(num1, num2), loc=num3, borderaxespad=num4, ncol=5, frameon=False, markerscale=0.5)
num1, num2, num3, num4 = 0.05, 1, 3, 0
plt.legend(bbox_to_anchor=(num1, num2), loc=num3, borderaxespad=num4, ncol=5, frameon=False, handlelength=1,
handletextpad=0.45, columnspacing=1)
plt.ylim([60, 100])
plt.show()
@ -367,9 +461,9 @@ def plot_FNR2(y_data):
'savefig.dpi': 600,
'xtick.direction': 'in',
'ytick.direction': 'in',
'xtick.labelsize': 15,
'ytick.labelsize': 15,
'legend.fontsize': 12,
'xtick.labelsize': 18,
'ytick.labelsize': 20,
'legend.fontsize': 18,
}
plt.rcParams.update(parameters)
plt.figure()
@ -392,12 +486,16 @@ def plot_FNR2(y_data):
# plt.tick_params(bottom=False, top=False, left=True, right=False, direction='in', pad=1)
plt.xticks([])
plt.ylabel('False Negative Rate(%)', fontsize=18)
plt.ylabel('False Negative Rate(%)', fontsize=22)
# plt.xlabel('Time', fontsize=5)
# plt.tight_layout()
num1, num2, num3, num4 = 0.16, 1, 3, 0
plt.legend(bbox_to_anchor=(num1, num2), loc=num3, borderaxespad=num4, ncol=5, frameon=False, markerscale=0.5)
plt.xlabel('Methods', fontsize=22)
# plt.tight_layout()
num1, num2, num3, num4 = 0.08, 1, 3, 0
plt.legend(bbox_to_anchor=(num1, num2), loc=num3, borderaxespad=num4, ncol=5, frameon=False, handlelength=1,
handletextpad=0.45, columnspacing=1)
plt.ylim([0, 5])
plt.show()
@ -415,7 +513,7 @@ def plot_hot(data):
'fraction': 0.6,
'pad': 0.00
})
plt.title("Pred=0,Expc=0", size=18)
plt.title("Pred={},Expc={}".format(i, i), size=18)
plt.axis('off') # 去坐标轴
# 调整色带的标签：
cbar = cmap.collections[0].colorbar
@ -431,7 +529,7 @@ def plot_hot_one(data):
# 画热点图heatmap
# plt.figure(1, figsize=(9, 12))
plt.subplots(figsize=(14,2))
plt.subplots(figsize=(14, 2))
# cbar_kws设置色带，fraction设置大小，pad设置填充
cmap = sns.heatmap(data, annot=False, cmap='Blues', linecolor='black', linewidths=1,
cbar_kws={"orientation": "horizontal",
@ -451,18 +549,21 @@ def plot_hot_one(data):
def plot_mse(file_name="../others_idea/mse"):
mse=np.loadtxt(file_name,delimiter=",")
mse= mse[2000:2338]
mse = np.loadtxt(file_name, delimiter=",")
print(mse.shape)
mse.shape[0]*2/3
# mse = mse[2000:2300]
# mse = mse[1800:2150]
parameters = {
'figure.dpi': 600,
'figure.figsize': (7,5),
'figure.figsize': (2.8, 2),
'savefig.dpi': 600,
'xtick.direction': 'in',
'ytick.direction': 'in',
'xtick.labelsize': 15,
'ytick.labelsize': 15,
'legend.fontsize': 10,
'xtick.labelsize': 6,
'ytick.labelsize': 6,
'legend.fontsize': 5,
}
plt.rcParams.update(parameters)
plt.figure()
@ -471,27 +572,62 @@ def plot_mse(file_name="../others_idea/mse"):
classes = ['13/09/17', '14/09/17', '15/09/17', '16/09/17', '17/09/17']
# plt.xticks([index + 0.5 for index in indices], classes, rotation=25) # 设置横坐标方向，rotation=45为45度倾斜
plt.ylabel('MSE', fontsize=15)
plt.xlabel('Time', fontsize=15)
plt.ylabel('MSE', fontsize=8)
plt.xlabel('Time', fontsize=8)
plt.tight_layout()
plt.plot(mse)
plt.plot(mse, lw=0.5)
plt.show()
def plot_3d():
# 线
fig = plt.figure()
ax = fig.add_subplot(projection='3d')
data,label = read_data(file_name='G:\data\SCADA数据\jb4q_8_delete_total_zero.csv',isNew=False)
print(data)
print(data.shape)
x=range(data.shape[1])
y=range(data.shape[0])
color=[[0.16,0.14,0.13],
[0.89,0.09,0.05],
[0.12,0.56,1.00],
[0.01,0.66,0.62],
[0.63,0.04,0.83],
[0.63,0.32,0.18],
[0.20,0.63,0.79],
[0.50,0.16,0.16],
[0.61,0.4,0.12],
[1.00,0.38,0.00],
[0.53,0.81,0.92],
[0.13,0.55,0.13],
[1.00,0.89,0.52],
[0.44,0.50,0.41],
[0.20,0.63,0.79],
[0.00,0.78,0.55],
[1.00,0.39,0.28],
[0.25,0.41,0.88]]
# c颜色，marker：样式*雪花
ax.plot( xs=y,ys=x, zs=data, c=color, marker="*")
plt.show()
def test_result(file_name: str = result_file_name):
# result_data = np.recfromcsv(file_name)
result_data = np.loadtxt(file_name, delimiter=",")
result_data = np.array(result_data)
print(len(result_data))
theshold = len(result_data) - 1
theshold = len(result_data)
print(theshold)
print(theshold * 2 / 3)
theshold = theshold * 2 / 3-50
# 计算误报率和漏报率
positive_rate = result_data[:int(theshold * 2 / 3)][result_data[:int(theshold * 2 / 3)] < 0.66].__len__() / (
positive_rate = result_data[:int(theshold)][result_data[:int(theshold)] < 0.66].__len__() / (
theshold * 2 / 3)
negative_rate = result_data[int(theshold * 2 / 3):][result_data[int(theshold * 2 / 3):] > 0.66].__len__() / (
negative_rate = result_data[int(theshold):][result_data[int(theshold):] > 0.66].__len__() / (
theshold * 1 / 3)
print("误报率:", positive_rate)
print("漏报率", negative_rate)
@ -501,7 +637,7 @@ def test_result(file_name: str = result_file_name):
# result_data = np.concatenate([result_data, data], axis=0)
print(result_data)
print(result_data.shape)
plot_result(result_data)
plot_result_banda(result_data)
def test_mse(mse_file_name: str = mse_file_name, max_file_name: str = max_file_name):
@ -514,13 +650,12 @@ def test_mse(mse_file_name: str = mse_file_name, max_file_name: str = max_file_n
plot_MSE(mse_data, max_data)
def test_corr(file_name=source_path, N=10):
# needed_data, label = read_data(file_name=file_name, isNew=False)
# print(needed_data)
# print(needed_data.shape)
# np.save("corr.npy",needed_data)
needed_data=np.load("corr.npy")
needed_data = np.load("corr.npy")
# plot_original_data(needed_data)
person = plot_Corr(needed_data)
person = np.array(person)
@ -563,6 +698,7 @@ def test_model_weight_l(model_name=model_name):
pass
def test_model_weight_s(model_name=model_name):
print("===============第一次===================")
model = Joint_Monitoring_SE()
@ -581,7 +717,7 @@ def test_model_weight_s(model_name=model_name):
def test_model_visualization(model_name=file_name):
with open(file_name, 'r', encoding='utf-8') as f:
with open(model_name, 'r', encoding='utf-8') as f:
data = np.loadtxt(f, str, delimiter=',')
needed_data = data[1:, 1:].astype(dtype=np.float)
@ -592,8 +728,9 @@ def test_model_visualization(model_name=file_name):
if __name__ == '__main__':
# test_mse()
# test_result(file_name='E:\self_example\TensorFlow_eaxmple\Model_train_test\condition_monitoring\self_try\compare\mse\RNet_C\RNet_C_timestamp120_feature10_result2.csv')
test_mse()
# test_result(
# file_name='E:\self_example\TensorFlow_eaxmple\Model_train_test\condition_monitoring\self_try\compare\mse\ResNet\ResNet_banda_result1.csv')
# test_corr()
# acc()
# list = [3.77, 2.64, 2.35, 2.05, 1.76, 1.09, 0.757, 0.82, 1.1, 0.58, 0, 0.03, 0.02]
@ -601,16 +738,19 @@ if __name__ == '__main__':
# list=[99.99,98.95,99.95,96.1,95,99.65,76.25,72.64,75.87,68.74]
# plot_FNR1(list)
#
# #
# list=[3.43,1.99,1.92,2.17,1.63,1.81,1.78,1.8,0.6]
# plot_FNR2(list)
# 查看网络某一层的权重
# test_model_visualization()
file_name = "../self_try\compare\model\weight\RNet_L_banda_epoch17_0.0086_0.0092/weight"
test_model_weight_l(file_name)
# file_name = "../self_try\compare\model\weight\RNet_S_epoch3_2.47_1.63/weight"
# test_model_visualization(model_name = "E:\跑模型\论文写作/SE.txt")
# file_name = "../self_try\compare\model\weight\RNet_L_banda\RNet_L_banda_epoch18_0.01_0.0087/weight"
# test_model_weight_l(file_name)
# file_name = "../self_try\compare\model\weight\RNet_s_banda\RNet_S_banda_epoch4_0.0034067521/weight"
# test_model_weight_l(file_name)
#单独预测图
# plot_mse()
# 单独预测图
# plot_mse('E:\self_example\TensorFlow_eaxmple\Model_train_test\condition_monitoring\self_try\compare\mse\RNet_D/banda\RNet_D_banda_mse_predict1.csv')
#画3d图
# plot_3d()

62
TensorFlow_eaxmple/Model_train_test/condition_monitoring/self_try/Joint_Monitoring_banda.py
View File

@ -40,7 +40,7 @@ save_name = "../hard_model/weight/{0}_epoch16_0.0009_0.0014/weight".format(model
feature_num,
batch_size,
EPOCH)
save_step_two_name = "../hard_model/two_weight/{0}/weight".format(model_name,
save_step_two_name = "../hard_model/two_weight/{0}_epoch24_9875_9867/weight".format(model_name,
time_stamp,
feature_num,
batch_size,
@ -56,6 +56,8 @@ save_step_two_name = "../hard_model/two_weight/{0}/weight".format(model_name,
# feature_num,
# batch_size,
# EPOCH)
save_mse_name=r"./compare/mse/JM_banda/{0}_mse.csv".format(model_name)
'''文件名'''
file_name = "G:\data\SCADA数据\SCADA_已处理_粤水电达坂城2020.1月-5月\风机15.csv"
@ -488,7 +490,7 @@ def test(step_one_model, step_two_model, test_data, test_label1, test_label2):
print("val_loss:", val_loss)
def showResult(step_two_model: Joint_Monitoring, test_data, isPlot: bool = False):
def showResult(step_two_model: Joint_Monitoring, test_data, isPlot: bool = False,isSave:bool=True):
# 获取模型的所有参数的个数
# step_two_model.count_params()
total_result = []
@ -499,25 +501,41 @@ def showResult(step_two_model: Joint_Monitoring, test_data, isPlot: bool = False
total_result.append(output4)
total_result = np.reshape(total_result, [total_result.__len__(), -1])
total_result = np.reshape(total_result, [-1, ])
#误报率，漏报率，准确性的计算
if isSave:
np.savetxt(save_mse_name, total_result, delimiter=',')
if isPlot:
plt.figure(1, figsize=(6.0, 2.68))
plt.subplots_adjust(left=0.1, right=0.94, bottom=0.2, top=0.9, wspace=None,
hspace=None)
plt.tight_layout()
font1 = {'family': 'Times New Roman', 'weight': 'normal', 'size': 10} # 设置坐标标签的字体大小，字体
plt.scatter(list(range(total_result.shape[0])), total_result, c='black', s=10)
# 画出 y=1 这条水平线
plt.axhline(0.5, c='red', label='Failure threshold')
# 箭头指向上面的水平线
# plt.arrow(35000, 0.9, 33000, 0.75, head_width=0.02, head_length=0.1, shape="full", fc='red', ec='red',
# alpha=0.9, overhang=0.5)
# plt.text(35000, 0.9, "Truth Fault", fontsize=10, color='black', verticalalignment='top')
plt.text(test_data.shape[0] * 2 / 3+1000, 0.7, "Truth Fault", fontsize=10, color='red', verticalalignment='top')
plt.axvline(test_data.shape[0] * 2 / 3, c='blue', ls='-.')
plt.xlabel("time")
plt.ylabel("confience")
plt.xticks(range(6),('06/09/17','12/09/17','18/09/17','24/09/17','29/09/17')) # 设置x轴的标尺
plt.tick_params() #设置轴显示
plt.xlabel("time",fontdict=font1)
plt.ylabel("confience",fontdict=font1)
plt.text(total_result.shape[0] * 4 / 5, 0.6, "Fault", fontsize=10, color='black', verticalalignment='top',
horizontalalignment='center',
bbox={'facecolor': 'grey',
'pad': 10})
'pad': 10},fontdict=font1)
plt.text(total_result.shape[0] * 1 / 3, 0.4, "Norm", fontsize=10, color='black', verticalalignment='top',
horizontalalignment='center',
bbox={'facecolor': 'grey',
'pad': 10})
'pad': 10},fontdict=font1)
plt.grid()
# plt.ylim(0, 1)
# plt.xlim(-50, 1300)
@ -540,29 +558,29 @@ if __name__ == '__main__':
# train_step_one(train_data=train_data_healthy, train_label1=train_label1_healthy, train_label2=train_label2_healthy)
# 导入第一步已经训练好的模型,一个继续训练，一个只输出结果
step_one_model = Joint_Monitoring()
step_one_model.load_weights(save_name)
step_two_model = Joint_Monitoring()
step_two_model.load_weights(save_name)
# step_one_model = Joint_Monitoring()
# step_one_model.load_weights(save_name)
#
# step_two_model = Joint_Monitoring()
# step_two_model.load_weights(save_name)
#### TODO 第二步训练
### healthy_data.shape: (300333,120,10)
### unhealthy_data.shape: (16594,10)
healthy_size, _, _ = train_data_healthy.shape
unhealthy_size, _, _ = train_data_unhealthy.shape
train_data, train_label1, train_label2, test_data, test_label1, test_label2 = split_test_data(
healthy_data=train_data_healthy[healthy_size - 2 * unhealthy_size:, :, :],
healthy_label1=train_label1_healthy[healthy_size - 2 * unhealthy_size:, :],
healthy_label2=train_label2_healthy[healthy_size - 2 * unhealthy_size:, ], unhealthy_data=train_data_unhealthy,
unhealthy_label1=train_label1_unhealthy, unhealthy_label2=train_label2_unhealthy)
train_step_two(step_one_model=step_one_model, step_two_model=step_two_model,
train_data=train_data,
train_label1=train_label1, train_label2=np.expand_dims(train_label2, axis=-1))
# train_data, train_label1, train_label2, test_data, test_label1, test_label2 = split_test_data(
# healthy_data=train_data_healthy[healthy_size - 2 * unhealthy_size:, :, :],
# healthy_label1=train_label1_healthy[healthy_size - 2 * unhealthy_size:, :],
# healthy_label2=train_label2_healthy[healthy_size - 2 * unhealthy_size:, ], unhealthy_data=train_data_unhealthy,
# unhealthy_label1=train_label1_unhealthy, unhealthy_label2=train_label2_unhealthy)
# train_step_two(step_one_model=step_one_model, step_two_model=step_two_model,
# train_data=train_data,
# train_label1=train_label1, train_label2=np.expand_dims(train_label2, axis=-1))
### TODO 测试测试集
step_one_model = Joint_Monitoring()
step_one_model.load_weights(save_name)
# step_one_model = Joint_Monitoring()
# step_one_model.load_weights(save_name)
step_two_model = Joint_Monitoring()
step_two_model.load_weights(save_step_two_name)
# test(step_one_model=step_one_model, step_two_model=step_two_model, test_data=test_data, test_label1=test_label1,

118
TensorFlow_eaxmple/Model_train_test/condition_monitoring/self_try/compare/DCConv.py
View File

@ -2,16 +2,6 @@
# coding: utf-8
'''
@Author : dingjiawen
@Date : 2022/10/11 18:53
@Usage :
@Desc :
'''
# -*- coding: utf-8 -*-
# coding: utf-8
'''
@Author : dingjiawen
@ -20,9 +10,7 @@
@Desc :
'''
# -*- coding: utf-8 -*-
# coding: utf-8
import tensorflow as tf
import tensorflow.keras
import numpy as np
@ -30,7 +18,7 @@ import pandas as pd
import matplotlib.pyplot as plt
from model.DepthwiseCon1D.DepthwiseConv1D import DepthwiseConv1D
from model.Dynamic_channelAttention.Dynamic_channelAttention import DynamicChannelAttention
from condition_monitoring.data_deal import loadData
from condition_monitoring.data_deal import loadData_daban as loadData
from model.Joint_Monitoring.Joint_Monitoring3 import Joint_Monitoring
from model.CommonFunction.CommonFunction import *
@ -42,16 +30,16 @@ import random
'''超参数设置'''
time_stamp = 120
feature_num = 10
batch_size = 16
batch_size = 32
learning_rate = 0.001
EPOCH = 101
model_name = "DCConv"
model_name = "DCConv_banda"
'''EWMA超参数'''
K = 18
namuda = 0.01
'''保存名称'''
save_name = "./model/{0}_timestamp{1}_feature{2}.h5".format(model_name,
save_name = "./model/{0}_0.02_0.05.h5".format(model_name,
time_stamp,
feature_num,
batch_size,
@ -61,12 +49,12 @@ save_step_two_name = "../hard_model/two_weight/{0}_timestamp{1}_feature{2}_weigh
feature_num,
batch_size,
EPOCH)
save_mse_name = "./mse/{0}/{0}_timestamp{1}_feature{2}_mse.csv".format(model_name,
save_mse_name = "./mse/DCConv/banda/mse.csv".format(model_name,
time_stamp,
feature_num,
batch_size,
EPOCH)
save_max_name = "./mse/{0}/{0}_timestamp{1}_feature{2}_max.csv".format(model_name,
save_max_name = "./mse/DCConv/banda/max.csv".format(model_name,
time_stamp,
feature_num,
batch_size,
@ -83,7 +71,8 @@ save_max_name = "./mse/{0}/{0}_timestamp{1}_feature{2}_max.csv".format(model_nam
# batch_size,
# EPOCH)
'''文件名'''
file_name = "G:\data\SCADA数据\jb4q_8_delete_total_zero.csv"
'''文件名'''
file_name = "G:\data\SCADA数据\SCADA_已处理_粤水电达坂城2020.1月-5月\风机15.csv"
'''
文件说明：jb4q_8_delete_total_zero.csv是删除了只删除了全是0的列的文件
@ -92,9 +81,9 @@ file_name = "G:\data\SCADA数据\jb4q_8_delete_total_zero.csv"
'''
'''文件参数'''
# 最后正常的时间点
healthy_date = 415548
healthy_date = 96748
# 最后异常的时间点
unhealthy_date = 432153
unhealthy_date = 107116
# 异常容忍程度
unhealthy_patience = 5
@ -207,46 +196,7 @@ def EWMA(data, K=K, namuda=namuda):
pass
def get_MSE(data, label, new_model, isStandard: bool = True, isPlot: bool = True):
predicted_data = new_model.predict(data)
temp = np.abs(predicted_data - label)
temp1 = (temp - np.broadcast_to(np.mean(temp, axis=0), shape=predicted_data.shape))
temp2 = np.broadcast_to(np.sqrt(np.var(temp, axis=0)), shape=predicted_data.shape)
temp3 = temp1 / temp2
mse = np.sum((temp1 / temp2) ** 2, axis=1)
print("z:", mse)
print(mse.shape)
# mse=np.mean((predicted_data-label)**2,axis=1)
print("mse", mse)
if isStandard:
dims, = mse.shape
mean = np.mean(mse)
std = np.sqrt(np.var(mse))
max = mean + 3 * std
print("max:", max)
# min = mean-3*std
max = np.broadcast_to(max, shape=[dims, ])
# min = np.broadcast_to(min,shape=[dims,])
mean = np.broadcast_to(mean, shape=[dims, ])
if isPlot:
plt.figure(random.randint(1,9))
plt.plot(max)
plt.plot(mse)
plt.plot(mean)
# plt.plot(min)
plt.show()
else:
if isPlot:
plt.figure(random.randint(1, 9))
plt.plot(mse)
# plt.plot(min)
plt.show()
return mse
return mse, mean, max
# pass
def condition_monitoring_model():
@ -383,6 +333,48 @@ def DCConv_Model():
pass
def get_MSE(data, label, new_model, isStandard: bool = True, isPlot: bool = True, predictI: int = 1):
predicted_data = new_model.predict(data)
temp = np.abs(predicted_data - label)
temp1 = (temp - np.broadcast_to(np.mean(temp, axis=0), shape=predicted_data.shape))
temp2 = np.broadcast_to(np.sqrt(np.var(temp, axis=0)), shape=predicted_data.shape)
temp3 = temp1 / temp2
mse = np.sum((temp1 / temp2) ** 2, axis=1)
print("z:", mse)
print(mse.shape)
# mse=np.mean((predicted_data-label)**2,axis=1)
print("mse", mse)
if isStandard:
dims, = mse.shape
mean = np.mean(mse)
std = np.sqrt(np.var(mse))
max = mean + 3 * std
print("max:", max)
# min = mean-3*std
max = np.broadcast_to(max, shape=[dims, ])
# min = np.broadcast_to(min,shape=[dims,])
mean = np.broadcast_to(mean, shape=[dims, ])
if isPlot:
plt.figure(random.randint(1,9))
plt.plot(max)
plt.plot(mse)
plt.plot(mean)
# plt.plot(min)
plt.show()
else:
if isPlot:
plt.figure(random.randint(1, 9))
plt.plot(mse)
# plt.plot(min)
plt.show()
return mse
return mse, mean, max
# pass
# healthy_data是健康数据,用于确定阈值，all_data是完整的数据,用于模型出结果
def getResult(model: tf.keras.Model, healthy_data, healthy_label, unhealthy_data, unhealthy_label, isPlot: bool = False,
isSave: bool = True, predictI: int = 1):
@ -454,7 +446,7 @@ if __name__ == '__main__':
model.compile(optimizer=tf.optimizers.Adam(0.01), loss=tf.losses.mse)
model.summary()
early_stop = EarlyStopping(monitor='val_loss', min_delta=0.0001, patience=3, mode='min', verbose=1)
early_stop = EarlyStopping(monitor='val_loss', min_delta=0.0001, patience=20, mode='min', verbose=1)
checkpoint = tf.keras.callbacks.ModelCheckpoint(
filepath=save_name,
@ -463,8 +455,8 @@ if __name__ == '__main__':
save_best_only=True,
mode='min',
period=1)
# history = model.fit(train_data_healthy[:30000, :, :], train_label1_healthy[:30000, :], epochs=20,
#
# history = model.fit(train_data_healthy[:40000, :, :], train_label1_healthy[:40000, :], epochs=20,
# batch_size=32, validation_split=0.2, shuffle=False, verbose=1,
# callbacks=[checkpoint, early_stop])
# model.save(save_name)

148
TensorFlow_eaxmple/Model_train_test/condition_monitoring/self_try/compare/GRU-CNN.py
View File

@ -15,7 +15,7 @@ import pandas as pd
import matplotlib.pyplot as plt
from model.DepthwiseCon1D.DepthwiseConv1D import DepthwiseConv1D
from model.Dynamic_channelAttention.Dynamic_channelAttention import DynamicChannelAttention
from condition_monitoring.data_deal import loadData
from condition_monitoring.data_deal import loadData_daban as loadData
from model.Joint_Monitoring.Joint_Monitoring3 import Joint_Monitoring
from model.CommonFunction.CommonFunction import *
@ -30,13 +30,13 @@ feature_num = 10
batch_size = 16
learning_rate = 0.001
EPOCH = 101
model_name = "DCNN-GRU"
model_name = "CNN-GRU_banda"
'''EWMA超参数'''
K = 18
namuda = 0.01
'''保存名称'''
save_name = "./model/{0}_timestamp{1}_feature{2}.h5".format(model_name,
save_name = "./model/{0}_epoch1_0.023_0.04.h5".format(model_name,
time_stamp,
feature_num,
batch_size,
@ -47,6 +47,17 @@ save_step_two_name = "../hard_model/two_weight/{0}_timestamp{1}_feature{2}_weigh
batch_size,
EPOCH)
save_mse_name = "./mse/CNN_GRU/banda/mse.csv".format(model_name,
time_stamp,
feature_num,
batch_size,
EPOCH)
save_max_name = "./mse/CNN_GRU/banda/max.csv".format(model_name,
time_stamp,
feature_num,
batch_size,
EPOCH)
# save_name = "../model/joint/{0}_timestamp{1}_feature{2}.h5".format(model_name,
# time_stamp,
# feature_num,
@ -58,7 +69,8 @@ save_step_two_name = "../hard_model/two_weight/{0}_timestamp{1}_feature{2}_weigh
# batch_size,
# EPOCH)
'''文件名'''
file_name = "G:\data\SCADA数据\jb4q_8_delete_total_zero.csv"
'''文件名'''
file_name = "G:\data\SCADA数据\SCADA_已处理_粤水电达坂城2020.1月-5月\风机15.csv"
'''
文件说明：jb4q_8_delete_total_zero.csv是删除了只删除了全是0的列的文件
@ -67,13 +79,14 @@ file_name = "G:\data\SCADA数据\jb4q_8_delete_total_zero.csv"
'''
'''文件参数'''
# 最后正常的时间点
healthy_date = 415548
healthy_date = 96748
# 最后异常的时间点
unhealthy_date = 432153
unhealthy_date = 107116
# 异常容忍程度
unhealthy_patience = 5
def remove(data, time_stamp=time_stamp):
rows, cols = data.shape
print("remove_data.shape:", data.shape)
@ -182,46 +195,6 @@ def EWMA(data, K=K, namuda=namuda):
pass
def get_MSE(data, label, new_model, isStandard: bool = True, isPlot: bool = True):
predicted_data = new_model.predict(data)
temp = np.abs(predicted_data - label)
temp1 = (temp - np.broadcast_to(np.mean(temp, axis=0), shape=predicted_data.shape))
temp2 = np.broadcast_to(np.sqrt(np.var(temp, axis=0)), shape=predicted_data.shape)
temp3 = temp1 / temp2
mse = np.sum((temp1 / temp2) ** 2, axis=1)
print("z:", mse)
print(mse.shape)
# mse=np.mean((predicted_data-label)**2,axis=1)
print("mse", mse)
if isStandard:
dims, = mse.shape
mean = np.mean(mse)
std = np.sqrt(np.var(mse))
max = mean + 3 * std
print("max:", max)
# min = mean-3*std
max = np.broadcast_to(max, shape=[dims, ])
# min = np.broadcast_to(min,shape=[dims,])
mean = np.broadcast_to(mean, shape=[dims, ])
if isPlot:
plt.figure(random.randint(1,9))
plt.plot(max)
plt.plot(mse)
plt.plot(mean)
# plt.plot(min)
plt.show()
else:
if isPlot:
plt.figure(random.randint(1, 9))
plt.plot(mse)
# plt.plot(min)
plt.show()
return mse
return mse, mean, max
# pass
def condition_monitoring_model():
@ -357,9 +330,52 @@ def CNN_GRU_Model():
pass
def get_MSE(data, label, new_model, isStandard: bool = True, isPlot: bool = True, predictI: int = 1):
predicted_data = new_model.predict(data)
temp = np.abs(predicted_data - label)
temp1 = (temp - np.broadcast_to(np.mean(temp, axis=0), shape=predicted_data.shape))
temp2 = np.broadcast_to(np.sqrt(np.var(temp, axis=0)), shape=predicted_data.shape)
temp3 = temp1 / temp2
mse = np.sum((temp1 / temp2) ** 2, axis=1)
print("z:", mse)
print(mse.shape)
# mse=np.mean((predicted_data-label)**2,axis=1)
print("mse", mse)
if isStandard:
dims, = mse.shape
mean = np.mean(mse)
std = np.sqrt(np.var(mse))
max = mean + 3 * std
print("max:", max)
# min = mean-3*std
max = np.broadcast_to(max, shape=[dims, ])
# min = np.broadcast_to(min,shape=[dims,])
mean = np.broadcast_to(mean, shape=[dims, ])
if isPlot:
plt.figure(random.randint(1,9))
plt.plot(max)
plt.plot(mse)
plt.plot(mean)
# plt.plot(min)
plt.show()
else:
if isPlot:
plt.figure(random.randint(1, 9))
plt.plot(mse)
# plt.plot(min)
plt.show()
return mse
return mse, mean, max
# pass
# healthy_data是健康数据,用于确定阈值，all_data是完整的数据,用于模型出结果
def getResult(model: tf.keras.Model, healthy_data, healthy_label, unhealthy_data, unhealthy_label, isPlot: bool = False,
isSave: bool = False):
isSave: bool = True, predictI: int = 1):
# TODO 计算MSE确定阈值
# TODO 计算MSE确定阈值
mse, mean, max = get_MSE(healthy_data, healthy_label, model)
@ -368,7 +384,7 @@ def getResult(model: tf.keras.Model, healthy_data, healthy_label, unhealthy_data
total, = mse.shape
faultNum = 0
faultList = []
faultNum=mse[mse[:]>max[0]].__len__()
faultNum = mse[mse[:] > max[0]].__len__()
# for i in range(total):
# if (mse[i] > max[i]):
# faultNum += 1
@ -379,21 +395,37 @@ def getResult(model: tf.keras.Model, healthy_data, healthy_label, unhealthy_data
# 漏报率计算
missNum = 0
missList = []
mse1 = get_MSE(unhealthy_data, unhealthy_label, model, isStandard=False)
total_mse = np.concatenate([mse, mse1], axis=0)
total_max = np.broadcast_to(max[0], shape=[total_mse.shape[0], ])
# min = np.broadcast_to(min,shape=[dims,])
total_mean = np.broadcast_to(mean[0], shape=[total_mse.shape[0], ])
if isSave:
save_mse_name1 = save_mse_name
save_max_name1 = save_max_name
np.savetxt(save_mse_name1, total_mse, delimiter=',')
np.savetxt(save_max_name1, total_max, delimiter=',')
all, = mse1.shape
flag=True
for i in range(all):
if (mse1[i] < max[0] and flag) :
missNum += 1
missList.append(mse1[i])
elif(mse1[i]>max[0]):
print(i)
flag=False
print("all:",all)
missNum = mse1[mse1[:] < max[0]].__len__()
print("all:", all)
miss_rate = missNum / all
print("漏报率:", miss_rate)
plt.figure(random.randint(1, 100))
plt.plot(total_max)
plt.plot(total_mse)
plt.plot(total_mean)
# plt.plot(min)
plt.show()
pass

35
TensorFlow_eaxmple/Model_train_test/condition_monitoring/self_try/compare/GRU.py
View File

@ -16,7 +16,7 @@ import pandas as pd
import matplotlib.pyplot as plt
from model.DepthwiseCon1D.DepthwiseConv1D import DepthwiseConv1D
from model.Dynamic_channelAttention.Dynamic_channelAttention import DynamicChannelAttention
from condition_monitoring.data_deal import loadData
from condition_monitoring.data_deal import loadData_daban as loadData
from model.Joint_Monitoring.Joint_Monitoring3 import Joint_Monitoring
from model.CommonFunction.CommonFunction import *
@ -28,16 +28,16 @@ import random
'''超参数设置'''
time_stamp = 120
feature_num = 10
batch_size = 16
batch_size = 32
learning_rate = 0.001
EPOCH = 101
model_name = "GRU"
model_name = "GRU_banda"
'''EWMA超参数'''
K = 18
namuda = 0.01
'''保存名称'''
save_name = "./model/{0}_timestamp{1}_feature{2}_epoch1_loss0.005.h5".format(model_name,
save_name = "./model/{0}_epoch3_0.0138_0.0147.h5".format(model_name,
time_stamp,
feature_num,
batch_size,
@ -47,12 +47,12 @@ save_step_two_name = "../hard_model/two_weight/{0}_timestamp{1}_feature{2}_weigh
feature_num,
batch_size,
EPOCH)
save_mse_name = "./mse/{0}/{0}_timestamp{1}_feature{2}_mse.csv".format(model_name,
save_mse_name = "./mse/GRU/banda/mse.csv".format(model_name,
time_stamp,
feature_num,
batch_size,
EPOCH)
save_max_name = "./mse/{0}/{0}_timestamp{1}_feature{2}_max.csv".format(model_name,
save_max_name = "./mse/GRU/banda/max.csv".format(model_name,
time_stamp,
feature_num,
batch_size,
@ -70,7 +70,8 @@ save_max_name = "./mse/{0}/{0}_timestamp{1}_feature{2}_max.csv".format(model_nam
# batch_size,
# EPOCH)
'''文件名'''
file_name = "G:\data\SCADA数据\jb4q_8_delete_total_zero.csv"
'''文件名'''
file_name = "G:\data\SCADA数据\SCADA_已处理_粤水电达坂城2020.1月-5月\风机15.csv"
'''
文件说明：jb4q_8_delete_total_zero.csv是删除了只删除了全是0的列的文件
@ -79,9 +80,9 @@ file_name = "G:\data\SCADA数据\jb4q_8_delete_total_zero.csv"
'''
'''文件参数'''
# 最后正常的时间点
healthy_date = 415548
healthy_date = 96748
# 最后异常的时间点
unhealthy_date = 432153
unhealthy_date = 107116
# 异常容忍程度
unhealthy_patience = 5
@ -437,17 +438,17 @@ if __name__ == '__main__':
# 单次测试
model = GRU_Model()
model.compile(optimizer=tf.optimizers.Adam(0.01), loss=tf.losses.mse)
model.compile(optimizer=tf.optimizers.Adam(0.001), loss=tf.losses.mse)
model.summary()
early_stop = EarlyStopping(monitor='val_loss', min_delta=0.0001, patience=5, mode='min', verbose=1)
# checkpoint = tf.keras.callbacks.ModelCheckpoint(
# filepath=save_name,
# monitor='val_loss',
# verbose=1,
# save_best_only=True,
# mode='min',
# period=1)
checkpoint = tf.keras.callbacks.ModelCheckpoint(
filepath=save_name,
monitor='val_loss',
verbose=1,
save_best_only=True,
mode='min',
period=1)
# history = model.fit(train_data_healthy[:30000,:,:], train_label1_healthy[:30000,:], epochs=10,
# batch_size=32,validation_split=0.2,shuffle=False, verbose=1,callbacks=[checkpoint,early_stop])

102
TensorFlow_eaxmple/Model_train_test/condition_monitoring/self_try/compare/RNet-C.py
View File

@ -19,7 +19,7 @@ import pandas as pd
import matplotlib.pyplot as plt
from model.DepthwiseCon1D.DepthwiseConv1D import DepthwiseConv1D
from model.Dynamic_channelAttention.Dynamic_channelAttention import DynamicChannelAttention
from condition_monitoring.data_deal import loadData
from condition_monitoring.data_deal import loadData_daban as loadData
from model.Joint_Monitoring.compare.RNet_C import Joint_Monitoring
from model.CommonFunction.CommonFunction import *
@ -33,51 +33,50 @@ feature_num = 10
batch_size = 16
learning_rate = 0.001
EPOCH = 101
model_name = "RNet_C"
model_name = "RNet_C_banda"
'''EWMA超参数'''
K = 18
namuda = 0.01
'''保存名称'''
save_name = "./model/weight/{0}/weight".format(model_name,
time_stamp,
feature_num,
batch_size,
EPOCH)
save_step_two_name = "./model/two_weight/{0}_weight_epoch5_9991_9995/weight".format(model_name,
time_stamp,
feature_num,
batch_size,
EPOCH)
time_stamp,
feature_num,
batch_size,
EPOCH)
save_step_two_name = "./model/two_weight/{0}_epoch15_986_985/weight".format(model_name,
time_stamp,
feature_num,
batch_size,
EPOCH)
save_mse_name = "./mse/RNet_C/{0}_timestamp{1}_feature{2}_result2.csv".format(model_name,
time_stamp,
feature_num,
batch_size,
EPOCH)
save_max_name = "./mse/RNet_C/{0}_timestamp{1}_feature{2}_max.csv".format(model_name,
time_stamp,
feature_num,
batch_size,
EPOCH)
save_mse_name = "./mse/{0}/{0}_result1.csv".format(model_name,
time_stamp,
feature_num,
batch_size,
EPOCH)
save_max_name = "./mse/{0}/{0}_max.csv".format(model_name,
time_stamp,
feature_num,
batch_size,
EPOCH)
'''文件名'''
file_name = "G:\data\SCADA数据\jb4q_8_delete_total_zero.csv"
file_name = "G:\data\SCADA数据\SCADA_已处理_粤水电达坂城2020.1月-5月\风机15.csv"
'''
文件说明：jb4q_8_delete_total_zero.csv是删除了只删除了全是0的列的文件
文件从0:415548行均是正常值(2019/7.30 00:00:00 - 2019/9/18 11:14:00)
从415549:432153行均是异常值(2019/9/18 11:21:01 - 2021/1/18 00:00:00)
文件从0:96748行均是正常值(2019/12.30 00:00:00 - 2020/3/11 05:58:00)
从96748:107116行均是异常值(2020/3/11 05:58:01 - 2021/3/18 11:04:00)
'''
'''文件参数'''
# 最后正常的时间点
healthy_date = 415548
healthy_date = 96748
# 最后异常的时间点
unhealthy_date = 432153
unhealthy_date = 107116
# 异常容忍程度
unhealthy_patience = 5
# 画图相关设置
font1 = {'family': 'Times New Roman', 'weight': 'normal', 'size': 10} # 设置坐标标签的字体大小，字体
@ -243,9 +242,6 @@ def EWMA(data, K=K, namuda=namuda):
pass
def condition_monitoring_model():
input = tf.keras.Input(shape=[time_stamp, feature_num])
conv1 = tf.keras.layers.Conv1D(filters=256, kernel_size=1)(input)
@ -461,7 +457,7 @@ def test(step_one_model, step_two_model, test_data, test_label1, test_label2):
print("val_loss:", val_loss)
def showResult(step_two_model: Joint_Monitoring, test_data, isPlot: bool = False,isSave:bool=True):
def showResult(step_two_model: Joint_Monitoring, test_data, isPlot: bool = False, isSave: bool = True):
# 获取模型的所有参数的个数
# step_two_model.count_params()
total_result = []
@ -473,11 +469,9 @@ def showResult(step_two_model: Joint_Monitoring, test_data, isPlot: bool = False
total_result = np.reshape(total_result, [total_result.__len__(), -1])
total_result = np.reshape(total_result, [-1, ])
#误报率，漏报率，准确性的计算
# 误报率，漏报率，准确性的计算
if isSave:
np.savetxt(save_mse_name, total_result, delimiter=',')
if isPlot:
plt.figure(1, figsize=(6.0, 2.68))
@ -492,20 +486,21 @@ def showResult(step_two_model: Joint_Monitoring, test_data, isPlot: bool = False
# 箭头指向上面的水平线
# plt.arrow(35000, 0.9, 33000, 0.75, head_width=0.02, head_length=0.1, shape="full", fc='red', ec='red',
# alpha=0.9, overhang=0.5)
plt.text(test_data.shape[0] * 2 / 3+1000, 0.7, "Truth Fault", fontsize=10, color='red', verticalalignment='top')
plt.text(test_data.shape[0] * 2 / 3 + 1000, 0.7, "Truth Fault", fontsize=10, color='red',
verticalalignment='top')
plt.axvline(test_data.shape[0] * 2 / 3, c='blue', ls='-.')
plt.xticks(range(6),('06/09/17','12/09/17','18/09/17','24/09/17','29/09/17')) # 设置x轴的标尺
plt.tick_params() #设置轴显示
plt.xlabel("time",fontdict=font1)
plt.ylabel("confience",fontdict=font1)
plt.xticks(range(6), ('06/09/17', '12/09/17', '18/09/17', '24/09/17', '29/09/17')) # 设置x轴的标尺
plt.tick_params() # 设置轴显示
plt.xlabel("time", fontdict=font1)
plt.ylabel("confience", fontdict=font1)
plt.text(total_result.shape[0] * 4 / 5, 0.6, "Fault", fontsize=10, color='black', verticalalignment='top',
horizontalalignment='center',
bbox={'facecolor': 'grey',
'pad': 10},fontdict=font1)
'pad': 10}, fontdict=font1)
plt.text(total_result.shape[0] * 1 / 3, 0.4, "Norm", fontsize=10, color='black', verticalalignment='top',
horizontalalignment='center',
bbox={'facecolor': 'grey',
'pad': 10},fontdict=font1)
'pad': 10}, fontdict=font1)
plt.grid()
# plt.ylim(0, 1)
@ -598,7 +593,7 @@ def getResult(model: tf.keras.Model, healthy_data, healthy_label, unhealthy_data
mseList, meanList, maxList = get_MSE(healthy_data, healthy_label, model)
mse1List, _, _ = get_MSE(unhealthy_data, unhealthy_label, model, isStandard=False)
for mse, mean, max, mse1,j in zip(mseList, meanList, maxList, mse1List,range(3)):
for mse, mean, max, mse1, j in zip(mseList, meanList, maxList, mse1List, range(3)):
# 误报率的计算
total, = mse.shape
@ -634,12 +629,11 @@ def getResult(model: tf.keras.Model, healthy_data, healthy_label, unhealthy_data
total_mean = np.broadcast_to(mean[0], shape=[total_mse.shape[0], ])
if isSave:
save_mse_name1=save_mse_name[:-4]+"_predict"+str(j+1)+".csv"
save_max_name1=save_max_name[:-4]+"_predict"+str(j+1)+".csv"
np.savetxt(save_mse_name1,total_mse, delimiter=',')
np.savetxt(save_max_name1,total_max, delimiter=',')
save_mse_name1 = save_mse_name[:-4] + "_predict" + str(j + 1) + ".csv"
save_max_name1 = save_max_name[:-4] + "_predict" + str(j + 1) + ".csv"
np.savetxt(save_mse_name1, total_mse, delimiter=',')
np.savetxt(save_max_name1, total_max, delimiter=',')
plt.figure(random.randint(1, 100))
plt.plot(total_max)
@ -651,8 +645,8 @@ def getResult(model: tf.keras.Model, healthy_data, healthy_label, unhealthy_data
if __name__ == '__main__':
# total_data = loadData.execute(N=feature_num, file_name=file_name)
total_data = np.load("G:\data\SCADA数据\靖边8号处理后的数据\原始10SCADA数据/total_data.npy")
total_data = loadData.execute(N=feature_num, file_name=file_name)
# total_data = np.load("G:\data\SCADA数据\靖边8号处理后的数据\原始10SCADA数据/total_data.npy")
total_data = normalization(data=total_data)
train_data_healthy, train_label1_healthy, train_label2_healthy = get_training_data_overlapping(
total_data[:healthy_date, :], is_Healthy=True)
@ -665,10 +659,10 @@ if __name__ == '__main__':
# train_step_one(train_data=train_data_healthy, train_label1=train_label1_healthy, train_label2=train_label2_healthy)
# 导入第一步已经训练好的模型,一个继续训练，一个只输出结果
# step_one_model = Joint_Monitoring()
# # step_one_model.load_weights(save_name)
# #
# step_two_model = Joint_Monitoring()
step_one_model = Joint_Monitoring()
# step_one_model.load_weights(save_name)
#
step_two_model = Joint_Monitoring()
# step_two_model.load_weights(save_name)
#### TODO 第二步训练
@ -714,4 +708,4 @@ if __name__ == '__main__':
# showResult(step_two_model, test_data=all_data[:32], isPlot=True)
showResult(step_two_model, test_data=all_data, isPlot=True)
pass
pass

21
TensorFlow_eaxmple/Model_train_test/condition_monitoring/self_try/compare/RNet-D.py
View File

@ -16,8 +16,8 @@ import pandas as pd
import matplotlib.pyplot as plt
from model.DepthwiseCon1D.DepthwiseConv1D import DepthwiseConv1D
from model.Dynamic_channelAttention.Dynamic_channelAttention import DynamicChannelAttention
from condition_monitoring.data_deal import loadData
from model.Joint_Monitoring.Joint_Monitoring3 import Joint_Monitoring
from condition_monitoring.data_deal import loadData_daban as loadData
from model.Joint_Monitoring.Joint_Monitoring_banda import Joint_Monitoring
from model.CommonFunction.CommonFunction import *
from sklearn.model_selection import train_test_split
@ -27,15 +27,15 @@ import random
'''超参数设置'''
time_stamp = 120
feature_num = 10
batch_size = 16
batch_size = 32
learning_rate = 0.001
EPOCH = 101
model_name = "RNet_D"
model_name = "RNet_D_banda"
'''EWMA超参数'''
K = 18
namuda = 0.01
'''保存名称'''
save_name = "E:\self_example\TensorFlow_eaxmple\Model_train_test\condition_monitoring\hard_model\weight\joint_timestamp120_feature10_weight_epoch11_0.0077/weight"
save_name = "E:\self_example\TensorFlow_eaxmple\Model_train_test\condition_monitoring\hard_model\weight/banda_joint_epoch16_0.0009_0.0014/weight"
# save_name = "../hard_model/weight/{0}_timestamp{1}_feature{2}_weight_epoch11_0.0077/weight".format(model_name,
# time_stamp,
# feature_num,
@ -47,12 +47,12 @@ save_step_two_name = "../hard_model/two_weight/{0}_timestamp{1}_feature{2}_weigh
batch_size,
EPOCH)
save_mse_name = "./mse/RNet_D/{0}_timestamp{1}_feature{2}_mse.csv".format(model_name,
save_mse_name = "./mse/RNet_D/banda/{0}_mse.csv".format(model_name,
time_stamp,
feature_num,
batch_size,
EPOCH)
save_max_name = "./mse/RNet_D/{0}_timestamp{1}_feature{2}_max.csv".format(model_name,
save_max_name = "./mse/RNet_D/banda/{0}_max.csv".format(model_name,
time_stamp,
feature_num,
batch_size,
@ -69,7 +69,8 @@ save_max_name = "./mse/RNet_D/{0}_timestamp{1}_feature{2}_max.csv".format(model_
# batch_size,
# EPOCH)
'''文件名'''
file_name = "G:\data\SCADA数据\jb4q_8_delete_total_zero.csv"
'''文件名'''
file_name = "G:\data\SCADA数据\SCADA_已处理_粤水电达坂城2020.1月-5月\风机15.csv"
'''
文件说明：jb4q_8_delete_total_zero.csv是删除了只删除了全是0的列的文件
@ -78,9 +79,9 @@ file_name = "G:\data\SCADA数据\jb4q_8_delete_total_zero.csv"
'''
'''文件参数'''
# 最后正常的时间点
healthy_date = 415548
healthy_date = 96748
# 最后异常的时间点
unhealthy_date = 432153
unhealthy_date = 107116
# 异常容忍程度
unhealthy_patience = 5

39
TensorFlow_eaxmple/Model_train_test/condition_monitoring/self_try/compare/RNet-L.py
View File

@ -38,27 +38,27 @@ K = 18
namuda = 0.01
'''保存名称'''
save_name = "./model/weight/{0}/weight".format(model_name,
time_stamp,
feature_num,
batch_size,
EPOCH)
save_name = "./model/weight/{0}/RNet_L_banda_epoch18_0.01_0.0087/weight".format(model_name,
time_stamp,
feature_num,
batch_size,
EPOCH)
save_step_two_name = "./model/two_weight/{0}_timestamp{1}_feature{2}_weight/weight".format(model_name,
time_stamp,
feature_num,
batch_size,
EPOCH)
save_mse_name = "./mse/{0}/{0}_timestamp{1}_feature{2}_mse.csv".format(model_name,
time_stamp,
feature_num,
batch_size,
EPOCH)
save_max_name = "./mse/{0}/{0}_timestamp{1}_feature{2}_max.csv".format(model_name,
time_stamp,
feature_num,
batch_size,
EPOCH)
save_mse_name = "./mse/RNet_L/banda/{0}_mse.csv".format(model_name,
time_stamp,
feature_num,
batch_size,
EPOCH)
save_max_name = "./mse/RNet_L/banda/{0}_max.csv".format(model_name,
time_stamp,
feature_num,
batch_size,
EPOCH)
# save_name = "../model/joint/{0}_timestamp{1}_feature{2}.h5".format(model_name,
# time_stamp,
@ -642,7 +642,7 @@ if __name__ == '__main__':
# train_step_one(train_data=train_data_healthy[:256, :, :], train_label1=train_label1_healthy[:256, :],
# train_label2=train_label2_healthy[:256, ])
#### 模型训练
train_step_one(train_data=train_data_healthy, train_label1=train_label1_healthy, train_label2=train_label2_healthy)
# train_step_one(train_data=train_data_healthy, train_label1=train_label1_healthy, train_label2=train_label2_healthy)
# 导入第一步已经训练好的模型,一个继续训练，一个只输出结果
step_one_model = Joint_Monitoring()
@ -670,12 +670,5 @@ if __name__ == '__main__':
healthy_label=train_label1_healthy[healthy_size - 2 * unhealthy_size:, :],
unhealthy_data=train_data_unhealthy, unhealthy_label=train_label1_unhealthy, isSave=True)
# ###TODO 展示全部的结果
# all_data, _, _ = get_training_data_overlapping(
# total_data[healthy_size - 2 * unhealthy_size:unhealthy_date, :], is_Healthy=True)
# # all_data = np.concatenate([])
# # 单次测试
# # showResult(step_two_model, test_data=all_data[:32], isPlot=True)
# showResult(step_two_model, test_data=all_data, isPlot=True)
pass

25
TensorFlow_eaxmple/Model_train_test/condition_monitoring/self_try/compare/RNet-MSE.py
View File

@ -16,7 +16,7 @@ import pandas as pd
import matplotlib.pyplot as plt
from model.DepthwiseCon1D.DepthwiseConv1D import DepthwiseConv1D
from model.Dynamic_channelAttention.Dynamic_channelAttention import DynamicChannelAttention
from condition_monitoring.data_deal import loadData
from condition_monitoring.data_deal import loadData_daban as loadData
from model.Joint_Monitoring.compare.RNet_MSE import Joint_Monitoring
from model.CommonFunction.CommonFunction import *
@ -27,16 +27,16 @@ import random
'''超参数设置'''
time_stamp = 120
feature_num = 10
batch_size = 16
batch_size = 32
learning_rate = 0.001
EPOCH = 101
model_name = "RNet_MSE"
model_name = "RNet_MSE_banda"
'''EWMA超参数'''
K = 18
namuda = 0.01
'''保存名称'''
save_name = "./model/weight/{0}_weight/weight".format(model_name,
save_name = "./model/weight/{0}/RNet_MSE_banda_epoch17_0.0030094008/weight".format(model_name,
time_stamp,
feature_num,
batch_size,
@ -48,12 +48,12 @@ save_step_two_name = "./model/two_weight/{0}_timestamp{1}_feature{2}_weight/weig
EPOCH)
save_mse_name = "./mse/{0}/{0}_timestamp{1}_feature{2}_mse.csv".format(model_name,
save_mse_name = "./mse/RNet_MSE/banda/{0}_mse.csv".format(model_name,
time_stamp,
feature_num,
batch_size,
EPOCH)
save_max_name = "./mse/{0}/{0}_timestamp{1}_feature{2}_max.csv".format(model_name,
save_max_name = "./mse/RNet_MSE/banda/{0}_max.csv".format(model_name,
time_stamp,
feature_num,
batch_size,
@ -69,7 +69,8 @@ save_max_name = "./mse/{0}/{0}_timestamp{1}_feature{2}_max.csv".format(model_nam
# batch_size,
# EPOCH)
'''文件名'''
file_name = "G:\data\SCADA数据\jb4q_8_delete_total_zero.csv"
'''文件名'''
file_name = "G:\data\SCADA数据\SCADA_已处理_粤水电达坂城2020.1月-5月\风机15.csv"
'''
文件说明：jb4q_8_delete_total_zero.csv是删除了只删除了全是0的列的文件
@ -78,9 +79,9 @@ file_name = "G:\data\SCADA数据\jb4q_8_delete_total_zero.csv"
'''
'''文件参数'''
# 最后正常的时间点
healthy_date = 415548
healthy_date = 96748
# 最后异常的时间点
unhealthy_date = 432153
unhealthy_date = 107116
# 异常容忍程度
unhealthy_patience = 5
@ -348,7 +349,7 @@ def train_step_one(train_data, train_label1, train_label2):
k = k + 1
val_loss, val_accuracy = model.get_val_loss(val_data=val_data, val_label1=val_label1, val_label2=val_label2,
is_first_time=True)
SaveBestModel(model=model, save_name=save_name, history_loss=history_val_loss, loss_value=val_loss.numpy())
SaveBestModel(model=model, save_name=save_name, history_loss=history_val_loss, loss_value=val_loss.numpy(),epoch=epoch,is_all=True)
# SaveBestH5Model(model=model, save_name=save_name, history_loss=history_val_loss, loss_value=val_loss.numpy())
history_val_loss.append(val_loss)
history_loss.append(loss_value.numpy())
@ -625,7 +626,7 @@ def getResult(model: tf.keras.Model, healthy_data, healthy_label, unhealthy_data
if __name__ == '__main__':
total_data = loadData.execute(N=feature_num, file_name=file_name)
np.save("G:\data\SCADA数据/total_data.npy",total_data)
# np.save("G:\data\SCADA数据/total_data.npy",total_data)
total_data = normalization(data=total_data)
train_data_healthy, train_label1_healthy, train_label2_healthy = get_training_data_overlapping(
total_data[:healthy_date, :], is_Healthy=True)
@ -633,7 +634,7 @@ if __name__ == '__main__':
total_data[healthy_date - time_stamp + unhealthy_patience:unhealthy_date, :],
is_Healthy=False)
#### TODO 第一步训练
# 单次测试
# # 单次测试
# train_step_one(train_data=train_data_healthy[:256, :, :], train_label1=train_label1_healthy[:256, :],
# train_label2=train_label2_healthy[:256, ])
#### 模型训练

34
TensorFlow_eaxmple/Model_train_test/condition_monitoring/self_try/compare/RNet-S.py
View File

@ -16,7 +16,7 @@ import pandas as pd
import matplotlib.pyplot as plt
from model.DepthwiseCon1D.DepthwiseConv1D import DepthwiseConv1D
from model.Dynamic_channelAttention.Dynamic_channelAttention import DynamicChannelAttention
from condition_monitoring.data_deal import loadData
from condition_monitoring.data_deal import loadData_daban as loadData
from model.Joint_Monitoring.compare.RNet_S import Joint_Monitoring
from model.CommonFunction.CommonFunction import *
@ -27,16 +27,16 @@ import random
'''超参数设置'''
time_stamp = 120
feature_num = 10
batch_size = 16
batch_size = 32
learning_rate = 0.001
EPOCH = 101
model_name = "RNet_S"
model_name = "RNet_S_banda"
'''EWMA超参数'''
K = 18
namuda = 0.01
'''保存名称'''
# save_name = "E:\self_example\TensorFlow_eaxmple\Model_train_test\condition_monitoring\hard_model\weight\joint_timestamp120_feature10_weight_epoch11_0.0077/weight"
save_name = "./model/weight/{0}_epoch3_2.47_1.63/weight".format(model_name,
save_name = "./model/weight/{0}/RNet_S_banda_epoch17_0.0017712703/weight".format(model_name,
time_stamp,
feature_num,
batch_size,
@ -47,12 +47,12 @@ save_step_two_name = "../hard_model/two_weight/{0}_timestamp{1}_feature{2}_weigh
batch_size,
EPOCH)
save_mse_name = "./mse/{0}/{0}_timestamp{1}_feature{2}_mse.csv".format(model_name,
save_mse_name = "./mse/RNet_S/banda/{0}_mse.csv".format(model_name,
time_stamp,
feature_num,
batch_size,
EPOCH)
save_max_name = "./mse/{0}/{0}_timestamp{1}_feature{2}_max.csv".format(model_name,
save_max_name = "./mse/RNet_S/banda/{0}_max.csv".format(model_name,
time_stamp,
feature_num,
batch_size,
@ -69,7 +69,8 @@ save_max_name = "./mse/{0}/{0}_timestamp{1}_feature{2}_max.csv".format(model_nam
# batch_size,
# EPOCH)
'''文件名'''
file_name = "G:\data\SCADA数据\jb4q_8_delete_total_zero.csv"
'''文件名'''
file_name = "G:\data\SCADA数据\SCADA_已处理_粤水电达坂城2020.1月-5月\风机15.csv"
'''
文件说明：jb4q_8_delete_total_zero.csv是删除了只删除了全是0的列的文件
@ -78,9 +79,9 @@ file_name = "G:\data\SCADA数据\jb4q_8_delete_total_zero.csv"
'''
'''文件参数'''
# 最后正常的时间点
healthy_date = 415548
healthy_date = 96748
# 最后异常的时间点
unhealthy_date = 432153
unhealthy_date = 107116
# 异常容忍程度
unhealthy_patience = 5
@ -360,7 +361,7 @@ def train_step_one(train_data, train_label1, train_label2):
k = k + 1
val_loss, val_accuracy = model.get_val_loss(val_data=val_data, val_label1=val_label1, val_label2=val_label2,
is_first_time=True)
SaveBestModel(model=model, save_name=save_name, history_loss=history_val_loss, loss_value=val_loss.numpy())
SaveBestModel(model=model, save_name=save_name, history_loss=history_val_loss, loss_value=val_loss.numpy(),epoch=epoch,is_all=True)
# SaveBestH5Model(model=model, save_name=save_name, history_loss=history_val_loss, loss_value=val_loss.numpy())
history_val_loss.append(val_loss)
history_loss.append(loss_value.numpy())
@ -635,8 +636,8 @@ def getResult(model: tf.keras.Model, healthy_data, healthy_label, unhealthy_data
if __name__ == '__main__':
# total_data = loadData.execute(N=feature_num, file_name=file_name)
total_data=np.load("G:\data\SCADA数据\靖边8号处理后的数据\原始10SCADA数据/total_data.npy")
total_data = loadData.execute(N=feature_num, file_name=file_name)
# total_data=np.load("G:\data\SCADA数据\靖边8号处理后的数据\原始10SCADA数据/total_data.npy")
total_data = normalization(data=total_data)
train_data_healthy, train_label1_healthy, train_label2_healthy = get_training_data_overlapping(
total_data[:healthy_date, :], is_Healthy=True)
@ -645,7 +646,7 @@ if __name__ == '__main__':
is_Healthy=False)
#### TODO 第一步训练
# 单次测试
# train_step_one(train_data=train_data_healthy[:32, :, :], train_label1=train_label1_healthy[:32, :],train_label2=train_label2_healthy[:32, ])
# train_step_one(train_data=train_data_healthy[:128, :, :], train_label1=train_label1_healthy[:128, :],train_label2=train_label2_healthy[:128, ])
# train_step_one(train_data=train_data_healthy, train_label1=train_label1_healthy, train_label2=train_label2_healthy)
# 导入第一步已经训练好的模型,一个继续训练，一个只输出结果
@ -672,13 +673,6 @@ if __name__ == '__main__':
healthy_label=train_label1_healthy[healthy_size - 2 * unhealthy_size:, :],
unhealthy_data=train_data_unhealthy, unhealthy_label=train_label1_unhealthy, isSave=True)
###TODO 展示全部的结果
# all_data, _, _ = get_training_data_overlapping(
# total_data[healthy_size - 2 * unhealthy_size:unhealthy_date, :], is_Healthy=True)
# all_data = np.concatenate([])
# 单次测试
# showResult(step_two_model, test_data=all_data[:32], isPlot=True)
# showResult(step_two_model, test_data=all_data, isPlot=True)
pass

115
TensorFlow_eaxmple/Model_train_test/condition_monitoring/self_try/compare/RNet.py
View File

@ -16,7 +16,7 @@ import pandas as pd
import matplotlib.pyplot as plt
from model.DepthwiseCon1D.DepthwiseConv1D import DepthwiseConv1D
from model.Dynamic_channelAttention.Dynamic_channelAttention import DynamicChannelAttention
from condition_monitoring.data_deal import loadData
from condition_monitoring.data_deal import loadData_daban as loadData
from model.Joint_Monitoring.compare.RNet import Joint_Monitoring
from model.CommonFunction.CommonFunction import *
@ -36,16 +36,26 @@ K = 18
namuda = 0.01
'''保存名称'''
save_name = "./model/weight/{0}/weight".format(model_name,
time_stamp,
feature_num,
batch_size,
EPOCH)
save_name = "./model/weight/{0}/RNet_banda_epoch4_0.0070075933/weight".format(model_name,
time_stamp,
feature_num,
batch_size,
EPOCH)
save_step_two_name = "./model/two_weight/{0}_timestamp{1}_feature{2}_weight/weight".format(model_name,
time_stamp,
feature_num,
batch_size,
EPOCH)
save_mse_name = "./mse/RNet/banda/{0}_mse.csv".format(model_name,
time_stamp,
feature_num,
batch_size,
EPOCH)
save_max_name = "./mse/RNet/banda/{0}_max.csv".format(model_name,
time_stamp,
feature_num,
batch_size,
EPOCH)
# save_name = "../model/joint/{0}_timestamp{1}_feature{2}.h5".format(model_name,
# time_stamp,
@ -132,57 +142,7 @@ def get_training_data_overlapping(data, time_stamp: int = time_stamp, is_Healthy
return train_data, train_label, train_label2
# RepConv重参数化卷积
def RepConv(input_tensor, k=3):
_, _, output_dim = input_tensor.shape
conv1 = tf.keras.layers.Conv1D(filters=output_dim, kernel_size=k, strides=1, padding='SAME')(input_tensor)
b1 = tf.keras.layers.BatchNormalization()(conv1)
conv2 = tf.keras.layers.Conv1D(filters=output_dim, kernel_size=1, strides=1, padding='SAME')(input_tensor)
b2 = tf.keras.layers.BatchNormalization()(conv2)
b3 = tf.keras.layers.BatchNormalization()(input_tensor)
out = tf.keras.layers.Add()([b1, b2, b3])
out = tf.nn.relu(out)
return out
# RepBlock模块
def RepBlock(input_tensor, num: int = 3):
for i in range(num):
input_tensor = RepConv(input_tensor)
return input_tensor
# GAP 全局平均池化
def Global_avg_channelAttention(input_tensor):
_, length, channel = input_tensor.shape
DWC1 = DepthwiseConv1D(kernel_size=1, padding='SAME')(input_tensor)
GAP = tf.keras.layers.GlobalAvgPool1D()(DWC1)
c1 = tf.keras.layers.Conv1D(filters=channel, kernel_size=1, padding='SAME')(GAP)
s1 = tf.nn.sigmoid(c1)
output = tf.multiply(input_tensor, s1)
return output
# GDP 全局动态池化
def Global_Dynamic_channelAttention(input_tensor):
_, length, channel = input_tensor.shape
DWC1 = DepthwiseConv1D(kernel_size=1, padding='SAME')(input_tensor)
# GAP
GAP = tf.keras.layers.GlobalAvgPool1D()(DWC1)
c1 = tf.keras.layers.Conv1D(filters=channel, kernel_size=1, padding='SAME')(GAP)
s1 = tf.nn.sigmoid(c1)
# GMP
GMP = tf.keras.layers.GlobalMaxPool1D()(DWC1)
c2 = tf.keras.layers.Conv1D(filters=channel, kernel_size=1, padding='SAME')(GMP)
s3 = tf.nn.sigmoid(c2)
output = tf.multiply(input_tensor, s1)
return output
# 归一化
@ -338,7 +298,8 @@ def train_step_one(train_data, train_label1, train_label2):
k = k + 1
val_loss, val_accuracy = model.get_val_loss(val_data=val_data, val_label1=val_label1, val_label2=val_label2,
is_first_time=True)
SaveBestModel(model=model, save_name=save_name, history_loss=history_val_loss, loss_value=val_loss.numpy())
SaveBestModel(model=model, save_name=save_name, history_loss=history_val_loss, loss_value=val_loss.numpy(),
epoch=epoch, is_all=True)
# SaveBestH5Model(model=model, save_name=save_name, history_loss=history_val_loss, loss_value=val_loss.numpy())
history_val_loss.append(val_loss)
history_loss.append(loss_value.numpy())
@ -501,12 +462,12 @@ def get_MSE(data, label, new_model, isStandard: bool = True, isPlot: bool = True
predict_data = 0
predict_data = predicted_data1
mseList=[]
meanList=[]
maxList=[]
mseList = []
meanList = []
maxList = []
for i in range(1,4):
print("i:",i)
for i in range(1, 4):
print("i:", i)
if i == 1:
predict_data = predicted_data1
elif i == 2:
@ -519,7 +480,7 @@ def get_MSE(data, label, new_model, isStandard: bool = True, isPlot: bool = True
temp3 = temp1 / temp2
mse = np.sum((temp1 / temp2) ** 2, axis=1)
print("mse.shape:",mse.shape)
print("mse.shape:", mse.shape)
# mse=np.mean((predicted_data-label)**2,axis=1)
# print("mse", mse)
mseList.append(mse)
@ -549,7 +510,6 @@ def get_MSE(data, label, new_model, isStandard: bool = True, isPlot: bool = True
# plt.plot(min)
plt.show()
return mseList, meanList, maxList
# pass
@ -558,11 +518,11 @@ def get_MSE(data, label, new_model, isStandard: bool = True, isPlot: bool = True
def getResult(model: tf.keras.Model, healthy_data, healthy_label, unhealthy_data, unhealthy_label, isPlot: bool = False,
isSave: bool = False, predictI: int = 1):
# TODO 计算MSE确定阈值
# plt.ion()
mseList, meanList, maxList = get_MSE(healthy_data, healthy_label, model)
mse1List,_,_ = get_MSE(unhealthy_data, unhealthy_label, model, isStandard=False)
mse1List, _, _ = get_MSE(unhealthy_data, unhealthy_label, model, isStandard=False)
for mse,mean,max,mse1 in zip(mseList, meanList, maxList,mse1List):
for mse, mean, max, mse1, j in zip(mseList, meanList, maxList, mse1List, range(3)):
# 误报率的计算
total, = mse.shape
@ -588,16 +548,23 @@ def getResult(model: tf.keras.Model, healthy_data, healthy_label, unhealthy_data
miss_rate = missNum / all
print("漏报率:", miss_rate)
#总体图
print("mse:",mse)
print("mse1:",mse1)
# 总体图
# print("mse:", mse)
# print("mse1:", mse1)
print("============================================")
total_mse = np.concatenate([mse, mse1], axis=0)
total_max = np.broadcast_to(max[0], shape=[total_mse.shape[0], ])
# min = np.broadcast_to(min,shape=[dims,])
total_mean = np.broadcast_to(mean[0], shape=[total_mse.shape[0], ])
plt.figure(random.randint(1, 9))
if isSave:
save_mse_name1 = save_mse_name[:-4] + "_predict" + str(j + 1) + ".csv"
save_max_name1 = save_max_name[:-4] + "_predict" + str(j + 1) + ".csv"
np.savetxt(save_mse_name1, total_mse, delimiter=',')
np.savetxt(save_max_name1, total_max, delimiter=',')
plt.figure(random.randint(1, 100))
plt.plot(total_max)
plt.plot(total_mse)
plt.plot(total_mean)
@ -618,7 +585,7 @@ if __name__ == '__main__':
# 单次测试
# train_step_one(train_data=train_data_healthy[:256, :, :], train_label1=train_label1_healthy[:256, :],
# train_label2=train_label2_healthy[:256, ])
#### 模型训练
# #### 模型训练
# train_step_one(train_data=train_data_healthy, train_label1=train_label1_healthy, train_label2=train_label2_healthy)
# 导入第一步已经训练好的模型,一个继续训练，一个只输出结果
@ -640,11 +607,11 @@ if __name__ == '__main__':
# :],
# healthy_label=train_label1_healthy[
# healthy_size - 2 * unhealthy_size:healthy_size - 2 * unhealthy_size + 200, :],
# unhealthy_data=train_data_unhealthy[:200,:], unhealthy_label=train_label1_unhealthy[:200,:])
# unhealthy_data=train_data_unhealthy[:200,:], unhealthy_label=train_label1_unhealthy[:200,:], isSave=True)
getResult(step_one_model, healthy_data=train_data_healthy[healthy_size - 2 * unhealthy_size:, :],
healthy_label=train_label1_healthy[healthy_size - 2 * unhealthy_size:, :],
unhealthy_data=train_data_unhealthy, unhealthy_label=train_label1_unhealthy)
unhealthy_data=train_data_unhealthy, unhealthy_label=train_label1_unhealthy, isSave=True)
# ###TODO 展示全部的结果
# all_data, _, _ = get_training_data_overlapping(

51
TensorFlow_eaxmple/Model_train_test/condition_monitoring/self_try/compare/resnet_18.py
View File

@ -6,33 +6,33 @@ from sklearn.manifold import TSNE
from sklearn.metrics import confusion_matrix
import seaborn as sns
from sklearn.model_selection import train_test_split
from condition_monitoring.data_deal import loadData
from condition_monitoring.data_deal import loadData_daban as loadData
from keras.callbacks import EarlyStopping
'''超参数设置'''
time_stamp = 120
feature_num = 10
batch_size = 16
batch_size = 32
learning_rate = 0.001
EPOCH = 101
model_name = "ResNet"
model_name = "ResNet_banda"
'''EWMA超参数'''
K = 18
namuda = 0.01
'''保存名称'''
save_name = "./model/{0}_9990_9998.h5".format(model_name,
save_name = "./model/{0}.h5".format(model_name,
time_stamp,
feature_num,
batch_size,
EPOCH)
save_step_two_name = "./model/two_weight/{0}_weight_epoch6_99899_9996/weight".format(model_name,
save_step_two_name = "./model/two_weight/{0}/weight".format(model_name,
time_stamp,
feature_num,
batch_size,
EPOCH)
save_mse_name = "./mse/ResNet/{0}_timestamp{1}_feature{2}_result.csv".format(model_name,
save_mse_name = "./mse/ResNet/{0}_result.csv".format(model_name,
time_stamp,
feature_num,
batch_size,
@ -42,13 +42,14 @@ save_max_name = "./mse/ResNet/{0}_timestamp{1}_feature{2}_max.csv".format(model_
feature_num,
batch_size,
EPOCH)
save_mse_name1 = "./mse/ResNet/{0}_timestamp{1}_feature{2}_totalresult.csv".format(model_name,
save_mse_name1 = "./mse/ResNet/{0}_result1.csv".format(model_name,
time_stamp,
feature_num,
batch_size,
EPOCH)
'''文件名'''
file_name = "G:\data\SCADA数据\jb4q_8_delete_total_zero.csv"
'''文件名'''
file_name = "G:\data\SCADA数据\SCADA_已处理_粤水电达坂城2020.1月-5月\风机15.csv"
'''
文件说明：jb4q_8_delete_total_zero.csv是删除了只删除了全是0的列的文件
@ -57,9 +58,9 @@ file_name = "G:\data\SCADA数据\jb4q_8_delete_total_zero.csv"
'''
'''文件参数'''
# 最后正常的时间点
healthy_date = 415548
healthy_date = 96748
# 最后异常的时间点
unhealthy_date = 432153
unhealthy_date = 107116
# 异常容忍程度
unhealthy_patience = 5
@ -357,8 +358,8 @@ def plot_confusion_matrix_accuracy(cls, true_labels, predict_labels):
if __name__ == '__main__':
# # 数据读入
#
# total_data = loadData.execute(N=feature_num, file_name=file_name)
total_data = np.load("G:\data\SCADA数据\靖边8号处理后的数据\原始10SCADA数据/total_data.npy")
total_data = loadData.execute(N=feature_num, file_name=file_name)
# total_data = np.load("G:\data\SCADA数据\靖边8号处理后的数据\原始10SCADA数据/total_data.npy")
total_data = normalization(data=total_data)
train_data_healthy, train_label1_healthy, train_label2_healthy = get_training_data_overlapping(
total_data[:healthy_date, :], is_Healthy=True)
@ -381,27 +382,29 @@ if __name__ == '__main__':
model.compile(optimizer=tf.optimizers.Adam(), loss=tf.losses.binary_crossentropy,
metrics=['acc'])
model.summary()
# early_stop = EarlyStopping(monitor='val_loss', min_delta=0.0001, patience=5, mode='min', verbose=1)
#
# checkpoint = tf.keras.callbacks.ModelCheckpoint(
# filepath=save_name,
# monitor='val_loss',
# verbose=1,
# save_best_only=False,
# mode='min',
# period=1)
#
early_stop = EarlyStopping(monitor='val_loss', min_delta=0.0001, patience=5, mode='min', verbose=1)
checkpoint = tf.keras.callbacks.ModelCheckpoint(
filepath=save_name,
monitor='val_loss',
verbose=1,
save_best_only=False,
mode='min',
period=1)
# history = model.fit(train_data, train_label, epochs=20, batch_size=16, validation_data=(test_data, test_label),
# callbacks=[checkpoint, early_stop])
# model.save("./model/ResNet.h5")
model = tf.keras.models.load_model("./model/ResNet_2_9906.h5")
model = tf.keras.models.load_model("model/ResNet_banda/ResNet_banda_epoch12_9942.h5")
# 结果展示
healthy_size, _, _ = train_data_healthy.shape
unhealthy_size, _, _ = train_data_unhealthy.shape
# all_data, _, _ = get_training_data_overlapping(
# total_data[0:, :], is_Healthy=True)
all_data, _, _ = get_training_data_overlapping(
total_data[0:, :], is_Healthy=True)
total_data[healthy_size - 2 * unhealthy_size:unhealthy_date, :], is_Healthy=True)
showResult(model, test_data=all_data, isPlot=True)