基于深度残差收缩网络的故障诊断(简要解读)

abandon
2周前 阅读 15 点赞 1

论文题目:Deep residual shrinkage networks for fault diagnosis

题目翻译:基于深度残差收缩网络的故障诊断

期刊:IEEE Transactions on Industrial Informatics

面向问题:强噪声(或冗余信息繁多)

创新:①在深度残差网络中引入软阈值化,②并且自动设置阈值

本质:在注意力机制下实现特征的软阈值化

深度残差收缩网络的基本模块:

深度残差收缩网络的整体结构:

深度残差收缩网络的示例代码:

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
Created on Sat Dec 28 23:24:05 2019
Implemented using TensorFlow 1.0.1 and Keras 2.2.1
 
M. Zhao, S. Zhong, X. Fu, et al., Deep Residual Shrinkage Networks for Fault Diagnosis, 
IEEE Transactions on Industrial Informatics, 2019, DOI: 10.1109/TII.2019.2943898
@author: super_9527
"""


from __future__ import print_function
import keras
import numpy as np
from keras.datasets import mnist
from keras.layers import Dense, Conv2D, BatchNormalization, Activation
from keras.layers import AveragePooling2D, Input, GlobalAveragePooling2D
from keras.optimizers import Adam
from keras.regularizers import l2
from keras import backend as K
from keras.models import Model
from keras.layers.core import Lambda
K.set_learning_phase(1)


# Input image dimensions
img_rows, img_cols = 28, 28


# The data, split between train and test sets
(x_train, y_train), (x_test, y_test) = mnist.load_data()


if K.image_data_format() == 'channels_first':
    x_train = x_train.reshape(x_train.shape[0], 1, img_rows, img_cols)
    x_test = x_test.reshape(x_test.shape[0], 1, img_rows, img_cols)
    input_shape = (1, img_rows, img_cols)
else:
    x_train = x_train.reshape(x_train.shape[0], img_rows, img_cols, 1)
    x_test = x_test.reshape(x_test.shape[0], img_rows, img_cols, 1)
    input_shape = (img_rows, img_cols, 1)


# Noised data
x_train = x_train.astype('float32') / 255. + 0.5*np.random.random([x_train.shape[0], img_rows, img_cols, 1])
x_test = x_test.astype('float32') / 255. + 0.5*np.random.random([x_test.shape[0], img_rows, img_cols, 1])
print('x_train shape:', x_train.shape)
print(x_train.shape[0], 'train samples')
print(x_test.shape[0], 'test samples')


# convert class vectors to binary class matrices
y_train = keras.utils.to_categorical(y_train, 10)
y_test = keras.utils.to_categorical(y_test, 10)




def abs_backend(inputs):
    return K.abs(inputs)


def expand_dim_backend(inputs):
    return K.expand_dims(K.expand_dims(inputs,1),1)


def sign_backend(inputs):
    return K.sign(inputs)


def pad_backend(inputs, in_channels, out_channels):
    pad_dim = (out_channels - in_channels)//2
    return K.spatial_3d_padding(inputs, padding = ((0,0),(0,0),(pad_dim,pad_dim)))


# Residual Shrinakge Block
def residual_shrinkage_block(incoming, nb_blocks, out_channels, downsample=False,
                             downsample_strides=2):
    
    residual = incoming
    in_channels = incoming.get_shape().as_list()[-1]
    
    for i in range(nb_blocks):
        
        identity = residual
        
        if not downsample:
            downsample_strides = 1
        
        residual = BatchNormalization()(residual)
        residual = Activation('relu')(residual)
        residual = Conv2D(out_channels, 3, strides=(downsample_strides, downsample_strides), 
                          padding='same', kernel_initializer='he_normal', 
                          kernel_regularizer=l2(1e-4))(residual)
        
        residual = BatchNormalization()(residual)
        residual = Activation('relu')(residual)
        residual = Conv2D(out_channels, 3, padding='same', kernel_initializer='he_normal', 
                          kernel_regularizer=l2(1e-4))(residual)
        
        # Calculate global means
        residual_abs = Lambda(abs_backend)(residual)
        abs_mean = GlobalAveragePooling2D()(residual_abs)
        
        # Calculate scaling coefficients
        scales = Dense(out_channels, activation=None, kernel_initializer='he_normal', 
                       kernel_regularizer=l2(1e-4))(abs_mean)
        scales = BatchNormalization()(scales)
        scales = Activation('relu')(scales)
        scales = Dense(out_channels, activation='sigmoid', kernel_regularizer=l2(1e-4))(scales)
        scales = Lambda(expand_dim_backend)(scales)
        
        # Calculate thresholds
        thres = keras.layers.multiply([abs_mean, scales])
        
        # Soft thresholding
        sub = keras.layers.subtract([residual_abs, thres])
        zeros = keras.layers.subtract([sub, sub])
        n_sub = keras.layers.maximum([sub, zeros])
        residual = keras.layers.multiply([Lambda(sign_backend)(residual), n_sub])
        
        # Downsampling (it is important to use the pooL-size of (1, 1))
        if downsample_strides > 1:
            identity = AveragePooling2D(pool_size=(1,1), strides=(2,2))(identity)
            
        # Zero_padding to match channels (it is important to use zero padding rather than 1by1 convolution)
        if in_channels != out_channels:
            identity = Lambda(pad_backend)(identity, in_channels, out_channels)
        
        residual = keras.layers.add([residual, identity])
    
    return residual




# define and train a model
inputs = Input(shape=input_shape)
net = Conv2D(8, 3, padding='same', kernel_initializer='he_normal', kernel_regularizer=l2(1e-4))(inputs)
net = residual_shrinkage_block(net, 1, 8, downsample=True)
net = BatchNormalization()(net)
net = Activation('relu')(net)
net = GlobalAveragePooling2D()(net)
outputs = Dense(10, activation='softmax', kernel_initializer='he_normal', kernel_regularizer=l2(1e-4))(net)
model = Model(inputs=inputs, outputs=outputs)
model.compile(loss='categorical_crossentropy', optimizer=Adam(), metrics=['accuracy'])
model.fit(x_train, y_train, batch_size=100, epochs=5, verbose=1, validation_data=(x_test, y_test))


# get results
K.set_learning_phase(0)
DRSN_train_score = model.evaluate(x_train, y_train, batch_size=100, verbose=0)
print('Train loss:', DRSN_train_score[0])
print('Train accuracy:', DRSN_train_score[1])
DRSN_test_score = model.evaluate(x_test, y_test, batch_size=100, verbose=0)
print('Test loss:', DRSN_test_score[0])
print('Test accuracy:', DRSN_test_score[1])

M. Zhao, S. Zhong, X. Fu, et al., Deep Residual Shrinkage Networks for Fault Diagnosis, IEEE Transactions on Industrial Informatics, 2019, DOI: 10.1109/TII.2019.2943898

https://ieeexplore.ieee.org/document/8850096

| 1
登录后可评论,马上登录吧~
评论 ( 0 )

还没有人评论...

相关推荐