Intelligent Reliability and Maintainability of Energy Infrastructure Assets by He Li & Weiwen Peng & Sidum Adumene & Mohammad Yazdi

Author:He Li & Weiwen Peng & Sidum Adumene & Mohammad Yazdi
Language: eng
Format: epub
ISBN: 9783031299629
Publisher: Springer Nature Switzerland

5.2 Methods

5.2.1 Convolutional Siamese Networks Based on Global Average Pooling

This chapter designs Convolutional Siamese Networks with Global Average Pooling (GAPCSN) for fault diagnosis of planetary gearboxes under insufficient data conditions. The network uses the convolutional neural network for the feature extraction, and Euclidean distance is applied for the similarity metric.

Convolutional neural networks mainly consist of feature extraction and classification modules [23–25]. The feature extraction module extracts the features from the input data and consists of a convolutional layer, a pooling layer, and an activation layer [23]. The classification module identifies and classifies the extracted features, which consist of fully connected layers.

The convolutional neural network's convolutional layer uses weight sharing for convolutional computation. Its mathematical model is:

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.