公开(公告)号：US20240184678A1

公开(公告)日：2024-06-06

申请号：US17797133

申请日：2021-10-27

Applicant: Hefei University of Technology

Inventor： Qiang Zhang ,  Ting Huang ,  Shanlin Yang ,  Xianghong Hu ,  Chunhui Wang ,  Yuanhang Wang ,  Xiaojian Ding

IPC: G06F11/22

CPC classification number: G06F11/2263

Abstract: A deep learning fault diagnosis method includes the following steps: a fault diagnosis data set X is processed based on sliding window processing, to obtain a picture-like sample data set {tilde over (X)}, and obtain an attention matrix A of the picture-like sample data set {tilde over (X)}; and a 2D-CNN model is constructed to process the picture-like sample data set {tilde over (X)} to obtain a corresponding feature map F, and in the meantime, the feature map F is processed based on channel-oriented average pooling and channel-oriented maximum pooling to obtain an output P1 of the average pooling and an output P2 of the maximum pooling, and a weight matrix W is obtained based on the attention matrix A, the output P1 of the average pooling, and the output P2 of the maximum pooling, so that an output of the model is a feature map {tilde over (F)} based on an attention mechanism, where {tilde over (F)}=WF.

公开(公告)号：US12174715B2

公开(公告)日：2024-12-24

申请号：US17797133

申请日：2021-10-27

Applicant: Hefei University of Technology

Inventor： Qiang Zhang ,  Ting Huang ,  Shanlin Yang ,  Xianghong Hu ,  Chunhui Wang ,  Yuanhang Wang ,  Xiaojian Ding

IPC: G06F11/22

Abstract: A deep learning fault diagnosis method includes the following steps: a fault diagnosis data set X is processed based on sliding window processing, to obtain a picture-like sample data set {tilde over (X)}, and obtain an attention matrix A of the picture-like sample data set {tilde over (X)}; and a 2D-CNN model is constructed to process the picture-like sample data set {tilde over (X)} to obtain a corresponding feature map F, and in the meantime, the feature map F is processed based on channel-oriented average pooling and channel-oriented maximum pooling to obtain an output P1 of the average pooling and an output P2 of the maximum pooling, and a weight matrix W is obtained based on the attention matrix A, the output P1 of the average pooling, and the output P2 of the maximum pooling, so that an output of the model is a feature map {tilde over (F)} based on an attention mechanism, where {tilde over (F)}=WF.