航空发动机调制信号的坐标注意力时频特征增强方法

张晗; 李丹; 杜朝辉; 马猛

doi:10.7652/xjtuxb202611020

您当前的位置：

首页 >

文章列表页 >

航空发动机调制信号的坐标注意力时频特征增强方法

更新时间：2026-11-09

- 航空发动机调制信号的坐标注意力时频特征增强方法
- A Coordinate Attention-Based Time-Frequency Feature Enhancement Method for Modulated Signals of Aero-Engines
- 西安交通大学学报 2026年60卷第11期页码：224-235
- 作者机构：
  
  长安大学机械工程与智能装备学院，710064，西安
  长安大学道路施工技术与装备教育部重点实验室，710064，西安
  西北工业大学航海学院，710072，西安
  西安交通大学机械工程学院，710049，西安
- 作者简介：
  
  张晗（1989—），女，副教授，博士生导师。
- 基金信息：
  
  国家自然科学基金资助项目(52275085;52575140);陕西省重点研发计划资助项目(2024GX-YBXM-265)
- DOI：10.7652/xjtuxb202611020
  中图分类号： V233.1+4
- 收稿：2026-03-05，
  
  网络首发：2026-06-10，
  
  纸质出版：2026-11-10
- 稿件说明：
移动端阅览
张晗, 李丹, 杜朝辉, 等. 航空发动机调制信号的坐标注意力时频特征增强方法[J/OL]. 西安交通大学学报,2026,60 (11):224-235. https://doi.org/10.7652/xjtuxb202611020.

ZHANG Han, LI Dan, DU Zhaohui, et al. A Coordinate Attention-Based Time-Frequency Feature Enhancement Method for Modulated Signals of Aero-Engines[J/OL]. Journal of Xi'an Jiaotong University,2026,60 (11):224-235. https://doi.org/10.7652/xjtuxb202611020.
张晗, 李丹, 杜朝辉, 等. 航空发动机调制信号的坐标注意力时频特征增强方法[J/OL]. 西安交通大学学报,2026,60 (11):224-235. https://doi.org/10.7652/xjtuxb202611020. DOI：

ZHANG Han, LI Dan, DU Zhaohui, et al. A Coordinate Attention-Based Time-Frequency Feature Enhancement Method for Modulated Signals of Aero-Engines[J/OL]. Journal of Xi'an Jiaotong University,2026,60 (11):224-235. https://doi.org/10.7652/xjtuxb202611020. DOI：

摘要

针对航空发动机行星齿轮调制信号在噪声背景下存在时频脊线模糊、弱调制特征易淹没的问题，提出了一种基于坐标注意力机制的轻量化时频特征增强网络模型（TF_CoordAtt）。以短时傅里叶变换（STFT）时频图为输入，构建时频去噪-坐标注意力增强两阶段网络结构，利用深度可分离卷积与双尺度空洞卷积提取多尺度时频特征，并结合坐标注意力机制分别沿时间轴与频率轴编码方向位置信息，实现对调制脊线及故障敏感区域的自适应增强。为保持增强后时频结构连续性，设计基于时间-频率方向一阶梯度约束的结构保持损失函数，减弱了脊线断裂与局部模糊现象。调幅调频仿真实验结果表明：TF_CoordAtt网络在低参数量条件下，峰值信噪比达到19.87 dB，且能够获得连续清晰的时频脊线；与ResNet-6网络相比，模型参数量由11.6万减少至2.6万，体现出较好的轻量化与训练效率优势。航空发动机齿轮毂试验结果进一步验证了该方法对啮合频率及故障调制边频结构的有效增强能力。

Abstract

To address the problems that under noisy conditions

time-frequency ridges are blurred and weak modulation features are easily submerged in noise when modulated signals of aero-engine planetary gears are processed

a lightweight time-frequency feature enhancement network based on coordinate attention

termed TF_CoordAtt

is proposed.The proposed method takes short-time Fourier transform (STFT) time-frequency representations as input and constructs a two-stage framework consisting of time-frequency denoising and coordinate-attention-based enhancement. Depthwise separable convolutions and dual-scale dilated convolutions are employed to extract multi-scale time-frequency features

while the coordinate attention mechanism encodes directional positional information along both temporal and frequency axes to adaptively enhance modulation ridges and fault-sensitive regions.To maintain the continuity of the enhanced time-frequency structure

a structure-preserving loss function based on the first-order gradient constraint along the temporal and frequency directions is designed

effectively alleviating ridge discontinuities and local blurring artifacts during enhancement.Experimental results on amplitude-and frequency-modulated simulated signals demonstrate that TF_CoordAtt achieves a peak signal-to-noise ratio of 19.87 dB with a lightweight architecture while producing clearer and more continuous time-frequency ridges. Compared with the ResNet-6 network

the proposed method reduces the number of model parameters from 116000 to 26000

exhibiting notable advantages in lightweight design and training efficiency. The experimental results on an aero-engine gear hub further validate the effectiveness of the proposed method in enhancing the meshing frequency and the fault modulation sideband structure.

关键词

Keywords

references

Fedele L,Di Vito L,Ramundo F E.Increasing efficiency in an aeronautical engine through maintenance evaluation and upgrades:analysis of the reliability and performance improvements under financial issues [J]. Energies,2020,13(12):3059.

Duan Xiaohui,Feng Zhipeng.Time-varying filtering for nonstationary signal analysis of rotating machinery:principle and applications [J]. Mechanical Systems and Signal Processing,2023,192:110204.

Sun Bin,Li Hongkun,Wang Junxiang,et al.Differential time-frequency mode decomposition and its application in rotating machinery fault diagnosis [J]. Mechanical Systems and Signal Processing,2025,229:112571.

陈江涛,薛海,白永亮.基于HMI-POA-VMD的列车轴箱轴承故障诊断[J].北京交通大学学报,2024, 48(6):113-121.

Chen Jiangtao,Xue Hai,Bai Yongliang.Fault diagnosis of train axle box bearing based on HMI-POA-VMD[J]. Journal of Beijing Jiaotong University,2024,48 (6):113-121.

李彬,徐怡杭,罗杰.采用残差神经网络的无人机遥控信号识别监测算法[J].西安交通大学学报,2021 , 55(12):146-154.

Li Bin,Xu Yihang,Luo Jie.Recognition and monitoring algorithm for UAV remote-control signals using residual neural networks[J].Journal of Xi'an Jiaotong University,2021,55(12):146-154.

赵柄锡,冀大伟,袁奇,等.采用时域与时频域联合特征空间的转子系统碰磨故障诊断[J].西安交通大学学报,2020,54(1):75-84.

Zhao Bingxi,Ji Dawei,Yuan Qi,et al.Rubbing fault diagnosis of rotor system based on combined feature space in time and time-frequency domains[J].Journal of Xi'an Jiaotong University,2020,54(1):75-84.

徐洲常,王林军,刘洋,等.采用改进回归型支持向量机的滚动轴承剩余寿命预测方法[J].西安交通大学学报,2022,56(3):197-205.

Xu Zhouchang,Wang Linjun,Liu Yang,et al.A prediction method for remaining life of rolling bearing using improved regression support vector machine[J]. Journal of Xi'an Jiaotong University,2022,56 (3):197-205.

张弘斌,袁奇,赵柄锡,等.采用多通道样本和深度卷积神经网络的轴承故障诊断方法[J].西安交通大学学报,2020,54(8):58-66.

Zhang Hongbin,Yuan Qi,Zhao Bingxi,et al.Bearing fault diagnosis with multi-channel sample and deep convolutional neural network[J].Journal of Xi'an Jiaotong University,2020,54(8):58-66.

陶唐飞,周文洁,况佳臣,等.融合多小波分解的深度卷积神经网络轴承故障诊断方法[J].西安交通大学学报,2024,58(5):31-41.

Tao Tangfei, Zhou Wenjie,Kuang Jiachen, et al. Bearing fault diagnosis method of deep convolutional neural network based on multiwavelet decomposition[J]. Journal of Xi'an Jiaotong University,2024,58 (5):31-41.

姜祖华,张坤,杨欢,等.参数化S谱幅值调制及轴承故障诊断应用[J].西安交通大学学报,2024,58(3):117-128.

Jiang Zuhua,Zhang Kun,Yang Huan,et al.Parameterized S spectral amplitude modulation and its applications to bearing fault diagnosis[J].Journal of Xi'an Jiaotong University,2024,58(3):117-128.

张龙,肖逸文,周神赐,等.一种双阻尼小波赋能的可解释卷积神经网络在轴承故障诊断中的应用[J].西安交通大学学报,2025,59(3):210-221 .

Zhang Long,Xiao Yiwen,Zhou Shenci,et al.Application of interpretable convolutional neural network enabled by bi-damped wavelets in bearing fault diagnosis[J]. Journal of Xi'an Jiaotong University,2025,59 (3):210-221 .

Nguyen T D,Nguyen H C,Pham D H,et al.A distinguished deep learning method for gear fault classification using time-frequency representation [J]. Discover Applied Sciences,2024,6(7):340.

Pan Pingping,Zhang Yunjian,Deng Zhenmiao,et al. TFA-Net:a deep learning-based time-frequency analysis tool [J].IEEE Transactions on Neural Networks and Learning Systems,2023,34(11):9274-9286.

Zhao Dezun,Shao Depei,Cui Lingli.CTNet:a datadriven time-frequency technique for wind turbines fault diagnosis under time-varying speeds [J].ISA Transactions,2024,154:335-351.

Chen Tao,Chen Qiming,Zheng Qian,et al.Adaptive multi-scale TF-net for high-resolution time-frequency representations [J].Signal Processing,2024,214:109247.

Chen Qian,Dong Xingjian,Tu Guowei,et al.TFN:an interpretable neural network with time-frequency transform embedded for intelligent fault diagnosis [J]. Mechanical Systems and Signal Processing, 2024, 207:110952.

Deprem Z,Çetin A E.High-resolution time-frequency representation with generative adversarial networks [J]. Signal,Image and Video Processing,2023 ,17 (3):849-854.

张玺君,尚继洋.采用多尺度自适应选择卷积神经网络的轴承故障诊断研究[J].西安交通大学学报, 2024,58(2):127-135.

Zhang Xijun,Shang Jiyang.Bearing fault diagnosis based on multi-scale adaptive selective convolutional neural network[J].Journal of Xi'an Jiaotong University,2024,58(2):127-135.

邓飞跃,郑守禧,郝如江.一种轻量化尺度感知调制Swin Transformer模型的轴箱轴承故障诊断方法[J].西安交通大学学报,2024,58(9):83-93.

Deng Feiyue,Zheng Shouxi,Hao Rujiang.A lightweight scale-aware modulation Swin Transformer network for axlebox bearing fault diagnosis[J].Journal of Xi'an Jiaotong University,2024,58(9):83-93.

Peng Yihang, Zhang Jianjie, Liu Songpeng, et al. YConvFormer:a lightweight and robust transformer for gearbox fault diagnosis with time-frequency fusion [J]. Sensors,2025,25(15):4862.

冯志鹏,赵镭镭,褚福磊.行星齿轮箱故障诊断的幅值解调分析方法[J].中国电机工程学报,2013, 33(8):107-111.

Feng Zhipeng, Zhao Leilei, Chu Fulei. Amplitude demodulation analysis for fault diagnosis of planetary gearboxes[J].Proceedings of the CSEE,2013,33 (8):107-111.

冯志鹏,赵镭镭,褚福磊.行星齿轮箱齿轮局部故障振动频谱特征[J].中国电机工程学报,2013,33(5):119-127.

Feng Zhipeng,Zhao Leilei,Chu Fulei.Vibration spectral characteristics of localized gear fault of planetary gearboxes[J].Proceedings of the CSEE,2013,33(5):119-127.

冯志鹏,褚福磊.行星齿轮箱齿轮分布式故障振动频谱特征[J].中国电机工程学报,2013,33 (2):118-125.

Feng Zhipeng,Chu Fulei.Vibration spectral characteristics of distributed gear fault of planetary gearboxes[J]. Proceedings of the CSEE,2013,33(2):118-125.

Hou Qibin,Zhou Daquan,Feng Jiashi.Coordinate attention for efficient mobile network design [C]//2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).Piscataway,NJ,USA:IEEE,2021:13708-13717.

Abrahamyan L,Truong A M,Philips W,et al.Gradient variance loss for structure-enhanced image superresolution [C]//ICASSP 2022—2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).Piscataway,NJ,USA:IEEE, 2022:3219-3223.

Zhao Dezun,Shao Depei,Wang Tianyang,et al.Timefrequency self-similarity enhancement network and its application in wind turbines fault analysis [J].Advanced Engineering Informatics,2025,65:103322.

Zhang Han,Chen Xuefeng,Chen Wei,et al.Collaborative sparse classification for aero-engine's gear hub crack diagnosis [J].Mechanical Systems and Signal Processing,2020,141 :106426.

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

厌氧消化1号模型在光发酵制氢过程的应用与优化

涡轮叶片前缘冲击/气膜复合冷却特性实验研究

采用多状态柔性合成的离网型风光氢氨系统两阶段分布鲁棒调度

面向永磁同步电机无位置传感器驱动的新型K∞等效滑模控制

采用多传感器融合的空间全方位旋转角度测量方法