基于LSTM的多尺度并行空洞卷积AUV推进器故障诊断

杜文辽; 杨寅; 田淑侠; 王宏超; 巩晓赟; 高艺源; 赵斌斌; 侯冬冬

doi:10.12382/bgxb.2024.1063

您当前的位置：

首页 >

文章列表页 >

基于LSTM的多尺度并行空洞卷积AUV推进器故障诊断

更新时间：2026-04-24

- 基于LSTM的多尺度并行空洞卷积AUV推进器故障诊断
- MSPDC-LSTM-based Fault Diagnosis of AUV Thruster
- 兵工学报 2026年47卷第2期页码：241063
- 作者机构：
  
  郑州轻工业大学机电工程学院，河南郑州 450002
  中国船舶集团有限公司第七一三研究所，河南郑州 450015
- 作者简介：
  
  通信作者邮箱：2013071@zzuli.edu.cn
- 基金信息：
  
  国家自然科学基金项目(52275138;52475172;52505581);河南省重点研发专项项目(231111221100);河南省高校科技创新团队支持计划项目(25IRTSTHN024);机械系统与振动全国重点实验室课题资助项目(MSV202502)
- DOI：10.12382/bgxb.2024.1063
  中图分类号： TH165+.3
- 收稿：2024-11-24，
  
  网络首发：2025-12-25，
  
  纸质出版：2026-02-28
- 稿件说明：
移动端阅览
杜文辽, 杨寅, 田淑侠, 等. 基于LSTM的多尺度并行空洞卷积AUV推进器故障诊断[J]. 兵工学报, 2026,47(2):241063.

DU Wenliao, YANG Yin, TIAN Shuxia, et al. MSPDC-LSTM-based Fault Diagnosis of AUV Thruster[J]. Acta Armamentarii, 2026, 47(2): 241063.
杜文辽, 杨寅, 田淑侠, 等. 基于LSTM的多尺度并行空洞卷积AUV推进器故障诊断[J]. 兵工学报, 2026,47(2):241063. DOI： 10.12382/bgxb.2024.1063.

DU Wenliao, YANG Yin, TIAN Shuxia, et al. MSPDC-LSTM-based Fault Diagnosis of AUV Thruster[J]. Acta Armamentarii, 2026, 47(2): 241063. DOI： 10.12382/bgxb.2024.1063.

摘要

自主式水下航行器（Autonomous Underwater Vehicle，AUV）推进器的运行信号表现出强耦合性和非线性特性。针对常规智能故障诊断方法面临的计算复杂度高和诊断精度低的挑战，提出了一种基于长短期记忆网络的多尺度并行空洞卷积的故障诊断方法。为有效捕捉推进器故障的多尺度空间特征，在多尺度网络中引入并行空洞卷积模块。该模块通过并行处理不同尺度的特征，并在不增加参数的情况下扩大感受野，从而提高模型在处理复杂数据时的表现和效率。结合长短期记忆模块增强网络对振动信号时序信息特征的捕捉能力，进一步提高处理复杂序列数据的准确性。通过深度提取信号的时空特性，增强特征学习能力，显著提升诊断的鲁棒性。为验证新方法的可行性，使用AUV实验数据进行实验。研究结果表明，新方法的平均准确率达到99.67%，在2dB强噪声干扰下依然实现了平均95.97%的准确率，可有效识别和判断推进器故障，证明了该方法的有效性和鲁棒性。

Abstract

The operational signals of autonomous underwater vehicle(AUV)propeller exhibit strong coupling and nonlinear characteristics. In the aspect of fault diagnosis of propuller the conventional intelligent fault diagnosis methods has high computational complexity and low diagnostic accuracy. A fault diagnosis method based on multi-scale parallel dilated convolution and long short-term memory(MSPDC-LSTM)is proposed. In order to effectively capture the multi-scale spatial features of propeller faults

a parallel dilated convolution module is introduced in the multi-scale network. This module processes the features at different scales in parallel and expands the receptive field without increasing the number of parameters

thereby enhancing the performance and efficiency of model when handling complex data. The network's ability to capture the temporal information characteristics of vibration signals is improved by integrating the LSTM module

further increasing the accuracy of processing the complex sequential data. The feature learning capability was enhanced by deeply extracting the spatiotemporal characteristics of the signals

significantly improving the robustness of fault diagnosis. To validate the feasibility of the proposed method

the experiments are conducted using AUV experimental data. The results show that the proposed method achieves an average accuracy of 99.67%

and even under strong noise interference at 2 dB

it still reaches an average accuracy of 95.97%. This demonstrates that the method can effectively identify and diagnose the failures of AUV propeller

thereby proving its effectiveness and robustness.

关键词

Keywords

references

闫敬,关新平.海上无人系统跨域集群发展现状及其关键技术[J].自动化学报, 2025, 51(4):744-761.

YAN J, GUAN X P. Development status and key technologies of cross-domain clustering of marine unmanned systems[J]. Acta Automatica Sinica, 2025, 51(4):744-761. (in Chinese)

潘云伟,李敏,曾祥光,等.基于人工势场和改进强化学习的AUV避障和航迹规划[J].兵工学报, 2025, 46(4):72-83.

PAN Y W, LI M, ZENG X G, et al. AUV obstacle avoidance and trajectory planning based on artificial potential field and improved reinforcement learning[J]. Acta Armamentarii, 2025, 46 (4):72-83. (in Chinese)

周国华,李林锋,吴轲娜,等.基于L-SHADE算法的AUV载体磁干扰参数辨识的数值模拟[J].兵工学报, 2024, 45 (8):2678-2687.

ZHOUGH,LI L F, WU K N, et al. Numerical simulation of magnetic interference parameter identification of AUV carrier based on L-SHADE algorithm[J]. Acta Armamentarii, 2024, 45(8):2678-2687. (in Chinese)

SUN Y S, LI Y M, ZHANG G C. Actuator fault diagnosis of autonomous underwater vehicle based on improved Elman neural network[J]. Journal of Central South University,2016,23 (4):808-816.

高立娥,刘卫东,路尧. X字型舵水下航行器故障诊断观测器设计与仿真研究[J].弹箭与制导学报, 2008, (4):222-224.

GAO L, LIU W, LU Y. Failure Diagnose Observer Design and Simulation for X-type Rudder Plane of Underwater Vehicle[J]. Journal of Projectiles, Rockets, Missiles and Guidance, 2008, (4):222-224. (in Chinese)

徐德民,刘富樯,张立川,等.基于改进连续-离散无迹卡尔曼滤波的水下航行器故障诊断[J].西北工业大学学报, 2014,32 (5):756-760.

XU D M, LIU F Q, ZHANG L C, et al. Fault diagnosis of underwater vehicle based on improved continuous-discrete unscented Kalman filter[J]. Journal of Northwestern Polytechnical University, 2014, 32(5):756-760. (in Chinese)

陈云赛,褚振忠,刘坤,等.深海潜水器推进器故障诊断技术研究进展[J].推进技术, 2020, 41(11):2465-2474.

CHEN Y S, CHU Z Z, LIU K, et al. Research progress on fault diagnosis technology of thrusters for deep-sea submersibles[J]. Propulsion Technology, 2020, 41 (11):2465-2474. (in Chinese)

TSAI C M, WANG C S, CHUNG Y J, et al. Multi-sensor fault diagnosis of underwater thruster propeller based on deep learning[J]. Sensors, 2021, 21: 7187.

YEO S J, CHOI W S, HONG S Y, et al. Enhanced convolutional neural network for in situ AUV thruster health monitoring using acoustic signals[J]. Sensors, 2022, 22(18) :7073.

于大程,朱晨光,张铭钧.自主式水下机器人推进器弱故障辨识方法[J].哈尔滨工程大学学报, 2020, 41(8):1223-1229.

YU D C, ZHU C G, ZHANG M J. The autonomous underwater vehicle thruster weak fault identification based on gray relational theory[J]. Journal of Harbin Engineering University, 2020, 41 (8):1223-1229. (in Chinese)

ZHANG J M, WU J, CHU Z Z. Multi-fault diagnosis for autonomous underwater vehicle based on fuzzy weighted support vector domain description[J]. China Ocean Engineering, 2014, 28(5):599-616.

LIANG H P, ZHAO X Q. Rolling bearing fault diagnosis based on one-dimensional dilated convolution network with residual connection[J]. IEEE Access, 2021, 9:31078-31091.

CHEN Y M, WANG Y Z, YU Y, et al. A fault diagnosis method for the autonomous underwater vehicle via meta-self-attention multi-scale CNN[J]. Journal of Marine Science and Engineering, 2023, 11(6):1121.

ZHANG Z T, ZHANG X F, YAN T H, et al. Data-driven fault detection of AUV rudder system:a mixture model approach[J]. Machines, 2023, 11(5):551.

DU W L, XIONG Z H, ZHU P X, et al. Enhancing underwater thruster anomaly detection with support vector glow encoding description[J]. Ocean Engineering, 2024, 314:119655.

XIA S X, ZHOU X F, SHI H B, et al. A fault diagnosis method based on attention mechanism with application in Qianlong-2 autonomous underwater vehicle[J]. Ocean Engineering, 2021, 233:109049.

刘宏哲,陶相如,徐成,等.一种优化多尺度特征融合的人体姿态估计方法[J].机械工程学报, 2024, 60(16):306-313.

LIU H Z, TAO X R, XU C, et al. Human pose estimation method based on optimized multi-scale feature fusion[J]. Journal of Mechanical Engineering, 2024, 60 (16): 306-313. (in Chinese)

陈刚,王国新,明振军,等.基于DBSCAN聚类和LSTM网络的装甲车辆集群轨迹预测方法[J].兵工学报,2024, 45 (12):4295-4310.

CHEN G, WANG G X, MING Z J, et al. Trajectory prediction method of armored vehicle swarm based on DBSCAN clustering and LSTM network[J]. Acta Armamentarii, 2024, 45 (12):4295-4310. (in Chinese)

QIN H H, PAN J F, LI J, et al. Fault diagnosis method of rolling bearing based on CBAM_ResNet and ACON activation function[J]. Applied Sciences, 2023, 13(13):7593.

WANG H C, XUE G Q, DU W L. An adaptive model for time-varying speed fault diagnosis under strong noise interference[J]. Journal of Mechanical Science and Technology, 2024, 38 (6) :2831-2844.

WANG H C, XUE G Q, YU L, et al. Time-varying speed fault diagnosis based on dual-channel parallel multi-scale information[J]. Journal of Mechanical Science and Technology, 2024, 38 (11):5961-5978.

WANG H C, XUE G Q, YU L, et al. Fault diagnosis of AUV propeller based on interference suppression CLSTM-CNN-Attention[C]∥Proceedings of the 2024 Prognostics and System Health Management Conference. Stockholm, Sweden: IEEE, 2024.

浏览量

下载量

CNKI被引量

文章被引用时，请邮件提醒。

提交

工具集

关联资源

基于改进CNN-LSTM的开关闩故障诊断方法

基于LSTM-GBSVDD模型的飞行轨迹异常检测方法

基于KLDA-IDBO-BP的装甲车发动机故障诊断

基于频率切片小波变换和注意力增强ConvNeXt模型的行星齿轮箱故障诊断