针对高速列车减振器故障特征手工提取困难导致故障识别困难的难题,提出了基于一维残差卷积注意力(1DRCA)的故障诊断算法,对高速列车中抗蛇行减振器4种状态进行识别。首先,构建卷积层进行特征提取,利用卷积块注意力模块在通道和空间维度上进行自适应特征优化;然后,建立残差神经网络模型,利用残差信息调整权值参数;最后,通过试验证明了该方法对于抗蛇行减振器的4种状态的故障识别是可行的,可以准确地识别正常、启动不良、对称速率故障和锯齿波故障,所提出的方法的平均准确率达到99%左右。为了进一步证明所提出模型的泛化性,采用滚动轴承的故障数据来验证了所提出模型的有效性和准确性,结果表明所提的模型较好地实现滚动轴承不同故障状态的诊断。
Abstract
To address the challenge of difficulty in fault identification caused by manually extracting fault characteristics from high-speed train dampeners, a fault diagnosis algorithm based on one-dimensional residual convolutional attention (1DRCA) is proposed. This algorithm is designed to recognize the four states in the yaw damper of high-speed trains. Firstly, a convolutional layer is constructed for feature extraction, utilizing the convolutional block attention module to adaptively optimize features in both channel and spatial dimensions. Additionally, a residual neural network model is established, adjusting weight parameters using residual information. The experimental validation demonstrates the feasibility of this method for recognizing the four states of the yaw damper, accurately identifying normal, poor startup, symmetry rate faults, and sawtooth faults, with the proposed method achieving an average accuracy rate of approximately 99%. To further substantiate the model's generalization capability, fault data from rolling bearings is employed to validate the effectiveness and accuracy of the proposed model. The results indicate that the proposed model effectively diagnoses different fault states in rolling bearings.
关键词
振动测量 /
抗蛇行减振器 /
故障诊断 /
残差卷积 /
注意力机制
Key words
vibration measurement /
yaw damper /
fault diagnosis /
residual convolution /
attention mechanism
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
［1］黄彩虹, 曾京, 宋春元. 高速列车抗蛇行减振器的简化物理参数模型［J］. 铁道学报, 2021, 43(7): 47-56.
HUANG C H, ZENG J, SONG C Y. Simplified physical parameter model of high-speed train yaw damper［J］. Journal of the China Railway Society, 2021, 43(7): 47-56.
［2］姚远, 程俊, 张名扬, 等. 高速列车抗蛇行减振器作用机制与频变刚度应用研究［J］. 振动工程学报, 2022, 35(6): 1461-1470.
YAO Y, CHENG J, ZHANG M Y, et al. Mechanism analysis of yaw damper in high-speed train and frequency-dependent stiffness application［J］. Journal of Vibration Engineering, 2022: 35(6): 1461-1470.
［3］大庭拓也, 蔡千华. 基于振动分析的新干线转向架的状态监视［J］. 国外铁道车辆, 2013, 50(3): 32-40.
TAKUYA O, CAI Q H. The condition monitoring for shinkansen bogies based on vibration analysis［J］. Foreign Rolling Stock, 2013, 50(3): 32-40.
［4］岑潮宇, 代亮成, 池茂儒, 等.基于精细复合多尺度散布熵的抗蛇行减振器故障诊断［J/OL］.铁道科学与工程学报:1-11. DOI: 10.19713/j.cnki.43-1423/u.T20232079.
CENG C Y, DAI L C, CHI M R, et al. Yaw damper fault diagnosis based on refined composite multiscale dispersion entropy［J/OL］. Journal of Railway Science and Engineering: 1-11. DOI: 10.19713/j.cnki.43-1423/u. T20232079.
［5］郑航, 李刚, 李德仓. 基于PELCD样本熵的抗蛇行减振器故障诊断［J］. 电力机车与城轨车辆, 2023, 46(6): 34-41.
ZHENG H, LI G, LI D C. Fault diagnosis of anti-yaw damper based on PELCD sample entropy［J］. Electric Locomotives & Mass Transit Vehicles, 2023, 46(06): 34-41.
［6］秦娜, 金炜东, 黄进, 等. 基于EEMD样本熵的高速列车转向架故障特征提取［J］. 西南交通大学学报, 2014, 49(1): 27-32.
QIN N, JIN W D, HUANG J, et al. Feature extraction of high speed train bogie based on ensemble empirical mode decomposition and sample entropy［J］. Journal of Southwest Jiaotong University, 2014, 49(1): 27-32.
［7］ZHANG W, JIA M, ZHU L, et al. Comprehensive overview on computational intelligence techniques for machinery condition monitoring and fault diagnosis［J］. Chinese Journal of Mechanical Engineering, 2017, 30(4): 782-795.
［8］侯文擎, 叶鸣, 李巍华. 基于改进堆叠降噪自编码的滚动轴承故障分类［J］. 机械工程学报, 2018, 54(7): 87-96.
HOU W Q, YE M, LI W H. Rolling element bearing fault classification using improved stacked de-noising auto-encoders［J］. Journal of Mechanical Engineering, 2018, 54(7): 87-96.
［9］金海龙, 马吴旭, 孟宗, 等. 基于改进1DCNN-GRU的滚动轴承故障诊断［J］. 计量学报, 2023, 44(9): 1423-1428.
JIN H L, MA W X, MENG Z, et al. An improved 1DCNN-GRU for rolling bearing fault diagnosis［J］. Acta Metrologica Sinica, 2023, 44(9): 1423-1428.
［10］贾美霞, 韩宝坤, 王金瑞, 等. 基于迁移堆栈自编码器的轴承故障诊断方法［J］. 噪声与振动控制, 2021, 41(6): 84-89.
JIA M X, HAN B D, WANG J R, et al. Bearing fault diagnosis method based on transfer learning and stacked auto-encoders［J］. Noise and Vibration Control, 2021, 41(6): 84-89.
［11］侯东晓, 周子安, 程荣财, 等. 基于GADF-TL-ResNeXt的滚动轴承故障诊断方法［J］. 计量学报, 2023, 44(10): 1534-1542.
HOU D X, ZHOU Z A, CHENG R C, et al. Based on the GADF-TL-ResNeXt rolling bearing fault diagnosis method［J］. Acta Methrologica Sinica, 2023, 44(10): 1534-1542.
［12］陈保家, 陈学力, 沈保明, 等. CNN-LSTM深度神经网络在滚动轴承故障诊断中的应用［J］. 西安交通大学学报, 2021, 55(6): 28-36.
CHEN B J, CHEN X L, SHEN B M, et al. An application of convolution neural network and long short-term memory in rolling bearing fault diagnosis［J］. Journal of Southwest Jiaotong University, 2021, 55(6): 28-36.
［13］周兴康, 余建波. 基于深度一维残差卷积自编码网络的齿轮箱故障诊断［J］. 机械工程学报, 2020, 56(7): 96-108.
ZHOU X K, YU J B. Gearbox Fault Diagnosis Based on One-dimension residual convolutional auto-encoder［J］. Journal of Mechanical Engineering, 2020, 56(7): 96-108.
［14］陈广, 马闻达, 孙泽明, 等.基于卷积神经网络的高速列车抗蛇行减振器故障诊断［J］. 机床与液压, 2023, 51(8): 194-199.
CHEN G, MA W D, SUN Z M, et al. Fault diagnosis of yaw damper in high-speed train based on convolutional neural network［J］. Machine Tool & Hydraulics, 2023, 51(8): 194-199.
［15］CHEN X, ZHANG B, GAO D. Bearing fault diagnosis base on multi-scale CNN and LSTM model［J］. Journal of Intelligent Manufacturing, 2021, 32: 971-987.
［16］ZHANG W, LI C H, PENG G L, et al. A deep convolutional neural network with new training methods for bearing fault diagnosis under noisy environment and different working load［J］. Mechanical systems and signal processing, 2018, 100: 439-453.
［17］WANG H, LIU Z, PENG D, et al. Understanding and learning discriminant features based on multiattention 1DCNN for wheelset bearing fault diagnosis［J］. IEEE Transactions on Industrial Informatics, 2019, 16(9): 5735-5745.
PDF(894 KB)
