Fault diagnosis of axial piston pumps with multi-sensor data and convolutional neural network

Abstract

Axial piston pumps have wide applications in hydraulic systems for power transmission. Their condition monitoring and fault diagnosis are essential in ensuring the safety and reliability of the entire hydraulic system. Vibration and discharge pressure signals are two common signals used for the fault diagnosis of axial piston pumps because of their sensitivity to pump health conditions. However, most of the previous fault diagnosis methods only used vibration or pressure signal, and literatures related to multi-sensor data fusion for the pump fault diagnosis are limited. This paper presents an end-to-end multi-sensor data fusion method for the fault diagnosis of axial piston pumps. The vibration and pressure signals under different pump health conditions are fused into RGB images and then recognized by a convolutional neural network. Experiments were performed on an axial piston pump to confirm the effectiveness of the proposed method. Results show that the proposed multi-sensor data fusion method greatly improves the fault diagnosis of axial piston pumps in terms of accuracy and robustness and has better diagnostic performance than other existing diagnosis methods.

Graphical abstract

Keywords

axial piston pump / fault diagnosis / convolutional neural network / multi-sensor data fusion

References

[1] ChaoQ, ZhangJ H, XuB, WangQ N, LyuF, LiK. Integrated slipper retainer mechanism to eliminate slipper wear in high-speed axial piston pumps. Frontiers of Mechanical Engineering, 2022, 17( 1): 1– 13

[2] ChaoQ, XuZ, TaoJ F, LiuC L, ZhaiJ. Cavitation in a high-speed aviation axial piston pump over a wide range of fluid temperatures. Proceedings of the Institution of Mechanical Engineers, Part A: Journal of Power and Energy, 2022, 236( 4): 727– 737

[3] MaradeyLázaro J G, BorrásPinilla C. Detection and classification of wear fault in axial piston pumps: using ANNs and pressure signals. In: Burgos D A T, Vejar M A, Pozo F, eds. Pattern Recognition Applications in Engineering. Hershey: IGI Global, 2020, 286– 316

[4] XiaS Q, ZhangJ H, YeS G, XuB, HuangW D, XiangJ W. A spare support vector machine based fault detection strategy on key lubricating interfaces of axial piston pumps. IEEE Access, 2019, 7 : 178177– 178186

[5] LanY, HuJ W, HuangJ H, NiuL K, ZengX H, XiongX Y, WuB. Fault diagnosis on slipper abrasion of axial piston pump based on extreme learning machine. Measurement, 2018, 124 : 378– 385

[6] GuoR, ZhaoZ Q, WangT, LiuG H, ZhaoJ Y, GaoD R. Degradation state recognition of piston pump based on ICEEMDAN and XGBoost. Applied Sciences, 2020, 10( 18): 6593

[7] KellerN, SciancaleporeA, VaccaA. Demonstrating a condition monitoring process for axial piston pumps with damaged valve plates. International Journal of Fluid Power, 2022, 23( 2): 205– 236

[8] WangS H, XiangJ W, ZhongY T, TangH S. A data indicator-based deep belief networks to detect multiple faults in axial piston pumps. Mechanical Systems and Signal Processing, 2018, 112 : 154– 170

[9] ChaoQ, TaoJ F, WeiX L, WangY H, MengL H, LiuC L. Cavitation intensity recognition for high-speed axial piston pumps using 1-D convolutional neural networks with multi-channel inputs of vibration signals. Alexandria Engineering Journal, 2020, 59( 6): 4463– 4473

[10] ChaoQ, TaoJ F, WeiX L, LiuC L. Identification of cavitation intensity for high-speed aviation hydraulic pumps using 2D convolutional neural networks with an input of RGB-based vibration data. Measurement Science and Technology, 2020, 31( 10): 105102

[11] WangS H, XiangJ W. A minimum entropy deconvolution-enhanced convolutional neural networks for fault diagnosis of axial piston pumps. Soft Computing, 2020, 24( 4): 2983– 2997

[12] TangS N, YuanS Q, ZhuY, LiG P. An integrated deep learning method towards fault diagnosis of hydraulic axial piston pump. Sensors, 2020, 20( 22): 6576

[13] ChaoQ, GaoH H, TaoJ F, WangY H, ZhouJ, LiuC L. Adaptive decision-level fusion strategy for the fault diagnosis of axial piston pumps using multiple channels of vibration signals. Science China Technological Sciences, 2022, 65( 2): 470– 480

[14] TangS N, ZhuY, Yuan S Q. Intelligent fault diagnosis of hydraulic piston pump based on deep learning and Bayesian optimization. ISA Transactions, 2022 (in press)

[15] TangS N, ZhuY, YuanS Q. A novel adaptive convolutional neural network for fault diagnosis of hydraulic piston pump with acoustic images. Advanced Engineering Informatics, 2022, 52 : 101554

[16] LuC Q, WangS P, MakisV. Fault severity recognition of aviation piston pump based on feature extraction of EEMD paving and optimized support vector regression model. Aerospace Science and Technology, 2017, 67 : 105– 117

[17] WangY D, ZhuY, WangQ L, YuanS Q, TangS N, ZhengZ J. Effective component extraction for hydraulic pump pressure signal based on fast empirical mode decomposition and relative entropy. AIP Advances, 2020, 10( 7): 075103

[18] LuC Q, WangS P, ZhangC. Fault diagnosis of hydraulic piston pumps based on a two-step EMD method and fuzzy C-means clustering. Proceedings of the Institution of Mechanical Engineers, Part C: Journal of Mechanical Engineering Science, 2016, 230( 16): 2913– 2928

[19] YuH, LiH R, LiY L. Vibration signal fusion using improved empirical wavelet transform and variance contribution rate for weak fault detection of hydraulic pumps. ISA Transactions, 2020, 107 : 385– 401

[20] YuH, LiH R, LiY L, LiY F. A novel improved full vector spectrum algorithm and its application in multi-sensor data fusion for hydraulic pumps. Measurement, 2019, 133 : 145– 161

[21] SafizadehM S, LatifiS K. Using multi-sensor data fusion for vibration fault diagnosis of rolling element bearings by accelerometer and load cell. Information Fusion, 2014, 18 : 1– 8

[22] XiaM, LiT, XuL, LiuL Z, de SilvaC W. Fault diagnosis for rotating machinery using multiple sensors and convolutional neural networks. IEEE/ASME Transactions on Mechatronics, 2018, 23( 1): 101– 110

[23] WangH Q, LiS, SongL Y, CuiL L. A novel convolutional neural network based fault recognition method via image fusion of multi-vibration-signals. Computers in Industry, 2019, 105 : 182– 190

[24] GongW F, ChenH, ZhangZ H, ZhangM L, WangR H, GuanC, WangQ. A novel deep learning method for intelligent fault diagnosis of rotating machinery based on improved CNN-SVM and multichannel data fusion. Sensors, 2019, 19( 7): 1693

[25] WangJ J, FuP L, ZhangL B, GaoR X, ZhaoR. Multilevel information fusion for induction motor fault diagnosis. IEEE/ASME Transactions on Mechatronics, 2019, 24( 5): 2139– 2150

[26] ChenH P, HuN Q, ChengZ, ZhangL, ZhangY. A deep convolutional neural network based fusion method of two-direction vibration signal data for health state identification of planetary gearboxes. Measurement, 2019, 146 : 268– 278

[27] AzamfarM, SinghJ, Bravo-ImazI, LeeJ. Multisensor data fusion for gearbox fault diagnosis using 2-D convolutional neural network and motor current signature analysis. Mechanical Systems and Signal Processing, 2020, 144 : 106861

[28] KolarD, LisjakD, PająkM, PavkovićD. Fault diagnosis of rotary machines using deep convolutional neural network with wide three axis vibration signal input. Sensors, 2020, 20( 14): 4017

[29] YanX S, SunZ, ZhaoJ J, ShiZ G, ZhangC A. Fault diagnosis of rotating machinery equipped with multiple sensors using space-time fragments. Journal of Sound and Vibration, 2019, 456 : 49– 64

[30] ChaoQ, ZhangJ H, XuB, HuangH P, PanM. A review of high-speed electro-hydrostatic actuator pumps in aerospace applications: challenges and solutions. Journal of Mechanical Design, 2019, 141( 5): 050801

[31] MaJ M, ChenJ, LiJ, LiQ L, RenC Y. Wear analysis of swash plate/slipper pair of axis piston hydraulic pump. Tribology International, 2015, 90 : 467– 472

[32] HuangJ H, YanZ, QuanL, LanY, GaoY S. Characteristics of delivery pressure in the axial piston pump with combination of variable displacement and variable speed. Proceedings of the Institution of Mechanical Engineers, Part I: Journal of Systems and Control Engineering, 2015, 229( 7): 599– 613

[33] ChaoQ, TaoJ F, LeiJ B, WeiX L, LiuC L, WangY H, MengL H. Fast scaling approach based on cavitation conditions to estimate the speed limitation for axial piston pump design. Frontiers of Mechanical Engineering, 2021, 16( 1): 176– 185

[34] ChaconR, IvantysynovaM. Virtual prototyping of axial piston machines: numerical method and experimental validation. Energies, 2019, 12( 9): 1674

[35] DahlG E, SainathT N, HintonG E. Improving deep neural networks for LVCSR using rectified linear units and dropout. In: Proceedings of 2013 IEEE International Conference on Acoustics, Speech and Signal Processing. Vancouver: IEEE, 2013, 8609– 8613

[36] StankovicL, DakovićM, ThayaparanT. Time–Frequency Signal Analysis with Applications. Boston: Artech House, 2013

[37] LeCunY, BottouL, BengioY, HaffnerP. Gradient-based learning applied to document recognition. Proceedings of the IEEE, 1998, 86( 11): 2278– 2324

[38] ZhangW, LiC H, PengG L, ChenY H, ZhangZ J. A deep convolutional neural network with new training methods for bearing fault diagnosis under noisy environment and different working load. Mechanical Systems and Signal Processing, 2018, 100 : 439– 453

[39] LiuX C, ZhouQ C, ZhaoJ, ShenH H, XiongX L. Fault diagnosis of rotating machinery under noisy environment conditions based on a 1-D convolutional autoencoder and 1-D convolutional neural network. Sensors, 2019, 19( 4): 972

[40] TangS N, ZhuY, YuanS Q, LiG P. Intelligent diagnosis towards hydraulic axial piston pump using a novel integrated CNN model. Sensors, 2020, 20( 24): 7152

[41] JiangW L, WangC Y, ZouJ Y, ZhangS Q. Application of deep learning in fault diagnosis of rotating machinery. Processes, 2021, 9( 6): 919

Nomenclature

Abbreviations
1D	One-dimensional
2D	Two-dimensional
CNN	Convolutional neural network
SNR	Signal-to-noise ratio
STFT	Short-time Fourier transform
Variables
ahw	Feature map element at pixel (h, w) in the pooling window
A_k^l	The kth feature map at layer l
{\mathit{B}}_k^l	Bias of the kth group filter at layer l
c	Index of channels for input feature maps or the group filters
C	Total number of filter channels
f(·)	Activation function
H	Pooling window height
i^{\prime }	Height index of element pixels
j	Imaginary unit
j^{\prime }	Width index of element pixels
J	Loss function
k	Index of group filters or output feature maps
l	Index of network layers
L	Total layer number
m, n	Indices of discrete sampling points
N	Size of Hanning window
p_k^{l+1}	Maximum element in the pooling window
q	The qth class
Q	Total classification number
s	Index of samples
S	Total number of samples
t	Time
x(τ)	Vibration signal
x(s)	The sth sample
{\mathit{X}}_c^{l-1}	The cth-channel component of the input feature map at layer (l – 1)
{\mathit{X}}_k^l	The kth output feature map at layer l
y(s)	Predicted label
w(τ), w*(τ)	Window function and its conjugated form
W	Pooling window width
{\mathit{W}}_{c,k}^l	The cth-channel component of the kth group filter weight at layer l
η	Learning rate
{\mathit{\theta }}^L	Trainable parameters at the last layer L
θnew, θold	Trainable parameters after and before update, respectively
\tau	Time variable of integration
\omega	Angular frequency

Acknowledgements

This study was supported by the National Key R&D Program of China (Grant No. 2018YFB1702503), the Open Foundation of the State Key Laboratory of Fluid Power and Mechatronic Systems, China (Grant No. GZKF-202108), the National Postdoctoral Program for Innovative Talents, China (Grant No. BX20200210), the China Postdoctoral Science Foundation (Grant No. 2019M660086), and Shanghai Municipal Science and Technology Major Project, China (Grant No. 2021SHZDZX0102).

RIGHTS & PERMISSIONS

2022 Higher Education Press 2022