Degradation State Recognition of Piston Pump Based on ICEEMDAN and XGBoost

被引:45
作者
Guo, Rui [1 ,2 ]
Zhao, Zhiqian [1 ,3 ]
Wang, Tao [1 ,4 ]
Liu, Guangheng [1 ,4 ]
Zhao, Jingyi [2 ,4 ]
Gao, Dianrong [3 ]
机构
[1] Yanshan Univ, Hebei Prov Key Lab Heavy Machinery Fluid Power Tr, Qinhuangdao 066004, Hebei, Peoples R China
[2] Zhejiang Univ, State Key Lab Fluid Power & Mechatron Syst, Hangzhou 310027, Peoples R China
[3] Yanshan Univ, Key Lab Adv Forging & Stamping Technol & Sci, Qinhuangdao 066004, Hebei, Peoples R China
[4] Yanshan Univ, Hebei Key Lab Special Delivery Equipment, Qinhuangdao 066004, Hebei, Peoples R China
来源
APPLIED SCIENCES-BASEL | 2020年 / 10卷 / 18期
基金
中国国家自然科学基金;
关键词
piston pump; degraded state recognition; slipper; improved complete ensemble empirical mode decomposition with adaptive noise; principal component analysis; eXtreme gradient boosting; EMPIRICAL MODE DECOMPOSITION; FAULT-DIAGNOSIS; CLASSIFICATION; IDENTIFICATION; COMBINATION; ALGORITHM; MACHINE;
D O I
10.3390/app10186593
中图分类号
O6 [化学];
学科分类号
0703 ;
摘要
Under different degradation conditions, the complexity of natural oscillation of the piston pump will change. Given the difference of the characteristic values of the vibration signal under different degradation states, this paper presents a degradation state recognition method based on improved complete ensemble empirical mode decomposition with adaptive noise (ICEEMDAN) and eXtreme gradient boosting (XGBoost) to improve the accuracy of state recognition. Firstly, ICEEMDAN is proposed to alleviate the mode mixing phenomenon, which decomposes the vibration signal and obtain the intrinsic mode functions (IMFs) with less noise and more physical meaning, and subsequently the optimal IMF is found by using the correlation coefficient method. Then, the time domain, frequency domain, and entropy of the effective IMF are calculated, and the new characteristic values which can represent the degradation state are selected by principal component analysis (PCA) that it realizes dimension reduction. Finally, the above-mentioned characteristic indexes are used as the input of the XGBoost algorithm to achieve the recognition of the degradation state. In this paper, the vibration signals of four different degradation states are generated and analyzed through the piston pump slipper degradation experiment. By comparing the proposed method with different state recognition algorithms, it can be seen that the method based on ICEEMDAN and XGBoost is accurate and efficient, the average accuracy rate can reach more than 99%. Therefore, this method can more accurately describe the degradation state of the piston pump and has a highly practical application value.
引用
收藏
页数:17
相关论文
共 54 条
[1]   Classification of repeated measurements data using tree-based ensemble methods [J].
Adler, Werner ;
Potapov, Sergej ;
Lausen, Berthold .
COMPUTATIONAL STATISTICS, 2011, 26 (02) :355-369
[2]   Significant wave height forecasting via an extreme learning machine model integrated with improved complete ensemble empirical mode decomposition [J].
Ali, Mumtaz ;
Prasad, Ramendra .
RENEWABLE & SUSTAINABLE ENERGY REVIEWS, 2019, 104 :281-295
[3]   Evaluation of Tree-Based Ensemble Machine Learning Models in Predicting Stock Price Direction of Movement [J].
Ampomah, Ernest Kwame ;
Qin, Zhiguang ;
Nyame, Gabriel .
INFORMATION, 2020, 11 (06)
[4]   A flexible algorithm for fault diagnosis in a centrifugal pump with corrupted data and noise based on ANN and support vector machine with hyper-parameters optimization [J].
Azadeh, A. ;
Saberi, M. ;
Kazem, A. ;
Ebrahimipour, V. ;
Nourmohammadzadeh, A. ;
Saberi, Z. .
APPLIED SOFT COMPUTING, 2013, 13 (03) :1478-1485
[5]   Permutation entropy: A natural complexity measure for time series [J].
Bandt, C ;
Pompe, B .
PHYSICAL REVIEW LETTERS, 2002, 88 (17) :4
[6]   Diagnostic and decision support systems by identification of abnormal events: Application to helicopters [J].
Bect, Pierre ;
Simeu-Abazi, Zineb ;
Maisonneuve, Pierre Lois .
AEROSPACE SCIENCE AND TECHNOLOGY, 2015, 46 :339-350
[7]   Advanced machine learning techniques for building performance simulation: a comparative analysis [J].
Chakraborty, Debaditya ;
Elzarka, Hazem .
JOURNAL OF BUILDING PERFORMANCE SIMULATION, 2019, 12 (02) :193-207
[8]   Circular object recognition based on shape parameters [J].
Chen Aijun ;
Li Jinzong ;
Zhu Bing .
JOURNAL OF SYSTEMS ENGINEERING AND ELECTRONICS, 2007, 18 (02) :199-204
[9]   Virtual metrology of semiconductor PVD process based on combination of tree-based ensemble model [J].
Chen, Ching-Hsien ;
Zhao, Wei-Dong ;
Pang, Timothy ;
Lin, Yi-Zheng .
ISA TRANSACTIONS, 2020, 103 :192-202
[10]   XGBoost: A Scalable Tree Boosting System [J].
Chen, Tianqi ;
Guestrin, Carlos .
KDD'16: PROCEEDINGS OF THE 22ND ACM SIGKDD INTERNATIONAL CONFERENCE ON KNOWLEDGE DISCOVERY AND DATA MINING, 2016, :785-794