Abstract: A fault diagnosis method based on improved extreme learning machine (IELM) is proposed to solve the
weakness (weak generalization ability, low diagnostic rate) of traditional fault diagnosis with feedforward neural
network algorithm. This method fuses signal feature vectors, extracts six parameters as the principal component
analysis (PCA) variables, and calculates correlation coefficient matrix among the variables. The weight values
of control parameters in the extreme learning model are dynamically adjusted according to the test samples’
constantly changing. Consequently, the weight fixed drawback in the original model can be remedied. A fault
simulation experiment platform for wind turbine drive system is built, eight kinds of fault modes are diagnosed by
the improved extreme learning model, and the result is compared with that of other machine learning methods.
The experiment indicates that the method can enhance the accuracy and generalization ability of diagnosis, and
increase the computing speed. It is convenient for engineering application.
WU Bin1* (吴斌), XI Lifeng2 (奚立峰), FAN Sixia1 (范思遐), ZHAN Jian1 (占健)
. Fault Diagnosis for Wind Turbine Based on Improved Extreme Learning Machine[J]. Journal of Shanghai Jiaotong University(Science), 2017
, 22(4)
: 466
-473
.
DOI: 10.1007/s12204-017-1849-x
[1] WU C Y, LIU J, PENG F Q, et al. Gearbox fault diagnosisusing adaptive zero phase time-varying filterbased on multi-scale chirplet sparse signal decomposition[J]. Chinese Journal of Mechanical Engineering,2013, 26(4): 831-838.
[2] CHEN X M, YU D J, LI R. Analysis of gearbox compoundfault vibration signal using morphological componentanalysis [J]. Journal of Mechanical Engineering,2014, 50(3): 108-115 (in Chinese).
[3] CHEN G H, QIE L F, ZHANG A J, et al. ImprovedCICA algorithm used for single channel compoundfault diagnosis of rolling bearings [J]. Chinese Journalof Mechanical Engineering, 2016, 29(1): 204-211.
[4] FELDMAN M. Non-linear free vibration identificationvia the Hilbert transform [J]. Journal of Sound andVibration, 1997, 208(3): 475-489.
[5] CUI L L, MA C Q, ZHANG F B, et al. Quantitativediagnosis of fault severity trend of rolling elementbearings [J]. Chinese Journal of Mechanical Engineering,2015, 28(6): 1254-1260.
[6] LI Q, ATLAS L. Coherent modulation filtering forspeech [C]//Proceedings of the IEEE InternationalConference on Acoustics, Speech and Signal Processing.Las Vegas, NV: IEEE, 2008: 4481-4484.
[7] WANG Y B. Fault diagnosis of mine ventilator basedon neural network [D]. Shanghai: School of ElectronicInformation and Electrical Engineering, Shanghai JiaoTong University, 2012 (in Chinese).
[8] JAOUHER B A, NADER F, LOTFI S, et al. Applicationof empirical mode decomposition and artificialneural network for automatic bearing fault diagnosisbased on vibration signals [J]. Applied Acoustics, 2015,89: 16-27.
[9] LI W H, WENG S L, ZHANG S H. A firefly neuralnetwork and its application in bearing fault diagnosis[J]. Journal of Mechanical Engineering, 2015, 51(7):99-106 (in Chinese).
[10] LI N, WANG L G, JIA M T, et al. Fault intelligentdiagnosis system for fan based on information fusion[J]. Journal of Central South University (Science andTechnology), 2013, 44(7): 2861-2866 (in Chinese).
[11] HE Y, WANG G H, GUAN X. Information fusion theoryand application [M]. Beijing: Publishing House ofelectronics industry, 2010: 17-42 (in Chinese).
[12] XU X G, WANG S L, LIU J L. Mechanical fault diagnosisof fan based on wavelet packet energy analysisand improved support vector machine [J]. Journal ofChinese Society of Power Engineering, 2013, 33(8):606-612 (in Chinese).
[13] NIKOLAOU N G, ANTONIADIS I A. Rolling elementbearing fault diagnosis using wavelet packets [J]. NDT& E International, 2002, 35(3): 197-205.
[14] WANG Z Y, CHEN J, XIAO W B, et al. Fault diagnosisof rolling element bearing based on constrainedindependent component analysis [J]. Journal of Vibrationand Shock, 2012, 31(9): 118-122 (in Chinese).
[15] WANG Y, XU G H, LIANG L, et al. Detection ofweak transient signals based on wavelet packet transformand manifold learning for rolling element bearingfault diagnosis [J]. Mechanical Systems and SignalProcessing, 2015, 54/55: 259-276.
[16] HUANG G B, ZHU Q Y, SIEWC K. Extreme learningmachine: Theory and applications [J]. Neuroputing,2006, 70: 489-451.
[17] YUAN J H, ZHANG L W, WANG Y, et al. Study oftransformers fault diagnosis based on extreme learningmachine [J]. Electrical Measurement & Instrumentation,2013, 50(12): 21-26.
[18] RONG H J, HUANG G B, SUNDARARAJANN, et al.Online sequential fuzzy extreme learning machine forfunction approximation and classification problems [J].IEEE Transactions on Systems Man and CyberneticsPart B, 2009, 39(4): 1067-1072.
[19] YANG Y M. Researches on extreme learning theory forsystem identification and applications [D]. Changsha:College of Electrical and Information Engineering, HunanUniversity, 2013: 19-22 (in Chinese).
[20] WANG H L, HE X, LU J H, et al. Analog circuit onlinefault diagnosis based on fix-size sequence extremelearning machine [J]. Chinese Journal of Scientific Instrument,2014, 35(4): 738-744 (in Chinese).
