Real-time SVD-based detection of multiple combined faults in induction motors

https://doi.org/10.1016/j.compeleceng.2013.12.020Get rights and content

Highlights

  • Novel methodology for induction motor condition monitoring is proposed.

  • Singular value decomposition, statistical analysis and an artificial neural network.

  • One-broken-bar, unbalance, faulty bearing and their combinations are identified.

  • One-phase of the startup-transient current of the motor is analyzed in real time.

Abstract

Early detection of induction motor faults has been a main subject of investigation for many years. Several approaches have been proposed for identifying one or more faults treated in an isolated way. Multiple combined faults on induction motors represent a big challenge since the reliable diagnosis of a faulty condition under the presence of two or more simultaneous faults is really difficult. This work introduces a novel methodology that merges singular value decomposition, statistical analysis, and artificial neural networks for multiple combined fault identification. Obtained results demonstrate its high effectiveness on detecting faulty bearings, unbalance, broken rotor bars, and all their possible combinations. The developed field programmable gate array-based implementation offers a portable low-cost solution for online classification of the rotating machine condition in real time. Thanks to its generalized nature, the introduced approach can be extended for detecting multiple combined faults under different working conditions by a proper calibration.

Introduction

Induction motors are key elements of the electrical and processing industry. Their utilization is widespread because of their low price, stiffness, and reliability. Therefore, early detection of induction motor faults has been a matter of research for many years [1], [2], [3], since unexpected failures cause interruptions on production lines with severe consequences in cost, product quality, and safety [4], [5], [6]. Several approaches have been proposed for induction fault detection; unfortunately, most of them focus on identifying one single fault (i.e. an isolated fault) or multiple faults (i.e. different faults treated in an isolated way). In rotating machines more than 40% of the faults are bearing related [7], around 10% are rotor faults [8], and unbalance is inside the 12% of other faults [9]; moreover, in a real rotary machine two or more faults can be present at the same time; so that, one faulty condition could interfere on the detection of another one misleading to a wrong decision about the operational condition of the motor. The identification of multiple-combined faults on induction motors still represents a big challenge for condition monitoring, but it has been rarely considered despite it can be a very usual situation, because a reliable diagnosis of a faulty condition under the presence of two or more simultaneous faults is really difficult [10].

The main contribution of this work is a novel, non invasive methodology that merges singular value decomposition (SVD), statistical analysis, and an artificial neural network (ANN) for multiple combined fault identification in induction motors, outperforming most of the previously proposed approaches, which identify single faults (i.e. an isolated fault) or multiple faults (i.e. different faults treated in an isolated way). A field programmable gate array (FPGA)-based implementation of the propose methodology is developed to offer an online deterministic technique suited for real-time identification of single or multiple combined faults in early stages by analyzing one-phase of the steady-state current, unlike common approaches for induction motor fault detection that rely on the offline examination of current or/and vibration signatures to identify single isolated faults. Obtained results show the high performance of the proposed technique, implemented in a low-cost FPGA device from two different vendors, for identifying in a quantitative and automatic way outer-race bearing defect, unbalance, broken rotor bars, and all their possible combinations in an induction machine.

The remainder of this paper is organized as follows: the problem formulation is given in Section 2. Section 3 describes the proposed methodology. The experimental setup and obtained results for testing the effectiveness of the proposed methodology are presented in Section 4. Conclusions and remarks are given in Section 5.

Section snippets

Problem formulation

Many approaches for induction motor fault detection have been proposed; the aim in most of them is to provide computationally-effective diagnostic tools able to identify different faults by analyzing a minimum set of parameters [11]. Common signal processing techniques for induction motor condition monitoring are the fast Fourier transform (FFT) and the wavelet transform. For instance, in [12], an automatic computerized system is presented for diagnosing rotor bars in induction motor utilizing

Singular value decomposition

Singular value decomposition (SVD) of an m × n matrix A is defined byA=UΣVT,where U and V are m × m and n × n orthogonal matrices, respectively, i.e. UUT = Im, and VVT = In where I is the identity matrix. Σ is a diagonal matrix such that Σ = diag(σ1, σ2,  , σn), where σ1, σ2,  , σn are the singular values of A [31]. For m > n, at least m  n singular values are zero. If r = rank(A) and r < n, r of the singular values are non-zero. Matrix U contains m left singular vectors, and matrix V contains n right singular

Experimental setup

The steady-state current signal is used to find the multiple combined faults and classify the operational conditions since this signal describes the dynamic characteristics of the induction motor. Fig. 5a shows the experiment setup where different 1-hp three-phase induction motors (model WEG 00136APE48T) are used to test the performance of the proposed methodology identifying the single and multiple combined fault conditions treated in this work. The tested motors have 2 poles, 28 bars, and

Conclusions

Early fault detection in induction motor has been a subject of great interest for researchers in recent years. Most of the current techniques focus on the identification of faulty conditions treated in an isolated way; since, the correct classification of one fault, when another or more faults are present, is really difficult. In this regard, a novel methodology for multiple combined fault detection in induction motors was proposed and experimentally tested in this work. Obtained results showed

Acknowledgment

This work was supported in part by the National Council on Science and Technology (CONACYT), Mexico, under Scholarships Number: 336730.

M. Hernandez-Vargas received the B.E. degree from the FIMEE Universidad de Guanajuato, Mexico, in 2004, and the M.C. degree from the Universidad Autonoma de Queretaro, Mexico, in 2006. Currently, he is pursuing the Ph.D. degree at the Universidad de Guanajuato, Mexico. His fields of interest include real-time digital signal processing on FPGAs for applications in mechatronics.

References (43)

  • B. Liang et al.

    Application of power spectrum, cepstrum, higher order spectrum and neural network analyses for induction motor fault diagnosis

    Mech Syst Signal Process

    (2013)
  • M. Barakat et al.

    Self adaptive growing neural network classifier for faults detection and diagnosis

    Neurocomputing

    (2011)
  • E. Cabal-Yepez et al.

    Novel hardware processing unit for dynamic on-line entropy estimation of discrete time information

    Digit Signal Process

    (2010)
  • A. Katidiotis et al.

    Performance evaluation of artificial neural network-based learning schemes for cognitive radio systems

    Comput Electr Eng

    (2010)
  • O. Deperlioglu et al.

    An educational tool for artificial neural networks

    Comput Electr Eng

    (2011)
  • E. Cabal-Yepez et al.

    FPGA-based entropy neural processor for online detection of multiple combined faults on induction motors

    Mech Syst Signal Process

    (2012)
  • J. Seshadrinath et al.

    Investigation of vibration signatures for multiple fault diagnosis in variable frequency drives using complex wavelets

    IEEE Trans Power Electr

    (2014)
  • A. Soualhi et al.

    Detection and diagnosis of faults in induction motor using an improved artificial ant clustering technique

    IEEE Trans Ind Electron

    (2013)
  • J. Antonino-Daviu et al.

    Scale invariant feature extraction algorithm for the automatic diagnosis of rotor asymmetries in induction motors

    IEEE Trans Ind Inf

    (2013)
  • W. Cao et al.

    Overview of electric motor technologies used for more electric aircraft (MEA)

    IEEE Trans Ind Electron

    (2012)
  • A.V. Rocha et al.

    Increasing long-belt-conveyor availability by using fault-resilient medium-voltage AC drives

    IEEE Trans Ind Appl

    (2012)
  • Cited by (0)

    M. Hernandez-Vargas received the B.E. degree from the FIMEE Universidad de Guanajuato, Mexico, in 2004, and the M.C. degree from the Universidad Autonoma de Queretaro, Mexico, in 2006. Currently, he is pursuing the Ph.D. degree at the Universidad de Guanajuato, Mexico. His fields of interest include real-time digital signal processing on FPGAs for applications in mechatronics.

    E. Cabal-Yepez received the Ph.D. degree from the University of Sussex in the United Kingdom. The B.E. and M.E. degrees from FIMEE Universidad de Guanajuato, Mexico, where he is currently working as a full-time professor and doing research work focused on Digital Signal and Image Processing, FPGAs, and Embedded Systems for real-time processing.

    A. Garcia-Perez received the B.E. and M.E. degrees in electronics from the University of Guanajuato, Mexico, and the Ph.D. degree in electrical engineering from the University of Texas, Dallas, USA. He is currently a Titular Professor with the Department of Electronic Engineering, University of Guanajuato. His fields of interest include digital signal processing.

    Reviews processed and recommended for publication to Editor-in-Chief by Guest Editor Dr. Zhihong Man.

    View full text