Research ArticleFault Detection for the Scraper Chain Based on VibrationAnalysis Using the Adaptive Optimal KernelTime-Frequency Representation

Xing Zhang 12 Wei Li 12 Zhencai Zhu12 Shanguo Yang12 and Fan Jiang 12

1School of Mechatronic Engineering China University of Mining and Technology Xuzhou 221116 China2Jiangsu Key Laboratory of Mine Mechanical and Electrical Equipment China University of Mining and TechnologyXuzhou 221116 China

Correspondence should be addressed to Wei Li liwei_cmee163com

Received 14 April 2019 Accepted 13 June 2019 Published 24 July 2019

Academic Editor Chao Tao

Copyright copy 2019 Xing Zhang et al is is an open access article distributed under the Creative Commons Attribution Licensewhich permits unrestricted use distribution and reproduction in any medium provided the original work is properly cited

A scraper conveyor is a key component of large-scale mechanized coal mining equipment and its failure patterns are mainlycaused by chain jam and chain fracture Due to the difficulties with direct measurement for multiple performance parameters ofthe scraper chain this paper deals with a novel strategy for fault detection of the scraper chain based on vibration analysis of thechute First a chute vibration model (CVM) is applied for modal analysis and the hammer impact test (HIT) is conducted tovalidate the accuracy of the CVM second the measuring points for vibration analysis of the chute are determined based on themodal assurance criterion (MAC) and third to simulate the actual vibration properties of the chute a dynamic transmissionsystem model (DTSM) is constructed based on finite element modeling e fixed-point experimental testing (FPET) is thenconducted to indicate the correctness of simulation results Subsequently the DTSM-based vibration responses of the chute underdifferent operating conditions are obtained In this paper the proposed strategy is employed to determine the occurrence of chainfaults by amplitude comparisons while failure patterns are distinguished by the adaptive optimal kernel time-frequencyrepresentation (AOKR)

1 Introduction

A scraper conveyor is a piece of continuous mechanicaltransport equipment with a chain-type traction fortransporting bulk materials in coal mining face [1] AsFigure 1 describes the transmission system of the scraperconveyor is a complex system with multiple bodies thatmainly comprise sprockets scrapers scraper chains andchutes It operates as a closed loop by linking the scraperchains together and the scraper chains run around thesprockets In chain assembly failures most malfunctionsmainly involve jam and fracture of the scraper chains Maincauses for the jam and fracture failures of the scraper chainsinclude the following aspects uneven loads on the chutes[2] overlong laying length of the scraper conveyor [3 4]and complex interactions between the core components ofthe chain assembly [5 6] In order to monitor the working

state of the scraper conveyor there is an urgent need forfault detection of the scraper chains

Since the performance of the scraper chains can directlydetermine the reliability and stability of the scraper con-veyor to a great extent such issues about the dynamiccharacteristics of the scraper chains have been examined inseveral previous studies Generally speaking research ap-proaches include two major categories ie model-basedmethods and experimental methods For example Nie et al[7] presented an Euler method-based approach to model alarge-scale scraper conveyor and evaluate the dynamiccharacteristics of the transmission system Jiang et al [8]established rigid and rigid-flexible coupling models of thechain assembly to analyze their complex dynamic behaviorMyszkowski et al [9] designed a mobile measuring systemto measure essential operating parameters to study thesystem behavior of the chain-driven machines Wang et al

HindawiShock and VibrationVolume 2019 Article ID 6986240 14 pageshttpsdoiorg10115520196986240

[10] developed a dynamic tension test system for moni-toring the dynamic tension of a heavy scraper conveyorbased on microstrain detection Obviously previousstudies have largely concentrated on direct measurementand analysis for parameters of the moving scraper chainHowever the mentioned applications of model-basedmethods stop proceeding at the level of idealized hy-pothesis Currently due to installation difficulties ofmultiple sensors most of the experimental methods are notsuitable for actual engineering Considering limitations ofharsh conditions in coal mines the traditional mobileparameter measurement of the scraper chain is difficult anda fix-point measurement method is preferred Moreoverthe scraper chain moves continuously along the chutes andthe dynamic and vibration properties of the chutes caneffectively reflect the working state of the scraper chainerefore the fix-point vibration analysis of the chute isperformed in our work which lays a foundation for faultdetection of the scraper chain

In the recent scientific literature scholars have con-ducted preliminary research on fault detection of thescraper conveyor Most of the latest studies focus ontheoretical design more than practical strategy such asfuzzy fault tree [11] Bayesian network [12] fuzzy neuralnetwork [13] and rough set theory [14] Comparativelyspeaking fault detection strategy based on vibrationanalysis has attracted much attention over the past decadeZhang et al [15] detected crack faults of a lubricant-levelviewport and upper viewport of a high-speed railway traingear box through vibration tests Parloo et al [16] pre-sented a sensitivity-based damage assessment technique forautonomous condition monitoring of a slat track ofa commercial Airbus A320 aircraft Sakaris et al [17] in-troduced the random-vibration-based method for damagedetection and precise localization on a lab-scale air-craft stabilizer structure To date few studies have been

performed on fault detection for the scraper chain based onvibration analysis

For the purpose of fault detection for the scraper chainthe novelty of our study is represented by vibrationanalysis of the chute under various working conditions Inthis paper modal analysis is performed in two waysnumerically based on the CVM and experimentally basedon the HIT According to the results of model analysis theMAC is applied for the optimal placement of accelerationsensors and three measuring points on the chute arechosen for vibration detection Sequentially the DTSM isbuilt and serves to study the vibration properties of theactual chute and the correctness of the DTSM is verifiedby the experimental study of the FPET Furthermore thetransient responses of different measuring points on thefourth chute of the DTSM are obtained under threeworking conditions which are normal condition chainjam and chain fracture respectively Besides the effects ofdifferent external loads on vibration properties are alsodiscussed We propose the novel strategy that allows forfault detection of the scraper chain which include twoaspects on the one hand the occurrence of chain faults canbe detected by amplitude comparisons of the vibrationsignals on the other hand the fault patterns can be dis-tinguished by the AOKR

e main contributions of our work can be summarizedas follows (1) an optimal placement scheme of accelerationsensors is proposed based on the MAC to determine themeasuring points on the chute for vibration analysis (2) aDTSM is constructed through finite element modeling tosimulate the working state of the actual chute under differentoperating conditions and (3) to detect chain faults of thescraper chain a novel strategy is proposed by a combinationof amplitude comparisons and the AOKR

e paper is organized in the following way Section 2provides the theory basis of the proposed fault detection

1 drive motor

2 sprocketcomponents

3 scraperchains

4 scrapers 5 chutes

6 gears7 transitionaltrough

1

2 3 4 5

67

Figure 1 Transmission system of the scraper conveyor

2 Shock and Vibration

strategy Section 3 illustrates modal analysis of the chute andpresents details of the construction of the DTSM e op-timal placement scheme of acceleration sensors is presentedin Section 4 and then the proposed strategy of fault de-tection is implemented Finally Section 5 provides someconcluding remarks

2 Research Methodology of theProposed Strategy

21ModalAssuranceCriterion emodal analysis theory isadapted to analyze and evaluate intrinsic dynamic propertiesof the mechanical structure and it also serves as the premiseof vibration analysis [18] e vibration properties of astructure are usually described by modal parameters whichis the most fundamental content of vibration analysis Sincemost of the linear systems can be discretized into an elasticsystem with n degrees of freedom (DOFs) we can expressthe motion differential equation of the chute by n co-ordinates as

MδPrime(t) + DPδprime(t) + Kδ(t) Bf(t) (1)

y Cdδ(t) + Cvδprime(t) + Df(t) (2)

where M DP K and B isin Rntimesn denote the mass matrixdamping matrix stiff matrix and sensor position matrixrespectively f(t) and δ(t) isin Rntimes1 denote the force vectorand displacement vector respectively y isin Rmtimes1 denotes themeasurement vector and m denotes the number of thesensors Cd Cv and D denote the output coefficient ma-trices To perform modal analysis of the chute the dampingparameters will not affect the characteristics of the naturalfrequency and its corresponding vibration mode so wesuppose the influence of the damping coefficient can beignored When f(t) 0 equation (1) takes the linear andhomogeneous form

MδPrime(t) + Kδ(t) 0 (3)

To obtain the natural frequencies andmode shapes of thechute the solution of equation (3) is equivalent to solvingthe generalized eigenvalues and eigenvectors e vibra-tional shape is formed by the superposition of multiplemodes and then δ(t) can be represented as

δ(t) 1113944n

i1ϕiηi ϕη (4)

where ϕ ϕ1 ϕ2 ϕnminus1 ϕn1113858 1113859 and η η1 η2 1113858

ηnminus1 ηn]T denote the modal shape matrix and modal co-ordinate matrix respectively ϕi and ηi represent the modalvector and modal coordinate of mode i respectively Whenconducting modal experiments the research usually focuseson the first nt (nt lt n) modes en equations (1) and (2)can be expressed as

ηPrimei + 2ζ iωiηi

prime + ω2i ηi ϕT

i Bf(t) Γif(t)

i 1 2 3 nt( 1113857

yd 1113944

nt

i1Cdϕiηi + 1113944

nt

i1Cvϕiηiprime + Df(t)

1113944

nt

i1Cdiηi + 1113944

nt

i1Cviηiprime + Df(t)

(5)

where Cdi Cvi and Γi denote the influence coefficient vectorsof displacement velocity and sensors respectively ζ i and ωi

denote the modal damping ratio and frequency Limited bystructural configuration of the chute and the field condi-tions the installation of sensors can directly determine thevalidity of experimental data While considering the eco-nomic problem the layout scheme should also ensure thatthe dynamic characteristic information of the structure canfully be obtained With the determination of Γi more in-dependent and accurate modal information can bemeasuredby a limited number of sensors In our study the modalassurance criterion (MAC) [19] is applied to determine theoptimal installation positions and most reasonable numberof sensors e MAC has strong applicability in evaluatingthe angle between different vibration vectors and the in-fluence of the mass matrix and stiffness matrix of thestructure can be neglected e elements of the MACmatrixtake the following form

MACij ϕT

i middot ϕj1113872 11138732

ϕTi middot ϕi1113872 1113873 ϕT

j middot ϕj1113872 1113873isin [0 1] (ine j) (6)

where ϕi and ϕj can be obtained by using equation (4)eoretically the natural vibration modes of different nodesare orthogonal to each other

However the actual measured modal vectors are difficultto guarantee the orthogonality e placement of sensorsmust ensure a large space angle between the modal vectors ofthe measuring points so as to retain the original modelfeatures to the greatest extent e numerical variation rangesof the off-diagonal element MACij represent the followingstatements for MACij 0 the modal vectors are orthogonalto each other for MACij lt 025 the modal vectors are easilydistinguishable for MACij 1 the space angle between themodal vectors is 0 and the modal vectors are indistinguish-able According to the MAC a smaller MACij makes it easierto distinguish different modal vectors which also indicates abetter performance of the optimal placement scheme Hencethe installation position and number of sensors should bedetermined to minimize the off-diagonal elements of theMAC matrix e minimum value is given by

f max MACij

11138681113868111386811138681113868

11138681113868111386811138681113868 (ine j) (7)

22AdaptiveOptimalKernelTime-FrequencyRepresentatione vibration signals caused by chain faults belong tononstationary random signals Based on time-frequency

Shock and Vibration 3

analysis theory the main tools dedicated to the study ofnonstationary signals are available A commonly used time-frequency distribution is the WignerndashVille distribution(WVD) which has found many successful applications indifferent areas [20 21] e input signal in the time domainis denoted by using s(t) and the definition of WVD can begiven by

DWV(tω) 1113946 s t +τ2

1113874 1113875slowast

tminusτ2

1113874 1113875 eminusjωτ dτ (8)

where ω denotes the frequency variable and slowast(t) is theconjugate function of s(t) en equation (8) can be con-verted into the time-frequency distribution function as

A(θ τ) 1113946 s t +τ2

1113874 1113875slowast

tminusτ2

1113874 1113875 ejθt dt (9)

where θ and τ denote the fuzzy-domain variables Based onequations (8) and (9) we get

DWV(tω) 12π

1113946 1113946 A(θ τ)eminusjθtminusjωτ

dθ dτ

A(θ τ) 1113946 1113946 DWV(tω)ejθt+jωτ

dt dω

(10)

Developed from WVD the Cohen class bilinear distri-butions are the commonly used time-frequency analysismethods and can be expressed as

P(tω) 12π

1113946 1113946 A(θ τ)ϕ(θ τ)eminusjθtminusjωτ

dθ dτ (11)

where A(θ τ) denotes the ambiguity function and ϕ(θ τ)

denotes the kernel function Bilinear time-frequency anal-ysis such as WVD exhibits a high time-frequency resolutionwhereas it is contaminated by serious cross-terms en theAOKR is applied to suppress the cross-term interferencescaused by multicomponents [22] meanwhile the modifiedambiguity function is defined as

A1(t θ τ) 1113946 slowast

uminusτ2

1113874 1113875ωlowast uminus tminusτ2

1113874 1113875s u +τ2

1113874 1113875

times ω uminus t +τ2

1113874 1113875ejθu

du

(12)

where ω(u) denotes the symmetric window function Forany A1(t θ τ) we can get the corresponding adaptive op-timal kernel function ϕopt(t θ τ) Under the premise ofensuring high resolution the AOKR shows good perfor-mance to analyze vibration signals of the chute under dif-ferent working conditions e adaptive optimal kerneltime-frequency distribution is expressed as

PAOK(tω) 12π

1113946 1113946 A1(t θ τ)ϕopt(t θ τ)eminusjθtminusjωτ

dθ dτ

(13)

3 Finite Element Modeling

31 Modal Analysis e CVM shown in Figure 2(a) isestablished in the software ANSYS and designed to befaithful to the actual device In fact the CVM is a typical

multibody dynamics system that consists of the chute platchains vertical chains and scrapers the dimension ofscraper chains is φ48times152 (mm) and the CVM contains773782 elements and 1117993 nodes Referring toFigure 2(a) to describe the interactions between differentcomponents the kinematic restriction mainly includes thefollowing two aspects kinematic constraints and contactrelations e inertial coordinate system I works as thereference of the system frame and is applied to determine thelocation of different components and contact pairs occurbetween two bodies moving correspondingly to each othere vibration signals of the chute can be expressed as thesum of natural vibration modes of each order and the lowerorder vibration modes have a greater correlation with thevibration properties To obtain intrinsic properties of thechute we extract the first 6-order natural frequencies of theCVM for modal analysis which would facilitate furtherstudies on the optimal placement of the acceleration sensors

e natural frequencies of the chute are obtained byfinite element analysis (FEA) of the CVM and experimentalmodal analysis (EMA) through the HIT As shown inFigure 2(b) the HIT is performed with the use of an impacthammer (LC02) and an acceleration sensor (TST120A1000)which is conducted for vibration measurements of the actualchute e exciting point is impacted by using the impacthammer to trigger vibration signals and then the acceler-ation sensor mounted on the surface of the shovel coal boardis used to pick up the vibration response Besides thewireless acquisition device is utilized for collection andtransmission of the experimental data which is detailed inSection 42 us we get the exciting force and vibrationresponse as shown in Figure 3

rough model simulation we get the modal displace-ment contours of the first 6-order natural vibrationmodes ofthe chute (Figure 4) According to the HIT natural fre-quencies of the actual chute are extracted from the exper-imental modal curve and each peak of the response curve iscorresponding to a vibration mode As listed in Table 1comparisons between natural frequencies of the first 6-ordernatural vibration modes based on FEA and EMA are madeIn addition the error ψi is defined as ψi |(αi minus βi)βi|

(i 1 sim 6) where αi and βi denote the natural frequencyderived from FEA and EMA respectively In Table 1 theerrors show low values for each mode In particular themaximum error between the simulation results and ex-perimental results is 1330 and the minimum ψi is almost232 e mentioned errors indicate the superiority of theestablished CVM

32 Construction of the DTSM In actual operation thescraper chain is driven by using drive motors and movescontinuously along the chutes and its reliability is directlyrelated to the working performance of the transmissionsystem Figure 5 describes the dynamic transmission systemmodel (DTSM) built by using the transient analysis modulein ANSYS e primary goal of construction of the DTSM isto achieve an accurate dynamic model which would facil-itate further studies on the vibration behavior of the actual

4 Shock and Vibration

X

Y

ZI

Vertical chain

O

Scraper

System frameChute

Plat chain

Contact Contact

Fixed joint

Contact Fixed jo

int

(a)

Scraper

Scraper chainsAcceleration

sensor

Chute

Exciting point

Impact hammer

Shovelcoal board

(b)

Figure 2 Descriptions of modal analysis based on (a) the CVM and (b) the HIT

350

250300

200150100

500

ndash50

Time (s)

Exci

ting

forc

e (N

)

0 05 1

(a)

0 05 1

15

1

05

0

ndash05

ndash1

Acce

lera

tion

(ms

2 )

Time (s)

(b)

Figure 3 Exciting force and the vibration response

Mode 1

(1729 0612)(2114 04954)

(2578 08114) (3096 08938)

(314 04933)

(3511 01923)

Acce

lera

tion

(ms

2 ) Mode 2

Mode 3

Mode 4

Mode 5

Mode 6

0

05

1

15

2

Frequency (Hz)0 50 100 150 200 250 300 350 400 450 500

Experimental modal

Figure 4 Vibration modes of the chute

Shock and Vibration 5

chute without carrying out physical tests e simulationparameters and kinematic restriction of the DTSM areconsistent with the CVM in Figure 2(a) To illustrate thisissue in more detail the frictional contacts are set betweenthe vertical chains and the chutes and between the scrapersand chutes Correspondingly the bonded contacts are setbetween the plat chains and the vertical chains e dynamicmodel mainly contains two prominent parts the chain as-sembly and the chutes Wherein the chain assembly iscomposed of the scrapers and scraper chains and can beequally divided into multiple segments of length ΔL thenumber of the chutes is 7 and they are marked as shown inFigure 5 A translational joint is set between scraper 1 andthe middle plate based on which the chain assembly can runalong the chutes at a transmission speed of Vi Moreover thepretightening force and external load of the dynamic modelare defined as F1 and Wi respectively Here we setF1 79612 kN In Figure 5 the pretightening force is ap-plied at the two ends of the chain assembly which can ensurethat the scraper chain remains tight during the movemente external load Wi is applied on the upper surface of thescrapers and middle plates and the material density inthe chute under the full-load condition is W0 whereW0 694 kgm

4 AOKR-Based Fault Detection Strategy of theScraper Chain

Considering that the scraper conveyor is susceptible tofrequent loading excessive bending and artificial mis-conduct in actual engineering different patterns of chainfaults usually act on the scraper chain e failure patternswill cause abnormal vibration of the chute Actuallydirect measurement and analysis for parameters of themoving scraper chain is difficult which raises a contra-diction between practical engineering requirements andthe installation limitations of multiple sensors In order todetect chain faults promptly the fix-point vibrationmeasurement of the measuring points for the chute isconsidered instead of mobile parameter measurement forthe scraper chain In this section the DTSM is utilized toacquire the vibration properties of the actual chute undervarious working conditions and the FPETof the vibrationsignals is conducted to validate the accuracy of the dy-namic model e fundamental idea of our proposedstrategy is to determine the occurrence of chain faultsby amplitude comparisons and then fault signals areanalyzed through the AOKR to distinguish the faultpatterns

Table 1 Frequency response of vibration signals

Mode no FEA (Hz) EMA (Hz) ψi ()

1 17754 1729 2682 18329 2114 13303 28745 2578 11504 30242 3096 2325 32868 3140 4686 34197 3511 260

Position 1 Position 2

F1

F1

F1

F1

Translationaljoint

Scraper 1

Chain breakage

Chain jamContact point 1

∆L

Middle plateChain assembly

Wi

1 3 4 5 7

Vi

Vi

Figure 5 Simulation process of the DTSM

6 Shock and Vibration

41 Optimal Placement of the Acceleration Sensors Beforeconducting the FPET the installation scheme of the ac-celeration sensors should be determined to ensure thevalidity of the measured vibration signals and the economyof the experimental process Based on modal analysis of theFEA (Figure 2(a)) we extract 20 nodes from the chute as theprimary measuring points which are labeled in Figure 6(a)According to the theory of MAC the total modal dis-placements of the selected 20 points serve as the inputs ofequations (6) and (7) In Figure 6(b) the curve revealschange rules of the minimumMACij for different number ofsensors

To illustrate the optimal placement scheme of acceler-ation sensors in more detail the optimum installation po-sitions for different number of sensors are shown in Table 2Accordingly the vibration response of the actual chuteshould be recorded by 3 acceleration sensors and the op-timum installation positions are set at nodes 3 9 and 16respectively

42 Experimental Evaluations According to Section 32 thesimulation process of the DTSM is performed under normalcondition without a load and the transmission speed Vi isset as V0 in Figure 7(a) and the scraper 1 moves fromposition 1 to position 2 In our study nodes 9 16 and 3 ofthe fourth chute are chosen as the detecting points for vi-bration analysis For convenient expression in the followingwe mark the detecting nodes as measuring points 1 2 and3 respectively Taking measuring point 1 as the researchexample for description of the transmission process theoriginal signal is presented in Figure 7(b) e vibrationresponse presents a three-stage change ie the accelerationphase the steady phase and the deceleration phase Asmentioned above in the steady phase the vibration signalcan be treated as multiple segments with a time span of ΔTwhich corresponds to the chain assembly with a length of ΔL(Figure 5)

e field FPET of multiple monitoring points iscarried out to obtain the vibration properties of the actualchute and evaluate the dynamic performance of theDTSM e specifications of the experimental scraperconveyor correspond to SGB12003600 manufactured byLianyungang Tianming Equipment Co Ltd and therunning speed of the scraper chain is 10 ms In practicalengineering the actual double-drive transmission systemis driven by two drive motors e basic performanceparameters are shown in Table 3 As shown in Figure 8the experimental system mainly contains three acceler-ation sensors (TST120A1000) a wireless acquisitiondevice (TST5925EV) a wireless receiver and an on-sitePC e acceleration sensors are used for fixed-pointmeasurements by detecting the vibration responses ofmeasuring points 1 2 and 3 on the chute e wirelessacquisition device is used to collect experimental data inreal time In addition the wireless receiver is intended forremote data transmissione vibration data are stored inthe on-site PC and a sampling frequency of 1000 Hz istaken

Figure 9 shows the experimental and simulation signalsof measuring points 1 2 and 3 in the time span ΔT Inthe time domain the simulation results and the experi-mental results are similar which reflects the adaptability ofthe DTSM Considering the nonlinear and time-varyingcharacteristics of the vibration signals the AOKR is used fortime-frequency analysis

e vibration signal of measuring point 2 in the timespan ΔT is taken as the reference for time-frequency

1 2 3 4 5 6 7 8 9 10

11 12 13 14 15 16 17 18 19 20

(a)

Number of the sensors

Min

imum

val

ue o

f MAC

ij

2 3 4 5 6 7 8 9 10

028

0275

027

0265

026

0255

025

0245

(b)

Figure 6 Description of the optimum placement scheme (a) theprimary measuring points (b) the minimum MACij

Table 2 Measuring points placement based on the MAC

Number f Measuring points2 02515 8 163 02453 3 9 164 02454 1 3 9 165 02460 1 3 9 16 206 02483 1 3 9 10 11 167 02525 1 2 3 9 10 11 168 02644 1 3 8 9 10 11 15 169 02664 1 2 3 8 9 10 11 15 1610 02767 1 2 3 8 9 10 11 15 16 20

Shock and Vibration 7

analysis and then the time-frequency representations ofmeasuring points 1 and 3 in the same time period aredemonstrated Figure 10(a) shows the time-frequencyrepresentation of the experimental signals for differentmeasuring points and the power spectrum is alsodepicted It presents that the vibration signal of the chutehas a plurality of components ere are two major areaswith strong frequency responses of the vibration signalsfor different measuring points for measuring point 1 thefrequency range is 230ndash310Hz and the time ranges are015ndash030 s and 045ndash075 s for measuring point 2 thefrequency range is 230ndash340Hz and the time ranges are020ndash040 s and 055ndash070 s and for measuring point 3the frequency range is 225ndash330Hz and the time ranges are040ndash050 s and 070ndash090 s Considering the chain

assembly in Figure 5 there are two scrapers for length ΔLIn the actual transmission process the two major areaswith strong frequency responses are caused while the twoscrapers passing through the measuring points As shownbasic frequencies of the measuring points are 239 278 and249Hz respectively In addition the peak of the vibrationpower spectrum is also presented and the maximumvalues of the measuring points are 1588 1561 and1499m2s

Similarly the time-frequency representation of thesimulation signals with two major frequency responses isdiscussed As Figure 10(b) indicates for measuring points1 2 and 3 the vibration energies are concentrated atthe frequency ranges 240ndash340 200ndash310 and 200ndash320Hzrespectively Correspondingly for measuring point 1 the

t1 t3

t2

Chain jamoccurs

15

1

05

0

Time (s)

Velo

city

(ms

)

0 05 1 15 2 25 3 35 4 45 5 55 6

V0V1

(a)

Constant speed

Acceleration DecelerationΔT

Time (s)0 1 2 3 4 5 6

543210

ndash1ndash2ndash3ndash4ndash5

Acce

lera

tion

(ms

2 )

(b)

Figure 7 Transmission speed setting and vibration signal under normal condition

Table 3 Basic operating condition of the scraper conveyor

Model Chain size (mm) Conveyor width (mm) Chain speed (ms) Transport capacity (th) Transport length (m)SGZ12003600 φ48times152 1750times1180 0sim189 3700 360

Gear

Gear

Sprocket

Drive motor On-site PCWireless receiver

Transitionaltrough Chute

Accelerationsensor

Wirelessacquisition device

2

3

1

Figure 8 Experimental setup of the FPET

8 Shock and Vibration

time ranges are 010ndash025 and 050ndash065 s en the vi-bration energy of measuring point 2 is concentrated at thetime ranges 015ndash045 and 060ndash075 s and the time rangesof measuring point 3 are 040ndash050 and 070ndash085 s Forthe simulation signals the basic frequencies of the mea-suring points are 281 249 and 258Hz respectively emaximum values of vibration energies are 1527 1489 and1532m2s respectively Considering the experimental re-sults and the simulation results separately good agreementbetween the main parameters of different measuring pointsis obtained It also indicates that in the same time periodthe vibration signal has a delay characteristic for measuringpoints 1 2 and 3 is is possibly because that themeasuring points are at different positions of the chuteMoreover for different measuring points the experimentalsignals and the simulation signals show good consistency intime-frequency characteristics Hence the establishedDTSM can efficiently simulate the actual productionenvironment

43 Fault Detection of the Scraper Chain Based on theestablished DTSM two typical failure patterns of thescraper chain are discussed namely chain jam and chainfracture ree external load conditions are set that isempty load half-load and full-load conditions e valuesof Wi (Figure 5) are defined as 0 12W0 and W0 re-spectively As designed in Figure 5 when chain jam occursthe transmission speed Vi is set as V1 in Figure 7(a) and the

failure time range is t1 minus t3(t3 25 s) the chain fault istriggered at t1 2 sWithin the time range t1 minus t2(t2 225 s)the transmission speed Vi decreases from 1 to 0ms And thevalue of Vi increases from 0 to 1ms within the time ranget2 minus t3 e whole process lasts 05 s the chain assembly istightened and the scraper chains are jammed When chainfracture occurs the transmission speed Vi is set as V0 inFigure 7(a) Moreover as shown in Figure 5 the contactconstraint between two contacting scraper chains at the labeledcontact point 1 is removed As a result the two contactingscraper chains will be separated In order to ensure the accuracyof fault setting the chain fracture is also triggered at t1 2 sand the contact constraint is removed in 05 s which is con-sistent with chain jam

Considering both failure patterns of the scraper chainour research focuses on the steady phase of the operationprocess within the time range 05 to 55 s In the steadyphase the vibration signals are obtained simultaneously atmeasuring points 1 2 and 3 Taking measuring point1 as the case study the vibration signals for chain jam andchain fracture under empty load condition are presentedin Figures 11(a) and 11(b) respectively After faultstriggering the vibration signals of the two failure patternsshow a sudden increase after a short time delay Sub-sequently the signals exhibit unstable fluctuations at thetime range 24ndash28 s In order to obtain a more detaileddescription of the detection results the maximum am-plitudes of the vibration signals under different workingconditions are discussed ie normal condition chain jam

3210

ndash1ndash2ndash3

3210

ndash1ndash2ndash3

3210

ndash1ndash2ndash3

Time (s)

Acce

lera

tion

(ms

2 )

0 01 02 03 04 05 06 07 08 09

1 2 3

(a)

3210

ndash1ndash2ndash3

3210

ndash1ndash2ndash3

3210

ndash1ndash2ndash3

Time (s)Ac

cele

ratio

n (m

s2 )

0 01 02 03 04 05 06 07 08 09

123

(b)

Figure 9 Vibration signals of different measuring points in ΔT (a) experimental results (b) simulation results

Shock and Vibration 9

Point

1

2

3

(a) (b)

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

0

Time (s) Spectrum0 03 06 09

500

400

300

200

100

00 50 100 150

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

00 03 06 09

500

400

300

200

100

00 50 100 150

Time (s) Spectrum

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

00 03 06 09

500

400

300

200

100

00 50 100 150

Time (s) Spectrum

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

00 03 06 09

500

400

300

200

100

00 50 100 150

Time (s) Spectrum

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

0

Time (s) Spectrum0 03 06 09

500

400

300

200

100

00 50 100 150

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

00 03 06 09

500

400

300

200

100

00 50 100 150

Time (s) Spectrum

Figure 10 Time-frequency representation of vibration signals (a) experimental results (b) simulation results

Chain jamoccurs

Acce

lera

tion

(ms

2 ) Acce

lera

tion

(ms

2 )

10

5

0

ndash5

ndash10

Time (s)0 05 10 15 20 25 30 35 40 45 50

10

5

0

ndash5

ndash1024 25 26 27 28

(a)

Figure 11 Continued

10 Shock and Vibration

and chain fracture Wherein the empty load half-loadand full-load conditions are considered e maximumamplitudes of the vibration signals at measuring points1 2 and 3 are depicted in Figures 12(a)ndash12(c)respectively

Considering measuring point 1 the maximum am-plitudes of the vibration signals under empty load conditionare 289 718 and 698ms2 under normal condition chainjam and chain fracture respectively Similarly under half-load condition the maximum amplitudes are 324 937 and956ms2 Moreover under full-load condition the maxi-mum amplitudes are 446 1106 and 1082ms2 Withdifferent external loads the maximum amplitudes of thevibration signals for fault conditions are obviously higherthan those for normal condition and the difference betweenthe amplitudes of the two typical failure patterns is small Fordifferent fault conditions with the increase of the externalloads the maximum amplitudes show trends to increasee above statistical results are also applicable to measuringpoints 2 and 3 erefore chain faults can easily be de-tected by comparing the maximum amplitudes of the vi-bration signals whereas the fault patterns are difficult toidentify According to the nonstationary and nonlinearcharacteristics of fault signals the AOKR is utilized to an-alyze the vibration signals and classify failure patterns of thescraper chain Within 15 s after faults triggering the vi-bration signals at the three measuring points with differentexternal loads are processed Wherein for chain jam andchain fracture under empty load condition the time-fre-quency representations of vibration signals are presented inFigures 13(a) and 13(b) respectively

e frequency components and frequency ranges ofthe same fault pattern are similar for different measuringpoints As Figure 13(a) describes the bright color between0 and 50Hz indicates one high energy area caused bychain jam en chain fracture can easily be distinguishedaccording to the appearance of two high energy areasbetween 100 and 200Hz as shown in Figure 13(b)

Observing the spectrum results a more detailed de-scription is given When chain jam occurs for measuringpoints 1 2 and 3 the high energy areas occur ap-proximately at 05 075 and 09 s respectively Mean-while for chain fracture the high energy areas include twomain frequency components and are approximatelyconcentrated at the time ranges 050ndash070 075ndash085 and10ndash115 s respectively Hence there is a delay charac-teristic of the fault occurrence which is well in accordancewith the conclusions in Section 42 In order to explore theinfluence of external loads on fault characteristics thedetailed differences of the fault patterns at measuringpoint 2 are depicted in Figure 14 In fact the externalload has a great influence on the fault severity of both thefailure patterns Observing the spectrum results thebright areas vary with the external loads With increasingexternal load the frequency ranges of the high energyareas become larger Wherein for chain jam under emptyhalf- and full-load conditions the frequency ranges areapproximately 0ndash50 0ndash150 and 0ndash250 Hz respectivelyMeanwhile for chain fracture the frequency rangesare approximately 80ndash200 50ndash250 and 50ndash350Hzrespectively

In this part three working conditions of the scraperchain are investigated above including normal conditionchain jam and chain fracture e vibration signals ofmeasuring points 1 2 and 3 on the detecting chute areanalyzed and the effects of the external loads on the vi-bration characteristics are discussed Based on the aboveanalysis the occurrence of chain faults can easily be de-termined through amplitude comparisons of the originalvibration signals However the observation confirms thesimilarity of the time domain waveforms of fault signals forchain jam and chain fracture ese two patterns of failuresremain to be different through further processing by theAOKR and the fault patterns can be distinguished accordingto the number of high energy areas of the time-frequencyrepresentation of vibration signals In conclusion the

Chain fractureoccurs

Acce

lera

tion

(ms

2 ) Acce

lera

tion

(ms

2 )

10

5

0

ndash5

ndash10

Time (s)0 05 10 15 20 25 30 35 40 45 50

10

5

0

ndash5

ndash1024 25 26 27 28

(b)

Figure 11 Vibration signals of the measuring point 1 for (a) chain jam and (b) chain fracture

Shock and Vibration 11

Point 1 2 3

(a)

(b)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

Figure 13 Time-frequency representation of vibration signals under empty load (a) chain jam (b) chain fracture

12

10

8

6

4

2

0Max

imum

acce

lera

tion

ampl

itude

(ms

2 )

Normal conditionChain jamChain fracture

0 12 W0 W0

(a)

Normal conditionChain jamChain fracture

12

10

8

6

4

2

0Max

imum

acce

lera

tion

ampl

itude

(ms

2 )

0 12 W0 W0

(b)

Normal conditionChain jamChain fracture

12

10

8

6

4

2

0Max

imum

acce

lera

tion

ampl

itude

(ms

2 )

0 12 W0 W0

(c)

Figure 12 Statistical results of the maximum vibration amplitude at different measuring points (a) 1 (b) 2 (c) 3

12 Shock and Vibration

proposed detection strategy is effective at detecting theoccurrence of chain faults and identifying the failure pat-terns under different operating conditions

5 Conclusions

During the actual operation the working state of thescraper chain can reflect the dynamic performance of thescraper conveyor To address the difficulties with directsensor measurement for parameters of the moving scraperchain a novel strategy for fault detection of the scraperchain based on vibration analysis of the chute was pro-posed Based on modal analysis and the MAC the mea-suring points of vibration signals on the chute weredetermined To fit the actual behavior of the transmissionprocess the DTSM was presented based on finite elementmodeling and the correctness of the dynamic model wasverified by comparison with the FPET en the vibrationproperties of the measuring points on the chute undernormal condition chain jam and chain fracture werediscussed Moreover the occurrence of chain faults weredetermined by comparing the amplitudes of the vibrationsignal in the time domain while the AOKR was utilizedfor time-frequency representation of vibration signals anddistinguishing the two typical failure patterns Further-more the strategy verification based on experimental datawill be taken into consideration in the near future

Data Availability

e data used to support the findings of this study areavailable from the corresponding author upon request

Conflicts of Interest

e authors declare that they have no conflicts of interest

Acknowledgments

is work was supported by the Key Project of NationalNatural Science Foundation of China (U1510205) NaturalScience Foundation of Jiangsu Province (No BK20160251)Xuzhou Research program (KC14H0138) FundamentalResearch Funds for the Central Universities (2014Y05) andProject Funded by the Priority Academic Program De-velopment of Jiangsu Higher Education Institutions(PAPD)

References

[1] C D Brown ldquoDesign build and test of a longwall armouredface conveyorrdquo Longwall Mining 2002

[2] M Dolipski P Cheluszka E Remiorz and P SobotaldquoFollow-up chain tension in an armoured face conveyornadazne napinanie lancucha zgrzebłowego W przenosnikuscianowymrdquo Archives of Mining Sciences vol 60 no 1pp 25ndash38 2015

[3] L A Morley J L Kohler and H M Smolnikar ldquoA model forpredicting motor load for an armored face-conveyor driverdquoIEEE Transactions on Industry Applications vol 24 no 4pp 649ndash659 1988

[4] A A Ordin and A A Metelrsquokov ldquoAnalysis of longwall faceoutput in screw-type cutter-loader-and-scraper conveyorsystem in underground mining of flat-lying coal bedsrdquoJournal of Mining Science vol 51 no 6 pp 1173ndash1179 2015

[5] B He G Li H Shi et al ldquoDynamic behaviour modelling andsimulation of the chain transmission system for an armouredface conveyorrdquo in Proceedings of the IEEE 10th InternationalConference on Computer-Aided Industrial Design and Con-ceptual Design CAID amp CD 2009 pp 1000ndash1004 BeijingChina November 2009

[6] R Nie B He P Yuan L Zhang and G Li ldquoNovel approachto and implementation of design and analysis of armored faceconveyor power trainrdquo Science China Technological Sciencesvol 58 no 12 pp 2153ndash2168 2015

Loads

(a)

(b)

Empty load Half-load Full-load

500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)

Figure 14 Time-frequency representation of vibration signals at measuring point 2 with different external loads (a) chain jam (b) chainfracture

Shock and Vibration 13

[7] R Nie B He L Zhang and G Li ldquoModelling of thetransmission system in conveying equipment based on Eulermethod with applicationrdquo Proceedings of the Institution ofMechanical Engineers Part K Journal of Multi-body Dy-namics vol 228 no 3 pp 294ndash306 2014

[8] S B Jiang X Zhang K D Gao J Gao Q Y Wang andK Hidenori ldquoMulti-body dynamics and vibration analysis ofchain assembly in armoured face conveyorrdquo InternationalJournal of Simulation Modelling vol 16 no 3 pp 458ndash4702017

[9] M Myszkowski and D Loehning ldquoChain force measure-ments on armoured face conveyors and coal plows in heavy-duty longwallsrdquo CIM Bulletin vol 94 no 1054 pp 72ndash752001

[10] H Wang Q Zhang and F Xie ldquoDynamic tension test andintelligent coordinated control system of a heavy scraperconveyorrdquo IET Science Measurement and Technology vol 11no 7 pp 871ndash877 2017

[11] S Sen M X Min and Y Z She ldquoDiagnosis of coal scraperconveyor based on Fuzzy Fault treerdquo in Proceedings of the2015 Seventh International Conference on Measuring Tech-nology and Mechatronics Automation (ICMTMA) pp 392ndash395 IEEE Nanchang China June 2015

[12] S-s Xue X-c Li and X-y Xu ldquoFault tree and Bayesiannetwork based scraper conveyer fault diagnosisrdquo in Pro-ceedings of the 22nd International Conference on IndustrialEngineering and Engineering Management 2015 pp 783ndash795Atlantis Press Paris France January 2016

[13] X Gong X Ma Y Zhang et al ldquoApplication of fuzzy neuralnetwork in fault diagnosis for scraper conveyor vibrationrdquo inProceedings of the 2013 IEEE International Conference onInformation and Automation (ICIA) pp 1135ndash1140 IEEEYinchuan China August 2013

[14] Y Zhang X Ma Y Jianxiang et al ldquoFuzzy neural networkfault diagnosis and online vibration monitoring system for thecoal scraper conveyor based on rough set theoryrdquo in Pro-ceedings of the 2013 32nd Chinese Control Conference (CCC)pp 6134ndash6138 IEEE Xirsquoan China July 2013

[15] B Zhang A C C Tan and J-h Lin ldquoGearbox fault diagnosisof high-speed railway trainrdquo Engineering Failure Analysisvol 66 pp 407ndash420 2016

[16] E Parloo P Verboven P Guillaume and M Van OvermeireldquoAutonomous structural health monitoring-part ii vibration-based in-operation damage assessmentrdquo Mechanical Systemsand Signal Processing vol 16 no 4 pp 659ndash675 2002

[17] C S Sakaris J S Sakellariou and S D Fassois ldquoRandom-vibration-based damage detection and precise localization ona lab-scale aircraft stabilizer structure via the GeneralizedFunctional Model Based Methodrdquo Structural Health Moni-toring An International Journal vol 16 no 5 pp 594ndash6102017

[18] Y Zhang W Song M Karimi C-H Chi and A KudreykoldquoFractional autoregressive integrated moving average andfinite-element modal the forecast of tire vibration trendrdquoIEEE Access vol 6 pp 40137ndash40142 2018

[19] M Pastor M Binda and T Harcarik ldquoModal assurancecriterionrdquo Procedia Engineering vol 48 pp 543ndash548 2012

[20] W J Staszewski K Worden and G R Tomlinson ldquoTime-frequency analysis in gearbox fault detection using the Wigner-ville distribution and pattern recognitionrdquo Mechanical Systemsand Signal Processing vol 11 no 5 pp 673ndash692 1997

[21] J-D Wu and P-H Chiang ldquoApplication of Wigner-Villedistribution and probability neural network for scooter engine

fault diagnosisrdquo Expert Systems with Applications vol 36no 2 pp 2187ndash2199 2009

[22] Z Feng and M Liang ldquoFault diagnosis of wind turbineplanetary gearbox under nonstationary conditions viaadaptive optimal kernel time-frequency analysisrdquo RenewableEnergy vol 66 pp 468ndash477 2014

14 Shock and Vibration

International Journal of

AerospaceEngineeringHindawiwwwhindawicom Volume 2018

RoboticsJournal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Active and Passive Electronic Components

VLSI Design

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Shock and Vibration

Hindawiwwwhindawicom Volume 2018

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawiwwwhindawicom

Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Control Scienceand Engineering

Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom

Journal ofEngineeringVolume 2018

SensorsJournal of

Hindawiwwwhindawicom Volume 2018

International Journal of

RotatingMachinery

Hindawiwwwhindawicom Volume 2018

Modelling ampSimulationin EngineeringHindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Navigation and Observation

International Journal of

Hindawi

wwwhindawicom Volume 2018

Advances in

Multimedia

Submit your manuscripts atwwwhindawicom

strategy Section 3 illustrates modal analysis of the chute andpresents details of the construction of the DTSM e op-timal placement scheme of acceleration sensors is presentedin Section 4 and then the proposed strategy of fault de-tection is implemented Finally Section 5 provides someconcluding remarks

2 Research Methodology of theProposed Strategy

21ModalAssuranceCriterion emodal analysis theory isadapted to analyze and evaluate intrinsic dynamic propertiesof the mechanical structure and it also serves as the premiseof vibration analysis [18] e vibration properties of astructure are usually described by modal parameters whichis the most fundamental content of vibration analysis Sincemost of the linear systems can be discretized into an elasticsystem with n degrees of freedom (DOFs) we can expressthe motion differential equation of the chute by n co-ordinates as

MδPrime(t) + DPδprime(t) + Kδ(t) Bf(t) (1)

y Cdδ(t) + Cvδprime(t) + Df(t) (2)

where M DP K and B isin Rntimesn denote the mass matrixdamping matrix stiff matrix and sensor position matrixrespectively f(t) and δ(t) isin Rntimes1 denote the force vectorand displacement vector respectively y isin Rmtimes1 denotes themeasurement vector and m denotes the number of thesensors Cd Cv and D denote the output coefficient ma-trices To perform modal analysis of the chute the dampingparameters will not affect the characteristics of the naturalfrequency and its corresponding vibration mode so wesuppose the influence of the damping coefficient can beignored When f(t) 0 equation (1) takes the linear andhomogeneous form

MδPrime(t) + Kδ(t) 0 (3)

To obtain the natural frequencies andmode shapes of thechute the solution of equation (3) is equivalent to solvingthe generalized eigenvalues and eigenvectors e vibra-tional shape is formed by the superposition of multiplemodes and then δ(t) can be represented as

δ(t) 1113944n

i1ϕiηi ϕη (4)

where ϕ ϕ1 ϕ2 ϕnminus1 ϕn1113858 1113859 and η η1 η2 1113858

ηnminus1 ηn]T denote the modal shape matrix and modal co-ordinate matrix respectively ϕi and ηi represent the modalvector and modal coordinate of mode i respectively Whenconducting modal experiments the research usually focuseson the first nt (nt lt n) modes en equations (1) and (2)can be expressed as

ηPrimei + 2ζ iωiηi

prime + ω2i ηi ϕT

i Bf(t) Γif(t)

i 1 2 3 nt( 1113857

yd 1113944

nt

i1Cdϕiηi + 1113944

nt

i1Cvϕiηiprime + Df(t)

1113944

nt

i1Cdiηi + 1113944

nt

i1Cviηiprime + Df(t)

(5)

where Cdi Cvi and Γi denote the influence coefficient vectorsof displacement velocity and sensors respectively ζ i and ωi

denote the modal damping ratio and frequency Limited bystructural configuration of the chute and the field condi-tions the installation of sensors can directly determine thevalidity of experimental data While considering the eco-nomic problem the layout scheme should also ensure thatthe dynamic characteristic information of the structure canfully be obtained With the determination of Γi more in-dependent and accurate modal information can bemeasuredby a limited number of sensors In our study the modalassurance criterion (MAC) [19] is applied to determine theoptimal installation positions and most reasonable numberof sensors e MAC has strong applicability in evaluatingthe angle between different vibration vectors and the in-fluence of the mass matrix and stiffness matrix of thestructure can be neglected e elements of the MACmatrixtake the following form

MACij ϕT

i middot ϕj1113872 11138732

ϕTi middot ϕi1113872 1113873 ϕT

j middot ϕj1113872 1113873isin [0 1] (ine j) (6)

where ϕi and ϕj can be obtained by using equation (4)eoretically the natural vibration modes of different nodesare orthogonal to each other

However the actual measured modal vectors are difficultto guarantee the orthogonality e placement of sensorsmust ensure a large space angle between the modal vectors ofthe measuring points so as to retain the original modelfeatures to the greatest extent e numerical variation rangesof the off-diagonal element MACij represent the followingstatements for MACij 0 the modal vectors are orthogonalto each other for MACij lt 025 the modal vectors are easilydistinguishable for MACij 1 the space angle between themodal vectors is 0 and the modal vectors are indistinguish-able According to the MAC a smaller MACij makes it easierto distinguish different modal vectors which also indicates abetter performance of the optimal placement scheme Hencethe installation position and number of sensors should bedetermined to minimize the off-diagonal elements of theMAC matrix e minimum value is given by

f max MACij

11138681113868111386811138681113868

11138681113868111386811138681113868 (ine j) (7)

22AdaptiveOptimalKernelTime-FrequencyRepresentatione vibration signals caused by chain faults belong tononstationary random signals Based on time-frequency

Shock and Vibration 3

analysis theory the main tools dedicated to the study ofnonstationary signals are available A commonly used time-frequency distribution is the WignerndashVille distribution(WVD) which has found many successful applications indifferent areas [20 21] e input signal in the time domainis denoted by using s(t) and the definition of WVD can begiven by

DWV(tω) 1113946 s t +τ2

1113874 1113875slowast

tminusτ2

1113874 1113875 eminusjωτ dτ (8)

where ω denotes the frequency variable and slowast(t) is theconjugate function of s(t) en equation (8) can be con-verted into the time-frequency distribution function as

A(θ τ) 1113946 s t +τ2

1113874 1113875slowast

tminusτ2

1113874 1113875 ejθt dt (9)

where θ and τ denote the fuzzy-domain variables Based onequations (8) and (9) we get

DWV(tω) 12π

1113946 1113946 A(θ τ)eminusjθtminusjωτ

dθ dτ

A(θ τ) 1113946 1113946 DWV(tω)ejθt+jωτ

dt dω

(10)

Developed from WVD the Cohen class bilinear distri-butions are the commonly used time-frequency analysismethods and can be expressed as

P(tω) 12π

1113946 1113946 A(θ τ)ϕ(θ τ)eminusjθtminusjωτ

dθ dτ (11)

where A(θ τ) denotes the ambiguity function and ϕ(θ τ)

denotes the kernel function Bilinear time-frequency anal-ysis such as WVD exhibits a high time-frequency resolutionwhereas it is contaminated by serious cross-terms en theAOKR is applied to suppress the cross-term interferencescaused by multicomponents [22] meanwhile the modifiedambiguity function is defined as

A1(t θ τ) 1113946 slowast

uminusτ2

1113874 1113875ωlowast uminus tminusτ2

1113874 1113875s u +τ2

1113874 1113875

times ω uminus t +τ2

1113874 1113875ejθu

du

(12)

where ω(u) denotes the symmetric window function Forany A1(t θ τ) we can get the corresponding adaptive op-timal kernel function ϕopt(t θ τ) Under the premise ofensuring high resolution the AOKR shows good perfor-mance to analyze vibration signals of the chute under dif-ferent working conditions e adaptive optimal kerneltime-frequency distribution is expressed as

PAOK(tω) 12π

1113946 1113946 A1(t θ τ)ϕopt(t θ τ)eminusjθtminusjωτ

dθ dτ

(13)

3 Finite Element Modeling

31 Modal Analysis e CVM shown in Figure 2(a) isestablished in the software ANSYS and designed to befaithful to the actual device In fact the CVM is a typical

multibody dynamics system that consists of the chute platchains vertical chains and scrapers the dimension ofscraper chains is φ48times152 (mm) and the CVM contains773782 elements and 1117993 nodes Referring toFigure 2(a) to describe the interactions between differentcomponents the kinematic restriction mainly includes thefollowing two aspects kinematic constraints and contactrelations e inertial coordinate system I works as thereference of the system frame and is applied to determine thelocation of different components and contact pairs occurbetween two bodies moving correspondingly to each othere vibration signals of the chute can be expressed as thesum of natural vibration modes of each order and the lowerorder vibration modes have a greater correlation with thevibration properties To obtain intrinsic properties of thechute we extract the first 6-order natural frequencies of theCVM for modal analysis which would facilitate furtherstudies on the optimal placement of the acceleration sensors

e natural frequencies of the chute are obtained byfinite element analysis (FEA) of the CVM and experimentalmodal analysis (EMA) through the HIT As shown inFigure 2(b) the HIT is performed with the use of an impacthammer (LC02) and an acceleration sensor (TST120A1000)which is conducted for vibration measurements of the actualchute e exciting point is impacted by using the impacthammer to trigger vibration signals and then the acceler-ation sensor mounted on the surface of the shovel coal boardis used to pick up the vibration response Besides thewireless acquisition device is utilized for collection andtransmission of the experimental data which is detailed inSection 42 us we get the exciting force and vibrationresponse as shown in Figure 3

rough model simulation we get the modal displace-ment contours of the first 6-order natural vibrationmodes ofthe chute (Figure 4) According to the HIT natural fre-quencies of the actual chute are extracted from the exper-imental modal curve and each peak of the response curve iscorresponding to a vibration mode As listed in Table 1comparisons between natural frequencies of the first 6-ordernatural vibration modes based on FEA and EMA are madeIn addition the error ψi is defined as ψi |(αi minus βi)βi|

(i 1 sim 6) where αi and βi denote the natural frequencyderived from FEA and EMA respectively In Table 1 theerrors show low values for each mode In particular themaximum error between the simulation results and ex-perimental results is 1330 and the minimum ψi is almost232 e mentioned errors indicate the superiority of theestablished CVM

32 Construction of the DTSM In actual operation thescraper chain is driven by using drive motors and movescontinuously along the chutes and its reliability is directlyrelated to the working performance of the transmissionsystem Figure 5 describes the dynamic transmission systemmodel (DTSM) built by using the transient analysis modulein ANSYS e primary goal of construction of the DTSM isto achieve an accurate dynamic model which would facil-itate further studies on the vibration behavior of the actual

4 Shock and Vibration

X

Y

ZI

Vertical chain

O

Scraper

System frameChute

Plat chain

Contact Contact

Fixed joint

Contact Fixed jo

int

(a)

Scraper

Scraper chainsAcceleration

sensor

Chute

Exciting point

Impact hammer

Shovelcoal board

(b)

Figure 2 Descriptions of modal analysis based on (a) the CVM and (b) the HIT

350

250300

200150100

500

ndash50

Time (s)

Exci

ting

forc

e (N

)

0 05 1

(a)

0 05 1

15

1

05

0

ndash05

ndash1

Acce

lera

tion

(ms

2 )

Time (s)

(b)

Figure 3 Exciting force and the vibration response

Mode 1

(1729 0612)(2114 04954)

(2578 08114) (3096 08938)

(314 04933)

(3511 01923)

Acce

lera

tion

(ms

2 ) Mode 2

Mode 3

Mode 4

Mode 5

Mode 6

0

05

1

15

2

Frequency (Hz)0 50 100 150 200 250 300 350 400 450 500

Experimental modal

Figure 4 Vibration modes of the chute

Shock and Vibration 5

chute without carrying out physical tests e simulationparameters and kinematic restriction of the DTSM areconsistent with the CVM in Figure 2(a) To illustrate thisissue in more detail the frictional contacts are set betweenthe vertical chains and the chutes and between the scrapersand chutes Correspondingly the bonded contacts are setbetween the plat chains and the vertical chains e dynamicmodel mainly contains two prominent parts the chain as-sembly and the chutes Wherein the chain assembly iscomposed of the scrapers and scraper chains and can beequally divided into multiple segments of length ΔL thenumber of the chutes is 7 and they are marked as shown inFigure 5 A translational joint is set between scraper 1 andthe middle plate based on which the chain assembly can runalong the chutes at a transmission speed of Vi Moreover thepretightening force and external load of the dynamic modelare defined as F1 and Wi respectively Here we setF1 79612 kN In Figure 5 the pretightening force is ap-plied at the two ends of the chain assembly which can ensurethat the scraper chain remains tight during the movemente external load Wi is applied on the upper surface of thescrapers and middle plates and the material density inthe chute under the full-load condition is W0 whereW0 694 kgm

4 AOKR-Based Fault Detection Strategy of theScraper Chain

Considering that the scraper conveyor is susceptible tofrequent loading excessive bending and artificial mis-conduct in actual engineering different patterns of chainfaults usually act on the scraper chain e failure patternswill cause abnormal vibration of the chute Actuallydirect measurement and analysis for parameters of themoving scraper chain is difficult which raises a contra-diction between practical engineering requirements andthe installation limitations of multiple sensors In order todetect chain faults promptly the fix-point vibrationmeasurement of the measuring points for the chute isconsidered instead of mobile parameter measurement forthe scraper chain In this section the DTSM is utilized toacquire the vibration properties of the actual chute undervarious working conditions and the FPETof the vibrationsignals is conducted to validate the accuracy of the dy-namic model e fundamental idea of our proposedstrategy is to determine the occurrence of chain faultsby amplitude comparisons and then fault signals areanalyzed through the AOKR to distinguish the faultpatterns

Table 1 Frequency response of vibration signals

Mode no FEA (Hz) EMA (Hz) ψi ()

1 17754 1729 2682 18329 2114 13303 28745 2578 11504 30242 3096 2325 32868 3140 4686 34197 3511 260

Position 1 Position 2

F1

F1

F1

F1

Translationaljoint

Scraper 1

Chain breakage

Chain jamContact point 1

∆L

Middle plateChain assembly

Wi

1 3 4 5 7

Vi

Vi

Figure 5 Simulation process of the DTSM

6 Shock and Vibration

41 Optimal Placement of the Acceleration Sensors Beforeconducting the FPET the installation scheme of the ac-celeration sensors should be determined to ensure thevalidity of the measured vibration signals and the economyof the experimental process Based on modal analysis of theFEA (Figure 2(a)) we extract 20 nodes from the chute as theprimary measuring points which are labeled in Figure 6(a)According to the theory of MAC the total modal dis-placements of the selected 20 points serve as the inputs ofequations (6) and (7) In Figure 6(b) the curve revealschange rules of the minimumMACij for different number ofsensors

To illustrate the optimal placement scheme of acceler-ation sensors in more detail the optimum installation po-sitions for different number of sensors are shown in Table 2Accordingly the vibration response of the actual chuteshould be recorded by 3 acceleration sensors and the op-timum installation positions are set at nodes 3 9 and 16respectively

42 Experimental Evaluations According to Section 32 thesimulation process of the DTSM is performed under normalcondition without a load and the transmission speed Vi isset as V0 in Figure 7(a) and the scraper 1 moves fromposition 1 to position 2 In our study nodes 9 16 and 3 ofthe fourth chute are chosen as the detecting points for vi-bration analysis For convenient expression in the followingwe mark the detecting nodes as measuring points 1 2 and3 respectively Taking measuring point 1 as the researchexample for description of the transmission process theoriginal signal is presented in Figure 7(b) e vibrationresponse presents a three-stage change ie the accelerationphase the steady phase and the deceleration phase Asmentioned above in the steady phase the vibration signalcan be treated as multiple segments with a time span of ΔTwhich corresponds to the chain assembly with a length of ΔL(Figure 5)

e field FPET of multiple monitoring points iscarried out to obtain the vibration properties of the actualchute and evaluate the dynamic performance of theDTSM e specifications of the experimental scraperconveyor correspond to SGB12003600 manufactured byLianyungang Tianming Equipment Co Ltd and therunning speed of the scraper chain is 10 ms In practicalengineering the actual double-drive transmission systemis driven by two drive motors e basic performanceparameters are shown in Table 3 As shown in Figure 8the experimental system mainly contains three acceler-ation sensors (TST120A1000) a wireless acquisitiondevice (TST5925EV) a wireless receiver and an on-sitePC e acceleration sensors are used for fixed-pointmeasurements by detecting the vibration responses ofmeasuring points 1 2 and 3 on the chute e wirelessacquisition device is used to collect experimental data inreal time In addition the wireless receiver is intended forremote data transmissione vibration data are stored inthe on-site PC and a sampling frequency of 1000 Hz istaken

Figure 9 shows the experimental and simulation signalsof measuring points 1 2 and 3 in the time span ΔT Inthe time domain the simulation results and the experi-mental results are similar which reflects the adaptability ofthe DTSM Considering the nonlinear and time-varyingcharacteristics of the vibration signals the AOKR is used fortime-frequency analysis

e vibration signal of measuring point 2 in the timespan ΔT is taken as the reference for time-frequency

1 2 3 4 5 6 7 8 9 10

11 12 13 14 15 16 17 18 19 20

(a)

Number of the sensors

Min

imum

val

ue o

f MAC

ij

2 3 4 5 6 7 8 9 10

028

0275

027

0265

026

0255

025

0245

(b)

Figure 6 Description of the optimum placement scheme (a) theprimary measuring points (b) the minimum MACij

Table 2 Measuring points placement based on the MAC

Number f Measuring points2 02515 8 163 02453 3 9 164 02454 1 3 9 165 02460 1 3 9 16 206 02483 1 3 9 10 11 167 02525 1 2 3 9 10 11 168 02644 1 3 8 9 10 11 15 169 02664 1 2 3 8 9 10 11 15 1610 02767 1 2 3 8 9 10 11 15 16 20

Shock and Vibration 7

analysis and then the time-frequency representations ofmeasuring points 1 and 3 in the same time period aredemonstrated Figure 10(a) shows the time-frequencyrepresentation of the experimental signals for differentmeasuring points and the power spectrum is alsodepicted It presents that the vibration signal of the chutehas a plurality of components ere are two major areaswith strong frequency responses of the vibration signalsfor different measuring points for measuring point 1 thefrequency range is 230ndash310Hz and the time ranges are015ndash030 s and 045ndash075 s for measuring point 2 thefrequency range is 230ndash340Hz and the time ranges are020ndash040 s and 055ndash070 s and for measuring point 3the frequency range is 225ndash330Hz and the time ranges are040ndash050 s and 070ndash090 s Considering the chain

assembly in Figure 5 there are two scrapers for length ΔLIn the actual transmission process the two major areaswith strong frequency responses are caused while the twoscrapers passing through the measuring points As shownbasic frequencies of the measuring points are 239 278 and249Hz respectively In addition the peak of the vibrationpower spectrum is also presented and the maximumvalues of the measuring points are 1588 1561 and1499m2s

Similarly the time-frequency representation of thesimulation signals with two major frequency responses isdiscussed As Figure 10(b) indicates for measuring points1 2 and 3 the vibration energies are concentrated atthe frequency ranges 240ndash340 200ndash310 and 200ndash320Hzrespectively Correspondingly for measuring point 1 the

t1 t3

t2

Chain jamoccurs

15

1

05

0

Time (s)

Velo

city

(ms

)

0 05 1 15 2 25 3 35 4 45 5 55 6

V0V1

(a)

Constant speed

Acceleration DecelerationΔT

Time (s)0 1 2 3 4 5 6

543210

ndash1ndash2ndash3ndash4ndash5

Acce

lera

tion

(ms

2 )

(b)

Figure 7 Transmission speed setting and vibration signal under normal condition

Table 3 Basic operating condition of the scraper conveyor

Model Chain size (mm) Conveyor width (mm) Chain speed (ms) Transport capacity (th) Transport length (m)SGZ12003600 φ48times152 1750times1180 0sim189 3700 360

Gear

Gear

Sprocket

Drive motor On-site PCWireless receiver

Transitionaltrough Chute

Accelerationsensor

Wirelessacquisition device

2

3

1

Figure 8 Experimental setup of the FPET

8 Shock and Vibration

time ranges are 010ndash025 and 050ndash065 s en the vi-bration energy of measuring point 2 is concentrated at thetime ranges 015ndash045 and 060ndash075 s and the time rangesof measuring point 3 are 040ndash050 and 070ndash085 s Forthe simulation signals the basic frequencies of the mea-suring points are 281 249 and 258Hz respectively emaximum values of vibration energies are 1527 1489 and1532m2s respectively Considering the experimental re-sults and the simulation results separately good agreementbetween the main parameters of different measuring pointsis obtained It also indicates that in the same time periodthe vibration signal has a delay characteristic for measuringpoints 1 2 and 3 is is possibly because that themeasuring points are at different positions of the chuteMoreover for different measuring points the experimentalsignals and the simulation signals show good consistency intime-frequency characteristics Hence the establishedDTSM can efficiently simulate the actual productionenvironment

43 Fault Detection of the Scraper Chain Based on theestablished DTSM two typical failure patterns of thescraper chain are discussed namely chain jam and chainfracture ree external load conditions are set that isempty load half-load and full-load conditions e valuesof Wi (Figure 5) are defined as 0 12W0 and W0 re-spectively As designed in Figure 5 when chain jam occursthe transmission speed Vi is set as V1 in Figure 7(a) and the

failure time range is t1 minus t3(t3 25 s) the chain fault istriggered at t1 2 sWithin the time range t1 minus t2(t2 225 s)the transmission speed Vi decreases from 1 to 0ms And thevalue of Vi increases from 0 to 1ms within the time ranget2 minus t3 e whole process lasts 05 s the chain assembly istightened and the scraper chains are jammed When chainfracture occurs the transmission speed Vi is set as V0 inFigure 7(a) Moreover as shown in Figure 5 the contactconstraint between two contacting scraper chains at the labeledcontact point 1 is removed As a result the two contactingscraper chains will be separated In order to ensure the accuracyof fault setting the chain fracture is also triggered at t1 2 sand the contact constraint is removed in 05 s which is con-sistent with chain jam

Considering both failure patterns of the scraper chainour research focuses on the steady phase of the operationprocess within the time range 05 to 55 s In the steadyphase the vibration signals are obtained simultaneously atmeasuring points 1 2 and 3 Taking measuring point1 as the case study the vibration signals for chain jam andchain fracture under empty load condition are presentedin Figures 11(a) and 11(b) respectively After faultstriggering the vibration signals of the two failure patternsshow a sudden increase after a short time delay Sub-sequently the signals exhibit unstable fluctuations at thetime range 24ndash28 s In order to obtain a more detaileddescription of the detection results the maximum am-plitudes of the vibration signals under different workingconditions are discussed ie normal condition chain jam

3210

ndash1ndash2ndash3

3210

ndash1ndash2ndash3

3210

ndash1ndash2ndash3

Time (s)

Acce

lera

tion

(ms

2 )

0 01 02 03 04 05 06 07 08 09

1 2 3

(a)

3210

ndash1ndash2ndash3

3210

ndash1ndash2ndash3

3210

ndash1ndash2ndash3

Time (s)Ac

cele

ratio

n (m

s2 )

0 01 02 03 04 05 06 07 08 09

123

(b)

Figure 9 Vibration signals of different measuring points in ΔT (a) experimental results (b) simulation results

Shock and Vibration 9

Point

1

2

3

(a) (b)

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

0

Time (s) Spectrum0 03 06 09

500

400

300

200

100

00 50 100 150

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

00 03 06 09

500

400

300

200

100

00 50 100 150

Time (s) Spectrum

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

00 03 06 09

500

400

300

200

100

00 50 100 150

Time (s) Spectrum

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

00 03 06 09

500

400

300

200

100

00 50 100 150

Time (s) Spectrum

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

0

Time (s) Spectrum0 03 06 09

500

400

300

200

100

00 50 100 150

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

00 03 06 09

500

400

300

200

100

00 50 100 150

Time (s) Spectrum

Figure 10 Time-frequency representation of vibration signals (a) experimental results (b) simulation results

Chain jamoccurs

Acce

lera

tion

(ms

2 ) Acce

lera

tion

(ms

2 )

10

5

0

ndash5

ndash10

Time (s)0 05 10 15 20 25 30 35 40 45 50

10

5

0

ndash5

ndash1024 25 26 27 28

(a)

Figure 11 Continued

10 Shock and Vibration

and chain fracture Wherein the empty load half-loadand full-load conditions are considered e maximumamplitudes of the vibration signals at measuring points1 2 and 3 are depicted in Figures 12(a)ndash12(c)respectively

Considering measuring point 1 the maximum am-plitudes of the vibration signals under empty load conditionare 289 718 and 698ms2 under normal condition chainjam and chain fracture respectively Similarly under half-load condition the maximum amplitudes are 324 937 and956ms2 Moreover under full-load condition the maxi-mum amplitudes are 446 1106 and 1082ms2 Withdifferent external loads the maximum amplitudes of thevibration signals for fault conditions are obviously higherthan those for normal condition and the difference betweenthe amplitudes of the two typical failure patterns is small Fordifferent fault conditions with the increase of the externalloads the maximum amplitudes show trends to increasee above statistical results are also applicable to measuringpoints 2 and 3 erefore chain faults can easily be de-tected by comparing the maximum amplitudes of the vi-bration signals whereas the fault patterns are difficult toidentify According to the nonstationary and nonlinearcharacteristics of fault signals the AOKR is utilized to an-alyze the vibration signals and classify failure patterns of thescraper chain Within 15 s after faults triggering the vi-bration signals at the three measuring points with differentexternal loads are processed Wherein for chain jam andchain fracture under empty load condition the time-fre-quency representations of vibration signals are presented inFigures 13(a) and 13(b) respectively

e frequency components and frequency ranges ofthe same fault pattern are similar for different measuringpoints As Figure 13(a) describes the bright color between0 and 50Hz indicates one high energy area caused bychain jam en chain fracture can easily be distinguishedaccording to the appearance of two high energy areasbetween 100 and 200Hz as shown in Figure 13(b)

Observing the spectrum results a more detailed de-scription is given When chain jam occurs for measuringpoints 1 2 and 3 the high energy areas occur ap-proximately at 05 075 and 09 s respectively Mean-while for chain fracture the high energy areas include twomain frequency components and are approximatelyconcentrated at the time ranges 050ndash070 075ndash085 and10ndash115 s respectively Hence there is a delay charac-teristic of the fault occurrence which is well in accordancewith the conclusions in Section 42 In order to explore theinfluence of external loads on fault characteristics thedetailed differences of the fault patterns at measuringpoint 2 are depicted in Figure 14 In fact the externalload has a great influence on the fault severity of both thefailure patterns Observing the spectrum results thebright areas vary with the external loads With increasingexternal load the frequency ranges of the high energyareas become larger Wherein for chain jam under emptyhalf- and full-load conditions the frequency ranges areapproximately 0ndash50 0ndash150 and 0ndash250 Hz respectivelyMeanwhile for chain fracture the frequency rangesare approximately 80ndash200 50ndash250 and 50ndash350Hzrespectively

In this part three working conditions of the scraperchain are investigated above including normal conditionchain jam and chain fracture e vibration signals ofmeasuring points 1 2 and 3 on the detecting chute areanalyzed and the effects of the external loads on the vi-bration characteristics are discussed Based on the aboveanalysis the occurrence of chain faults can easily be de-termined through amplitude comparisons of the originalvibration signals However the observation confirms thesimilarity of the time domain waveforms of fault signals forchain jam and chain fracture ese two patterns of failuresremain to be different through further processing by theAOKR and the fault patterns can be distinguished accordingto the number of high energy areas of the time-frequencyrepresentation of vibration signals In conclusion the

Chain fractureoccurs

Acce

lera

tion

(ms

2 ) Acce

lera

tion

(ms

2 )

10

5

0

ndash5

ndash10

Time (s)0 05 10 15 20 25 30 35 40 45 50

10

5

0

ndash5

ndash1024 25 26 27 28

(b)

Figure 11 Vibration signals of the measuring point 1 for (a) chain jam and (b) chain fracture

Shock and Vibration 11

Point 1 2 3

(a)

(b)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

Figure 13 Time-frequency representation of vibration signals under empty load (a) chain jam (b) chain fracture

12

10

8

6

4

2

0Max

imum

acce

lera

tion

ampl

itude

(ms

2 )

Normal conditionChain jamChain fracture

0 12 W0 W0

(a)

Normal conditionChain jamChain fracture

12

10

8

6

4

2

0Max

imum

acce

lera

tion

ampl

itude

(ms

2 )

0 12 W0 W0

(b)

Normal conditionChain jamChain fracture

12

10

8

6

4

2

0Max

imum

acce

lera

tion

ampl

itude

(ms

2 )

0 12 W0 W0

(c)

Figure 12 Statistical results of the maximum vibration amplitude at different measuring points (a) 1 (b) 2 (c) 3

12 Shock and Vibration

proposed detection strategy is effective at detecting theoccurrence of chain faults and identifying the failure pat-terns under different operating conditions

5 Conclusions

During the actual operation the working state of thescraper chain can reflect the dynamic performance of thescraper conveyor To address the difficulties with directsensor measurement for parameters of the moving scraperchain a novel strategy for fault detection of the scraperchain based on vibration analysis of the chute was pro-posed Based on modal analysis and the MAC the mea-suring points of vibration signals on the chute weredetermined To fit the actual behavior of the transmissionprocess the DTSM was presented based on finite elementmodeling and the correctness of the dynamic model wasverified by comparison with the FPET en the vibrationproperties of the measuring points on the chute undernormal condition chain jam and chain fracture werediscussed Moreover the occurrence of chain faults weredetermined by comparing the amplitudes of the vibrationsignal in the time domain while the AOKR was utilizedfor time-frequency representation of vibration signals anddistinguishing the two typical failure patterns Further-more the strategy verification based on experimental datawill be taken into consideration in the near future

Data Availability

e data used to support the findings of this study areavailable from the corresponding author upon request

Conflicts of Interest

e authors declare that they have no conflicts of interest

Acknowledgments

is work was supported by the Key Project of NationalNatural Science Foundation of China (U1510205) NaturalScience Foundation of Jiangsu Province (No BK20160251)Xuzhou Research program (KC14H0138) FundamentalResearch Funds for the Central Universities (2014Y05) andProject Funded by the Priority Academic Program De-velopment of Jiangsu Higher Education Institutions(PAPD)

References

[1] C D Brown ldquoDesign build and test of a longwall armouredface conveyorrdquo Longwall Mining 2002

[2] M Dolipski P Cheluszka E Remiorz and P SobotaldquoFollow-up chain tension in an armoured face conveyornadazne napinanie lancucha zgrzebłowego W przenosnikuscianowymrdquo Archives of Mining Sciences vol 60 no 1pp 25ndash38 2015

[3] L A Morley J L Kohler and H M Smolnikar ldquoA model forpredicting motor load for an armored face-conveyor driverdquoIEEE Transactions on Industry Applications vol 24 no 4pp 649ndash659 1988

[4] A A Ordin and A A Metelrsquokov ldquoAnalysis of longwall faceoutput in screw-type cutter-loader-and-scraper conveyorsystem in underground mining of flat-lying coal bedsrdquoJournal of Mining Science vol 51 no 6 pp 1173ndash1179 2015

[5] B He G Li H Shi et al ldquoDynamic behaviour modelling andsimulation of the chain transmission system for an armouredface conveyorrdquo in Proceedings of the IEEE 10th InternationalConference on Computer-Aided Industrial Design and Con-ceptual Design CAID amp CD 2009 pp 1000ndash1004 BeijingChina November 2009

[6] R Nie B He P Yuan L Zhang and G Li ldquoNovel approachto and implementation of design and analysis of armored faceconveyor power trainrdquo Science China Technological Sciencesvol 58 no 12 pp 2153ndash2168 2015

Loads

(a)

(b)

Empty load Half-load Full-load

500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)

Figure 14 Time-frequency representation of vibration signals at measuring point 2 with different external loads (a) chain jam (b) chainfracture

Shock and Vibration 13

[7] R Nie B He L Zhang and G Li ldquoModelling of thetransmission system in conveying equipment based on Eulermethod with applicationrdquo Proceedings of the Institution ofMechanical Engineers Part K Journal of Multi-body Dy-namics vol 228 no 3 pp 294ndash306 2014

[8] S B Jiang X Zhang K D Gao J Gao Q Y Wang andK Hidenori ldquoMulti-body dynamics and vibration analysis ofchain assembly in armoured face conveyorrdquo InternationalJournal of Simulation Modelling vol 16 no 3 pp 458ndash4702017

[9] M Myszkowski and D Loehning ldquoChain force measure-ments on armoured face conveyors and coal plows in heavy-duty longwallsrdquo CIM Bulletin vol 94 no 1054 pp 72ndash752001

[10] H Wang Q Zhang and F Xie ldquoDynamic tension test andintelligent coordinated control system of a heavy scraperconveyorrdquo IET Science Measurement and Technology vol 11no 7 pp 871ndash877 2017

[11] S Sen M X Min and Y Z She ldquoDiagnosis of coal scraperconveyor based on Fuzzy Fault treerdquo in Proceedings of the2015 Seventh International Conference on Measuring Tech-nology and Mechatronics Automation (ICMTMA) pp 392ndash395 IEEE Nanchang China June 2015

[12] S-s Xue X-c Li and X-y Xu ldquoFault tree and Bayesiannetwork based scraper conveyer fault diagnosisrdquo in Pro-ceedings of the 22nd International Conference on IndustrialEngineering and Engineering Management 2015 pp 783ndash795Atlantis Press Paris France January 2016

[13] X Gong X Ma Y Zhang et al ldquoApplication of fuzzy neuralnetwork in fault diagnosis for scraper conveyor vibrationrdquo inProceedings of the 2013 IEEE International Conference onInformation and Automation (ICIA) pp 1135ndash1140 IEEEYinchuan China August 2013

[14] Y Zhang X Ma Y Jianxiang et al ldquoFuzzy neural networkfault diagnosis and online vibration monitoring system for thecoal scraper conveyor based on rough set theoryrdquo in Pro-ceedings of the 2013 32nd Chinese Control Conference (CCC)pp 6134ndash6138 IEEE Xirsquoan China July 2013

[15] B Zhang A C C Tan and J-h Lin ldquoGearbox fault diagnosisof high-speed railway trainrdquo Engineering Failure Analysisvol 66 pp 407ndash420 2016

[16] E Parloo P Verboven P Guillaume and M Van OvermeireldquoAutonomous structural health monitoring-part ii vibration-based in-operation damage assessmentrdquo Mechanical Systemsand Signal Processing vol 16 no 4 pp 659ndash675 2002

[17] C S Sakaris J S Sakellariou and S D Fassois ldquoRandom-vibration-based damage detection and precise localization ona lab-scale aircraft stabilizer structure via the GeneralizedFunctional Model Based Methodrdquo Structural Health Moni-toring An International Journal vol 16 no 5 pp 594ndash6102017

[18] Y Zhang W Song M Karimi C-H Chi and A KudreykoldquoFractional autoregressive integrated moving average andfinite-element modal the forecast of tire vibration trendrdquoIEEE Access vol 6 pp 40137ndash40142 2018

[19] M Pastor M Binda and T Harcarik ldquoModal assurancecriterionrdquo Procedia Engineering vol 48 pp 543ndash548 2012

[20] W J Staszewski K Worden and G R Tomlinson ldquoTime-frequency analysis in gearbox fault detection using the Wigner-ville distribution and pattern recognitionrdquo Mechanical Systemsand Signal Processing vol 11 no 5 pp 673ndash692 1997

[21] J-D Wu and P-H Chiang ldquoApplication of Wigner-Villedistribution and probability neural network for scooter engine

fault diagnosisrdquo Expert Systems with Applications vol 36no 2 pp 2187ndash2199 2009

[22] Z Feng and M Liang ldquoFault diagnosis of wind turbineplanetary gearbox under nonstationary conditions viaadaptive optimal kernel time-frequency analysisrdquo RenewableEnergy vol 66 pp 468ndash477 2014

14 Shock and Vibration

International Journal of

AerospaceEngineeringHindawiwwwhindawicom Volume 2018

RoboticsJournal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Active and Passive Electronic Components

VLSI Design

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Shock and Vibration

Hindawiwwwhindawicom Volume 2018

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawiwwwhindawicom

Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Control Scienceand Engineering

Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom

Journal ofEngineeringVolume 2018

SensorsJournal of

Hindawiwwwhindawicom Volume 2018

International Journal of

RotatingMachinery

Hindawiwwwhindawicom Volume 2018

Modelling ampSimulationin EngineeringHindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Navigation and Observation

International Journal of

Hindawi

wwwhindawicom Volume 2018

Advances in

Multimedia

Submit your manuscripts atwwwhindawicom

analysis theory the main tools dedicated to the study ofnonstationary signals are available A commonly used time-frequency distribution is the WignerndashVille distribution(WVD) which has found many successful applications indifferent areas [20 21] e input signal in the time domainis denoted by using s(t) and the definition of WVD can begiven by

DWV(tω) 1113946 s t +τ2

1113874 1113875slowast

tminusτ2

1113874 1113875 eminusjωτ dτ (8)

where ω denotes the frequency variable and slowast(t) is theconjugate function of s(t) en equation (8) can be con-verted into the time-frequency distribution function as

A(θ τ) 1113946 s t +τ2

1113874 1113875slowast

tminusτ2

1113874 1113875 ejθt dt (9)

where θ and τ denote the fuzzy-domain variables Based onequations (8) and (9) we get

DWV(tω) 12π

1113946 1113946 A(θ τ)eminusjθtminusjωτ

dθ dτ

A(θ τ) 1113946 1113946 DWV(tω)ejθt+jωτ

dt dω

(10)

Developed from WVD the Cohen class bilinear distri-butions are the commonly used time-frequency analysismethods and can be expressed as

P(tω) 12π

1113946 1113946 A(θ τ)ϕ(θ τ)eminusjθtminusjωτ

dθ dτ (11)

where A(θ τ) denotes the ambiguity function and ϕ(θ τ)

denotes the kernel function Bilinear time-frequency anal-ysis such as WVD exhibits a high time-frequency resolutionwhereas it is contaminated by serious cross-terms en theAOKR is applied to suppress the cross-term interferencescaused by multicomponents [22] meanwhile the modifiedambiguity function is defined as

A1(t θ τ) 1113946 slowast

uminusτ2

1113874 1113875ωlowast uminus tminusτ2

1113874 1113875s u +τ2

1113874 1113875

times ω uminus t +τ2

1113874 1113875ejθu

du

(12)

where ω(u) denotes the symmetric window function Forany A1(t θ τ) we can get the corresponding adaptive op-timal kernel function ϕopt(t θ τ) Under the premise ofensuring high resolution the AOKR shows good perfor-mance to analyze vibration signals of the chute under dif-ferent working conditions e adaptive optimal kerneltime-frequency distribution is expressed as

PAOK(tω) 12π

1113946 1113946 A1(t θ τ)ϕopt(t θ τ)eminusjθtminusjωτ

dθ dτ

(13)

3 Finite Element Modeling

31 Modal Analysis e CVM shown in Figure 2(a) isestablished in the software ANSYS and designed to befaithful to the actual device In fact the CVM is a typical

multibody dynamics system that consists of the chute platchains vertical chains and scrapers the dimension ofscraper chains is φ48times152 (mm) and the CVM contains773782 elements and 1117993 nodes Referring toFigure 2(a) to describe the interactions between differentcomponents the kinematic restriction mainly includes thefollowing two aspects kinematic constraints and contactrelations e inertial coordinate system I works as thereference of the system frame and is applied to determine thelocation of different components and contact pairs occurbetween two bodies moving correspondingly to each othere vibration signals of the chute can be expressed as thesum of natural vibration modes of each order and the lowerorder vibration modes have a greater correlation with thevibration properties To obtain intrinsic properties of thechute we extract the first 6-order natural frequencies of theCVM for modal analysis which would facilitate furtherstudies on the optimal placement of the acceleration sensors

e natural frequencies of the chute are obtained byfinite element analysis (FEA) of the CVM and experimentalmodal analysis (EMA) through the HIT As shown inFigure 2(b) the HIT is performed with the use of an impacthammer (LC02) and an acceleration sensor (TST120A1000)which is conducted for vibration measurements of the actualchute e exciting point is impacted by using the impacthammer to trigger vibration signals and then the acceler-ation sensor mounted on the surface of the shovel coal boardis used to pick up the vibration response Besides thewireless acquisition device is utilized for collection andtransmission of the experimental data which is detailed inSection 42 us we get the exciting force and vibrationresponse as shown in Figure 3

rough model simulation we get the modal displace-ment contours of the first 6-order natural vibrationmodes ofthe chute (Figure 4) According to the HIT natural fre-quencies of the actual chute are extracted from the exper-imental modal curve and each peak of the response curve iscorresponding to a vibration mode As listed in Table 1comparisons between natural frequencies of the first 6-ordernatural vibration modes based on FEA and EMA are madeIn addition the error ψi is defined as ψi |(αi minus βi)βi|

(i 1 sim 6) where αi and βi denote the natural frequencyderived from FEA and EMA respectively In Table 1 theerrors show low values for each mode In particular themaximum error between the simulation results and ex-perimental results is 1330 and the minimum ψi is almost232 e mentioned errors indicate the superiority of theestablished CVM

32 Construction of the DTSM In actual operation thescraper chain is driven by using drive motors and movescontinuously along the chutes and its reliability is directlyrelated to the working performance of the transmissionsystem Figure 5 describes the dynamic transmission systemmodel (DTSM) built by using the transient analysis modulein ANSYS e primary goal of construction of the DTSM isto achieve an accurate dynamic model which would facil-itate further studies on the vibration behavior of the actual

4 Shock and Vibration

X

Y

ZI

Vertical chain

O

Scraper

System frameChute

Plat chain

Contact Contact

Fixed joint

Contact Fixed jo

int

(a)

Scraper

Scraper chainsAcceleration

sensor

Chute

Exciting point

Impact hammer

Shovelcoal board

(b)

Figure 2 Descriptions of modal analysis based on (a) the CVM and (b) the HIT

350

250300

200150100

500

ndash50

Time (s)

Exci

ting

forc

e (N

)

0 05 1

(a)

0 05 1

15

1

05

0

ndash05

ndash1

Acce

lera

tion

(ms

2 )

Time (s)

(b)

Figure 3 Exciting force and the vibration response

Mode 1

(1729 0612)(2114 04954)

(2578 08114) (3096 08938)

(314 04933)

(3511 01923)

Acce

lera

tion

(ms

2 ) Mode 2

Mode 3

Mode 4

Mode 5

Mode 6

0

05

1

15

2

Frequency (Hz)0 50 100 150 200 250 300 350 400 450 500

Experimental modal

Figure 4 Vibration modes of the chute

Shock and Vibration 5

chute without carrying out physical tests e simulationparameters and kinematic restriction of the DTSM areconsistent with the CVM in Figure 2(a) To illustrate thisissue in more detail the frictional contacts are set betweenthe vertical chains and the chutes and between the scrapersand chutes Correspondingly the bonded contacts are setbetween the plat chains and the vertical chains e dynamicmodel mainly contains two prominent parts the chain as-sembly and the chutes Wherein the chain assembly iscomposed of the scrapers and scraper chains and can beequally divided into multiple segments of length ΔL thenumber of the chutes is 7 and they are marked as shown inFigure 5 A translational joint is set between scraper 1 andthe middle plate based on which the chain assembly can runalong the chutes at a transmission speed of Vi Moreover thepretightening force and external load of the dynamic modelare defined as F1 and Wi respectively Here we setF1 79612 kN In Figure 5 the pretightening force is ap-plied at the two ends of the chain assembly which can ensurethat the scraper chain remains tight during the movemente external load Wi is applied on the upper surface of thescrapers and middle plates and the material density inthe chute under the full-load condition is W0 whereW0 694 kgm

4 AOKR-Based Fault Detection Strategy of theScraper Chain

Considering that the scraper conveyor is susceptible tofrequent loading excessive bending and artificial mis-conduct in actual engineering different patterns of chainfaults usually act on the scraper chain e failure patternswill cause abnormal vibration of the chute Actuallydirect measurement and analysis for parameters of themoving scraper chain is difficult which raises a contra-diction between practical engineering requirements andthe installation limitations of multiple sensors In order todetect chain faults promptly the fix-point vibrationmeasurement of the measuring points for the chute isconsidered instead of mobile parameter measurement forthe scraper chain In this section the DTSM is utilized toacquire the vibration properties of the actual chute undervarious working conditions and the FPETof the vibrationsignals is conducted to validate the accuracy of the dy-namic model e fundamental idea of our proposedstrategy is to determine the occurrence of chain faultsby amplitude comparisons and then fault signals areanalyzed through the AOKR to distinguish the faultpatterns

Table 1 Frequency response of vibration signals

Mode no FEA (Hz) EMA (Hz) ψi ()

1 17754 1729 2682 18329 2114 13303 28745 2578 11504 30242 3096 2325 32868 3140 4686 34197 3511 260

Position 1 Position 2

F1

F1

F1

F1

Translationaljoint

Scraper 1

Chain breakage

Chain jamContact point 1

∆L

Middle plateChain assembly

Wi

1 3 4 5 7

Vi

Vi

Figure 5 Simulation process of the DTSM

6 Shock and Vibration

41 Optimal Placement of the Acceleration Sensors Beforeconducting the FPET the installation scheme of the ac-celeration sensors should be determined to ensure thevalidity of the measured vibration signals and the economyof the experimental process Based on modal analysis of theFEA (Figure 2(a)) we extract 20 nodes from the chute as theprimary measuring points which are labeled in Figure 6(a)According to the theory of MAC the total modal dis-placements of the selected 20 points serve as the inputs ofequations (6) and (7) In Figure 6(b) the curve revealschange rules of the minimumMACij for different number ofsensors

To illustrate the optimal placement scheme of acceler-ation sensors in more detail the optimum installation po-sitions for different number of sensors are shown in Table 2Accordingly the vibration response of the actual chuteshould be recorded by 3 acceleration sensors and the op-timum installation positions are set at nodes 3 9 and 16respectively

42 Experimental Evaluations According to Section 32 thesimulation process of the DTSM is performed under normalcondition without a load and the transmission speed Vi isset as V0 in Figure 7(a) and the scraper 1 moves fromposition 1 to position 2 In our study nodes 9 16 and 3 ofthe fourth chute are chosen as the detecting points for vi-bration analysis For convenient expression in the followingwe mark the detecting nodes as measuring points 1 2 and3 respectively Taking measuring point 1 as the researchexample for description of the transmission process theoriginal signal is presented in Figure 7(b) e vibrationresponse presents a three-stage change ie the accelerationphase the steady phase and the deceleration phase Asmentioned above in the steady phase the vibration signalcan be treated as multiple segments with a time span of ΔTwhich corresponds to the chain assembly with a length of ΔL(Figure 5)

e field FPET of multiple monitoring points iscarried out to obtain the vibration properties of the actualchute and evaluate the dynamic performance of theDTSM e specifications of the experimental scraperconveyor correspond to SGB12003600 manufactured byLianyungang Tianming Equipment Co Ltd and therunning speed of the scraper chain is 10 ms In practicalengineering the actual double-drive transmission systemis driven by two drive motors e basic performanceparameters are shown in Table 3 As shown in Figure 8the experimental system mainly contains three acceler-ation sensors (TST120A1000) a wireless acquisitiondevice (TST5925EV) a wireless receiver and an on-sitePC e acceleration sensors are used for fixed-pointmeasurements by detecting the vibration responses ofmeasuring points 1 2 and 3 on the chute e wirelessacquisition device is used to collect experimental data inreal time In addition the wireless receiver is intended forremote data transmissione vibration data are stored inthe on-site PC and a sampling frequency of 1000 Hz istaken

Figure 9 shows the experimental and simulation signalsof measuring points 1 2 and 3 in the time span ΔT Inthe time domain the simulation results and the experi-mental results are similar which reflects the adaptability ofthe DTSM Considering the nonlinear and time-varyingcharacteristics of the vibration signals the AOKR is used fortime-frequency analysis

e vibration signal of measuring point 2 in the timespan ΔT is taken as the reference for time-frequency

1 2 3 4 5 6 7 8 9 10

11 12 13 14 15 16 17 18 19 20

(a)

Number of the sensors

Min

imum

val

ue o

f MAC

ij

2 3 4 5 6 7 8 9 10

028

0275

027

0265

026

0255

025

0245

(b)

Figure 6 Description of the optimum placement scheme (a) theprimary measuring points (b) the minimum MACij

Table 2 Measuring points placement based on the MAC

Number f Measuring points2 02515 8 163 02453 3 9 164 02454 1 3 9 165 02460 1 3 9 16 206 02483 1 3 9 10 11 167 02525 1 2 3 9 10 11 168 02644 1 3 8 9 10 11 15 169 02664 1 2 3 8 9 10 11 15 1610 02767 1 2 3 8 9 10 11 15 16 20

Shock and Vibration 7

analysis and then the time-frequency representations ofmeasuring points 1 and 3 in the same time period aredemonstrated Figure 10(a) shows the time-frequencyrepresentation of the experimental signals for differentmeasuring points and the power spectrum is alsodepicted It presents that the vibration signal of the chutehas a plurality of components ere are two major areaswith strong frequency responses of the vibration signalsfor different measuring points for measuring point 1 thefrequency range is 230ndash310Hz and the time ranges are015ndash030 s and 045ndash075 s for measuring point 2 thefrequency range is 230ndash340Hz and the time ranges are020ndash040 s and 055ndash070 s and for measuring point 3the frequency range is 225ndash330Hz and the time ranges are040ndash050 s and 070ndash090 s Considering the chain

assembly in Figure 5 there are two scrapers for length ΔLIn the actual transmission process the two major areaswith strong frequency responses are caused while the twoscrapers passing through the measuring points As shownbasic frequencies of the measuring points are 239 278 and249Hz respectively In addition the peak of the vibrationpower spectrum is also presented and the maximumvalues of the measuring points are 1588 1561 and1499m2s

Similarly the time-frequency representation of thesimulation signals with two major frequency responses isdiscussed As Figure 10(b) indicates for measuring points1 2 and 3 the vibration energies are concentrated atthe frequency ranges 240ndash340 200ndash310 and 200ndash320Hzrespectively Correspondingly for measuring point 1 the

t1 t3

t2

Chain jamoccurs

15

1

05

0

Time (s)

Velo

city

(ms

)

0 05 1 15 2 25 3 35 4 45 5 55 6

V0V1

(a)

Constant speed

Acceleration DecelerationΔT

Time (s)0 1 2 3 4 5 6

543210

ndash1ndash2ndash3ndash4ndash5

Acce

lera

tion

(ms

2 )

(b)

Figure 7 Transmission speed setting and vibration signal under normal condition

Table 3 Basic operating condition of the scraper conveyor

Model Chain size (mm) Conveyor width (mm) Chain speed (ms) Transport capacity (th) Transport length (m)SGZ12003600 φ48times152 1750times1180 0sim189 3700 360

Gear

Gear

Sprocket

Drive motor On-site PCWireless receiver

Transitionaltrough Chute

Accelerationsensor

Wirelessacquisition device

2

3

1

Figure 8 Experimental setup of the FPET

8 Shock and Vibration

time ranges are 010ndash025 and 050ndash065 s en the vi-bration energy of measuring point 2 is concentrated at thetime ranges 015ndash045 and 060ndash075 s and the time rangesof measuring point 3 are 040ndash050 and 070ndash085 s Forthe simulation signals the basic frequencies of the mea-suring points are 281 249 and 258Hz respectively emaximum values of vibration energies are 1527 1489 and1532m2s respectively Considering the experimental re-sults and the simulation results separately good agreementbetween the main parameters of different measuring pointsis obtained It also indicates that in the same time periodthe vibration signal has a delay characteristic for measuringpoints 1 2 and 3 is is possibly because that themeasuring points are at different positions of the chuteMoreover for different measuring points the experimentalsignals and the simulation signals show good consistency intime-frequency characteristics Hence the establishedDTSM can efficiently simulate the actual productionenvironment

43 Fault Detection of the Scraper Chain Based on theestablished DTSM two typical failure patterns of thescraper chain are discussed namely chain jam and chainfracture ree external load conditions are set that isempty load half-load and full-load conditions e valuesof Wi (Figure 5) are defined as 0 12W0 and W0 re-spectively As designed in Figure 5 when chain jam occursthe transmission speed Vi is set as V1 in Figure 7(a) and the

failure time range is t1 minus t3(t3 25 s) the chain fault istriggered at t1 2 sWithin the time range t1 minus t2(t2 225 s)the transmission speed Vi decreases from 1 to 0ms And thevalue of Vi increases from 0 to 1ms within the time ranget2 minus t3 e whole process lasts 05 s the chain assembly istightened and the scraper chains are jammed When chainfracture occurs the transmission speed Vi is set as V0 inFigure 7(a) Moreover as shown in Figure 5 the contactconstraint between two contacting scraper chains at the labeledcontact point 1 is removed As a result the two contactingscraper chains will be separated In order to ensure the accuracyof fault setting the chain fracture is also triggered at t1 2 sand the contact constraint is removed in 05 s which is con-sistent with chain jam

Considering both failure patterns of the scraper chainour research focuses on the steady phase of the operationprocess within the time range 05 to 55 s In the steadyphase the vibration signals are obtained simultaneously atmeasuring points 1 2 and 3 Taking measuring point1 as the case study the vibration signals for chain jam andchain fracture under empty load condition are presentedin Figures 11(a) and 11(b) respectively After faultstriggering the vibration signals of the two failure patternsshow a sudden increase after a short time delay Sub-sequently the signals exhibit unstable fluctuations at thetime range 24ndash28 s In order to obtain a more detaileddescription of the detection results the maximum am-plitudes of the vibration signals under different workingconditions are discussed ie normal condition chain jam

3210

ndash1ndash2ndash3

3210

ndash1ndash2ndash3

3210

ndash1ndash2ndash3

Time (s)

Acce

lera

tion

(ms

2 )

0 01 02 03 04 05 06 07 08 09

1 2 3

(a)

3210

ndash1ndash2ndash3

3210

ndash1ndash2ndash3

3210

ndash1ndash2ndash3

Time (s)Ac

cele

ratio

n (m

s2 )

0 01 02 03 04 05 06 07 08 09

123

(b)

Figure 9 Vibration signals of different measuring points in ΔT (a) experimental results (b) simulation results

Shock and Vibration 9

Point

1

2

3

(a) (b)

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

0

Time (s) Spectrum0 03 06 09

500

400

300

200

100

00 50 100 150

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

00 03 06 09

500

400

300

200

100

00 50 100 150

Time (s) Spectrum

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

00 03 06 09

500

400

300

200

100

00 50 100 150

Time (s) Spectrum

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

00 03 06 09

500

400

300

200

100

00 50 100 150

Time (s) Spectrum

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

0

Time (s) Spectrum0 03 06 09

500

400

300

200

100

00 50 100 150

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

00 03 06 09

500

400

300

200

100

00 50 100 150

Time (s) Spectrum

Figure 10 Time-frequency representation of vibration signals (a) experimental results (b) simulation results

Chain jamoccurs

Acce

lera

tion

(ms

2 ) Acce

lera

tion

(ms

2 )

10

5

0

ndash5

ndash10

Time (s)0 05 10 15 20 25 30 35 40 45 50

10

5

0

ndash5

ndash1024 25 26 27 28

(a)

Figure 11 Continued

10 Shock and Vibration

and chain fracture Wherein the empty load half-loadand full-load conditions are considered e maximumamplitudes of the vibration signals at measuring points1 2 and 3 are depicted in Figures 12(a)ndash12(c)respectively

Considering measuring point 1 the maximum am-plitudes of the vibration signals under empty load conditionare 289 718 and 698ms2 under normal condition chainjam and chain fracture respectively Similarly under half-load condition the maximum amplitudes are 324 937 and956ms2 Moreover under full-load condition the maxi-mum amplitudes are 446 1106 and 1082ms2 Withdifferent external loads the maximum amplitudes of thevibration signals for fault conditions are obviously higherthan those for normal condition and the difference betweenthe amplitudes of the two typical failure patterns is small Fordifferent fault conditions with the increase of the externalloads the maximum amplitudes show trends to increasee above statistical results are also applicable to measuringpoints 2 and 3 erefore chain faults can easily be de-tected by comparing the maximum amplitudes of the vi-bration signals whereas the fault patterns are difficult toidentify According to the nonstationary and nonlinearcharacteristics of fault signals the AOKR is utilized to an-alyze the vibration signals and classify failure patterns of thescraper chain Within 15 s after faults triggering the vi-bration signals at the three measuring points with differentexternal loads are processed Wherein for chain jam andchain fracture under empty load condition the time-fre-quency representations of vibration signals are presented inFigures 13(a) and 13(b) respectively

e frequency components and frequency ranges ofthe same fault pattern are similar for different measuringpoints As Figure 13(a) describes the bright color between0 and 50Hz indicates one high energy area caused bychain jam en chain fracture can easily be distinguishedaccording to the appearance of two high energy areasbetween 100 and 200Hz as shown in Figure 13(b)

Observing the spectrum results a more detailed de-scription is given When chain jam occurs for measuringpoints 1 2 and 3 the high energy areas occur ap-proximately at 05 075 and 09 s respectively Mean-while for chain fracture the high energy areas include twomain frequency components and are approximatelyconcentrated at the time ranges 050ndash070 075ndash085 and10ndash115 s respectively Hence there is a delay charac-teristic of the fault occurrence which is well in accordancewith the conclusions in Section 42 In order to explore theinfluence of external loads on fault characteristics thedetailed differences of the fault patterns at measuringpoint 2 are depicted in Figure 14 In fact the externalload has a great influence on the fault severity of both thefailure patterns Observing the spectrum results thebright areas vary with the external loads With increasingexternal load the frequency ranges of the high energyareas become larger Wherein for chain jam under emptyhalf- and full-load conditions the frequency ranges areapproximately 0ndash50 0ndash150 and 0ndash250 Hz respectivelyMeanwhile for chain fracture the frequency rangesare approximately 80ndash200 50ndash250 and 50ndash350Hzrespectively

In this part three working conditions of the scraperchain are investigated above including normal conditionchain jam and chain fracture e vibration signals ofmeasuring points 1 2 and 3 on the detecting chute areanalyzed and the effects of the external loads on the vi-bration characteristics are discussed Based on the aboveanalysis the occurrence of chain faults can easily be de-termined through amplitude comparisons of the originalvibration signals However the observation confirms thesimilarity of the time domain waveforms of fault signals forchain jam and chain fracture ese two patterns of failuresremain to be different through further processing by theAOKR and the fault patterns can be distinguished accordingto the number of high energy areas of the time-frequencyrepresentation of vibration signals In conclusion the

Chain fractureoccurs

Acce

lera

tion

(ms

2 ) Acce

lera

tion

(ms

2 )

10

5

0

ndash5

ndash10

Time (s)0 05 10 15 20 25 30 35 40 45 50

10

5

0

ndash5

ndash1024 25 26 27 28

(b)

Figure 11 Vibration signals of the measuring point 1 for (a) chain jam and (b) chain fracture

Shock and Vibration 11

Point 1 2 3

(a)

(b)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

Figure 13 Time-frequency representation of vibration signals under empty load (a) chain jam (b) chain fracture

12

10

8

6

4

2

0Max

imum

acce

lera

tion

ampl

itude

(ms

2 )

Normal conditionChain jamChain fracture

0 12 W0 W0

(a)

Normal conditionChain jamChain fracture

12

10

8

6

4

2

0Max

imum

acce

lera

tion

ampl

itude

(ms

2 )

0 12 W0 W0

(b)

Normal conditionChain jamChain fracture

12

10

8

6

4

2

0Max

imum

acce

lera

tion

ampl

itude

(ms

2 )

0 12 W0 W0

(c)

Figure 12 Statistical results of the maximum vibration amplitude at different measuring points (a) 1 (b) 2 (c) 3

12 Shock and Vibration

proposed detection strategy is effective at detecting theoccurrence of chain faults and identifying the failure pat-terns under different operating conditions

5 Conclusions

During the actual operation the working state of thescraper chain can reflect the dynamic performance of thescraper conveyor To address the difficulties with directsensor measurement for parameters of the moving scraperchain a novel strategy for fault detection of the scraperchain based on vibration analysis of the chute was pro-posed Based on modal analysis and the MAC the mea-suring points of vibration signals on the chute weredetermined To fit the actual behavior of the transmissionprocess the DTSM was presented based on finite elementmodeling and the correctness of the dynamic model wasverified by comparison with the FPET en the vibrationproperties of the measuring points on the chute undernormal condition chain jam and chain fracture werediscussed Moreover the occurrence of chain faults weredetermined by comparing the amplitudes of the vibrationsignal in the time domain while the AOKR was utilizedfor time-frequency representation of vibration signals anddistinguishing the two typical failure patterns Further-more the strategy verification based on experimental datawill be taken into consideration in the near future

Data Availability

e data used to support the findings of this study areavailable from the corresponding author upon request

Conflicts of Interest

e authors declare that they have no conflicts of interest

Acknowledgments

is work was supported by the Key Project of NationalNatural Science Foundation of China (U1510205) NaturalScience Foundation of Jiangsu Province (No BK20160251)Xuzhou Research program (KC14H0138) FundamentalResearch Funds for the Central Universities (2014Y05) andProject Funded by the Priority Academic Program De-velopment of Jiangsu Higher Education Institutions(PAPD)

References

[1] C D Brown ldquoDesign build and test of a longwall armouredface conveyorrdquo Longwall Mining 2002

[2] M Dolipski P Cheluszka E Remiorz and P SobotaldquoFollow-up chain tension in an armoured face conveyornadazne napinanie lancucha zgrzebłowego W przenosnikuscianowymrdquo Archives of Mining Sciences vol 60 no 1pp 25ndash38 2015

[3] L A Morley J L Kohler and H M Smolnikar ldquoA model forpredicting motor load for an armored face-conveyor driverdquoIEEE Transactions on Industry Applications vol 24 no 4pp 649ndash659 1988

[4] A A Ordin and A A Metelrsquokov ldquoAnalysis of longwall faceoutput in screw-type cutter-loader-and-scraper conveyorsystem in underground mining of flat-lying coal bedsrdquoJournal of Mining Science vol 51 no 6 pp 1173ndash1179 2015

[5] B He G Li H Shi et al ldquoDynamic behaviour modelling andsimulation of the chain transmission system for an armouredface conveyorrdquo in Proceedings of the IEEE 10th InternationalConference on Computer-Aided Industrial Design and Con-ceptual Design CAID amp CD 2009 pp 1000ndash1004 BeijingChina November 2009

[6] R Nie B He P Yuan L Zhang and G Li ldquoNovel approachto and implementation of design and analysis of armored faceconveyor power trainrdquo Science China Technological Sciencesvol 58 no 12 pp 2153ndash2168 2015

Loads

(a)

(b)

Empty load Half-load Full-load

500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)

Figure 14 Time-frequency representation of vibration signals at measuring point 2 with different external loads (a) chain jam (b) chainfracture

Shock and Vibration 13

[7] R Nie B He L Zhang and G Li ldquoModelling of thetransmission system in conveying equipment based on Eulermethod with applicationrdquo Proceedings of the Institution ofMechanical Engineers Part K Journal of Multi-body Dy-namics vol 228 no 3 pp 294ndash306 2014

[8] S B Jiang X Zhang K D Gao J Gao Q Y Wang andK Hidenori ldquoMulti-body dynamics and vibration analysis ofchain assembly in armoured face conveyorrdquo InternationalJournal of Simulation Modelling vol 16 no 3 pp 458ndash4702017

[9] M Myszkowski and D Loehning ldquoChain force measure-ments on armoured face conveyors and coal plows in heavy-duty longwallsrdquo CIM Bulletin vol 94 no 1054 pp 72ndash752001

[10] H Wang Q Zhang and F Xie ldquoDynamic tension test andintelligent coordinated control system of a heavy scraperconveyorrdquo IET Science Measurement and Technology vol 11no 7 pp 871ndash877 2017

[11] S Sen M X Min and Y Z She ldquoDiagnosis of coal scraperconveyor based on Fuzzy Fault treerdquo in Proceedings of the2015 Seventh International Conference on Measuring Tech-nology and Mechatronics Automation (ICMTMA) pp 392ndash395 IEEE Nanchang China June 2015

[12] S-s Xue X-c Li and X-y Xu ldquoFault tree and Bayesiannetwork based scraper conveyer fault diagnosisrdquo in Pro-ceedings of the 22nd International Conference on IndustrialEngineering and Engineering Management 2015 pp 783ndash795Atlantis Press Paris France January 2016

[13] X Gong X Ma Y Zhang et al ldquoApplication of fuzzy neuralnetwork in fault diagnosis for scraper conveyor vibrationrdquo inProceedings of the 2013 IEEE International Conference onInformation and Automation (ICIA) pp 1135ndash1140 IEEEYinchuan China August 2013

[14] Y Zhang X Ma Y Jianxiang et al ldquoFuzzy neural networkfault diagnosis and online vibration monitoring system for thecoal scraper conveyor based on rough set theoryrdquo in Pro-ceedings of the 2013 32nd Chinese Control Conference (CCC)pp 6134ndash6138 IEEE Xirsquoan China July 2013

[15] B Zhang A C C Tan and J-h Lin ldquoGearbox fault diagnosisof high-speed railway trainrdquo Engineering Failure Analysisvol 66 pp 407ndash420 2016

[16] E Parloo P Verboven P Guillaume and M Van OvermeireldquoAutonomous structural health monitoring-part ii vibration-based in-operation damage assessmentrdquo Mechanical Systemsand Signal Processing vol 16 no 4 pp 659ndash675 2002

[17] C S Sakaris J S Sakellariou and S D Fassois ldquoRandom-vibration-based damage detection and precise localization ona lab-scale aircraft stabilizer structure via the GeneralizedFunctional Model Based Methodrdquo Structural Health Moni-toring An International Journal vol 16 no 5 pp 594ndash6102017

[18] Y Zhang W Song M Karimi C-H Chi and A KudreykoldquoFractional autoregressive integrated moving average andfinite-element modal the forecast of tire vibration trendrdquoIEEE Access vol 6 pp 40137ndash40142 2018

[19] M Pastor M Binda and T Harcarik ldquoModal assurancecriterionrdquo Procedia Engineering vol 48 pp 543ndash548 2012

[20] W J Staszewski K Worden and G R Tomlinson ldquoTime-frequency analysis in gearbox fault detection using the Wigner-ville distribution and pattern recognitionrdquo Mechanical Systemsand Signal Processing vol 11 no 5 pp 673ndash692 1997

[21] J-D Wu and P-H Chiang ldquoApplication of Wigner-Villedistribution and probability neural network for scooter engine

fault diagnosisrdquo Expert Systems with Applications vol 36no 2 pp 2187ndash2199 2009

[22] Z Feng and M Liang ldquoFault diagnosis of wind turbineplanetary gearbox under nonstationary conditions viaadaptive optimal kernel time-frequency analysisrdquo RenewableEnergy vol 66 pp 468ndash477 2014

14 Shock and Vibration

International Journal of

AerospaceEngineeringHindawiwwwhindawicom Volume 2018

RoboticsJournal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Active and Passive Electronic Components

VLSI Design

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Shock and Vibration

Hindawiwwwhindawicom Volume 2018

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawiwwwhindawicom

Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Control Scienceand Engineering

Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom

Journal ofEngineeringVolume 2018

SensorsJournal of

Hindawiwwwhindawicom Volume 2018

International Journal of

RotatingMachinery

Hindawiwwwhindawicom Volume 2018

Modelling ampSimulationin EngineeringHindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Navigation and Observation

International Journal of

Hindawi

wwwhindawicom Volume 2018

Advances in

Multimedia

Submit your manuscripts atwwwhindawicom

X

Y

ZI

Vertical chain

O

Scraper

System frameChute

Plat chain

Contact Contact

Fixed joint

Contact Fixed jo

int

(a)

Scraper

Scraper chainsAcceleration

sensor

Chute

Exciting point

Impact hammer

Shovelcoal board

(b)

Figure 2 Descriptions of modal analysis based on (a) the CVM and (b) the HIT

350

250300

200150100

500

ndash50

Time (s)

Exci

ting

forc

e (N

)

0 05 1

(a)

0 05 1

15

1

05

0

ndash05

ndash1

Acce

lera

tion

(ms

2 )

Time (s)

(b)

Figure 3 Exciting force and the vibration response

Mode 1

(1729 0612)(2114 04954)

(2578 08114) (3096 08938)

(314 04933)

(3511 01923)

Acce

lera

tion

(ms

2 ) Mode 2

Mode 3

Mode 4

Mode 5

Mode 6

0

05

1

15

2

Frequency (Hz)0 50 100 150 200 250 300 350 400 450 500

Experimental modal

Figure 4 Vibration modes of the chute

Shock and Vibration 5

chute without carrying out physical tests e simulationparameters and kinematic restriction of the DTSM areconsistent with the CVM in Figure 2(a) To illustrate thisissue in more detail the frictional contacts are set betweenthe vertical chains and the chutes and between the scrapersand chutes Correspondingly the bonded contacts are setbetween the plat chains and the vertical chains e dynamicmodel mainly contains two prominent parts the chain as-sembly and the chutes Wherein the chain assembly iscomposed of the scrapers and scraper chains and can beequally divided into multiple segments of length ΔL thenumber of the chutes is 7 and they are marked as shown inFigure 5 A translational joint is set between scraper 1 andthe middle plate based on which the chain assembly can runalong the chutes at a transmission speed of Vi Moreover thepretightening force and external load of the dynamic modelare defined as F1 and Wi respectively Here we setF1 79612 kN In Figure 5 the pretightening force is ap-plied at the two ends of the chain assembly which can ensurethat the scraper chain remains tight during the movemente external load Wi is applied on the upper surface of thescrapers and middle plates and the material density inthe chute under the full-load condition is W0 whereW0 694 kgm

4 AOKR-Based Fault Detection Strategy of theScraper Chain

Considering that the scraper conveyor is susceptible tofrequent loading excessive bending and artificial mis-conduct in actual engineering different patterns of chainfaults usually act on the scraper chain e failure patternswill cause abnormal vibration of the chute Actuallydirect measurement and analysis for parameters of themoving scraper chain is difficult which raises a contra-diction between practical engineering requirements andthe installation limitations of multiple sensors In order todetect chain faults promptly the fix-point vibrationmeasurement of the measuring points for the chute isconsidered instead of mobile parameter measurement forthe scraper chain In this section the DTSM is utilized toacquire the vibration properties of the actual chute undervarious working conditions and the FPETof the vibrationsignals is conducted to validate the accuracy of the dy-namic model e fundamental idea of our proposedstrategy is to determine the occurrence of chain faultsby amplitude comparisons and then fault signals areanalyzed through the AOKR to distinguish the faultpatterns

Table 1 Frequency response of vibration signals

Mode no FEA (Hz) EMA (Hz) ψi ()

1 17754 1729 2682 18329 2114 13303 28745 2578 11504 30242 3096 2325 32868 3140 4686 34197 3511 260

Position 1 Position 2

F1

F1

F1

F1

Translationaljoint

Scraper 1

Chain breakage

Chain jamContact point 1

∆L

Middle plateChain assembly

Wi

1 3 4 5 7

Vi

Vi

Figure 5 Simulation process of the DTSM

6 Shock and Vibration

41 Optimal Placement of the Acceleration Sensors Beforeconducting the FPET the installation scheme of the ac-celeration sensors should be determined to ensure thevalidity of the measured vibration signals and the economyof the experimental process Based on modal analysis of theFEA (Figure 2(a)) we extract 20 nodes from the chute as theprimary measuring points which are labeled in Figure 6(a)According to the theory of MAC the total modal dis-placements of the selected 20 points serve as the inputs ofequations (6) and (7) In Figure 6(b) the curve revealschange rules of the minimumMACij for different number ofsensors

To illustrate the optimal placement scheme of acceler-ation sensors in more detail the optimum installation po-sitions for different number of sensors are shown in Table 2Accordingly the vibration response of the actual chuteshould be recorded by 3 acceleration sensors and the op-timum installation positions are set at nodes 3 9 and 16respectively

42 Experimental Evaluations According to Section 32 thesimulation process of the DTSM is performed under normalcondition without a load and the transmission speed Vi isset as V0 in Figure 7(a) and the scraper 1 moves fromposition 1 to position 2 In our study nodes 9 16 and 3 ofthe fourth chute are chosen as the detecting points for vi-bration analysis For convenient expression in the followingwe mark the detecting nodes as measuring points 1 2 and3 respectively Taking measuring point 1 as the researchexample for description of the transmission process theoriginal signal is presented in Figure 7(b) e vibrationresponse presents a three-stage change ie the accelerationphase the steady phase and the deceleration phase Asmentioned above in the steady phase the vibration signalcan be treated as multiple segments with a time span of ΔTwhich corresponds to the chain assembly with a length of ΔL(Figure 5)

e field FPET of multiple monitoring points iscarried out to obtain the vibration properties of the actualchute and evaluate the dynamic performance of theDTSM e specifications of the experimental scraperconveyor correspond to SGB12003600 manufactured byLianyungang Tianming Equipment Co Ltd and therunning speed of the scraper chain is 10 ms In practicalengineering the actual double-drive transmission systemis driven by two drive motors e basic performanceparameters are shown in Table 3 As shown in Figure 8the experimental system mainly contains three acceler-ation sensors (TST120A1000) a wireless acquisitiondevice (TST5925EV) a wireless receiver and an on-sitePC e acceleration sensors are used for fixed-pointmeasurements by detecting the vibration responses ofmeasuring points 1 2 and 3 on the chute e wirelessacquisition device is used to collect experimental data inreal time In addition the wireless receiver is intended forremote data transmissione vibration data are stored inthe on-site PC and a sampling frequency of 1000 Hz istaken

Figure 9 shows the experimental and simulation signalsof measuring points 1 2 and 3 in the time span ΔT Inthe time domain the simulation results and the experi-mental results are similar which reflects the adaptability ofthe DTSM Considering the nonlinear and time-varyingcharacteristics of the vibration signals the AOKR is used fortime-frequency analysis

e vibration signal of measuring point 2 in the timespan ΔT is taken as the reference for time-frequency

1 2 3 4 5 6 7 8 9 10

11 12 13 14 15 16 17 18 19 20

(a)

Number of the sensors

Min

imum

val

ue o

f MAC

ij

2 3 4 5 6 7 8 9 10

028

0275

027

0265

026

0255

025

0245

(b)

Figure 6 Description of the optimum placement scheme (a) theprimary measuring points (b) the minimum MACij

Table 2 Measuring points placement based on the MAC

Number f Measuring points2 02515 8 163 02453 3 9 164 02454 1 3 9 165 02460 1 3 9 16 206 02483 1 3 9 10 11 167 02525 1 2 3 9 10 11 168 02644 1 3 8 9 10 11 15 169 02664 1 2 3 8 9 10 11 15 1610 02767 1 2 3 8 9 10 11 15 16 20

Shock and Vibration 7

analysis and then the time-frequency representations ofmeasuring points 1 and 3 in the same time period aredemonstrated Figure 10(a) shows the time-frequencyrepresentation of the experimental signals for differentmeasuring points and the power spectrum is alsodepicted It presents that the vibration signal of the chutehas a plurality of components ere are two major areaswith strong frequency responses of the vibration signalsfor different measuring points for measuring point 1 thefrequency range is 230ndash310Hz and the time ranges are015ndash030 s and 045ndash075 s for measuring point 2 thefrequency range is 230ndash340Hz and the time ranges are020ndash040 s and 055ndash070 s and for measuring point 3the frequency range is 225ndash330Hz and the time ranges are040ndash050 s and 070ndash090 s Considering the chain

assembly in Figure 5 there are two scrapers for length ΔLIn the actual transmission process the two major areaswith strong frequency responses are caused while the twoscrapers passing through the measuring points As shownbasic frequencies of the measuring points are 239 278 and249Hz respectively In addition the peak of the vibrationpower spectrum is also presented and the maximumvalues of the measuring points are 1588 1561 and1499m2s

Similarly the time-frequency representation of thesimulation signals with two major frequency responses isdiscussed As Figure 10(b) indicates for measuring points1 2 and 3 the vibration energies are concentrated atthe frequency ranges 240ndash340 200ndash310 and 200ndash320Hzrespectively Correspondingly for measuring point 1 the

t1 t3

t2

Chain jamoccurs

15

1

05

0

Time (s)

Velo

city

(ms

)

0 05 1 15 2 25 3 35 4 45 5 55 6

V0V1

(a)

Constant speed

Acceleration DecelerationΔT

Time (s)0 1 2 3 4 5 6

543210

ndash1ndash2ndash3ndash4ndash5

Acce

lera

tion

(ms

2 )

(b)

Figure 7 Transmission speed setting and vibration signal under normal condition

Table 3 Basic operating condition of the scraper conveyor

Model Chain size (mm) Conveyor width (mm) Chain speed (ms) Transport capacity (th) Transport length (m)SGZ12003600 φ48times152 1750times1180 0sim189 3700 360

Gear

Gear

Sprocket

Drive motor On-site PCWireless receiver

Transitionaltrough Chute

Accelerationsensor

Wirelessacquisition device

2

3

1

Figure 8 Experimental setup of the FPET

8 Shock and Vibration

time ranges are 010ndash025 and 050ndash065 s en the vi-bration energy of measuring point 2 is concentrated at thetime ranges 015ndash045 and 060ndash075 s and the time rangesof measuring point 3 are 040ndash050 and 070ndash085 s Forthe simulation signals the basic frequencies of the mea-suring points are 281 249 and 258Hz respectively emaximum values of vibration energies are 1527 1489 and1532m2s respectively Considering the experimental re-sults and the simulation results separately good agreementbetween the main parameters of different measuring pointsis obtained It also indicates that in the same time periodthe vibration signal has a delay characteristic for measuringpoints 1 2 and 3 is is possibly because that themeasuring points are at different positions of the chuteMoreover for different measuring points the experimentalsignals and the simulation signals show good consistency intime-frequency characteristics Hence the establishedDTSM can efficiently simulate the actual productionenvironment

43 Fault Detection of the Scraper Chain Based on theestablished DTSM two typical failure patterns of thescraper chain are discussed namely chain jam and chainfracture ree external load conditions are set that isempty load half-load and full-load conditions e valuesof Wi (Figure 5) are defined as 0 12W0 and W0 re-spectively As designed in Figure 5 when chain jam occursthe transmission speed Vi is set as V1 in Figure 7(a) and the

failure time range is t1 minus t3(t3 25 s) the chain fault istriggered at t1 2 sWithin the time range t1 minus t2(t2 225 s)the transmission speed Vi decreases from 1 to 0ms And thevalue of Vi increases from 0 to 1ms within the time ranget2 minus t3 e whole process lasts 05 s the chain assembly istightened and the scraper chains are jammed When chainfracture occurs the transmission speed Vi is set as V0 inFigure 7(a) Moreover as shown in Figure 5 the contactconstraint between two contacting scraper chains at the labeledcontact point 1 is removed As a result the two contactingscraper chains will be separated In order to ensure the accuracyof fault setting the chain fracture is also triggered at t1 2 sand the contact constraint is removed in 05 s which is con-sistent with chain jam

Considering both failure patterns of the scraper chainour research focuses on the steady phase of the operationprocess within the time range 05 to 55 s In the steadyphase the vibration signals are obtained simultaneously atmeasuring points 1 2 and 3 Taking measuring point1 as the case study the vibration signals for chain jam andchain fracture under empty load condition are presentedin Figures 11(a) and 11(b) respectively After faultstriggering the vibration signals of the two failure patternsshow a sudden increase after a short time delay Sub-sequently the signals exhibit unstable fluctuations at thetime range 24ndash28 s In order to obtain a more detaileddescription of the detection results the maximum am-plitudes of the vibration signals under different workingconditions are discussed ie normal condition chain jam

3210

ndash1ndash2ndash3

3210

ndash1ndash2ndash3

3210

ndash1ndash2ndash3

Time (s)

Acce

lera

tion

(ms

2 )

0 01 02 03 04 05 06 07 08 09

1 2 3

(a)

3210

ndash1ndash2ndash3

3210

ndash1ndash2ndash3

3210

ndash1ndash2ndash3

Time (s)Ac

cele

ratio

n (m

s2 )

0 01 02 03 04 05 06 07 08 09

123

(b)

Figure 9 Vibration signals of different measuring points in ΔT (a) experimental results (b) simulation results

Shock and Vibration 9

Point

1

2

3

(a) (b)

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

0

Time (s) Spectrum0 03 06 09

500

400

300

200

100

00 50 100 150

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

00 03 06 09

500

400

300

200

100

00 50 100 150

Time (s) Spectrum

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

00 03 06 09

500

400

300

200

100

00 50 100 150

Time (s) Spectrum

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

00 03 06 09

500

400

300

200

100

00 50 100 150

Time (s) Spectrum

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

0

Time (s) Spectrum0 03 06 09

500

400

300

200

100

00 50 100 150

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

00 03 06 09

500

400

300

200

100

00 50 100 150

Time (s) Spectrum

Figure 10 Time-frequency representation of vibration signals (a) experimental results (b) simulation results

Chain jamoccurs

Acce

lera

tion

(ms

2 ) Acce

lera

tion

(ms

2 )

10

5

0

ndash5

ndash10

Time (s)0 05 10 15 20 25 30 35 40 45 50

10

5

0

ndash5

ndash1024 25 26 27 28

(a)

Figure 11 Continued

10 Shock and Vibration

and chain fracture Wherein the empty load half-loadand full-load conditions are considered e maximumamplitudes of the vibration signals at measuring points1 2 and 3 are depicted in Figures 12(a)ndash12(c)respectively

Considering measuring point 1 the maximum am-plitudes of the vibration signals under empty load conditionare 289 718 and 698ms2 under normal condition chainjam and chain fracture respectively Similarly under half-load condition the maximum amplitudes are 324 937 and956ms2 Moreover under full-load condition the maxi-mum amplitudes are 446 1106 and 1082ms2 Withdifferent external loads the maximum amplitudes of thevibration signals for fault conditions are obviously higherthan those for normal condition and the difference betweenthe amplitudes of the two typical failure patterns is small Fordifferent fault conditions with the increase of the externalloads the maximum amplitudes show trends to increasee above statistical results are also applicable to measuringpoints 2 and 3 erefore chain faults can easily be de-tected by comparing the maximum amplitudes of the vi-bration signals whereas the fault patterns are difficult toidentify According to the nonstationary and nonlinearcharacteristics of fault signals the AOKR is utilized to an-alyze the vibration signals and classify failure patterns of thescraper chain Within 15 s after faults triggering the vi-bration signals at the three measuring points with differentexternal loads are processed Wherein for chain jam andchain fracture under empty load condition the time-fre-quency representations of vibration signals are presented inFigures 13(a) and 13(b) respectively

e frequency components and frequency ranges ofthe same fault pattern are similar for different measuringpoints As Figure 13(a) describes the bright color between0 and 50Hz indicates one high energy area caused bychain jam en chain fracture can easily be distinguishedaccording to the appearance of two high energy areasbetween 100 and 200Hz as shown in Figure 13(b)

Observing the spectrum results a more detailed de-scription is given When chain jam occurs for measuringpoints 1 2 and 3 the high energy areas occur ap-proximately at 05 075 and 09 s respectively Mean-while for chain fracture the high energy areas include twomain frequency components and are approximatelyconcentrated at the time ranges 050ndash070 075ndash085 and10ndash115 s respectively Hence there is a delay charac-teristic of the fault occurrence which is well in accordancewith the conclusions in Section 42 In order to explore theinfluence of external loads on fault characteristics thedetailed differences of the fault patterns at measuringpoint 2 are depicted in Figure 14 In fact the externalload has a great influence on the fault severity of both thefailure patterns Observing the spectrum results thebright areas vary with the external loads With increasingexternal load the frequency ranges of the high energyareas become larger Wherein for chain jam under emptyhalf- and full-load conditions the frequency ranges areapproximately 0ndash50 0ndash150 and 0ndash250 Hz respectivelyMeanwhile for chain fracture the frequency rangesare approximately 80ndash200 50ndash250 and 50ndash350Hzrespectively

In this part three working conditions of the scraperchain are investigated above including normal conditionchain jam and chain fracture e vibration signals ofmeasuring points 1 2 and 3 on the detecting chute areanalyzed and the effects of the external loads on the vi-bration characteristics are discussed Based on the aboveanalysis the occurrence of chain faults can easily be de-termined through amplitude comparisons of the originalvibration signals However the observation confirms thesimilarity of the time domain waveforms of fault signals forchain jam and chain fracture ese two patterns of failuresremain to be different through further processing by theAOKR and the fault patterns can be distinguished accordingto the number of high energy areas of the time-frequencyrepresentation of vibration signals In conclusion the

Chain fractureoccurs

Acce

lera

tion

(ms

2 ) Acce

lera

tion

(ms

2 )

10

5

0

ndash5

ndash10

Time (s)0 05 10 15 20 25 30 35 40 45 50

10

5

0

ndash5

ndash1024 25 26 27 28

(b)

Figure 11 Vibration signals of the measuring point 1 for (a) chain jam and (b) chain fracture

Shock and Vibration 11

Point 1 2 3

(a)

(b)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

Figure 13 Time-frequency representation of vibration signals under empty load (a) chain jam (b) chain fracture

12

10

8

6

4

2

0Max

imum

acce

lera

tion

ampl

itude

(ms

2 )

Normal conditionChain jamChain fracture

0 12 W0 W0

(a)

Normal conditionChain jamChain fracture

12

10

8

6

4

2

0Max

imum

acce

lera

tion

ampl

itude

(ms

2 )

0 12 W0 W0

(b)

Normal conditionChain jamChain fracture

12

10

8

6

4

2

0Max

imum

acce

lera

tion

ampl

itude

(ms

2 )

0 12 W0 W0

(c)

Figure 12 Statistical results of the maximum vibration amplitude at different measuring points (a) 1 (b) 2 (c) 3

12 Shock and Vibration

proposed detection strategy is effective at detecting theoccurrence of chain faults and identifying the failure pat-terns under different operating conditions

5 Conclusions

During the actual operation the working state of thescraper chain can reflect the dynamic performance of thescraper conveyor To address the difficulties with directsensor measurement for parameters of the moving scraperchain a novel strategy for fault detection of the scraperchain based on vibration analysis of the chute was pro-posed Based on modal analysis and the MAC the mea-suring points of vibration signals on the chute weredetermined To fit the actual behavior of the transmissionprocess the DTSM was presented based on finite elementmodeling and the correctness of the dynamic model wasverified by comparison with the FPET en the vibrationproperties of the measuring points on the chute undernormal condition chain jam and chain fracture werediscussed Moreover the occurrence of chain faults weredetermined by comparing the amplitudes of the vibrationsignal in the time domain while the AOKR was utilizedfor time-frequency representation of vibration signals anddistinguishing the two typical failure patterns Further-more the strategy verification based on experimental datawill be taken into consideration in the near future

Data Availability

e data used to support the findings of this study areavailable from the corresponding author upon request

Conflicts of Interest

e authors declare that they have no conflicts of interest

Acknowledgments

is work was supported by the Key Project of NationalNatural Science Foundation of China (U1510205) NaturalScience Foundation of Jiangsu Province (No BK20160251)Xuzhou Research program (KC14H0138) FundamentalResearch Funds for the Central Universities (2014Y05) andProject Funded by the Priority Academic Program De-velopment of Jiangsu Higher Education Institutions(PAPD)

References

[1] C D Brown ldquoDesign build and test of a longwall armouredface conveyorrdquo Longwall Mining 2002

[2] M Dolipski P Cheluszka E Remiorz and P SobotaldquoFollow-up chain tension in an armoured face conveyornadazne napinanie lancucha zgrzebłowego W przenosnikuscianowymrdquo Archives of Mining Sciences vol 60 no 1pp 25ndash38 2015

[3] L A Morley J L Kohler and H M Smolnikar ldquoA model forpredicting motor load for an armored face-conveyor driverdquoIEEE Transactions on Industry Applications vol 24 no 4pp 649ndash659 1988

[4] A A Ordin and A A Metelrsquokov ldquoAnalysis of longwall faceoutput in screw-type cutter-loader-and-scraper conveyorsystem in underground mining of flat-lying coal bedsrdquoJournal of Mining Science vol 51 no 6 pp 1173ndash1179 2015

[5] B He G Li H Shi et al ldquoDynamic behaviour modelling andsimulation of the chain transmission system for an armouredface conveyorrdquo in Proceedings of the IEEE 10th InternationalConference on Computer-Aided Industrial Design and Con-ceptual Design CAID amp CD 2009 pp 1000ndash1004 BeijingChina November 2009

[6] R Nie B He P Yuan L Zhang and G Li ldquoNovel approachto and implementation of design and analysis of armored faceconveyor power trainrdquo Science China Technological Sciencesvol 58 no 12 pp 2153ndash2168 2015

Loads

(a)

(b)

Empty load Half-load Full-load

500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)

Figure 14 Time-frequency representation of vibration signals at measuring point 2 with different external loads (a) chain jam (b) chainfracture

Shock and Vibration 13

[7] R Nie B He L Zhang and G Li ldquoModelling of thetransmission system in conveying equipment based on Eulermethod with applicationrdquo Proceedings of the Institution ofMechanical Engineers Part K Journal of Multi-body Dy-namics vol 228 no 3 pp 294ndash306 2014

[8] S B Jiang X Zhang K D Gao J Gao Q Y Wang andK Hidenori ldquoMulti-body dynamics and vibration analysis ofchain assembly in armoured face conveyorrdquo InternationalJournal of Simulation Modelling vol 16 no 3 pp 458ndash4702017

[9] M Myszkowski and D Loehning ldquoChain force measure-ments on armoured face conveyors and coal plows in heavy-duty longwallsrdquo CIM Bulletin vol 94 no 1054 pp 72ndash752001

[10] H Wang Q Zhang and F Xie ldquoDynamic tension test andintelligent coordinated control system of a heavy scraperconveyorrdquo IET Science Measurement and Technology vol 11no 7 pp 871ndash877 2017

[11] S Sen M X Min and Y Z She ldquoDiagnosis of coal scraperconveyor based on Fuzzy Fault treerdquo in Proceedings of the2015 Seventh International Conference on Measuring Tech-nology and Mechatronics Automation (ICMTMA) pp 392ndash395 IEEE Nanchang China June 2015

[12] S-s Xue X-c Li and X-y Xu ldquoFault tree and Bayesiannetwork based scraper conveyer fault diagnosisrdquo in Pro-ceedings of the 22nd International Conference on IndustrialEngineering and Engineering Management 2015 pp 783ndash795Atlantis Press Paris France January 2016

[13] X Gong X Ma Y Zhang et al ldquoApplication of fuzzy neuralnetwork in fault diagnosis for scraper conveyor vibrationrdquo inProceedings of the 2013 IEEE International Conference onInformation and Automation (ICIA) pp 1135ndash1140 IEEEYinchuan China August 2013

[14] Y Zhang X Ma Y Jianxiang et al ldquoFuzzy neural networkfault diagnosis and online vibration monitoring system for thecoal scraper conveyor based on rough set theoryrdquo in Pro-ceedings of the 2013 32nd Chinese Control Conference (CCC)pp 6134ndash6138 IEEE Xirsquoan China July 2013

[15] B Zhang A C C Tan and J-h Lin ldquoGearbox fault diagnosisof high-speed railway trainrdquo Engineering Failure Analysisvol 66 pp 407ndash420 2016

[16] E Parloo P Verboven P Guillaume and M Van OvermeireldquoAutonomous structural health monitoring-part ii vibration-based in-operation damage assessmentrdquo Mechanical Systemsand Signal Processing vol 16 no 4 pp 659ndash675 2002

[17] C S Sakaris J S Sakellariou and S D Fassois ldquoRandom-vibration-based damage detection and precise localization ona lab-scale aircraft stabilizer structure via the GeneralizedFunctional Model Based Methodrdquo Structural Health Moni-toring An International Journal vol 16 no 5 pp 594ndash6102017

[18] Y Zhang W Song M Karimi C-H Chi and A KudreykoldquoFractional autoregressive integrated moving average andfinite-element modal the forecast of tire vibration trendrdquoIEEE Access vol 6 pp 40137ndash40142 2018

[19] M Pastor M Binda and T Harcarik ldquoModal assurancecriterionrdquo Procedia Engineering vol 48 pp 543ndash548 2012

[20] W J Staszewski K Worden and G R Tomlinson ldquoTime-frequency analysis in gearbox fault detection using the Wigner-ville distribution and pattern recognitionrdquo Mechanical Systemsand Signal Processing vol 11 no 5 pp 673ndash692 1997

[21] J-D Wu and P-H Chiang ldquoApplication of Wigner-Villedistribution and probability neural network for scooter engine

fault diagnosisrdquo Expert Systems with Applications vol 36no 2 pp 2187ndash2199 2009

[22] Z Feng and M Liang ldquoFault diagnosis of wind turbineplanetary gearbox under nonstationary conditions viaadaptive optimal kernel time-frequency analysisrdquo RenewableEnergy vol 66 pp 468ndash477 2014

14 Shock and Vibration

International Journal of

AerospaceEngineeringHindawiwwwhindawicom Volume 2018

RoboticsJournal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Active and Passive Electronic Components

VLSI Design

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Shock and Vibration

Hindawiwwwhindawicom Volume 2018

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawiwwwhindawicom

Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Control Scienceand Engineering

Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom

Journal ofEngineeringVolume 2018

SensorsJournal of

Hindawiwwwhindawicom Volume 2018

International Journal of

RotatingMachinery

Hindawiwwwhindawicom Volume 2018

Modelling ampSimulationin EngineeringHindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Navigation and Observation

International Journal of

Hindawi

wwwhindawicom Volume 2018

Advances in

Multimedia

Submit your manuscripts atwwwhindawicom

chute without carrying out physical tests e simulationparameters and kinematic restriction of the DTSM areconsistent with the CVM in Figure 2(a) To illustrate thisissue in more detail the frictional contacts are set betweenthe vertical chains and the chutes and between the scrapersand chutes Correspondingly the bonded contacts are setbetween the plat chains and the vertical chains e dynamicmodel mainly contains two prominent parts the chain as-sembly and the chutes Wherein the chain assembly iscomposed of the scrapers and scraper chains and can beequally divided into multiple segments of length ΔL thenumber of the chutes is 7 and they are marked as shown inFigure 5 A translational joint is set between scraper 1 andthe middle plate based on which the chain assembly can runalong the chutes at a transmission speed of Vi Moreover thepretightening force and external load of the dynamic modelare defined as F1 and Wi respectively Here we setF1 79612 kN In Figure 5 the pretightening force is ap-plied at the two ends of the chain assembly which can ensurethat the scraper chain remains tight during the movemente external load Wi is applied on the upper surface of thescrapers and middle plates and the material density inthe chute under the full-load condition is W0 whereW0 694 kgm

4 AOKR-Based Fault Detection Strategy of theScraper Chain

Considering that the scraper conveyor is susceptible tofrequent loading excessive bending and artificial mis-conduct in actual engineering different patterns of chainfaults usually act on the scraper chain e failure patternswill cause abnormal vibration of the chute Actuallydirect measurement and analysis for parameters of themoving scraper chain is difficult which raises a contra-diction between practical engineering requirements andthe installation limitations of multiple sensors In order todetect chain faults promptly the fix-point vibrationmeasurement of the measuring points for the chute isconsidered instead of mobile parameter measurement forthe scraper chain In this section the DTSM is utilized toacquire the vibration properties of the actual chute undervarious working conditions and the FPETof the vibrationsignals is conducted to validate the accuracy of the dy-namic model e fundamental idea of our proposedstrategy is to determine the occurrence of chain faultsby amplitude comparisons and then fault signals areanalyzed through the AOKR to distinguish the faultpatterns

Table 1 Frequency response of vibration signals

Mode no FEA (Hz) EMA (Hz) ψi ()

1 17754 1729 2682 18329 2114 13303 28745 2578 11504 30242 3096 2325 32868 3140 4686 34197 3511 260

Position 1 Position 2

F1

F1

F1

F1

Translationaljoint

Scraper 1

Chain breakage

Chain jamContact point 1

∆L

Middle plateChain assembly

Wi

1 3 4 5 7

Vi

Vi

Figure 5 Simulation process of the DTSM

6 Shock and Vibration

41 Optimal Placement of the Acceleration Sensors Beforeconducting the FPET the installation scheme of the ac-celeration sensors should be determined to ensure thevalidity of the measured vibration signals and the economyof the experimental process Based on modal analysis of theFEA (Figure 2(a)) we extract 20 nodes from the chute as theprimary measuring points which are labeled in Figure 6(a)According to the theory of MAC the total modal dis-placements of the selected 20 points serve as the inputs ofequations (6) and (7) In Figure 6(b) the curve revealschange rules of the minimumMACij for different number ofsensors

To illustrate the optimal placement scheme of acceler-ation sensors in more detail the optimum installation po-sitions for different number of sensors are shown in Table 2Accordingly the vibration response of the actual chuteshould be recorded by 3 acceleration sensors and the op-timum installation positions are set at nodes 3 9 and 16respectively

42 Experimental Evaluations According to Section 32 thesimulation process of the DTSM is performed under normalcondition without a load and the transmission speed Vi isset as V0 in Figure 7(a) and the scraper 1 moves fromposition 1 to position 2 In our study nodes 9 16 and 3 ofthe fourth chute are chosen as the detecting points for vi-bration analysis For convenient expression in the followingwe mark the detecting nodes as measuring points 1 2 and3 respectively Taking measuring point 1 as the researchexample for description of the transmission process theoriginal signal is presented in Figure 7(b) e vibrationresponse presents a three-stage change ie the accelerationphase the steady phase and the deceleration phase Asmentioned above in the steady phase the vibration signalcan be treated as multiple segments with a time span of ΔTwhich corresponds to the chain assembly with a length of ΔL(Figure 5)

e field FPET of multiple monitoring points iscarried out to obtain the vibration properties of the actualchute and evaluate the dynamic performance of theDTSM e specifications of the experimental scraperconveyor correspond to SGB12003600 manufactured byLianyungang Tianming Equipment Co Ltd and therunning speed of the scraper chain is 10 ms In practicalengineering the actual double-drive transmission systemis driven by two drive motors e basic performanceparameters are shown in Table 3 As shown in Figure 8the experimental system mainly contains three acceler-ation sensors (TST120A1000) a wireless acquisitiondevice (TST5925EV) a wireless receiver and an on-sitePC e acceleration sensors are used for fixed-pointmeasurements by detecting the vibration responses ofmeasuring points 1 2 and 3 on the chute e wirelessacquisition device is used to collect experimental data inreal time In addition the wireless receiver is intended forremote data transmissione vibration data are stored inthe on-site PC and a sampling frequency of 1000 Hz istaken

Figure 9 shows the experimental and simulation signalsof measuring points 1 2 and 3 in the time span ΔT Inthe time domain the simulation results and the experi-mental results are similar which reflects the adaptability ofthe DTSM Considering the nonlinear and time-varyingcharacteristics of the vibration signals the AOKR is used fortime-frequency analysis

e vibration signal of measuring point 2 in the timespan ΔT is taken as the reference for time-frequency

1 2 3 4 5 6 7 8 9 10

11 12 13 14 15 16 17 18 19 20

(a)

Number of the sensors

Min

imum

val

ue o

f MAC

ij

2 3 4 5 6 7 8 9 10

028

0275

027

0265

026

0255

025

0245

(b)

Figure 6 Description of the optimum placement scheme (a) theprimary measuring points (b) the minimum MACij

Table 2 Measuring points placement based on the MAC

Number f Measuring points2 02515 8 163 02453 3 9 164 02454 1 3 9 165 02460 1 3 9 16 206 02483 1 3 9 10 11 167 02525 1 2 3 9 10 11 168 02644 1 3 8 9 10 11 15 169 02664 1 2 3 8 9 10 11 15 1610 02767 1 2 3 8 9 10 11 15 16 20

Shock and Vibration 7

analysis and then the time-frequency representations ofmeasuring points 1 and 3 in the same time period aredemonstrated Figure 10(a) shows the time-frequencyrepresentation of the experimental signals for differentmeasuring points and the power spectrum is alsodepicted It presents that the vibration signal of the chutehas a plurality of components ere are two major areaswith strong frequency responses of the vibration signalsfor different measuring points for measuring point 1 thefrequency range is 230ndash310Hz and the time ranges are015ndash030 s and 045ndash075 s for measuring point 2 thefrequency range is 230ndash340Hz and the time ranges are020ndash040 s and 055ndash070 s and for measuring point 3the frequency range is 225ndash330Hz and the time ranges are040ndash050 s and 070ndash090 s Considering the chain

assembly in Figure 5 there are two scrapers for length ΔLIn the actual transmission process the two major areaswith strong frequency responses are caused while the twoscrapers passing through the measuring points As shownbasic frequencies of the measuring points are 239 278 and249Hz respectively In addition the peak of the vibrationpower spectrum is also presented and the maximumvalues of the measuring points are 1588 1561 and1499m2s

Similarly the time-frequency representation of thesimulation signals with two major frequency responses isdiscussed As Figure 10(b) indicates for measuring points1 2 and 3 the vibration energies are concentrated atthe frequency ranges 240ndash340 200ndash310 and 200ndash320Hzrespectively Correspondingly for measuring point 1 the

t1 t3

t2

Chain jamoccurs

15

1

05

0

Time (s)

Velo

city

(ms

)

0 05 1 15 2 25 3 35 4 45 5 55 6

V0V1

(a)

Constant speed

Acceleration DecelerationΔT

Time (s)0 1 2 3 4 5 6

543210

ndash1ndash2ndash3ndash4ndash5

Acce

lera

tion

(ms

2 )

(b)

Figure 7 Transmission speed setting and vibration signal under normal condition

Table 3 Basic operating condition of the scraper conveyor

Model Chain size (mm) Conveyor width (mm) Chain speed (ms) Transport capacity (th) Transport length (m)SGZ12003600 φ48times152 1750times1180 0sim189 3700 360

Gear

Gear

Sprocket

Drive motor On-site PCWireless receiver

Transitionaltrough Chute

Accelerationsensor

Wirelessacquisition device

2

3

1

Figure 8 Experimental setup of the FPET

8 Shock and Vibration

time ranges are 010ndash025 and 050ndash065 s en the vi-bration energy of measuring point 2 is concentrated at thetime ranges 015ndash045 and 060ndash075 s and the time rangesof measuring point 3 are 040ndash050 and 070ndash085 s Forthe simulation signals the basic frequencies of the mea-suring points are 281 249 and 258Hz respectively emaximum values of vibration energies are 1527 1489 and1532m2s respectively Considering the experimental re-sults and the simulation results separately good agreementbetween the main parameters of different measuring pointsis obtained It also indicates that in the same time periodthe vibration signal has a delay characteristic for measuringpoints 1 2 and 3 is is possibly because that themeasuring points are at different positions of the chuteMoreover for different measuring points the experimentalsignals and the simulation signals show good consistency intime-frequency characteristics Hence the establishedDTSM can efficiently simulate the actual productionenvironment

43 Fault Detection of the Scraper Chain Based on theestablished DTSM two typical failure patterns of thescraper chain are discussed namely chain jam and chainfracture ree external load conditions are set that isempty load half-load and full-load conditions e valuesof Wi (Figure 5) are defined as 0 12W0 and W0 re-spectively As designed in Figure 5 when chain jam occursthe transmission speed Vi is set as V1 in Figure 7(a) and the

failure time range is t1 minus t3(t3 25 s) the chain fault istriggered at t1 2 sWithin the time range t1 minus t2(t2 225 s)the transmission speed Vi decreases from 1 to 0ms And thevalue of Vi increases from 0 to 1ms within the time ranget2 minus t3 e whole process lasts 05 s the chain assembly istightened and the scraper chains are jammed When chainfracture occurs the transmission speed Vi is set as V0 inFigure 7(a) Moreover as shown in Figure 5 the contactconstraint between two contacting scraper chains at the labeledcontact point 1 is removed As a result the two contactingscraper chains will be separated In order to ensure the accuracyof fault setting the chain fracture is also triggered at t1 2 sand the contact constraint is removed in 05 s which is con-sistent with chain jam

Considering both failure patterns of the scraper chainour research focuses on the steady phase of the operationprocess within the time range 05 to 55 s In the steadyphase the vibration signals are obtained simultaneously atmeasuring points 1 2 and 3 Taking measuring point1 as the case study the vibration signals for chain jam andchain fracture under empty load condition are presentedin Figures 11(a) and 11(b) respectively After faultstriggering the vibration signals of the two failure patternsshow a sudden increase after a short time delay Sub-sequently the signals exhibit unstable fluctuations at thetime range 24ndash28 s In order to obtain a more detaileddescription of the detection results the maximum am-plitudes of the vibration signals under different workingconditions are discussed ie normal condition chain jam

3210

ndash1ndash2ndash3

3210

ndash1ndash2ndash3

3210

ndash1ndash2ndash3

Time (s)

Acce

lera

tion

(ms

2 )

0 01 02 03 04 05 06 07 08 09

1 2 3

(a)

3210

ndash1ndash2ndash3

3210

ndash1ndash2ndash3

3210

ndash1ndash2ndash3

Time (s)Ac

cele

ratio

n (m

s2 )

0 01 02 03 04 05 06 07 08 09

123

(b)

Figure 9 Vibration signals of different measuring points in ΔT (a) experimental results (b) simulation results

Shock and Vibration 9

Point

1

2

3

(a) (b)

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

0

Time (s) Spectrum0 03 06 09

500

400

300

200

100

00 50 100 150

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

00 03 06 09

500

400

300

200

100

00 50 100 150

Time (s) Spectrum

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

00 03 06 09

500

400

300

200

100

00 50 100 150

Time (s) Spectrum

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

00 03 06 09

500

400

300

200

100

00 50 100 150

Time (s) Spectrum

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

0

Time (s) Spectrum0 03 06 09

500

400

300

200

100

00 50 100 150

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

00 03 06 09

500

400

300

200

100

00 50 100 150

Time (s) Spectrum

Figure 10 Time-frequency representation of vibration signals (a) experimental results (b) simulation results

Chain jamoccurs

Acce

lera

tion

(ms

2 ) Acce

lera

tion

(ms

2 )

10

5

0

ndash5

ndash10

Time (s)0 05 10 15 20 25 30 35 40 45 50

10

5

0

ndash5

ndash1024 25 26 27 28

(a)

Figure 11 Continued

10 Shock and Vibration

and chain fracture Wherein the empty load half-loadand full-load conditions are considered e maximumamplitudes of the vibration signals at measuring points1 2 and 3 are depicted in Figures 12(a)ndash12(c)respectively

Considering measuring point 1 the maximum am-plitudes of the vibration signals under empty load conditionare 289 718 and 698ms2 under normal condition chainjam and chain fracture respectively Similarly under half-load condition the maximum amplitudes are 324 937 and956ms2 Moreover under full-load condition the maxi-mum amplitudes are 446 1106 and 1082ms2 Withdifferent external loads the maximum amplitudes of thevibration signals for fault conditions are obviously higherthan those for normal condition and the difference betweenthe amplitudes of the two typical failure patterns is small Fordifferent fault conditions with the increase of the externalloads the maximum amplitudes show trends to increasee above statistical results are also applicable to measuringpoints 2 and 3 erefore chain faults can easily be de-tected by comparing the maximum amplitudes of the vi-bration signals whereas the fault patterns are difficult toidentify According to the nonstationary and nonlinearcharacteristics of fault signals the AOKR is utilized to an-alyze the vibration signals and classify failure patterns of thescraper chain Within 15 s after faults triggering the vi-bration signals at the three measuring points with differentexternal loads are processed Wherein for chain jam andchain fracture under empty load condition the time-fre-quency representations of vibration signals are presented inFigures 13(a) and 13(b) respectively

e frequency components and frequency ranges ofthe same fault pattern are similar for different measuringpoints As Figure 13(a) describes the bright color between0 and 50Hz indicates one high energy area caused bychain jam en chain fracture can easily be distinguishedaccording to the appearance of two high energy areasbetween 100 and 200Hz as shown in Figure 13(b)

Observing the spectrum results a more detailed de-scription is given When chain jam occurs for measuringpoints 1 2 and 3 the high energy areas occur ap-proximately at 05 075 and 09 s respectively Mean-while for chain fracture the high energy areas include twomain frequency components and are approximatelyconcentrated at the time ranges 050ndash070 075ndash085 and10ndash115 s respectively Hence there is a delay charac-teristic of the fault occurrence which is well in accordancewith the conclusions in Section 42 In order to explore theinfluence of external loads on fault characteristics thedetailed differences of the fault patterns at measuringpoint 2 are depicted in Figure 14 In fact the externalload has a great influence on the fault severity of both thefailure patterns Observing the spectrum results thebright areas vary with the external loads With increasingexternal load the frequency ranges of the high energyareas become larger Wherein for chain jam under emptyhalf- and full-load conditions the frequency ranges areapproximately 0ndash50 0ndash150 and 0ndash250 Hz respectivelyMeanwhile for chain fracture the frequency rangesare approximately 80ndash200 50ndash250 and 50ndash350Hzrespectively

In this part three working conditions of the scraperchain are investigated above including normal conditionchain jam and chain fracture e vibration signals ofmeasuring points 1 2 and 3 on the detecting chute areanalyzed and the effects of the external loads on the vi-bration characteristics are discussed Based on the aboveanalysis the occurrence of chain faults can easily be de-termined through amplitude comparisons of the originalvibration signals However the observation confirms thesimilarity of the time domain waveforms of fault signals forchain jam and chain fracture ese two patterns of failuresremain to be different through further processing by theAOKR and the fault patterns can be distinguished accordingto the number of high energy areas of the time-frequencyrepresentation of vibration signals In conclusion the

Chain fractureoccurs

Acce

lera

tion

(ms

2 ) Acce

lera

tion

(ms

2 )

10

5

0

ndash5

ndash10

Time (s)0 05 10 15 20 25 30 35 40 45 50

10

5

0

ndash5

ndash1024 25 26 27 28

(b)

Figure 11 Vibration signals of the measuring point 1 for (a) chain jam and (b) chain fracture

Shock and Vibration 11

Point 1 2 3

(a)

(b)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

Figure 13 Time-frequency representation of vibration signals under empty load (a) chain jam (b) chain fracture

12

10

8

6

4

2

0Max

imum

acce

lera

tion

ampl

itude

(ms

2 )

Normal conditionChain jamChain fracture

0 12 W0 W0

(a)

Normal conditionChain jamChain fracture

12

10

8

6

4

2

0Max

imum

acce

lera

tion

ampl

itude

(ms

2 )

0 12 W0 W0

(b)

Normal conditionChain jamChain fracture

12

10

8

6

4

2

0Max

imum

acce

lera

tion

ampl

itude

(ms

2 )

0 12 W0 W0

(c)

Figure 12 Statistical results of the maximum vibration amplitude at different measuring points (a) 1 (b) 2 (c) 3

12 Shock and Vibration

proposed detection strategy is effective at detecting theoccurrence of chain faults and identifying the failure pat-terns under different operating conditions

5 Conclusions

During the actual operation the working state of thescraper chain can reflect the dynamic performance of thescraper conveyor To address the difficulties with directsensor measurement for parameters of the moving scraperchain a novel strategy for fault detection of the scraperchain based on vibration analysis of the chute was pro-posed Based on modal analysis and the MAC the mea-suring points of vibration signals on the chute weredetermined To fit the actual behavior of the transmissionprocess the DTSM was presented based on finite elementmodeling and the correctness of the dynamic model wasverified by comparison with the FPET en the vibrationproperties of the measuring points on the chute undernormal condition chain jam and chain fracture werediscussed Moreover the occurrence of chain faults weredetermined by comparing the amplitudes of the vibrationsignal in the time domain while the AOKR was utilizedfor time-frequency representation of vibration signals anddistinguishing the two typical failure patterns Further-more the strategy verification based on experimental datawill be taken into consideration in the near future

Data Availability

e data used to support the findings of this study areavailable from the corresponding author upon request

Conflicts of Interest

e authors declare that they have no conflicts of interest

Acknowledgments

is work was supported by the Key Project of NationalNatural Science Foundation of China (U1510205) NaturalScience Foundation of Jiangsu Province (No BK20160251)Xuzhou Research program (KC14H0138) FundamentalResearch Funds for the Central Universities (2014Y05) andProject Funded by the Priority Academic Program De-velopment of Jiangsu Higher Education Institutions(PAPD)

References

[1] C D Brown ldquoDesign build and test of a longwall armouredface conveyorrdquo Longwall Mining 2002

[2] M Dolipski P Cheluszka E Remiorz and P SobotaldquoFollow-up chain tension in an armoured face conveyornadazne napinanie lancucha zgrzebłowego W przenosnikuscianowymrdquo Archives of Mining Sciences vol 60 no 1pp 25ndash38 2015

[3] L A Morley J L Kohler and H M Smolnikar ldquoA model forpredicting motor load for an armored face-conveyor driverdquoIEEE Transactions on Industry Applications vol 24 no 4pp 649ndash659 1988

[4] A A Ordin and A A Metelrsquokov ldquoAnalysis of longwall faceoutput in screw-type cutter-loader-and-scraper conveyorsystem in underground mining of flat-lying coal bedsrdquoJournal of Mining Science vol 51 no 6 pp 1173ndash1179 2015

[5] B He G Li H Shi et al ldquoDynamic behaviour modelling andsimulation of the chain transmission system for an armouredface conveyorrdquo in Proceedings of the IEEE 10th InternationalConference on Computer-Aided Industrial Design and Con-ceptual Design CAID amp CD 2009 pp 1000ndash1004 BeijingChina November 2009

[6] R Nie B He P Yuan L Zhang and G Li ldquoNovel approachto and implementation of design and analysis of armored faceconveyor power trainrdquo Science China Technological Sciencesvol 58 no 12 pp 2153ndash2168 2015

Loads

(a)

(b)

Empty load Half-load Full-load

500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)

Figure 14 Time-frequency representation of vibration signals at measuring point 2 with different external loads (a) chain jam (b) chainfracture

Shock and Vibration 13

[7] R Nie B He L Zhang and G Li ldquoModelling of thetransmission system in conveying equipment based on Eulermethod with applicationrdquo Proceedings of the Institution ofMechanical Engineers Part K Journal of Multi-body Dy-namics vol 228 no 3 pp 294ndash306 2014

[8] S B Jiang X Zhang K D Gao J Gao Q Y Wang andK Hidenori ldquoMulti-body dynamics and vibration analysis ofchain assembly in armoured face conveyorrdquo InternationalJournal of Simulation Modelling vol 16 no 3 pp 458ndash4702017

[9] M Myszkowski and D Loehning ldquoChain force measure-ments on armoured face conveyors and coal plows in heavy-duty longwallsrdquo CIM Bulletin vol 94 no 1054 pp 72ndash752001

[10] H Wang Q Zhang and F Xie ldquoDynamic tension test andintelligent coordinated control system of a heavy scraperconveyorrdquo IET Science Measurement and Technology vol 11no 7 pp 871ndash877 2017

[11] S Sen M X Min and Y Z She ldquoDiagnosis of coal scraperconveyor based on Fuzzy Fault treerdquo in Proceedings of the2015 Seventh International Conference on Measuring Tech-nology and Mechatronics Automation (ICMTMA) pp 392ndash395 IEEE Nanchang China June 2015

[12] S-s Xue X-c Li and X-y Xu ldquoFault tree and Bayesiannetwork based scraper conveyer fault diagnosisrdquo in Pro-ceedings of the 22nd International Conference on IndustrialEngineering and Engineering Management 2015 pp 783ndash795Atlantis Press Paris France January 2016

[13] X Gong X Ma Y Zhang et al ldquoApplication of fuzzy neuralnetwork in fault diagnosis for scraper conveyor vibrationrdquo inProceedings of the 2013 IEEE International Conference onInformation and Automation (ICIA) pp 1135ndash1140 IEEEYinchuan China August 2013

[14] Y Zhang X Ma Y Jianxiang et al ldquoFuzzy neural networkfault diagnosis and online vibration monitoring system for thecoal scraper conveyor based on rough set theoryrdquo in Pro-ceedings of the 2013 32nd Chinese Control Conference (CCC)pp 6134ndash6138 IEEE Xirsquoan China July 2013

[15] B Zhang A C C Tan and J-h Lin ldquoGearbox fault diagnosisof high-speed railway trainrdquo Engineering Failure Analysisvol 66 pp 407ndash420 2016

[16] E Parloo P Verboven P Guillaume and M Van OvermeireldquoAutonomous structural health monitoring-part ii vibration-based in-operation damage assessmentrdquo Mechanical Systemsand Signal Processing vol 16 no 4 pp 659ndash675 2002

[17] C S Sakaris J S Sakellariou and S D Fassois ldquoRandom-vibration-based damage detection and precise localization ona lab-scale aircraft stabilizer structure via the GeneralizedFunctional Model Based Methodrdquo Structural Health Moni-toring An International Journal vol 16 no 5 pp 594ndash6102017

[18] Y Zhang W Song M Karimi C-H Chi and A KudreykoldquoFractional autoregressive integrated moving average andfinite-element modal the forecast of tire vibration trendrdquoIEEE Access vol 6 pp 40137ndash40142 2018

[19] M Pastor M Binda and T Harcarik ldquoModal assurancecriterionrdquo Procedia Engineering vol 48 pp 543ndash548 2012

[20] W J Staszewski K Worden and G R Tomlinson ldquoTime-frequency analysis in gearbox fault detection using the Wigner-ville distribution and pattern recognitionrdquo Mechanical Systemsand Signal Processing vol 11 no 5 pp 673ndash692 1997

[21] J-D Wu and P-H Chiang ldquoApplication of Wigner-Villedistribution and probability neural network for scooter engine

fault diagnosisrdquo Expert Systems with Applications vol 36no 2 pp 2187ndash2199 2009

[22] Z Feng and M Liang ldquoFault diagnosis of wind turbineplanetary gearbox under nonstationary conditions viaadaptive optimal kernel time-frequency analysisrdquo RenewableEnergy vol 66 pp 468ndash477 2014

14 Shock and Vibration

International Journal of

AerospaceEngineeringHindawiwwwhindawicom Volume 2018

RoboticsJournal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Active and Passive Electronic Components

VLSI Design

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Shock and Vibration

Hindawiwwwhindawicom Volume 2018

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawiwwwhindawicom

Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Control Scienceand Engineering

Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom

Journal ofEngineeringVolume 2018

SensorsJournal of

Hindawiwwwhindawicom Volume 2018

International Journal of

RotatingMachinery

Hindawiwwwhindawicom Volume 2018

Modelling ampSimulationin EngineeringHindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Navigation and Observation

International Journal of

Hindawi

wwwhindawicom Volume 2018

Advances in

Multimedia

Submit your manuscripts atwwwhindawicom

41 Optimal Placement of the Acceleration Sensors Beforeconducting the FPET the installation scheme of the ac-celeration sensors should be determined to ensure thevalidity of the measured vibration signals and the economyof the experimental process Based on modal analysis of theFEA (Figure 2(a)) we extract 20 nodes from the chute as theprimary measuring points which are labeled in Figure 6(a)According to the theory of MAC the total modal dis-placements of the selected 20 points serve as the inputs ofequations (6) and (7) In Figure 6(b) the curve revealschange rules of the minimumMACij for different number ofsensors

To illustrate the optimal placement scheme of acceler-ation sensors in more detail the optimum installation po-sitions for different number of sensors are shown in Table 2Accordingly the vibration response of the actual chuteshould be recorded by 3 acceleration sensors and the op-timum installation positions are set at nodes 3 9 and 16respectively

42 Experimental Evaluations According to Section 32 thesimulation process of the DTSM is performed under normalcondition without a load and the transmission speed Vi isset as V0 in Figure 7(a) and the scraper 1 moves fromposition 1 to position 2 In our study nodes 9 16 and 3 ofthe fourth chute are chosen as the detecting points for vi-bration analysis For convenient expression in the followingwe mark the detecting nodes as measuring points 1 2 and3 respectively Taking measuring point 1 as the researchexample for description of the transmission process theoriginal signal is presented in Figure 7(b) e vibrationresponse presents a three-stage change ie the accelerationphase the steady phase and the deceleration phase Asmentioned above in the steady phase the vibration signalcan be treated as multiple segments with a time span of ΔTwhich corresponds to the chain assembly with a length of ΔL(Figure 5)

e field FPET of multiple monitoring points iscarried out to obtain the vibration properties of the actualchute and evaluate the dynamic performance of theDTSM e specifications of the experimental scraperconveyor correspond to SGB12003600 manufactured byLianyungang Tianming Equipment Co Ltd and therunning speed of the scraper chain is 10 ms In practicalengineering the actual double-drive transmission systemis driven by two drive motors e basic performanceparameters are shown in Table 3 As shown in Figure 8the experimental system mainly contains three acceler-ation sensors (TST120A1000) a wireless acquisitiondevice (TST5925EV) a wireless receiver and an on-sitePC e acceleration sensors are used for fixed-pointmeasurements by detecting the vibration responses ofmeasuring points 1 2 and 3 on the chute e wirelessacquisition device is used to collect experimental data inreal time In addition the wireless receiver is intended forremote data transmissione vibration data are stored inthe on-site PC and a sampling frequency of 1000 Hz istaken

Figure 9 shows the experimental and simulation signalsof measuring points 1 2 and 3 in the time span ΔT Inthe time domain the simulation results and the experi-mental results are similar which reflects the adaptability ofthe DTSM Considering the nonlinear and time-varyingcharacteristics of the vibration signals the AOKR is used fortime-frequency analysis

e vibration signal of measuring point 2 in the timespan ΔT is taken as the reference for time-frequency

1 2 3 4 5 6 7 8 9 10

11 12 13 14 15 16 17 18 19 20

(a)

Number of the sensors

Min

imum

val

ue o

f MAC

ij

2 3 4 5 6 7 8 9 10

028

0275

027

0265

026

0255

025

0245

(b)

Figure 6 Description of the optimum placement scheme (a) theprimary measuring points (b) the minimum MACij

Table 2 Measuring points placement based on the MAC

Number f Measuring points2 02515 8 163 02453 3 9 164 02454 1 3 9 165 02460 1 3 9 16 206 02483 1 3 9 10 11 167 02525 1 2 3 9 10 11 168 02644 1 3 8 9 10 11 15 169 02664 1 2 3 8 9 10 11 15 1610 02767 1 2 3 8 9 10 11 15 16 20

Shock and Vibration 7

analysis and then the time-frequency representations ofmeasuring points 1 and 3 in the same time period aredemonstrated Figure 10(a) shows the time-frequencyrepresentation of the experimental signals for differentmeasuring points and the power spectrum is alsodepicted It presents that the vibration signal of the chutehas a plurality of components ere are two major areaswith strong frequency responses of the vibration signalsfor different measuring points for measuring point 1 thefrequency range is 230ndash310Hz and the time ranges are015ndash030 s and 045ndash075 s for measuring point 2 thefrequency range is 230ndash340Hz and the time ranges are020ndash040 s and 055ndash070 s and for measuring point 3the frequency range is 225ndash330Hz and the time ranges are040ndash050 s and 070ndash090 s Considering the chain

assembly in Figure 5 there are two scrapers for length ΔLIn the actual transmission process the two major areaswith strong frequency responses are caused while the twoscrapers passing through the measuring points As shownbasic frequencies of the measuring points are 239 278 and249Hz respectively In addition the peak of the vibrationpower spectrum is also presented and the maximumvalues of the measuring points are 1588 1561 and1499m2s

Similarly the time-frequency representation of thesimulation signals with two major frequency responses isdiscussed As Figure 10(b) indicates for measuring points1 2 and 3 the vibration energies are concentrated atthe frequency ranges 240ndash340 200ndash310 and 200ndash320Hzrespectively Correspondingly for measuring point 1 the

t1 t3

t2

Chain jamoccurs

15

1

05

0

Time (s)

Velo

city

(ms

)

0 05 1 15 2 25 3 35 4 45 5 55 6

V0V1

(a)

Constant speed

Acceleration DecelerationΔT

Time (s)0 1 2 3 4 5 6

543210

ndash1ndash2ndash3ndash4ndash5

Acce

lera

tion

(ms

2 )

(b)

Figure 7 Transmission speed setting and vibration signal under normal condition

Table 3 Basic operating condition of the scraper conveyor

Model Chain size (mm) Conveyor width (mm) Chain speed (ms) Transport capacity (th) Transport length (m)SGZ12003600 φ48times152 1750times1180 0sim189 3700 360

Gear

Gear

Sprocket

Drive motor On-site PCWireless receiver

Transitionaltrough Chute

Accelerationsensor

Wirelessacquisition device

2

3

1

Figure 8 Experimental setup of the FPET

8 Shock and Vibration

time ranges are 010ndash025 and 050ndash065 s en the vi-bration energy of measuring point 2 is concentrated at thetime ranges 015ndash045 and 060ndash075 s and the time rangesof measuring point 3 are 040ndash050 and 070ndash085 s Forthe simulation signals the basic frequencies of the mea-suring points are 281 249 and 258Hz respectively emaximum values of vibration energies are 1527 1489 and1532m2s respectively Considering the experimental re-sults and the simulation results separately good agreementbetween the main parameters of different measuring pointsis obtained It also indicates that in the same time periodthe vibration signal has a delay characteristic for measuringpoints 1 2 and 3 is is possibly because that themeasuring points are at different positions of the chuteMoreover for different measuring points the experimentalsignals and the simulation signals show good consistency intime-frequency characteristics Hence the establishedDTSM can efficiently simulate the actual productionenvironment

43 Fault Detection of the Scraper Chain Based on theestablished DTSM two typical failure patterns of thescraper chain are discussed namely chain jam and chainfracture ree external load conditions are set that isempty load half-load and full-load conditions e valuesof Wi (Figure 5) are defined as 0 12W0 and W0 re-spectively As designed in Figure 5 when chain jam occursthe transmission speed Vi is set as V1 in Figure 7(a) and the

failure time range is t1 minus t3(t3 25 s) the chain fault istriggered at t1 2 sWithin the time range t1 minus t2(t2 225 s)the transmission speed Vi decreases from 1 to 0ms And thevalue of Vi increases from 0 to 1ms within the time ranget2 minus t3 e whole process lasts 05 s the chain assembly istightened and the scraper chains are jammed When chainfracture occurs the transmission speed Vi is set as V0 inFigure 7(a) Moreover as shown in Figure 5 the contactconstraint between two contacting scraper chains at the labeledcontact point 1 is removed As a result the two contactingscraper chains will be separated In order to ensure the accuracyof fault setting the chain fracture is also triggered at t1 2 sand the contact constraint is removed in 05 s which is con-sistent with chain jam

Considering both failure patterns of the scraper chainour research focuses on the steady phase of the operationprocess within the time range 05 to 55 s In the steadyphase the vibration signals are obtained simultaneously atmeasuring points 1 2 and 3 Taking measuring point1 as the case study the vibration signals for chain jam andchain fracture under empty load condition are presentedin Figures 11(a) and 11(b) respectively After faultstriggering the vibration signals of the two failure patternsshow a sudden increase after a short time delay Sub-sequently the signals exhibit unstable fluctuations at thetime range 24ndash28 s In order to obtain a more detaileddescription of the detection results the maximum am-plitudes of the vibration signals under different workingconditions are discussed ie normal condition chain jam

3210

ndash1ndash2ndash3

3210

ndash1ndash2ndash3

3210

ndash1ndash2ndash3

Time (s)

Acce

lera

tion

(ms

2 )

0 01 02 03 04 05 06 07 08 09

1 2 3

(a)

3210

ndash1ndash2ndash3

3210

ndash1ndash2ndash3

3210

ndash1ndash2ndash3

Time (s)Ac

cele

ratio

n (m

s2 )

0 01 02 03 04 05 06 07 08 09

123

(b)

Figure 9 Vibration signals of different measuring points in ΔT (a) experimental results (b) simulation results

Shock and Vibration 9

Point

1

2

3

(a) (b)

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

0

Time (s) Spectrum0 03 06 09

500

400

300

200

100

00 50 100 150

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

00 03 06 09

500

400

300

200

100

00 50 100 150

Time (s) Spectrum

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

00 03 06 09

500

400

300

200

100

00 50 100 150

Time (s) Spectrum

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

00 03 06 09

500

400

300

200

100

00 50 100 150

Time (s) Spectrum

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

0

Time (s) Spectrum0 03 06 09

500

400

300

200

100

00 50 100 150

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

00 03 06 09

500

400

300

200

100

00 50 100 150

Time (s) Spectrum

Figure 10 Time-frequency representation of vibration signals (a) experimental results (b) simulation results

Chain jamoccurs

Acce

lera

tion

(ms

2 ) Acce

lera

tion

(ms

2 )

10

5

0

ndash5

ndash10

Time (s)0 05 10 15 20 25 30 35 40 45 50

10

5

0

ndash5

ndash1024 25 26 27 28

(a)

Figure 11 Continued

10 Shock and Vibration

and chain fracture Wherein the empty load half-loadand full-load conditions are considered e maximumamplitudes of the vibration signals at measuring points1 2 and 3 are depicted in Figures 12(a)ndash12(c)respectively

Considering measuring point 1 the maximum am-plitudes of the vibration signals under empty load conditionare 289 718 and 698ms2 under normal condition chainjam and chain fracture respectively Similarly under half-load condition the maximum amplitudes are 324 937 and956ms2 Moreover under full-load condition the maxi-mum amplitudes are 446 1106 and 1082ms2 Withdifferent external loads the maximum amplitudes of thevibration signals for fault conditions are obviously higherthan those for normal condition and the difference betweenthe amplitudes of the two typical failure patterns is small Fordifferent fault conditions with the increase of the externalloads the maximum amplitudes show trends to increasee above statistical results are also applicable to measuringpoints 2 and 3 erefore chain faults can easily be de-tected by comparing the maximum amplitudes of the vi-bration signals whereas the fault patterns are difficult toidentify According to the nonstationary and nonlinearcharacteristics of fault signals the AOKR is utilized to an-alyze the vibration signals and classify failure patterns of thescraper chain Within 15 s after faults triggering the vi-bration signals at the three measuring points with differentexternal loads are processed Wherein for chain jam andchain fracture under empty load condition the time-fre-quency representations of vibration signals are presented inFigures 13(a) and 13(b) respectively

e frequency components and frequency ranges ofthe same fault pattern are similar for different measuringpoints As Figure 13(a) describes the bright color between0 and 50Hz indicates one high energy area caused bychain jam en chain fracture can easily be distinguishedaccording to the appearance of two high energy areasbetween 100 and 200Hz as shown in Figure 13(b)

Observing the spectrum results a more detailed de-scription is given When chain jam occurs for measuringpoints 1 2 and 3 the high energy areas occur ap-proximately at 05 075 and 09 s respectively Mean-while for chain fracture the high energy areas include twomain frequency components and are approximatelyconcentrated at the time ranges 050ndash070 075ndash085 and10ndash115 s respectively Hence there is a delay charac-teristic of the fault occurrence which is well in accordancewith the conclusions in Section 42 In order to explore theinfluence of external loads on fault characteristics thedetailed differences of the fault patterns at measuringpoint 2 are depicted in Figure 14 In fact the externalload has a great influence on the fault severity of both thefailure patterns Observing the spectrum results thebright areas vary with the external loads With increasingexternal load the frequency ranges of the high energyareas become larger Wherein for chain jam under emptyhalf- and full-load conditions the frequency ranges areapproximately 0ndash50 0ndash150 and 0ndash250 Hz respectivelyMeanwhile for chain fracture the frequency rangesare approximately 80ndash200 50ndash250 and 50ndash350Hzrespectively

In this part three working conditions of the scraperchain are investigated above including normal conditionchain jam and chain fracture e vibration signals ofmeasuring points 1 2 and 3 on the detecting chute areanalyzed and the effects of the external loads on the vi-bration characteristics are discussed Based on the aboveanalysis the occurrence of chain faults can easily be de-termined through amplitude comparisons of the originalvibration signals However the observation confirms thesimilarity of the time domain waveforms of fault signals forchain jam and chain fracture ese two patterns of failuresremain to be different through further processing by theAOKR and the fault patterns can be distinguished accordingto the number of high energy areas of the time-frequencyrepresentation of vibration signals In conclusion the

Chain fractureoccurs

Acce

lera

tion

(ms

2 ) Acce

lera

tion

(ms

2 )

10

5

0

ndash5

ndash10

Time (s)0 05 10 15 20 25 30 35 40 45 50

10

5

0

ndash5

ndash1024 25 26 27 28

(b)

Figure 11 Vibration signals of the measuring point 1 for (a) chain jam and (b) chain fracture

Shock and Vibration 11

Point 1 2 3

(a)

(b)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

Figure 13 Time-frequency representation of vibration signals under empty load (a) chain jam (b) chain fracture

12

10

8

6

4

2

0Max

imum

acce

lera

tion

ampl

itude

(ms

2 )

Normal conditionChain jamChain fracture

0 12 W0 W0

(a)

Normal conditionChain jamChain fracture

12

10

8

6

4

2

0Max

imum

acce

lera

tion

ampl

itude

(ms

2 )

0 12 W0 W0

(b)

Normal conditionChain jamChain fracture

12

10

8

6

4

2

0Max

imum

acce

lera

tion

ampl

itude

(ms

2 )

0 12 W0 W0

(c)

Figure 12 Statistical results of the maximum vibration amplitude at different measuring points (a) 1 (b) 2 (c) 3

12 Shock and Vibration

proposed detection strategy is effective at detecting theoccurrence of chain faults and identifying the failure pat-terns under different operating conditions

5 Conclusions

During the actual operation the working state of thescraper chain can reflect the dynamic performance of thescraper conveyor To address the difficulties with directsensor measurement for parameters of the moving scraperchain a novel strategy for fault detection of the scraperchain based on vibration analysis of the chute was pro-posed Based on modal analysis and the MAC the mea-suring points of vibration signals on the chute weredetermined To fit the actual behavior of the transmissionprocess the DTSM was presented based on finite elementmodeling and the correctness of the dynamic model wasverified by comparison with the FPET en the vibrationproperties of the measuring points on the chute undernormal condition chain jam and chain fracture werediscussed Moreover the occurrence of chain faults weredetermined by comparing the amplitudes of the vibrationsignal in the time domain while the AOKR was utilizedfor time-frequency representation of vibration signals anddistinguishing the two typical failure patterns Further-more the strategy verification based on experimental datawill be taken into consideration in the near future

Data Availability

e data used to support the findings of this study areavailable from the corresponding author upon request

Conflicts of Interest

e authors declare that they have no conflicts of interest

Acknowledgments

is work was supported by the Key Project of NationalNatural Science Foundation of China (U1510205) NaturalScience Foundation of Jiangsu Province (No BK20160251)Xuzhou Research program (KC14H0138) FundamentalResearch Funds for the Central Universities (2014Y05) andProject Funded by the Priority Academic Program De-velopment of Jiangsu Higher Education Institutions(PAPD)

References

[1] C D Brown ldquoDesign build and test of a longwall armouredface conveyorrdquo Longwall Mining 2002

[2] M Dolipski P Cheluszka E Remiorz and P SobotaldquoFollow-up chain tension in an armoured face conveyornadazne napinanie lancucha zgrzebłowego W przenosnikuscianowymrdquo Archives of Mining Sciences vol 60 no 1pp 25ndash38 2015

[3] L A Morley J L Kohler and H M Smolnikar ldquoA model forpredicting motor load for an armored face-conveyor driverdquoIEEE Transactions on Industry Applications vol 24 no 4pp 649ndash659 1988

[4] A A Ordin and A A Metelrsquokov ldquoAnalysis of longwall faceoutput in screw-type cutter-loader-and-scraper conveyorsystem in underground mining of flat-lying coal bedsrdquoJournal of Mining Science vol 51 no 6 pp 1173ndash1179 2015

[5] B He G Li H Shi et al ldquoDynamic behaviour modelling andsimulation of the chain transmission system for an armouredface conveyorrdquo in Proceedings of the IEEE 10th InternationalConference on Computer-Aided Industrial Design and Con-ceptual Design CAID amp CD 2009 pp 1000ndash1004 BeijingChina November 2009

[6] R Nie B He P Yuan L Zhang and G Li ldquoNovel approachto and implementation of design and analysis of armored faceconveyor power trainrdquo Science China Technological Sciencesvol 58 no 12 pp 2153ndash2168 2015

Loads

(a)

(b)

Empty load Half-load Full-load

500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)

Figure 14 Time-frequency representation of vibration signals at measuring point 2 with different external loads (a) chain jam (b) chainfracture

Shock and Vibration 13

[7] R Nie B He L Zhang and G Li ldquoModelling of thetransmission system in conveying equipment based on Eulermethod with applicationrdquo Proceedings of the Institution ofMechanical Engineers Part K Journal of Multi-body Dy-namics vol 228 no 3 pp 294ndash306 2014

[8] S B Jiang X Zhang K D Gao J Gao Q Y Wang andK Hidenori ldquoMulti-body dynamics and vibration analysis ofchain assembly in armoured face conveyorrdquo InternationalJournal of Simulation Modelling vol 16 no 3 pp 458ndash4702017

[9] M Myszkowski and D Loehning ldquoChain force measure-ments on armoured face conveyors and coal plows in heavy-duty longwallsrdquo CIM Bulletin vol 94 no 1054 pp 72ndash752001

[10] H Wang Q Zhang and F Xie ldquoDynamic tension test andintelligent coordinated control system of a heavy scraperconveyorrdquo IET Science Measurement and Technology vol 11no 7 pp 871ndash877 2017

[11] S Sen M X Min and Y Z She ldquoDiagnosis of coal scraperconveyor based on Fuzzy Fault treerdquo in Proceedings of the2015 Seventh International Conference on Measuring Tech-nology and Mechatronics Automation (ICMTMA) pp 392ndash395 IEEE Nanchang China June 2015

[12] S-s Xue X-c Li and X-y Xu ldquoFault tree and Bayesiannetwork based scraper conveyer fault diagnosisrdquo in Pro-ceedings of the 22nd International Conference on IndustrialEngineering and Engineering Management 2015 pp 783ndash795Atlantis Press Paris France January 2016

[13] X Gong X Ma Y Zhang et al ldquoApplication of fuzzy neuralnetwork in fault diagnosis for scraper conveyor vibrationrdquo inProceedings of the 2013 IEEE International Conference onInformation and Automation (ICIA) pp 1135ndash1140 IEEEYinchuan China August 2013

[14] Y Zhang X Ma Y Jianxiang et al ldquoFuzzy neural networkfault diagnosis and online vibration monitoring system for thecoal scraper conveyor based on rough set theoryrdquo in Pro-ceedings of the 2013 32nd Chinese Control Conference (CCC)pp 6134ndash6138 IEEE Xirsquoan China July 2013

[15] B Zhang A C C Tan and J-h Lin ldquoGearbox fault diagnosisof high-speed railway trainrdquo Engineering Failure Analysisvol 66 pp 407ndash420 2016

[16] E Parloo P Verboven P Guillaume and M Van OvermeireldquoAutonomous structural health monitoring-part ii vibration-based in-operation damage assessmentrdquo Mechanical Systemsand Signal Processing vol 16 no 4 pp 659ndash675 2002

[17] C S Sakaris J S Sakellariou and S D Fassois ldquoRandom-vibration-based damage detection and precise localization ona lab-scale aircraft stabilizer structure via the GeneralizedFunctional Model Based Methodrdquo Structural Health Moni-toring An International Journal vol 16 no 5 pp 594ndash6102017

[18] Y Zhang W Song M Karimi C-H Chi and A KudreykoldquoFractional autoregressive integrated moving average andfinite-element modal the forecast of tire vibration trendrdquoIEEE Access vol 6 pp 40137ndash40142 2018

[19] M Pastor M Binda and T Harcarik ldquoModal assurancecriterionrdquo Procedia Engineering vol 48 pp 543ndash548 2012

[20] W J Staszewski K Worden and G R Tomlinson ldquoTime-frequency analysis in gearbox fault detection using the Wigner-ville distribution and pattern recognitionrdquo Mechanical Systemsand Signal Processing vol 11 no 5 pp 673ndash692 1997

[21] J-D Wu and P-H Chiang ldquoApplication of Wigner-Villedistribution and probability neural network for scooter engine

fault diagnosisrdquo Expert Systems with Applications vol 36no 2 pp 2187ndash2199 2009

[22] Z Feng and M Liang ldquoFault diagnosis of wind turbineplanetary gearbox under nonstationary conditions viaadaptive optimal kernel time-frequency analysisrdquo RenewableEnergy vol 66 pp 468ndash477 2014

14 Shock and Vibration

International Journal of

AerospaceEngineeringHindawiwwwhindawicom Volume 2018

RoboticsJournal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Active and Passive Electronic Components

VLSI Design

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Shock and Vibration

Hindawiwwwhindawicom Volume 2018

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawiwwwhindawicom

Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Control Scienceand Engineering

Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom

Journal ofEngineeringVolume 2018

SensorsJournal of

Hindawiwwwhindawicom Volume 2018

International Journal of

RotatingMachinery

Hindawiwwwhindawicom Volume 2018

Modelling ampSimulationin EngineeringHindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Navigation and Observation

International Journal of

Hindawi

wwwhindawicom Volume 2018

Advances in

Multimedia

Submit your manuscripts atwwwhindawicom

analysis and then the time-frequency representations ofmeasuring points 1 and 3 in the same time period aredemonstrated Figure 10(a) shows the time-frequencyrepresentation of the experimental signals for differentmeasuring points and the power spectrum is alsodepicted It presents that the vibration signal of the chutehas a plurality of components ere are two major areaswith strong frequency responses of the vibration signalsfor different measuring points for measuring point 1 thefrequency range is 230ndash310Hz and the time ranges are015ndash030 s and 045ndash075 s for measuring point 2 thefrequency range is 230ndash340Hz and the time ranges are020ndash040 s and 055ndash070 s and for measuring point 3the frequency range is 225ndash330Hz and the time ranges are040ndash050 s and 070ndash090 s Considering the chain

assembly in Figure 5 there are two scrapers for length ΔLIn the actual transmission process the two major areaswith strong frequency responses are caused while the twoscrapers passing through the measuring points As shownbasic frequencies of the measuring points are 239 278 and249Hz respectively In addition the peak of the vibrationpower spectrum is also presented and the maximumvalues of the measuring points are 1588 1561 and1499m2s

Similarly the time-frequency representation of thesimulation signals with two major frequency responses isdiscussed As Figure 10(b) indicates for measuring points1 2 and 3 the vibration energies are concentrated atthe frequency ranges 240ndash340 200ndash310 and 200ndash320Hzrespectively Correspondingly for measuring point 1 the

t1 t3

t2

Chain jamoccurs

15

1

05

0

Time (s)

Velo

city

(ms

)

0 05 1 15 2 25 3 35 4 45 5 55 6

V0V1

(a)

Constant speed

Acceleration DecelerationΔT

Time (s)0 1 2 3 4 5 6

543210

ndash1ndash2ndash3ndash4ndash5

Acce

lera

tion

(ms

2 )

(b)

Figure 7 Transmission speed setting and vibration signal under normal condition

Table 3 Basic operating condition of the scraper conveyor

Model Chain size (mm) Conveyor width (mm) Chain speed (ms) Transport capacity (th) Transport length (m)SGZ12003600 φ48times152 1750times1180 0sim189 3700 360

Gear

Gear

Sprocket

Drive motor On-site PCWireless receiver

Transitionaltrough Chute

Accelerationsensor

Wirelessacquisition device

2

3

1

Figure 8 Experimental setup of the FPET

8 Shock and Vibration

time ranges are 010ndash025 and 050ndash065 s en the vi-bration energy of measuring point 2 is concentrated at thetime ranges 015ndash045 and 060ndash075 s and the time rangesof measuring point 3 are 040ndash050 and 070ndash085 s Forthe simulation signals the basic frequencies of the mea-suring points are 281 249 and 258Hz respectively emaximum values of vibration energies are 1527 1489 and1532m2s respectively Considering the experimental re-sults and the simulation results separately good agreementbetween the main parameters of different measuring pointsis obtained It also indicates that in the same time periodthe vibration signal has a delay characteristic for measuringpoints 1 2 and 3 is is possibly because that themeasuring points are at different positions of the chuteMoreover for different measuring points the experimentalsignals and the simulation signals show good consistency intime-frequency characteristics Hence the establishedDTSM can efficiently simulate the actual productionenvironment

43 Fault Detection of the Scraper Chain Based on theestablished DTSM two typical failure patterns of thescraper chain are discussed namely chain jam and chainfracture ree external load conditions are set that isempty load half-load and full-load conditions e valuesof Wi (Figure 5) are defined as 0 12W0 and W0 re-spectively As designed in Figure 5 when chain jam occursthe transmission speed Vi is set as V1 in Figure 7(a) and the

failure time range is t1 minus t3(t3 25 s) the chain fault istriggered at t1 2 sWithin the time range t1 minus t2(t2 225 s)the transmission speed Vi decreases from 1 to 0ms And thevalue of Vi increases from 0 to 1ms within the time ranget2 minus t3 e whole process lasts 05 s the chain assembly istightened and the scraper chains are jammed When chainfracture occurs the transmission speed Vi is set as V0 inFigure 7(a) Moreover as shown in Figure 5 the contactconstraint between two contacting scraper chains at the labeledcontact point 1 is removed As a result the two contactingscraper chains will be separated In order to ensure the accuracyof fault setting the chain fracture is also triggered at t1 2 sand the contact constraint is removed in 05 s which is con-sistent with chain jam

Considering both failure patterns of the scraper chainour research focuses on the steady phase of the operationprocess within the time range 05 to 55 s In the steadyphase the vibration signals are obtained simultaneously atmeasuring points 1 2 and 3 Taking measuring point1 as the case study the vibration signals for chain jam andchain fracture under empty load condition are presentedin Figures 11(a) and 11(b) respectively After faultstriggering the vibration signals of the two failure patternsshow a sudden increase after a short time delay Sub-sequently the signals exhibit unstable fluctuations at thetime range 24ndash28 s In order to obtain a more detaileddescription of the detection results the maximum am-plitudes of the vibration signals under different workingconditions are discussed ie normal condition chain jam

3210

ndash1ndash2ndash3

3210

ndash1ndash2ndash3

3210

ndash1ndash2ndash3

Time (s)

Acce

lera

tion

(ms

2 )

0 01 02 03 04 05 06 07 08 09

1 2 3

(a)

3210

ndash1ndash2ndash3

3210

ndash1ndash2ndash3

3210

ndash1ndash2ndash3

Time (s)Ac

cele

ratio

n (m

s2 )

0 01 02 03 04 05 06 07 08 09

123

(b)

Figure 9 Vibration signals of different measuring points in ΔT (a) experimental results (b) simulation results

Shock and Vibration 9

Point

1

2

3

(a) (b)

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

0

Time (s) Spectrum0 03 06 09

500

400

300

200

100

00 50 100 150

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

00 03 06 09

500

400

300

200

100

00 50 100 150

Time (s) Spectrum

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

00 03 06 09

500

400

300

200

100

00 50 100 150

Time (s) Spectrum

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

00 03 06 09

500

400

300

200

100

00 50 100 150

Time (s) Spectrum

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

0

Time (s) Spectrum0 03 06 09

500

400

300

200

100

00 50 100 150

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

00 03 06 09

500

400

300

200

100

00 50 100 150

Time (s) Spectrum

Figure 10 Time-frequency representation of vibration signals (a) experimental results (b) simulation results

Chain jamoccurs

Acce

lera

tion

(ms

2 ) Acce

lera

tion

(ms

2 )

10

5

0

ndash5

ndash10

Time (s)0 05 10 15 20 25 30 35 40 45 50

10

5

0

ndash5

ndash1024 25 26 27 28

(a)

Figure 11 Continued

10 Shock and Vibration

and chain fracture Wherein the empty load half-loadand full-load conditions are considered e maximumamplitudes of the vibration signals at measuring points1 2 and 3 are depicted in Figures 12(a)ndash12(c)respectively

Considering measuring point 1 the maximum am-plitudes of the vibration signals under empty load conditionare 289 718 and 698ms2 under normal condition chainjam and chain fracture respectively Similarly under half-load condition the maximum amplitudes are 324 937 and956ms2 Moreover under full-load condition the maxi-mum amplitudes are 446 1106 and 1082ms2 Withdifferent external loads the maximum amplitudes of thevibration signals for fault conditions are obviously higherthan those for normal condition and the difference betweenthe amplitudes of the two typical failure patterns is small Fordifferent fault conditions with the increase of the externalloads the maximum amplitudes show trends to increasee above statistical results are also applicable to measuringpoints 2 and 3 erefore chain faults can easily be de-tected by comparing the maximum amplitudes of the vi-bration signals whereas the fault patterns are difficult toidentify According to the nonstationary and nonlinearcharacteristics of fault signals the AOKR is utilized to an-alyze the vibration signals and classify failure patterns of thescraper chain Within 15 s after faults triggering the vi-bration signals at the three measuring points with differentexternal loads are processed Wherein for chain jam andchain fracture under empty load condition the time-fre-quency representations of vibration signals are presented inFigures 13(a) and 13(b) respectively

e frequency components and frequency ranges ofthe same fault pattern are similar for different measuringpoints As Figure 13(a) describes the bright color between0 and 50Hz indicates one high energy area caused bychain jam en chain fracture can easily be distinguishedaccording to the appearance of two high energy areasbetween 100 and 200Hz as shown in Figure 13(b)

Observing the spectrum results a more detailed de-scription is given When chain jam occurs for measuringpoints 1 2 and 3 the high energy areas occur ap-proximately at 05 075 and 09 s respectively Mean-while for chain fracture the high energy areas include twomain frequency components and are approximatelyconcentrated at the time ranges 050ndash070 075ndash085 and10ndash115 s respectively Hence there is a delay charac-teristic of the fault occurrence which is well in accordancewith the conclusions in Section 42 In order to explore theinfluence of external loads on fault characteristics thedetailed differences of the fault patterns at measuringpoint 2 are depicted in Figure 14 In fact the externalload has a great influence on the fault severity of both thefailure patterns Observing the spectrum results thebright areas vary with the external loads With increasingexternal load the frequency ranges of the high energyareas become larger Wherein for chain jam under emptyhalf- and full-load conditions the frequency ranges areapproximately 0ndash50 0ndash150 and 0ndash250 Hz respectivelyMeanwhile for chain fracture the frequency rangesare approximately 80ndash200 50ndash250 and 50ndash350Hzrespectively

In this part three working conditions of the scraperchain are investigated above including normal conditionchain jam and chain fracture e vibration signals ofmeasuring points 1 2 and 3 on the detecting chute areanalyzed and the effects of the external loads on the vi-bration characteristics are discussed Based on the aboveanalysis the occurrence of chain faults can easily be de-termined through amplitude comparisons of the originalvibration signals However the observation confirms thesimilarity of the time domain waveforms of fault signals forchain jam and chain fracture ese two patterns of failuresremain to be different through further processing by theAOKR and the fault patterns can be distinguished accordingto the number of high energy areas of the time-frequencyrepresentation of vibration signals In conclusion the

Chain fractureoccurs

Acce

lera

tion

(ms

2 ) Acce

lera

tion

(ms

2 )

10

5

0

ndash5

ndash10

Time (s)0 05 10 15 20 25 30 35 40 45 50

10

5

0

ndash5

ndash1024 25 26 27 28

(b)

Figure 11 Vibration signals of the measuring point 1 for (a) chain jam and (b) chain fracture

Shock and Vibration 11

Point 1 2 3

(a)

(b)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

Figure 13 Time-frequency representation of vibration signals under empty load (a) chain jam (b) chain fracture

12

10

8

6

4

2

0Max

imum

acce

lera

tion

ampl

itude

(ms

2 )

Normal conditionChain jamChain fracture

0 12 W0 W0

(a)

Normal conditionChain jamChain fracture

12

10

8

6

4

2

0Max

imum

acce

lera

tion

ampl

itude

(ms

2 )

0 12 W0 W0

(b)

Normal conditionChain jamChain fracture

12

10

8

6

4

2

0Max

imum

acce

lera

tion

ampl

itude

(ms

2 )

0 12 W0 W0

(c)

Figure 12 Statistical results of the maximum vibration amplitude at different measuring points (a) 1 (b) 2 (c) 3

12 Shock and Vibration

proposed detection strategy is effective at detecting theoccurrence of chain faults and identifying the failure pat-terns under different operating conditions

5 Conclusions

During the actual operation the working state of thescraper chain can reflect the dynamic performance of thescraper conveyor To address the difficulties with directsensor measurement for parameters of the moving scraperchain a novel strategy for fault detection of the scraperchain based on vibration analysis of the chute was pro-posed Based on modal analysis and the MAC the mea-suring points of vibration signals on the chute weredetermined To fit the actual behavior of the transmissionprocess the DTSM was presented based on finite elementmodeling and the correctness of the dynamic model wasverified by comparison with the FPET en the vibrationproperties of the measuring points on the chute undernormal condition chain jam and chain fracture werediscussed Moreover the occurrence of chain faults weredetermined by comparing the amplitudes of the vibrationsignal in the time domain while the AOKR was utilizedfor time-frequency representation of vibration signals anddistinguishing the two typical failure patterns Further-more the strategy verification based on experimental datawill be taken into consideration in the near future

Data Availability

e data used to support the findings of this study areavailable from the corresponding author upon request

Conflicts of Interest

e authors declare that they have no conflicts of interest

Acknowledgments

is work was supported by the Key Project of NationalNatural Science Foundation of China (U1510205) NaturalScience Foundation of Jiangsu Province (No BK20160251)Xuzhou Research program (KC14H0138) FundamentalResearch Funds for the Central Universities (2014Y05) andProject Funded by the Priority Academic Program De-velopment of Jiangsu Higher Education Institutions(PAPD)

References

[1] C D Brown ldquoDesign build and test of a longwall armouredface conveyorrdquo Longwall Mining 2002

[2] M Dolipski P Cheluszka E Remiorz and P SobotaldquoFollow-up chain tension in an armoured face conveyornadazne napinanie lancucha zgrzebłowego W przenosnikuscianowymrdquo Archives of Mining Sciences vol 60 no 1pp 25ndash38 2015

[3] L A Morley J L Kohler and H M Smolnikar ldquoA model forpredicting motor load for an armored face-conveyor driverdquoIEEE Transactions on Industry Applications vol 24 no 4pp 649ndash659 1988

[4] A A Ordin and A A Metelrsquokov ldquoAnalysis of longwall faceoutput in screw-type cutter-loader-and-scraper conveyorsystem in underground mining of flat-lying coal bedsrdquoJournal of Mining Science vol 51 no 6 pp 1173ndash1179 2015

[5] B He G Li H Shi et al ldquoDynamic behaviour modelling andsimulation of the chain transmission system for an armouredface conveyorrdquo in Proceedings of the IEEE 10th InternationalConference on Computer-Aided Industrial Design and Con-ceptual Design CAID amp CD 2009 pp 1000ndash1004 BeijingChina November 2009

[6] R Nie B He P Yuan L Zhang and G Li ldquoNovel approachto and implementation of design and analysis of armored faceconveyor power trainrdquo Science China Technological Sciencesvol 58 no 12 pp 2153ndash2168 2015

Loads

(a)

(b)

Empty load Half-load Full-load

500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)

Figure 14 Time-frequency representation of vibration signals at measuring point 2 with different external loads (a) chain jam (b) chainfracture

Shock and Vibration 13

[7] R Nie B He L Zhang and G Li ldquoModelling of thetransmission system in conveying equipment based on Eulermethod with applicationrdquo Proceedings of the Institution ofMechanical Engineers Part K Journal of Multi-body Dy-namics vol 228 no 3 pp 294ndash306 2014

[8] S B Jiang X Zhang K D Gao J Gao Q Y Wang andK Hidenori ldquoMulti-body dynamics and vibration analysis ofchain assembly in armoured face conveyorrdquo InternationalJournal of Simulation Modelling vol 16 no 3 pp 458ndash4702017

[9] M Myszkowski and D Loehning ldquoChain force measure-ments on armoured face conveyors and coal plows in heavy-duty longwallsrdquo CIM Bulletin vol 94 no 1054 pp 72ndash752001

[10] H Wang Q Zhang and F Xie ldquoDynamic tension test andintelligent coordinated control system of a heavy scraperconveyorrdquo IET Science Measurement and Technology vol 11no 7 pp 871ndash877 2017

[11] S Sen M X Min and Y Z She ldquoDiagnosis of coal scraperconveyor based on Fuzzy Fault treerdquo in Proceedings of the2015 Seventh International Conference on Measuring Tech-nology and Mechatronics Automation (ICMTMA) pp 392ndash395 IEEE Nanchang China June 2015

[12] S-s Xue X-c Li and X-y Xu ldquoFault tree and Bayesiannetwork based scraper conveyer fault diagnosisrdquo in Pro-ceedings of the 22nd International Conference on IndustrialEngineering and Engineering Management 2015 pp 783ndash795Atlantis Press Paris France January 2016

[13] X Gong X Ma Y Zhang et al ldquoApplication of fuzzy neuralnetwork in fault diagnosis for scraper conveyor vibrationrdquo inProceedings of the 2013 IEEE International Conference onInformation and Automation (ICIA) pp 1135ndash1140 IEEEYinchuan China August 2013

[14] Y Zhang X Ma Y Jianxiang et al ldquoFuzzy neural networkfault diagnosis and online vibration monitoring system for thecoal scraper conveyor based on rough set theoryrdquo in Pro-ceedings of the 2013 32nd Chinese Control Conference (CCC)pp 6134ndash6138 IEEE Xirsquoan China July 2013

[15] B Zhang A C C Tan and J-h Lin ldquoGearbox fault diagnosisof high-speed railway trainrdquo Engineering Failure Analysisvol 66 pp 407ndash420 2016

[16] E Parloo P Verboven P Guillaume and M Van OvermeireldquoAutonomous structural health monitoring-part ii vibration-based in-operation damage assessmentrdquo Mechanical Systemsand Signal Processing vol 16 no 4 pp 659ndash675 2002

[17] C S Sakaris J S Sakellariou and S D Fassois ldquoRandom-vibration-based damage detection and precise localization ona lab-scale aircraft stabilizer structure via the GeneralizedFunctional Model Based Methodrdquo Structural Health Moni-toring An International Journal vol 16 no 5 pp 594ndash6102017

[18] Y Zhang W Song M Karimi C-H Chi and A KudreykoldquoFractional autoregressive integrated moving average andfinite-element modal the forecast of tire vibration trendrdquoIEEE Access vol 6 pp 40137ndash40142 2018

[19] M Pastor M Binda and T Harcarik ldquoModal assurancecriterionrdquo Procedia Engineering vol 48 pp 543ndash548 2012

[20] W J Staszewski K Worden and G R Tomlinson ldquoTime-frequency analysis in gearbox fault detection using the Wigner-ville distribution and pattern recognitionrdquo Mechanical Systemsand Signal Processing vol 11 no 5 pp 673ndash692 1997

[21] J-D Wu and P-H Chiang ldquoApplication of Wigner-Villedistribution and probability neural network for scooter engine

fault diagnosisrdquo Expert Systems with Applications vol 36no 2 pp 2187ndash2199 2009

[22] Z Feng and M Liang ldquoFault diagnosis of wind turbineplanetary gearbox under nonstationary conditions viaadaptive optimal kernel time-frequency analysisrdquo RenewableEnergy vol 66 pp 468ndash477 2014

14 Shock and Vibration

International Journal of

AerospaceEngineeringHindawiwwwhindawicom Volume 2018

RoboticsJournal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Active and Passive Electronic Components

VLSI Design

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Shock and Vibration

Hindawiwwwhindawicom Volume 2018

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawiwwwhindawicom

Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Control Scienceand Engineering

Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom

Journal ofEngineeringVolume 2018

SensorsJournal of

Hindawiwwwhindawicom Volume 2018

International Journal of

RotatingMachinery

Hindawiwwwhindawicom Volume 2018

Modelling ampSimulationin EngineeringHindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Navigation and Observation

International Journal of

Hindawi

wwwhindawicom Volume 2018

Advances in

Multimedia

Submit your manuscripts atwwwhindawicom

time ranges are 010ndash025 and 050ndash065 s en the vi-bration energy of measuring point 2 is concentrated at thetime ranges 015ndash045 and 060ndash075 s and the time rangesof measuring point 3 are 040ndash050 and 070ndash085 s Forthe simulation signals the basic frequencies of the mea-suring points are 281 249 and 258Hz respectively emaximum values of vibration energies are 1527 1489 and1532m2s respectively Considering the experimental re-sults and the simulation results separately good agreementbetween the main parameters of different measuring pointsis obtained It also indicates that in the same time periodthe vibration signal has a delay characteristic for measuringpoints 1 2 and 3 is is possibly because that themeasuring points are at different positions of the chuteMoreover for different measuring points the experimentalsignals and the simulation signals show good consistency intime-frequency characteristics Hence the establishedDTSM can efficiently simulate the actual productionenvironment

43 Fault Detection of the Scraper Chain Based on theestablished DTSM two typical failure patterns of thescraper chain are discussed namely chain jam and chainfracture ree external load conditions are set that isempty load half-load and full-load conditions e valuesof Wi (Figure 5) are defined as 0 12W0 and W0 re-spectively As designed in Figure 5 when chain jam occursthe transmission speed Vi is set as V1 in Figure 7(a) and the

failure time range is t1 minus t3(t3 25 s) the chain fault istriggered at t1 2 sWithin the time range t1 minus t2(t2 225 s)the transmission speed Vi decreases from 1 to 0ms And thevalue of Vi increases from 0 to 1ms within the time ranget2 minus t3 e whole process lasts 05 s the chain assembly istightened and the scraper chains are jammed When chainfracture occurs the transmission speed Vi is set as V0 inFigure 7(a) Moreover as shown in Figure 5 the contactconstraint between two contacting scraper chains at the labeledcontact point 1 is removed As a result the two contactingscraper chains will be separated In order to ensure the accuracyof fault setting the chain fracture is also triggered at t1 2 sand the contact constraint is removed in 05 s which is con-sistent with chain jam

Considering both failure patterns of the scraper chainour research focuses on the steady phase of the operationprocess within the time range 05 to 55 s In the steadyphase the vibration signals are obtained simultaneously atmeasuring points 1 2 and 3 Taking measuring point1 as the case study the vibration signals for chain jam andchain fracture under empty load condition are presentedin Figures 11(a) and 11(b) respectively After faultstriggering the vibration signals of the two failure patternsshow a sudden increase after a short time delay Sub-sequently the signals exhibit unstable fluctuations at thetime range 24ndash28 s In order to obtain a more detaileddescription of the detection results the maximum am-plitudes of the vibration signals under different workingconditions are discussed ie normal condition chain jam

3210

ndash1ndash2ndash3

3210

ndash1ndash2ndash3

3210

ndash1ndash2ndash3

Time (s)

Acce

lera

tion

(ms

2 )

0 01 02 03 04 05 06 07 08 09

1 2 3

(a)

3210

ndash1ndash2ndash3

3210

ndash1ndash2ndash3

3210

ndash1ndash2ndash3

Time (s)Ac

cele

ratio

n (m

s2 )

0 01 02 03 04 05 06 07 08 09

123

(b)

Figure 9 Vibration signals of different measuring points in ΔT (a) experimental results (b) simulation results

Shock and Vibration 9

Point

1

2

3

(a) (b)

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

0

Time (s) Spectrum0 03 06 09

500

400

300

200

100

00 50 100 150

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

00 03 06 09

500

400

300

200

100

00 50 100 150

Time (s) Spectrum

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

00 03 06 09

500

400

300

200

100

00 50 100 150

Time (s) Spectrum

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

00 03 06 09

500

400

300

200

100

00 50 100 150

Time (s) Spectrum

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

0

Time (s) Spectrum0 03 06 09

500

400

300

200

100

00 50 100 150

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

00 03 06 09

500

400

300

200

100

00 50 100 150

Time (s) Spectrum

Figure 10 Time-frequency representation of vibration signals (a) experimental results (b) simulation results

Chain jamoccurs

Acce

lera

tion

(ms

2 ) Acce

lera

tion

(ms

2 )

10

5

0

ndash5

ndash10

Time (s)0 05 10 15 20 25 30 35 40 45 50

10

5

0

ndash5

ndash1024 25 26 27 28

(a)

Figure 11 Continued

10 Shock and Vibration

and chain fracture Wherein the empty load half-loadand full-load conditions are considered e maximumamplitudes of the vibration signals at measuring points1 2 and 3 are depicted in Figures 12(a)ndash12(c)respectively

Considering measuring point 1 the maximum am-plitudes of the vibration signals under empty load conditionare 289 718 and 698ms2 under normal condition chainjam and chain fracture respectively Similarly under half-load condition the maximum amplitudes are 324 937 and956ms2 Moreover under full-load condition the maxi-mum amplitudes are 446 1106 and 1082ms2 Withdifferent external loads the maximum amplitudes of thevibration signals for fault conditions are obviously higherthan those for normal condition and the difference betweenthe amplitudes of the two typical failure patterns is small Fordifferent fault conditions with the increase of the externalloads the maximum amplitudes show trends to increasee above statistical results are also applicable to measuringpoints 2 and 3 erefore chain faults can easily be de-tected by comparing the maximum amplitudes of the vi-bration signals whereas the fault patterns are difficult toidentify According to the nonstationary and nonlinearcharacteristics of fault signals the AOKR is utilized to an-alyze the vibration signals and classify failure patterns of thescraper chain Within 15 s after faults triggering the vi-bration signals at the three measuring points with differentexternal loads are processed Wherein for chain jam andchain fracture under empty load condition the time-fre-quency representations of vibration signals are presented inFigures 13(a) and 13(b) respectively

e frequency components and frequency ranges ofthe same fault pattern are similar for different measuringpoints As Figure 13(a) describes the bright color between0 and 50Hz indicates one high energy area caused bychain jam en chain fracture can easily be distinguishedaccording to the appearance of two high energy areasbetween 100 and 200Hz as shown in Figure 13(b)

Observing the spectrum results a more detailed de-scription is given When chain jam occurs for measuringpoints 1 2 and 3 the high energy areas occur ap-proximately at 05 075 and 09 s respectively Mean-while for chain fracture the high energy areas include twomain frequency components and are approximatelyconcentrated at the time ranges 050ndash070 075ndash085 and10ndash115 s respectively Hence there is a delay charac-teristic of the fault occurrence which is well in accordancewith the conclusions in Section 42 In order to explore theinfluence of external loads on fault characteristics thedetailed differences of the fault patterns at measuringpoint 2 are depicted in Figure 14 In fact the externalload has a great influence on the fault severity of both thefailure patterns Observing the spectrum results thebright areas vary with the external loads With increasingexternal load the frequency ranges of the high energyareas become larger Wherein for chain jam under emptyhalf- and full-load conditions the frequency ranges areapproximately 0ndash50 0ndash150 and 0ndash250 Hz respectivelyMeanwhile for chain fracture the frequency rangesare approximately 80ndash200 50ndash250 and 50ndash350Hzrespectively

In this part three working conditions of the scraperchain are investigated above including normal conditionchain jam and chain fracture e vibration signals ofmeasuring points 1 2 and 3 on the detecting chute areanalyzed and the effects of the external loads on the vi-bration characteristics are discussed Based on the aboveanalysis the occurrence of chain faults can easily be de-termined through amplitude comparisons of the originalvibration signals However the observation confirms thesimilarity of the time domain waveforms of fault signals forchain jam and chain fracture ese two patterns of failuresremain to be different through further processing by theAOKR and the fault patterns can be distinguished accordingto the number of high energy areas of the time-frequencyrepresentation of vibration signals In conclusion the

Chain fractureoccurs

Acce

lera

tion

(ms

2 ) Acce

lera

tion

(ms

2 )

10

5

0

ndash5

ndash10

Time (s)0 05 10 15 20 25 30 35 40 45 50

10

5

0

ndash5

ndash1024 25 26 27 28

(b)

Figure 11 Vibration signals of the measuring point 1 for (a) chain jam and (b) chain fracture

Shock and Vibration 11

Point 1 2 3

(a)

(b)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

Figure 13 Time-frequency representation of vibration signals under empty load (a) chain jam (b) chain fracture

12

10

8

6

4

2

0Max

imum

acce

lera

tion

ampl

itude

(ms

2 )

Normal conditionChain jamChain fracture

0 12 W0 W0

(a)

Normal conditionChain jamChain fracture

12

10

8

6

4

2

0Max

imum

acce

lera

tion

ampl

itude

(ms

2 )

0 12 W0 W0

(b)

Normal conditionChain jamChain fracture

12

10

8

6

4

2

0Max

imum

acce

lera

tion

ampl

itude

(ms

2 )

0 12 W0 W0

(c)

Figure 12 Statistical results of the maximum vibration amplitude at different measuring points (a) 1 (b) 2 (c) 3

12 Shock and Vibration

proposed detection strategy is effective at detecting theoccurrence of chain faults and identifying the failure pat-terns under different operating conditions

5 Conclusions

During the actual operation the working state of thescraper chain can reflect the dynamic performance of thescraper conveyor To address the difficulties with directsensor measurement for parameters of the moving scraperchain a novel strategy for fault detection of the scraperchain based on vibration analysis of the chute was pro-posed Based on modal analysis and the MAC the mea-suring points of vibration signals on the chute weredetermined To fit the actual behavior of the transmissionprocess the DTSM was presented based on finite elementmodeling and the correctness of the dynamic model wasverified by comparison with the FPET en the vibrationproperties of the measuring points on the chute undernormal condition chain jam and chain fracture werediscussed Moreover the occurrence of chain faults weredetermined by comparing the amplitudes of the vibrationsignal in the time domain while the AOKR was utilizedfor time-frequency representation of vibration signals anddistinguishing the two typical failure patterns Further-more the strategy verification based on experimental datawill be taken into consideration in the near future

Data Availability

e data used to support the findings of this study areavailable from the corresponding author upon request

Conflicts of Interest

e authors declare that they have no conflicts of interest

Acknowledgments

is work was supported by the Key Project of NationalNatural Science Foundation of China (U1510205) NaturalScience Foundation of Jiangsu Province (No BK20160251)Xuzhou Research program (KC14H0138) FundamentalResearch Funds for the Central Universities (2014Y05) andProject Funded by the Priority Academic Program De-velopment of Jiangsu Higher Education Institutions(PAPD)

References

[1] C D Brown ldquoDesign build and test of a longwall armouredface conveyorrdquo Longwall Mining 2002

[2] M Dolipski P Cheluszka E Remiorz and P SobotaldquoFollow-up chain tension in an armoured face conveyornadazne napinanie lancucha zgrzebłowego W przenosnikuscianowymrdquo Archives of Mining Sciences vol 60 no 1pp 25ndash38 2015

[3] L A Morley J L Kohler and H M Smolnikar ldquoA model forpredicting motor load for an armored face-conveyor driverdquoIEEE Transactions on Industry Applications vol 24 no 4pp 649ndash659 1988

[4] A A Ordin and A A Metelrsquokov ldquoAnalysis of longwall faceoutput in screw-type cutter-loader-and-scraper conveyorsystem in underground mining of flat-lying coal bedsrdquoJournal of Mining Science vol 51 no 6 pp 1173ndash1179 2015

[5] B He G Li H Shi et al ldquoDynamic behaviour modelling andsimulation of the chain transmission system for an armouredface conveyorrdquo in Proceedings of the IEEE 10th InternationalConference on Computer-Aided Industrial Design and Con-ceptual Design CAID amp CD 2009 pp 1000ndash1004 BeijingChina November 2009

[6] R Nie B He P Yuan L Zhang and G Li ldquoNovel approachto and implementation of design and analysis of armored faceconveyor power trainrdquo Science China Technological Sciencesvol 58 no 12 pp 2153ndash2168 2015

Loads

(a)

(b)

Empty load Half-load Full-load

500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)

Figure 14 Time-frequency representation of vibration signals at measuring point 2 with different external loads (a) chain jam (b) chainfracture

Shock and Vibration 13

[7] R Nie B He L Zhang and G Li ldquoModelling of thetransmission system in conveying equipment based on Eulermethod with applicationrdquo Proceedings of the Institution ofMechanical Engineers Part K Journal of Multi-body Dy-namics vol 228 no 3 pp 294ndash306 2014

[8] S B Jiang X Zhang K D Gao J Gao Q Y Wang andK Hidenori ldquoMulti-body dynamics and vibration analysis ofchain assembly in armoured face conveyorrdquo InternationalJournal of Simulation Modelling vol 16 no 3 pp 458ndash4702017

[9] M Myszkowski and D Loehning ldquoChain force measure-ments on armoured face conveyors and coal plows in heavy-duty longwallsrdquo CIM Bulletin vol 94 no 1054 pp 72ndash752001

[10] H Wang Q Zhang and F Xie ldquoDynamic tension test andintelligent coordinated control system of a heavy scraperconveyorrdquo IET Science Measurement and Technology vol 11no 7 pp 871ndash877 2017

[11] S Sen M X Min and Y Z She ldquoDiagnosis of coal scraperconveyor based on Fuzzy Fault treerdquo in Proceedings of the2015 Seventh International Conference on Measuring Tech-nology and Mechatronics Automation (ICMTMA) pp 392ndash395 IEEE Nanchang China June 2015

[12] S-s Xue X-c Li and X-y Xu ldquoFault tree and Bayesiannetwork based scraper conveyer fault diagnosisrdquo in Pro-ceedings of the 22nd International Conference on IndustrialEngineering and Engineering Management 2015 pp 783ndash795Atlantis Press Paris France January 2016

[13] X Gong X Ma Y Zhang et al ldquoApplication of fuzzy neuralnetwork in fault diagnosis for scraper conveyor vibrationrdquo inProceedings of the 2013 IEEE International Conference onInformation and Automation (ICIA) pp 1135ndash1140 IEEEYinchuan China August 2013

[14] Y Zhang X Ma Y Jianxiang et al ldquoFuzzy neural networkfault diagnosis and online vibration monitoring system for thecoal scraper conveyor based on rough set theoryrdquo in Pro-ceedings of the 2013 32nd Chinese Control Conference (CCC)pp 6134ndash6138 IEEE Xirsquoan China July 2013

[15] B Zhang A C C Tan and J-h Lin ldquoGearbox fault diagnosisof high-speed railway trainrdquo Engineering Failure Analysisvol 66 pp 407ndash420 2016

[16] E Parloo P Verboven P Guillaume and M Van OvermeireldquoAutonomous structural health monitoring-part ii vibration-based in-operation damage assessmentrdquo Mechanical Systemsand Signal Processing vol 16 no 4 pp 659ndash675 2002

[17] C S Sakaris J S Sakellariou and S D Fassois ldquoRandom-vibration-based damage detection and precise localization ona lab-scale aircraft stabilizer structure via the GeneralizedFunctional Model Based Methodrdquo Structural Health Moni-toring An International Journal vol 16 no 5 pp 594ndash6102017

[18] Y Zhang W Song M Karimi C-H Chi and A KudreykoldquoFractional autoregressive integrated moving average andfinite-element modal the forecast of tire vibration trendrdquoIEEE Access vol 6 pp 40137ndash40142 2018

[19] M Pastor M Binda and T Harcarik ldquoModal assurancecriterionrdquo Procedia Engineering vol 48 pp 543ndash548 2012

[20] W J Staszewski K Worden and G R Tomlinson ldquoTime-frequency analysis in gearbox fault detection using the Wigner-ville distribution and pattern recognitionrdquo Mechanical Systemsand Signal Processing vol 11 no 5 pp 673ndash692 1997

[21] J-D Wu and P-H Chiang ldquoApplication of Wigner-Villedistribution and probability neural network for scooter engine

fault diagnosisrdquo Expert Systems with Applications vol 36no 2 pp 2187ndash2199 2009

[22] Z Feng and M Liang ldquoFault diagnosis of wind turbineplanetary gearbox under nonstationary conditions viaadaptive optimal kernel time-frequency analysisrdquo RenewableEnergy vol 66 pp 468ndash477 2014

14 Shock and Vibration

International Journal of

AerospaceEngineeringHindawiwwwhindawicom Volume 2018

RoboticsJournal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Active and Passive Electronic Components

VLSI Design

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Shock and Vibration

Hindawiwwwhindawicom Volume 2018

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawiwwwhindawicom

Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Control Scienceand Engineering

Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom

Journal ofEngineeringVolume 2018

SensorsJournal of

Hindawiwwwhindawicom Volume 2018

International Journal of

RotatingMachinery

Hindawiwwwhindawicom Volume 2018

Modelling ampSimulationin EngineeringHindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Navigation and Observation

International Journal of

Hindawi

wwwhindawicom Volume 2018

Advances in

Multimedia

Submit your manuscripts atwwwhindawicom

Point

1

2

3

(a) (b)

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

0

Time (s) Spectrum0 03 06 09

500

400

300

200

100

00 50 100 150

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

00 03 06 09

500

400

300

200

100

00 50 100 150

Time (s) Spectrum

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

00 03 06 09

500

400

300

200

100

00 50 100 150

Time (s) Spectrum

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

00 03 06 09

500

400

300

200

100

00 50 100 150

Time (s) Spectrum

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

0

Time (s) Spectrum0 03 06 09

500

400

300

200

100

00 50 100 150

Freq

uenc

y (H

z)

Freq

uenc

y (H

z)

500

400

300

200

100

00 03 06 09

500

400

300

200

100

00 50 100 150

Time (s) Spectrum

Figure 10 Time-frequency representation of vibration signals (a) experimental results (b) simulation results

Chain jamoccurs

Acce

lera

tion

(ms

2 ) Acce

lera

tion

(ms

2 )

10

5

0

ndash5

ndash10

Time (s)0 05 10 15 20 25 30 35 40 45 50

10

5

0

ndash5

ndash1024 25 26 27 28

(a)

Figure 11 Continued

10 Shock and Vibration

and chain fracture Wherein the empty load half-loadand full-load conditions are considered e maximumamplitudes of the vibration signals at measuring points1 2 and 3 are depicted in Figures 12(a)ndash12(c)respectively

Considering measuring point 1 the maximum am-plitudes of the vibration signals under empty load conditionare 289 718 and 698ms2 under normal condition chainjam and chain fracture respectively Similarly under half-load condition the maximum amplitudes are 324 937 and956ms2 Moreover under full-load condition the maxi-mum amplitudes are 446 1106 and 1082ms2 Withdifferent external loads the maximum amplitudes of thevibration signals for fault conditions are obviously higherthan those for normal condition and the difference betweenthe amplitudes of the two typical failure patterns is small Fordifferent fault conditions with the increase of the externalloads the maximum amplitudes show trends to increasee above statistical results are also applicable to measuringpoints 2 and 3 erefore chain faults can easily be de-tected by comparing the maximum amplitudes of the vi-bration signals whereas the fault patterns are difficult toidentify According to the nonstationary and nonlinearcharacteristics of fault signals the AOKR is utilized to an-alyze the vibration signals and classify failure patterns of thescraper chain Within 15 s after faults triggering the vi-bration signals at the three measuring points with differentexternal loads are processed Wherein for chain jam andchain fracture under empty load condition the time-fre-quency representations of vibration signals are presented inFigures 13(a) and 13(b) respectively

e frequency components and frequency ranges ofthe same fault pattern are similar for different measuringpoints As Figure 13(a) describes the bright color between0 and 50Hz indicates one high energy area caused bychain jam en chain fracture can easily be distinguishedaccording to the appearance of two high energy areasbetween 100 and 200Hz as shown in Figure 13(b)

Observing the spectrum results a more detailed de-scription is given When chain jam occurs for measuringpoints 1 2 and 3 the high energy areas occur ap-proximately at 05 075 and 09 s respectively Mean-while for chain fracture the high energy areas include twomain frequency components and are approximatelyconcentrated at the time ranges 050ndash070 075ndash085 and10ndash115 s respectively Hence there is a delay charac-teristic of the fault occurrence which is well in accordancewith the conclusions in Section 42 In order to explore theinfluence of external loads on fault characteristics thedetailed differences of the fault patterns at measuringpoint 2 are depicted in Figure 14 In fact the externalload has a great influence on the fault severity of both thefailure patterns Observing the spectrum results thebright areas vary with the external loads With increasingexternal load the frequency ranges of the high energyareas become larger Wherein for chain jam under emptyhalf- and full-load conditions the frequency ranges areapproximately 0ndash50 0ndash150 and 0ndash250 Hz respectivelyMeanwhile for chain fracture the frequency rangesare approximately 80ndash200 50ndash250 and 50ndash350Hzrespectively

In this part three working conditions of the scraperchain are investigated above including normal conditionchain jam and chain fracture e vibration signals ofmeasuring points 1 2 and 3 on the detecting chute areanalyzed and the effects of the external loads on the vi-bration characteristics are discussed Based on the aboveanalysis the occurrence of chain faults can easily be de-termined through amplitude comparisons of the originalvibration signals However the observation confirms thesimilarity of the time domain waveforms of fault signals forchain jam and chain fracture ese two patterns of failuresremain to be different through further processing by theAOKR and the fault patterns can be distinguished accordingto the number of high energy areas of the time-frequencyrepresentation of vibration signals In conclusion the

Chain fractureoccurs

Acce

lera

tion

(ms

2 ) Acce

lera

tion

(ms

2 )

10

5

0

ndash5

ndash10

Time (s)0 05 10 15 20 25 30 35 40 45 50

10

5

0

ndash5

ndash1024 25 26 27 28

(b)

Figure 11 Vibration signals of the measuring point 1 for (a) chain jam and (b) chain fracture

Shock and Vibration 11

Point 1 2 3

(a)

(b)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

Figure 13 Time-frequency representation of vibration signals under empty load (a) chain jam (b) chain fracture

12

10

8

6

4

2

0Max

imum

acce

lera

tion

ampl

itude

(ms

2 )

Normal conditionChain jamChain fracture

0 12 W0 W0

(a)

Normal conditionChain jamChain fracture

12

10

8

6

4

2

0Max

imum

acce

lera

tion

ampl

itude

(ms

2 )

0 12 W0 W0

(b)

Normal conditionChain jamChain fracture

12

10

8

6

4

2

0Max

imum

acce

lera

tion

ampl

itude

(ms

2 )

0 12 W0 W0

(c)

Figure 12 Statistical results of the maximum vibration amplitude at different measuring points (a) 1 (b) 2 (c) 3

12 Shock and Vibration

proposed detection strategy is effective at detecting theoccurrence of chain faults and identifying the failure pat-terns under different operating conditions

5 Conclusions

During the actual operation the working state of thescraper chain can reflect the dynamic performance of thescraper conveyor To address the difficulties with directsensor measurement for parameters of the moving scraperchain a novel strategy for fault detection of the scraperchain based on vibration analysis of the chute was pro-posed Based on modal analysis and the MAC the mea-suring points of vibration signals on the chute weredetermined To fit the actual behavior of the transmissionprocess the DTSM was presented based on finite elementmodeling and the correctness of the dynamic model wasverified by comparison with the FPET en the vibrationproperties of the measuring points on the chute undernormal condition chain jam and chain fracture werediscussed Moreover the occurrence of chain faults weredetermined by comparing the amplitudes of the vibrationsignal in the time domain while the AOKR was utilizedfor time-frequency representation of vibration signals anddistinguishing the two typical failure patterns Further-more the strategy verification based on experimental datawill be taken into consideration in the near future

Data Availability

e data used to support the findings of this study areavailable from the corresponding author upon request

Conflicts of Interest

e authors declare that they have no conflicts of interest

Acknowledgments

is work was supported by the Key Project of NationalNatural Science Foundation of China (U1510205) NaturalScience Foundation of Jiangsu Province (No BK20160251)Xuzhou Research program (KC14H0138) FundamentalResearch Funds for the Central Universities (2014Y05) andProject Funded by the Priority Academic Program De-velopment of Jiangsu Higher Education Institutions(PAPD)

References

[1] C D Brown ldquoDesign build and test of a longwall armouredface conveyorrdquo Longwall Mining 2002

[2] M Dolipski P Cheluszka E Remiorz and P SobotaldquoFollow-up chain tension in an armoured face conveyornadazne napinanie lancucha zgrzebłowego W przenosnikuscianowymrdquo Archives of Mining Sciences vol 60 no 1pp 25ndash38 2015

[3] L A Morley J L Kohler and H M Smolnikar ldquoA model forpredicting motor load for an armored face-conveyor driverdquoIEEE Transactions on Industry Applications vol 24 no 4pp 649ndash659 1988

[4] A A Ordin and A A Metelrsquokov ldquoAnalysis of longwall faceoutput in screw-type cutter-loader-and-scraper conveyorsystem in underground mining of flat-lying coal bedsrdquoJournal of Mining Science vol 51 no 6 pp 1173ndash1179 2015

[5] B He G Li H Shi et al ldquoDynamic behaviour modelling andsimulation of the chain transmission system for an armouredface conveyorrdquo in Proceedings of the IEEE 10th InternationalConference on Computer-Aided Industrial Design and Con-ceptual Design CAID amp CD 2009 pp 1000ndash1004 BeijingChina November 2009

[6] R Nie B He P Yuan L Zhang and G Li ldquoNovel approachto and implementation of design and analysis of armored faceconveyor power trainrdquo Science China Technological Sciencesvol 58 no 12 pp 2153ndash2168 2015

Loads

(a)

(b)

Empty load Half-load Full-load

500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)

Figure 14 Time-frequency representation of vibration signals at measuring point 2 with different external loads (a) chain jam (b) chainfracture

Shock and Vibration 13

[7] R Nie B He L Zhang and G Li ldquoModelling of thetransmission system in conveying equipment based on Eulermethod with applicationrdquo Proceedings of the Institution ofMechanical Engineers Part K Journal of Multi-body Dy-namics vol 228 no 3 pp 294ndash306 2014

[8] S B Jiang X Zhang K D Gao J Gao Q Y Wang andK Hidenori ldquoMulti-body dynamics and vibration analysis ofchain assembly in armoured face conveyorrdquo InternationalJournal of Simulation Modelling vol 16 no 3 pp 458ndash4702017

[9] M Myszkowski and D Loehning ldquoChain force measure-ments on armoured face conveyors and coal plows in heavy-duty longwallsrdquo CIM Bulletin vol 94 no 1054 pp 72ndash752001

[10] H Wang Q Zhang and F Xie ldquoDynamic tension test andintelligent coordinated control system of a heavy scraperconveyorrdquo IET Science Measurement and Technology vol 11no 7 pp 871ndash877 2017

[11] S Sen M X Min and Y Z She ldquoDiagnosis of coal scraperconveyor based on Fuzzy Fault treerdquo in Proceedings of the2015 Seventh International Conference on Measuring Tech-nology and Mechatronics Automation (ICMTMA) pp 392ndash395 IEEE Nanchang China June 2015

[12] S-s Xue X-c Li and X-y Xu ldquoFault tree and Bayesiannetwork based scraper conveyer fault diagnosisrdquo in Pro-ceedings of the 22nd International Conference on IndustrialEngineering and Engineering Management 2015 pp 783ndash795Atlantis Press Paris France January 2016

[13] X Gong X Ma Y Zhang et al ldquoApplication of fuzzy neuralnetwork in fault diagnosis for scraper conveyor vibrationrdquo inProceedings of the 2013 IEEE International Conference onInformation and Automation (ICIA) pp 1135ndash1140 IEEEYinchuan China August 2013

[14] Y Zhang X Ma Y Jianxiang et al ldquoFuzzy neural networkfault diagnosis and online vibration monitoring system for thecoal scraper conveyor based on rough set theoryrdquo in Pro-ceedings of the 2013 32nd Chinese Control Conference (CCC)pp 6134ndash6138 IEEE Xirsquoan China July 2013

[15] B Zhang A C C Tan and J-h Lin ldquoGearbox fault diagnosisof high-speed railway trainrdquo Engineering Failure Analysisvol 66 pp 407ndash420 2016

[16] E Parloo P Verboven P Guillaume and M Van OvermeireldquoAutonomous structural health monitoring-part ii vibration-based in-operation damage assessmentrdquo Mechanical Systemsand Signal Processing vol 16 no 4 pp 659ndash675 2002

[17] C S Sakaris J S Sakellariou and S D Fassois ldquoRandom-vibration-based damage detection and precise localization ona lab-scale aircraft stabilizer structure via the GeneralizedFunctional Model Based Methodrdquo Structural Health Moni-toring An International Journal vol 16 no 5 pp 594ndash6102017

[18] Y Zhang W Song M Karimi C-H Chi and A KudreykoldquoFractional autoregressive integrated moving average andfinite-element modal the forecast of tire vibration trendrdquoIEEE Access vol 6 pp 40137ndash40142 2018

[19] M Pastor M Binda and T Harcarik ldquoModal assurancecriterionrdquo Procedia Engineering vol 48 pp 543ndash548 2012

[20] W J Staszewski K Worden and G R Tomlinson ldquoTime-frequency analysis in gearbox fault detection using the Wigner-ville distribution and pattern recognitionrdquo Mechanical Systemsand Signal Processing vol 11 no 5 pp 673ndash692 1997

[21] J-D Wu and P-H Chiang ldquoApplication of Wigner-Villedistribution and probability neural network for scooter engine

fault diagnosisrdquo Expert Systems with Applications vol 36no 2 pp 2187ndash2199 2009

[22] Z Feng and M Liang ldquoFault diagnosis of wind turbineplanetary gearbox under nonstationary conditions viaadaptive optimal kernel time-frequency analysisrdquo RenewableEnergy vol 66 pp 468ndash477 2014

14 Shock and Vibration

International Journal of

AerospaceEngineeringHindawiwwwhindawicom Volume 2018

RoboticsJournal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Active and Passive Electronic Components

VLSI Design

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Shock and Vibration

Hindawiwwwhindawicom Volume 2018

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawiwwwhindawicom

Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Control Scienceand Engineering

Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom

Journal ofEngineeringVolume 2018

SensorsJournal of

Hindawiwwwhindawicom Volume 2018

International Journal of

RotatingMachinery

Hindawiwwwhindawicom Volume 2018

Modelling ampSimulationin EngineeringHindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Navigation and Observation

International Journal of

Hindawi

wwwhindawicom Volume 2018

Advances in

Multimedia

Submit your manuscripts atwwwhindawicom

and chain fracture Wherein the empty load half-loadand full-load conditions are considered e maximumamplitudes of the vibration signals at measuring points1 2 and 3 are depicted in Figures 12(a)ndash12(c)respectively

Considering measuring point 1 the maximum am-plitudes of the vibration signals under empty load conditionare 289 718 and 698ms2 under normal condition chainjam and chain fracture respectively Similarly under half-load condition the maximum amplitudes are 324 937 and956ms2 Moreover under full-load condition the maxi-mum amplitudes are 446 1106 and 1082ms2 Withdifferent external loads the maximum amplitudes of thevibration signals for fault conditions are obviously higherthan those for normal condition and the difference betweenthe amplitudes of the two typical failure patterns is small Fordifferent fault conditions with the increase of the externalloads the maximum amplitudes show trends to increasee above statistical results are also applicable to measuringpoints 2 and 3 erefore chain faults can easily be de-tected by comparing the maximum amplitudes of the vi-bration signals whereas the fault patterns are difficult toidentify According to the nonstationary and nonlinearcharacteristics of fault signals the AOKR is utilized to an-alyze the vibration signals and classify failure patterns of thescraper chain Within 15 s after faults triggering the vi-bration signals at the three measuring points with differentexternal loads are processed Wherein for chain jam andchain fracture under empty load condition the time-fre-quency representations of vibration signals are presented inFigures 13(a) and 13(b) respectively

e frequency components and frequency ranges ofthe same fault pattern are similar for different measuringpoints As Figure 13(a) describes the bright color between0 and 50Hz indicates one high energy area caused bychain jam en chain fracture can easily be distinguishedaccording to the appearance of two high energy areasbetween 100 and 200Hz as shown in Figure 13(b)

Observing the spectrum results a more detailed de-scription is given When chain jam occurs for measuringpoints 1 2 and 3 the high energy areas occur ap-proximately at 05 075 and 09 s respectively Mean-while for chain fracture the high energy areas include twomain frequency components and are approximatelyconcentrated at the time ranges 050ndash070 075ndash085 and10ndash115 s respectively Hence there is a delay charac-teristic of the fault occurrence which is well in accordancewith the conclusions in Section 42 In order to explore theinfluence of external loads on fault characteristics thedetailed differences of the fault patterns at measuringpoint 2 are depicted in Figure 14 In fact the externalload has a great influence on the fault severity of both thefailure patterns Observing the spectrum results thebright areas vary with the external loads With increasingexternal load the frequency ranges of the high energyareas become larger Wherein for chain jam under emptyhalf- and full-load conditions the frequency ranges areapproximately 0ndash50 0ndash150 and 0ndash250 Hz respectivelyMeanwhile for chain fracture the frequency rangesare approximately 80ndash200 50ndash250 and 50ndash350Hzrespectively

In this part three working conditions of the scraperchain are investigated above including normal conditionchain jam and chain fracture e vibration signals ofmeasuring points 1 2 and 3 on the detecting chute areanalyzed and the effects of the external loads on the vi-bration characteristics are discussed Based on the aboveanalysis the occurrence of chain faults can easily be de-termined through amplitude comparisons of the originalvibration signals However the observation confirms thesimilarity of the time domain waveforms of fault signals forchain jam and chain fracture ese two patterns of failuresremain to be different through further processing by theAOKR and the fault patterns can be distinguished accordingto the number of high energy areas of the time-frequencyrepresentation of vibration signals In conclusion the

Chain fractureoccurs

Acce

lera

tion

(ms

2 ) Acce

lera

tion

(ms

2 )

10

5

0

ndash5

ndash10

Time (s)0 05 10 15 20 25 30 35 40 45 50

10

5

0

ndash5

ndash1024 25 26 27 28

(b)

Figure 11 Vibration signals of the measuring point 1 for (a) chain jam and (b) chain fracture

Shock and Vibration 11

Point 1 2 3

(a)

(b)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

Figure 13 Time-frequency representation of vibration signals under empty load (a) chain jam (b) chain fracture

12

10

8

6

4

2

0Max

imum

acce

lera

tion

ampl

itude

(ms

2 )

Normal conditionChain jamChain fracture

0 12 W0 W0

(a)

Normal conditionChain jamChain fracture

12

10

8

6

4

2

0Max

imum

acce

lera

tion

ampl

itude

(ms

2 )

0 12 W0 W0

(b)

Normal conditionChain jamChain fracture

12

10

8

6

4

2

0Max

imum

acce

lera

tion

ampl

itude

(ms

2 )

0 12 W0 W0

(c)

Figure 12 Statistical results of the maximum vibration amplitude at different measuring points (a) 1 (b) 2 (c) 3

12 Shock and Vibration

proposed detection strategy is effective at detecting theoccurrence of chain faults and identifying the failure pat-terns under different operating conditions

5 Conclusions

During the actual operation the working state of thescraper chain can reflect the dynamic performance of thescraper conveyor To address the difficulties with directsensor measurement for parameters of the moving scraperchain a novel strategy for fault detection of the scraperchain based on vibration analysis of the chute was pro-posed Based on modal analysis and the MAC the mea-suring points of vibration signals on the chute weredetermined To fit the actual behavior of the transmissionprocess the DTSM was presented based on finite elementmodeling and the correctness of the dynamic model wasverified by comparison with the FPET en the vibrationproperties of the measuring points on the chute undernormal condition chain jam and chain fracture werediscussed Moreover the occurrence of chain faults weredetermined by comparing the amplitudes of the vibrationsignal in the time domain while the AOKR was utilizedfor time-frequency representation of vibration signals anddistinguishing the two typical failure patterns Further-more the strategy verification based on experimental datawill be taken into consideration in the near future

Data Availability

e data used to support the findings of this study areavailable from the corresponding author upon request

Conflicts of Interest

e authors declare that they have no conflicts of interest

Acknowledgments

is work was supported by the Key Project of NationalNatural Science Foundation of China (U1510205) NaturalScience Foundation of Jiangsu Province (No BK20160251)Xuzhou Research program (KC14H0138) FundamentalResearch Funds for the Central Universities (2014Y05) andProject Funded by the Priority Academic Program De-velopment of Jiangsu Higher Education Institutions(PAPD)

References

[1] C D Brown ldquoDesign build and test of a longwall armouredface conveyorrdquo Longwall Mining 2002

[2] M Dolipski P Cheluszka E Remiorz and P SobotaldquoFollow-up chain tension in an armoured face conveyornadazne napinanie lancucha zgrzebłowego W przenosnikuscianowymrdquo Archives of Mining Sciences vol 60 no 1pp 25ndash38 2015

[3] L A Morley J L Kohler and H M Smolnikar ldquoA model forpredicting motor load for an armored face-conveyor driverdquoIEEE Transactions on Industry Applications vol 24 no 4pp 649ndash659 1988

[4] A A Ordin and A A Metelrsquokov ldquoAnalysis of longwall faceoutput in screw-type cutter-loader-and-scraper conveyorsystem in underground mining of flat-lying coal bedsrdquoJournal of Mining Science vol 51 no 6 pp 1173ndash1179 2015

[5] B He G Li H Shi et al ldquoDynamic behaviour modelling andsimulation of the chain transmission system for an armouredface conveyorrdquo in Proceedings of the IEEE 10th InternationalConference on Computer-Aided Industrial Design and Con-ceptual Design CAID amp CD 2009 pp 1000ndash1004 BeijingChina November 2009

[6] R Nie B He P Yuan L Zhang and G Li ldquoNovel approachto and implementation of design and analysis of armored faceconveyor power trainrdquo Science China Technological Sciencesvol 58 no 12 pp 2153ndash2168 2015

Loads

(a)

(b)

Empty load Half-load Full-load

500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)

Figure 14 Time-frequency representation of vibration signals at measuring point 2 with different external loads (a) chain jam (b) chainfracture

Shock and Vibration 13

[7] R Nie B He L Zhang and G Li ldquoModelling of thetransmission system in conveying equipment based on Eulermethod with applicationrdquo Proceedings of the Institution ofMechanical Engineers Part K Journal of Multi-body Dy-namics vol 228 no 3 pp 294ndash306 2014

[8] S B Jiang X Zhang K D Gao J Gao Q Y Wang andK Hidenori ldquoMulti-body dynamics and vibration analysis ofchain assembly in armoured face conveyorrdquo InternationalJournal of Simulation Modelling vol 16 no 3 pp 458ndash4702017

[9] M Myszkowski and D Loehning ldquoChain force measure-ments on armoured face conveyors and coal plows in heavy-duty longwallsrdquo CIM Bulletin vol 94 no 1054 pp 72ndash752001

[10] H Wang Q Zhang and F Xie ldquoDynamic tension test andintelligent coordinated control system of a heavy scraperconveyorrdquo IET Science Measurement and Technology vol 11no 7 pp 871ndash877 2017

[11] S Sen M X Min and Y Z She ldquoDiagnosis of coal scraperconveyor based on Fuzzy Fault treerdquo in Proceedings of the2015 Seventh International Conference on Measuring Tech-nology and Mechatronics Automation (ICMTMA) pp 392ndash395 IEEE Nanchang China June 2015

[12] S-s Xue X-c Li and X-y Xu ldquoFault tree and Bayesiannetwork based scraper conveyer fault diagnosisrdquo in Pro-ceedings of the 22nd International Conference on IndustrialEngineering and Engineering Management 2015 pp 783ndash795Atlantis Press Paris France January 2016

[13] X Gong X Ma Y Zhang et al ldquoApplication of fuzzy neuralnetwork in fault diagnosis for scraper conveyor vibrationrdquo inProceedings of the 2013 IEEE International Conference onInformation and Automation (ICIA) pp 1135ndash1140 IEEEYinchuan China August 2013

[14] Y Zhang X Ma Y Jianxiang et al ldquoFuzzy neural networkfault diagnosis and online vibration monitoring system for thecoal scraper conveyor based on rough set theoryrdquo in Pro-ceedings of the 2013 32nd Chinese Control Conference (CCC)pp 6134ndash6138 IEEE Xirsquoan China July 2013

[15] B Zhang A C C Tan and J-h Lin ldquoGearbox fault diagnosisof high-speed railway trainrdquo Engineering Failure Analysisvol 66 pp 407ndash420 2016

[16] E Parloo P Verboven P Guillaume and M Van OvermeireldquoAutonomous structural health monitoring-part ii vibration-based in-operation damage assessmentrdquo Mechanical Systemsand Signal Processing vol 16 no 4 pp 659ndash675 2002

[17] C S Sakaris J S Sakellariou and S D Fassois ldquoRandom-vibration-based damage detection and precise localization ona lab-scale aircraft stabilizer structure via the GeneralizedFunctional Model Based Methodrdquo Structural Health Moni-toring An International Journal vol 16 no 5 pp 594ndash6102017

[18] Y Zhang W Song M Karimi C-H Chi and A KudreykoldquoFractional autoregressive integrated moving average andfinite-element modal the forecast of tire vibration trendrdquoIEEE Access vol 6 pp 40137ndash40142 2018

[19] M Pastor M Binda and T Harcarik ldquoModal assurancecriterionrdquo Procedia Engineering vol 48 pp 543ndash548 2012

[20] W J Staszewski K Worden and G R Tomlinson ldquoTime-frequency analysis in gearbox fault detection using the Wigner-ville distribution and pattern recognitionrdquo Mechanical Systemsand Signal Processing vol 11 no 5 pp 673ndash692 1997

[21] J-D Wu and P-H Chiang ldquoApplication of Wigner-Villedistribution and probability neural network for scooter engine

fault diagnosisrdquo Expert Systems with Applications vol 36no 2 pp 2187ndash2199 2009

[22] Z Feng and M Liang ldquoFault diagnosis of wind turbineplanetary gearbox under nonstationary conditions viaadaptive optimal kernel time-frequency analysisrdquo RenewableEnergy vol 66 pp 468ndash477 2014

14 Shock and Vibration

International Journal of

AerospaceEngineeringHindawiwwwhindawicom Volume 2018

RoboticsJournal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Active and Passive Electronic Components

VLSI Design

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Shock and Vibration

Hindawiwwwhindawicom Volume 2018

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawiwwwhindawicom

Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Control Scienceand Engineering

Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom

Journal ofEngineeringVolume 2018

SensorsJournal of

Hindawiwwwhindawicom Volume 2018

International Journal of

RotatingMachinery

Hindawiwwwhindawicom Volume 2018

Modelling ampSimulationin EngineeringHindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Navigation and Observation

International Journal of

Hindawi

wwwhindawicom Volume 2018

Advances in

Multimedia

Submit your manuscripts atwwwhindawicom

Point 1 2 3

(a)

(b)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

Figure 13 Time-frequency representation of vibration signals under empty load (a) chain jam (b) chain fracture

12

10

8

6

4

2

0Max

imum

acce

lera

tion

ampl

itude

(ms

2 )

Normal conditionChain jamChain fracture

0 12 W0 W0

(a)

Normal conditionChain jamChain fracture

12

10

8

6

4

2

0Max

imum

acce

lera

tion

ampl

itude

(ms

2 )

0 12 W0 W0

(b)

Normal conditionChain jamChain fracture

12

10

8

6

4

2

0Max

imum

acce

lera

tion

ampl

itude

(ms

2 )

0 12 W0 W0

(c)

Figure 12 Statistical results of the maximum vibration amplitude at different measuring points (a) 1 (b) 2 (c) 3

12 Shock and Vibration

proposed detection strategy is effective at detecting theoccurrence of chain faults and identifying the failure pat-terns under different operating conditions

5 Conclusions

During the actual operation the working state of thescraper chain can reflect the dynamic performance of thescraper conveyor To address the difficulties with directsensor measurement for parameters of the moving scraperchain a novel strategy for fault detection of the scraperchain based on vibration analysis of the chute was pro-posed Based on modal analysis and the MAC the mea-suring points of vibration signals on the chute weredetermined To fit the actual behavior of the transmissionprocess the DTSM was presented based on finite elementmodeling and the correctness of the dynamic model wasverified by comparison with the FPET en the vibrationproperties of the measuring points on the chute undernormal condition chain jam and chain fracture werediscussed Moreover the occurrence of chain faults weredetermined by comparing the amplitudes of the vibrationsignal in the time domain while the AOKR was utilizedfor time-frequency representation of vibration signals anddistinguishing the two typical failure patterns Further-more the strategy verification based on experimental datawill be taken into consideration in the near future

Data Availability

e data used to support the findings of this study areavailable from the corresponding author upon request

Conflicts of Interest

e authors declare that they have no conflicts of interest

Acknowledgments

is work was supported by the Key Project of NationalNatural Science Foundation of China (U1510205) NaturalScience Foundation of Jiangsu Province (No BK20160251)Xuzhou Research program (KC14H0138) FundamentalResearch Funds for the Central Universities (2014Y05) andProject Funded by the Priority Academic Program De-velopment of Jiangsu Higher Education Institutions(PAPD)

References

[1] C D Brown ldquoDesign build and test of a longwall armouredface conveyorrdquo Longwall Mining 2002

[2] M Dolipski P Cheluszka E Remiorz and P SobotaldquoFollow-up chain tension in an armoured face conveyornadazne napinanie lancucha zgrzebłowego W przenosnikuscianowymrdquo Archives of Mining Sciences vol 60 no 1pp 25ndash38 2015

[3] L A Morley J L Kohler and H M Smolnikar ldquoA model forpredicting motor load for an armored face-conveyor driverdquoIEEE Transactions on Industry Applications vol 24 no 4pp 649ndash659 1988

[4] A A Ordin and A A Metelrsquokov ldquoAnalysis of longwall faceoutput in screw-type cutter-loader-and-scraper conveyorsystem in underground mining of flat-lying coal bedsrdquoJournal of Mining Science vol 51 no 6 pp 1173ndash1179 2015

[5] B He G Li H Shi et al ldquoDynamic behaviour modelling andsimulation of the chain transmission system for an armouredface conveyorrdquo in Proceedings of the IEEE 10th InternationalConference on Computer-Aided Industrial Design and Con-ceptual Design CAID amp CD 2009 pp 1000ndash1004 BeijingChina November 2009

[6] R Nie B He P Yuan L Zhang and G Li ldquoNovel approachto and implementation of design and analysis of armored faceconveyor power trainrdquo Science China Technological Sciencesvol 58 no 12 pp 2153ndash2168 2015

Loads

(a)

(b)

Empty load Half-load Full-load

500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)

Figure 14 Time-frequency representation of vibration signals at measuring point 2 with different external loads (a) chain jam (b) chainfracture

Shock and Vibration 13

[7] R Nie B He L Zhang and G Li ldquoModelling of thetransmission system in conveying equipment based on Eulermethod with applicationrdquo Proceedings of the Institution ofMechanical Engineers Part K Journal of Multi-body Dy-namics vol 228 no 3 pp 294ndash306 2014

[8] S B Jiang X Zhang K D Gao J Gao Q Y Wang andK Hidenori ldquoMulti-body dynamics and vibration analysis ofchain assembly in armoured face conveyorrdquo InternationalJournal of Simulation Modelling vol 16 no 3 pp 458ndash4702017

[9] M Myszkowski and D Loehning ldquoChain force measure-ments on armoured face conveyors and coal plows in heavy-duty longwallsrdquo CIM Bulletin vol 94 no 1054 pp 72ndash752001

[10] H Wang Q Zhang and F Xie ldquoDynamic tension test andintelligent coordinated control system of a heavy scraperconveyorrdquo IET Science Measurement and Technology vol 11no 7 pp 871ndash877 2017

[11] S Sen M X Min and Y Z She ldquoDiagnosis of coal scraperconveyor based on Fuzzy Fault treerdquo in Proceedings of the2015 Seventh International Conference on Measuring Tech-nology and Mechatronics Automation (ICMTMA) pp 392ndash395 IEEE Nanchang China June 2015

[12] S-s Xue X-c Li and X-y Xu ldquoFault tree and Bayesiannetwork based scraper conveyer fault diagnosisrdquo in Pro-ceedings of the 22nd International Conference on IndustrialEngineering and Engineering Management 2015 pp 783ndash795Atlantis Press Paris France January 2016

[13] X Gong X Ma Y Zhang et al ldquoApplication of fuzzy neuralnetwork in fault diagnosis for scraper conveyor vibrationrdquo inProceedings of the 2013 IEEE International Conference onInformation and Automation (ICIA) pp 1135ndash1140 IEEEYinchuan China August 2013

[14] Y Zhang X Ma Y Jianxiang et al ldquoFuzzy neural networkfault diagnosis and online vibration monitoring system for thecoal scraper conveyor based on rough set theoryrdquo in Pro-ceedings of the 2013 32nd Chinese Control Conference (CCC)pp 6134ndash6138 IEEE Xirsquoan China July 2013

[15] B Zhang A C C Tan and J-h Lin ldquoGearbox fault diagnosisof high-speed railway trainrdquo Engineering Failure Analysisvol 66 pp 407ndash420 2016

[16] E Parloo P Verboven P Guillaume and M Van OvermeireldquoAutonomous structural health monitoring-part ii vibration-based in-operation damage assessmentrdquo Mechanical Systemsand Signal Processing vol 16 no 4 pp 659ndash675 2002

[17] C S Sakaris J S Sakellariou and S D Fassois ldquoRandom-vibration-based damage detection and precise localization ona lab-scale aircraft stabilizer structure via the GeneralizedFunctional Model Based Methodrdquo Structural Health Moni-toring An International Journal vol 16 no 5 pp 594ndash6102017

[18] Y Zhang W Song M Karimi C-H Chi and A KudreykoldquoFractional autoregressive integrated moving average andfinite-element modal the forecast of tire vibration trendrdquoIEEE Access vol 6 pp 40137ndash40142 2018

[19] M Pastor M Binda and T Harcarik ldquoModal assurancecriterionrdquo Procedia Engineering vol 48 pp 543ndash548 2012

[20] W J Staszewski K Worden and G R Tomlinson ldquoTime-frequency analysis in gearbox fault detection using the Wigner-ville distribution and pattern recognitionrdquo Mechanical Systemsand Signal Processing vol 11 no 5 pp 673ndash692 1997

[21] J-D Wu and P-H Chiang ldquoApplication of Wigner-Villedistribution and probability neural network for scooter engine

fault diagnosisrdquo Expert Systems with Applications vol 36no 2 pp 2187ndash2199 2009

[22] Z Feng and M Liang ldquoFault diagnosis of wind turbineplanetary gearbox under nonstationary conditions viaadaptive optimal kernel time-frequency analysisrdquo RenewableEnergy vol 66 pp 468ndash477 2014

14 Shock and Vibration

International Journal of

AerospaceEngineeringHindawiwwwhindawicom Volume 2018

RoboticsJournal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Active and Passive Electronic Components

VLSI Design

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Shock and Vibration

Hindawiwwwhindawicom Volume 2018

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawiwwwhindawicom

Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Control Scienceand Engineering

Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom

Journal ofEngineeringVolume 2018

SensorsJournal of

Hindawiwwwhindawicom Volume 2018

International Journal of

RotatingMachinery

Hindawiwwwhindawicom Volume 2018

Modelling ampSimulationin EngineeringHindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Navigation and Observation

International Journal of

Hindawi

wwwhindawicom Volume 2018

Advances in

Multimedia

Submit your manuscripts atwwwhindawicom

proposed detection strategy is effective at detecting theoccurrence of chain faults and identifying the failure pat-terns under different operating conditions

5 Conclusions

During the actual operation the working state of thescraper chain can reflect the dynamic performance of thescraper conveyor To address the difficulties with directsensor measurement for parameters of the moving scraperchain a novel strategy for fault detection of the scraperchain based on vibration analysis of the chute was pro-posed Based on modal analysis and the MAC the mea-suring points of vibration signals on the chute weredetermined To fit the actual behavior of the transmissionprocess the DTSM was presented based on finite elementmodeling and the correctness of the dynamic model wasverified by comparison with the FPET en the vibrationproperties of the measuring points on the chute undernormal condition chain jam and chain fracture werediscussed Moreover the occurrence of chain faults weredetermined by comparing the amplitudes of the vibrationsignal in the time domain while the AOKR was utilizedfor time-frequency representation of vibration signals anddistinguishing the two typical failure patterns Further-more the strategy verification based on experimental datawill be taken into consideration in the near future

Data Availability

e data used to support the findings of this study areavailable from the corresponding author upon request

Conflicts of Interest

e authors declare that they have no conflicts of interest

Acknowledgments

is work was supported by the Key Project of NationalNatural Science Foundation of China (U1510205) NaturalScience Foundation of Jiangsu Province (No BK20160251)Xuzhou Research program (KC14H0138) FundamentalResearch Funds for the Central Universities (2014Y05) andProject Funded by the Priority Academic Program De-velopment of Jiangsu Higher Education Institutions(PAPD)

References

[1] C D Brown ldquoDesign build and test of a longwall armouredface conveyorrdquo Longwall Mining 2002

[2] M Dolipski P Cheluszka E Remiorz and P SobotaldquoFollow-up chain tension in an armoured face conveyornadazne napinanie lancucha zgrzebłowego W przenosnikuscianowymrdquo Archives of Mining Sciences vol 60 no 1pp 25ndash38 2015

[3] L A Morley J L Kohler and H M Smolnikar ldquoA model forpredicting motor load for an armored face-conveyor driverdquoIEEE Transactions on Industry Applications vol 24 no 4pp 649ndash659 1988

[4] A A Ordin and A A Metelrsquokov ldquoAnalysis of longwall faceoutput in screw-type cutter-loader-and-scraper conveyorsystem in underground mining of flat-lying coal bedsrdquoJournal of Mining Science vol 51 no 6 pp 1173ndash1179 2015

[5] B He G Li H Shi et al ldquoDynamic behaviour modelling andsimulation of the chain transmission system for an armouredface conveyorrdquo in Proceedings of the IEEE 10th InternationalConference on Computer-Aided Industrial Design and Con-ceptual Design CAID amp CD 2009 pp 1000ndash1004 BeijingChina November 2009

[6] R Nie B He P Yuan L Zhang and G Li ldquoNovel approachto and implementation of design and analysis of armored faceconveyor power trainrdquo Science China Technological Sciencesvol 58 no 12 pp 2153ndash2168 2015

Loads

(a)

(b)

Empty load Half-load Full-load

500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)500

400

300

200

100

00 05 10 15

Time (s)

Freq

uenc

y (H

z)

500

400

300

200

100

00 05 10 15

Time (s)

500

400

300

200

100

00 05 10 15

Time (s)

Figure 14 Time-frequency representation of vibration signals at measuring point 2 with different external loads (a) chain jam (b) chainfracture

Shock and Vibration 13

[7] R Nie B He L Zhang and G Li ldquoModelling of thetransmission system in conveying equipment based on Eulermethod with applicationrdquo Proceedings of the Institution ofMechanical Engineers Part K Journal of Multi-body Dy-namics vol 228 no 3 pp 294ndash306 2014

[8] S B Jiang X Zhang K D Gao J Gao Q Y Wang andK Hidenori ldquoMulti-body dynamics and vibration analysis ofchain assembly in armoured face conveyorrdquo InternationalJournal of Simulation Modelling vol 16 no 3 pp 458ndash4702017

[9] M Myszkowski and D Loehning ldquoChain force measure-ments on armoured face conveyors and coal plows in heavy-duty longwallsrdquo CIM Bulletin vol 94 no 1054 pp 72ndash752001

[10] H Wang Q Zhang and F Xie ldquoDynamic tension test andintelligent coordinated control system of a heavy scraperconveyorrdquo IET Science Measurement and Technology vol 11no 7 pp 871ndash877 2017

[11] S Sen M X Min and Y Z She ldquoDiagnosis of coal scraperconveyor based on Fuzzy Fault treerdquo in Proceedings of the2015 Seventh International Conference on Measuring Tech-nology and Mechatronics Automation (ICMTMA) pp 392ndash395 IEEE Nanchang China June 2015

[12] S-s Xue X-c Li and X-y Xu ldquoFault tree and Bayesiannetwork based scraper conveyer fault diagnosisrdquo in Pro-ceedings of the 22nd International Conference on IndustrialEngineering and Engineering Management 2015 pp 783ndash795Atlantis Press Paris France January 2016

[13] X Gong X Ma Y Zhang et al ldquoApplication of fuzzy neuralnetwork in fault diagnosis for scraper conveyor vibrationrdquo inProceedings of the 2013 IEEE International Conference onInformation and Automation (ICIA) pp 1135ndash1140 IEEEYinchuan China August 2013

[14] Y Zhang X Ma Y Jianxiang et al ldquoFuzzy neural networkfault diagnosis and online vibration monitoring system for thecoal scraper conveyor based on rough set theoryrdquo in Pro-ceedings of the 2013 32nd Chinese Control Conference (CCC)pp 6134ndash6138 IEEE Xirsquoan China July 2013

[15] B Zhang A C C Tan and J-h Lin ldquoGearbox fault diagnosisof high-speed railway trainrdquo Engineering Failure Analysisvol 66 pp 407ndash420 2016

[16] E Parloo P Verboven P Guillaume and M Van OvermeireldquoAutonomous structural health monitoring-part ii vibration-based in-operation damage assessmentrdquo Mechanical Systemsand Signal Processing vol 16 no 4 pp 659ndash675 2002

[17] C S Sakaris J S Sakellariou and S D Fassois ldquoRandom-vibration-based damage detection and precise localization ona lab-scale aircraft stabilizer structure via the GeneralizedFunctional Model Based Methodrdquo Structural Health Moni-toring An International Journal vol 16 no 5 pp 594ndash6102017

[18] Y Zhang W Song M Karimi C-H Chi and A KudreykoldquoFractional autoregressive integrated moving average andfinite-element modal the forecast of tire vibration trendrdquoIEEE Access vol 6 pp 40137ndash40142 2018

[19] M Pastor M Binda and T Harcarik ldquoModal assurancecriterionrdquo Procedia Engineering vol 48 pp 543ndash548 2012

[20] W J Staszewski K Worden and G R Tomlinson ldquoTime-frequency analysis in gearbox fault detection using the Wigner-ville distribution and pattern recognitionrdquo Mechanical Systemsand Signal Processing vol 11 no 5 pp 673ndash692 1997

[21] J-D Wu and P-H Chiang ldquoApplication of Wigner-Villedistribution and probability neural network for scooter engine

fault diagnosisrdquo Expert Systems with Applications vol 36no 2 pp 2187ndash2199 2009

[22] Z Feng and M Liang ldquoFault diagnosis of wind turbineplanetary gearbox under nonstationary conditions viaadaptive optimal kernel time-frequency analysisrdquo RenewableEnergy vol 66 pp 468ndash477 2014

14 Shock and Vibration

International Journal of

AerospaceEngineeringHindawiwwwhindawicom Volume 2018

RoboticsJournal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Active and Passive Electronic Components

VLSI Design

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Shock and Vibration

Hindawiwwwhindawicom Volume 2018

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawiwwwhindawicom

Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Control Scienceand Engineering

Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom

Journal ofEngineeringVolume 2018

SensorsJournal of

Hindawiwwwhindawicom Volume 2018

International Journal of

RotatingMachinery

Hindawiwwwhindawicom Volume 2018

Modelling ampSimulationin EngineeringHindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Navigation and Observation

International Journal of

Hindawi

wwwhindawicom Volume 2018

Advances in

Multimedia

Submit your manuscripts atwwwhindawicom

[7] R Nie B He L Zhang and G Li ldquoModelling of thetransmission system in conveying equipment based on Eulermethod with applicationrdquo Proceedings of the Institution ofMechanical Engineers Part K Journal of Multi-body Dy-namics vol 228 no 3 pp 294ndash306 2014

[8] S B Jiang X Zhang K D Gao J Gao Q Y Wang andK Hidenori ldquoMulti-body dynamics and vibration analysis ofchain assembly in armoured face conveyorrdquo InternationalJournal of Simulation Modelling vol 16 no 3 pp 458ndash4702017

[9] M Myszkowski and D Loehning ldquoChain force measure-ments on armoured face conveyors and coal plows in heavy-duty longwallsrdquo CIM Bulletin vol 94 no 1054 pp 72ndash752001

[10] H Wang Q Zhang and F Xie ldquoDynamic tension test andintelligent coordinated control system of a heavy scraperconveyorrdquo IET Science Measurement and Technology vol 11no 7 pp 871ndash877 2017

[11] S Sen M X Min and Y Z She ldquoDiagnosis of coal scraperconveyor based on Fuzzy Fault treerdquo in Proceedings of the2015 Seventh International Conference on Measuring Tech-nology and Mechatronics Automation (ICMTMA) pp 392ndash395 IEEE Nanchang China June 2015

[12] S-s Xue X-c Li and X-y Xu ldquoFault tree and Bayesiannetwork based scraper conveyer fault diagnosisrdquo in Pro-ceedings of the 22nd International Conference on IndustrialEngineering and Engineering Management 2015 pp 783ndash795Atlantis Press Paris France January 2016

[13] X Gong X Ma Y Zhang et al ldquoApplication of fuzzy neuralnetwork in fault diagnosis for scraper conveyor vibrationrdquo inProceedings of the 2013 IEEE International Conference onInformation and Automation (ICIA) pp 1135ndash1140 IEEEYinchuan China August 2013

[14] Y Zhang X Ma Y Jianxiang et al ldquoFuzzy neural networkfault diagnosis and online vibration monitoring system for thecoal scraper conveyor based on rough set theoryrdquo in Pro-ceedings of the 2013 32nd Chinese Control Conference (CCC)pp 6134ndash6138 IEEE Xirsquoan China July 2013

[15] B Zhang A C C Tan and J-h Lin ldquoGearbox fault diagnosisof high-speed railway trainrdquo Engineering Failure Analysisvol 66 pp 407ndash420 2016

[16] E Parloo P Verboven P Guillaume and M Van OvermeireldquoAutonomous structural health monitoring-part ii vibration-based in-operation damage assessmentrdquo Mechanical Systemsand Signal Processing vol 16 no 4 pp 659ndash675 2002

[17] C S Sakaris J S Sakellariou and S D Fassois ldquoRandom-vibration-based damage detection and precise localization ona lab-scale aircraft stabilizer structure via the GeneralizedFunctional Model Based Methodrdquo Structural Health Moni-toring An International Journal vol 16 no 5 pp 594ndash6102017

[18] Y Zhang W Song M Karimi C-H Chi and A KudreykoldquoFractional autoregressive integrated moving average andfinite-element modal the forecast of tire vibration trendrdquoIEEE Access vol 6 pp 40137ndash40142 2018

[19] M Pastor M Binda and T Harcarik ldquoModal assurancecriterionrdquo Procedia Engineering vol 48 pp 543ndash548 2012

[20] W J Staszewski K Worden and G R Tomlinson ldquoTime-frequency analysis in gearbox fault detection using the Wigner-ville distribution and pattern recognitionrdquo Mechanical Systemsand Signal Processing vol 11 no 5 pp 673ndash692 1997

[21] J-D Wu and P-H Chiang ldquoApplication of Wigner-Villedistribution and probability neural network for scooter engine

fault diagnosisrdquo Expert Systems with Applications vol 36no 2 pp 2187ndash2199 2009

[22] Z Feng and M Liang ldquoFault diagnosis of wind turbineplanetary gearbox under nonstationary conditions viaadaptive optimal kernel time-frequency analysisrdquo RenewableEnergy vol 66 pp 468ndash477 2014

14 Shock and Vibration

International Journal of

AerospaceEngineeringHindawiwwwhindawicom Volume 2018

RoboticsJournal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Active and Passive Electronic Components

VLSI Design

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Shock and Vibration

Hindawiwwwhindawicom Volume 2018

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawiwwwhindawicom

Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Control Scienceand Engineering

Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom

Journal ofEngineeringVolume 2018

SensorsJournal of

Hindawiwwwhindawicom Volume 2018

International Journal of

RotatingMachinery

Hindawiwwwhindawicom Volume 2018

Modelling ampSimulationin EngineeringHindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Navigation and Observation

International Journal of

Hindawi

wwwhindawicom Volume 2018

Advances in

Multimedia

Submit your manuscripts atwwwhindawicom

International Journal of

AerospaceEngineeringHindawiwwwhindawicom Volume 2018

RoboticsJournal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Active and Passive Electronic Components

VLSI Design

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Shock and Vibration

Hindawiwwwhindawicom Volume 2018

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawiwwwhindawicom

Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Control Scienceand Engineering

Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom

Journal ofEngineeringVolume 2018

SensorsJournal of

Hindawiwwwhindawicom Volume 2018

International Journal of

RotatingMachinery

Hindawiwwwhindawicom Volume 2018

Modelling ampSimulationin EngineeringHindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Navigation and Observation

International Journal of

Hindawi

wwwhindawicom Volume 2018

Advances in

Multimedia

Submit your manuscripts atwwwhindawicom