a r t i c l e i n f o a b s t r a c t

ten end the life ofin technique, safefciency and pro-nveyor are

nd Kicks [6with functions fatigue strength in shear loading of fabric conveyor belts. They concluded that the fatigue phenomebelts is explained by the decrease of shear strength, which is determined as an ultimate angle initiating the laminatiure of the central rubber layers [6].

1350-6307/$ - see front matter 2013 Elsevier Ltd. All rights reserved.

Corresponding author.E-mail address: gabriel.fedorko@tuke.sk (G. Fedorko).

Engineering Failure Analysis 36 (2014) 3038

Contents lists available at ScienceDirect

Engineering Failure Analysis

journal homepage: www.elsevier .com/locate /engfai lanalhttp://dx.doi.org/10.1016/j.engfailanal.2013.09.017motion, also is carrying components, support material load [2]. One of the main causes of damage and ofconveyor belts is their dynamic stress. Dynamic analysis is the key to decide whether the design is rationaland reliable in running, feasible in economy. It is very important to study dynamic properties, improve eductivity, guarantee conveyor safe, reliable and stable running [4]. The dynamic characteristics of a belt comined to a large extent by the properties of the belt [5].

Dynamic stress causes fatigue strength functions in shear loading of fabric conveyor belts. Kozhushko adeter-

] dealtnon ofon fail-1. Introduction

Belt conveyors are machines for continuous transport [1]. Belt conveyor is a commonly used equipment of continuoustransport; it has a high efciency and large conveying capacity, simpler construction, small amount of maintenance. Canbe achieved at different distances, different materials transportation [2]. Belt conveyor is widely used in mine, coal, chemicalindustry, ports, and power plants [3]. In the belt conveyor, as the conveyor belt is traction components, transmit power andArticle history:Received 10 July 2013Received in revised form 20 August 2013Accepted 20 September 2013Available online 2 October 2013

Keywords:Conveyor beltTest deviceExperimental measurementRegression modelThe most common case of conveyor belts damage is their puncture by falling sharp mate-rial. One of the ways, how to minimize this type of damage, is using of suitable type of con-veyor belt. Therefore, the analysis of conveyor belts on the part of their puncture resistanceis an important factor for their use in operation conditions. The aim of the paper is to deter-mine the dependence among the weight of sharp material falling on the conveyor belt,shatter height and force conditions in the conveyor belt on the base of experimental mea-surements by the help of regression mathematical model and to determine conditionsunder which the conveyor belt is damaged. The experimental results enable the operatorof a conveyor belt to set the shatter height and maximum weight of falling weight belowthe threshold values in order to prevent conveyor belt damage.

2013 Elsevier Ltd. All rights reserved.Failure analysis of belt conveyor damage caused by the fallingmaterial. Part I: Experimental measurements and regressionmodels

Gabriel Fedorko a,, Vieroslav Molnar a, Daniela Marasova a, Anna Grincova b, Miroslav Dovica c,Jozef Zivcak c, Teodor Toth c, Nikoleta Husakova a

a Faculty of Mining, Ecology, Process Control and Geotechnology, Technical University of Kosice, Park Komenskeho 14, 042 00 Kosice, Slovak Republicb Faculty of Electrical Engineering and Informatics, Technical University of Kosice, Letna 9, 042 00 Kosice, Slovak Republicc Faculty of Mechanical Engineering, Technical University of Kosice, Letna 9, 042 00 Kosice, Slovak Republic

For dynamic analysis of conveyor belts it has proved use of the nite element method (FEM). Hatala and Maras [9] in their

Therefore the effort of producers is to produce and interest of customers is to buy conveyor belt wit the greatest damage

G. Fedorko et al. / Engineering Failure Analysis 36 (2014) 3038 31resistance. The part of the article is a methodology, which can be used by conveyor belts tests associated with the assessmentof their damage resistance and breakdown creations in extreme cases.

2. Material and methods

2.1. Problem formulation

A conveyor belt changes its shape and size under the thumb of external load what means that it is deformed. Mechanicalproperties express the relation among the size of acting external force and caused deformation, which include elastic anddeformational characteristics characterized by the modulus of elasticity, modulus of deformability and strength properties,for example strength in compression, strength in tensile, strength in shear. Dimensional, functional, physical, mechanicaland special test methods are determined for the testing of basic properties of conveyor belts with textile cords and coverlayers from rubber.

Shock tests, which are the object of the research in this work, belong to special tests realized for purpose of overall assess-ment of conveyor belts intended for special use, primarily for bulk material conveying in hard operation conditions in open-cast quarries. They are very specic are of testing, and until now unmodied by any standards or regulations. The aim ofshock test is to determine the destructive resistance of a conveyor belt with regard to the conditions, in which the belt willwork (lumpiness of the conveying material, its shape and so on).

At present, there is no clear way of determining or evaluation of conveyor belts breakdown resistance. In practice, a con-veyor belt is often exposed to impact of sharp material on its surface, especially in the place of conveyor lling or on shifting.It occurs to their damage and in some cases conveyor belt is not able to the next operation. Its results are no small economicdamages which are related to the necessary of damaged belt exchange and created non-operating states. The paper is aimedat the problem of rubbertextile conveyor belt damage by falling material. The aim is to determine the dependence among aweight of material falling on the conveyor belt, shatter height and force ratios of conveyor belt by regression mathematicalmodel and to determine conditions by which it creates a damage of conveyor belt. For a more detailed study of the conveyorbelt after realization of experiments it was applied a method of industrial computed tomography.

2.2. Description of the experiment

Laboratory experiment is an appropriate means of determining the shock and tension forces during material impact on aconveyor belt, i.e. for determination, respectively verication of use properties of investigated object conveyor beltresearch focused on the application of the FEMnumericalmodeling stressstrain state in conveyor belts. Pascual et al. [10] pre-sented amethodology to compute dynamic stress distributions on large conveyor belts considering a viscous-dampingmodel.

The next author who dealt with the possibilities of FEM application within the frame of detailed research of the conveyorbelts problemwas Lodewijks [8,9]. Fedorko and Ivanco [10] researched by FEM the force ratios in conveyor belt of classic beltconveyor. By application of FEM for research of belt conveyors one of the key factor is determination of right material prop-erties of conveyor belts. This problem was solved by the work of Mazurkiewicz [11].

Conveyor belt is very important for belt conveying operator also on the part of economy. For this reason it is necessary topay increased attention to research of causes which cause its degradation and damage.

Romani [12] dealt with the ability to control the drive acceleration torque providing a smooth soft start while maintainingbelt tensions within specied safe limits which are critical for belt conveyors performance. Zimroz and Krl [13] deal withfailure analysis for condition monitoring. Mazurkiewicz [14] also dealt with the problem of conveyor belts damages.

One of the main reasons for conveyor belt wear is the stress caused by the impact of lumpy material [15].Impact of material on the conveyor belt is caused in many cases not only wear, but also conveyor belt damage. This prob-

lem was solved in the past not only in theory [1922], but also by testing in laboratory conditions [2326]. In theory, theeffort of authors is oriented to the creation of mathematical models with the aim to describe the properties of conveyor belt.Till now it was described mathematical apparatus for determination of a reliability of belt conveyors using the renewal the-ory [16,17]. Regression models in the area of conveyor belt breakdowns were solved only marginally [18,19]. The most atten-tion is in mathematical modeling by FEM [20,810].

Research of damaged conveyor belts can be realized directly during the operation of belt conveyors, what is often almostimpossible or strenuous for realization or it is possible to use special test devices. Aldrich et al. [21] realized online analysis ofcoal on a conveyor belt by use of machine vision and kernel methods. On the other hand Fiset and Dussault [22], Ballhaus[23,24], Flebbe [25] and Hardygora [26] dealt with analysis of the causes of belt conveyors damages in laboratory conditions.

Damage of conveyor belt leads to its gradual destruction and due this fact we can monitor huge economic losses for users.Cheng and Du [6] researched force state of belt conveyor during horizontal turning section to improve the design level ofbelt conveyor with horizontal turning. Kumar [7] presented the review of belt conveyor design modication and latest tech-nologies or methodologies used in different applications to reduce failures, maintenance cost and equipment related fatalaccidents occurs during operation.

resistance against rip. This experiment and its results enable to determine conveyor belt impact force and tension force atthe moment of sharp material simulated impact. Laboratory experiment realizes by the test device designed at the LogisticsInstitute of Industry and Transport.

2.3. Description of the test device

The test device simulates the impact of the sharp material on the conveyor belt. Principled it is a drop hammer withweight falling from a certain height, which is known to us on a conveyor belt. The device (Fig. 1) consists of a tower and testboard.

The construction of the tower is created from the metal truss construction, at the end of which is hoisting winch of thedrop hammer. The winch with the xed and free pulley create pulley block of the tower. The drop hammer can be lifted to amaximum height 2.6 m. It is led by two parallel lines, which are formed from single-strand steel ropes of closed construction.

The weight of the drop hammer may vary from 40 kg to 300 kg what is the capacity of the lifting winch. Adding or remov-ing of classic calibrated weight with the weights 520 kg realizes the change of its weight. Manipulation with the drop ham-mer is enabled by free pulley of the pulley block. By changing of the height of the drop hammer it is monitored the change ofdrop energy. The drop hammer is sinking by free fall. We attach impactors to the end of the drop hammer (Fig. 2). During thetest it was used impactor B, the Fig. 2. For the next experiments it was used impactors A, C, the Fig. 2 which are possible tosimulate other types of falling material. It was determined from the results of the measurements that the effect of the impac-tor type is negligible on the results.

The test board is a metal construction created by U proles in which are moving hydraulic jaws for clamping of the con-veyor belts sample. The construction of the clamping board was designed so that it creates solid unit resistant to impacts.

2.4. The test specimens

Sampling of conveyor belts by general applicable conditions is preceded by preparation of the test specimen. Sampling for

32 G. Fedorko et al. / Engineering Failure Analysis 36 (2014) 3038rubbertextile conveyor belts testing is realized at least 500 mm from the end of the produced conveyor belt. The length ofthe conveyor belts sample must be such that of this it is possible to prepare test specimens for all required tests.

The test specimen is die-cut or cut-out from the sample of the conveyor belt at the points of uniform distributed at thedirection of its width, and the minimum distance from the edges of the belt must be 100 mm. We recommend using for cut-ting out a knife with cuttingwedge angle not more than 18. For the purpose of laboratory experiments of which results arepresented in this paper, it is cut out the test specimen with dimensions 1400 600 mm from the sample of the rubbertex-tile conveyor belt (Fig. 3).

The test specimens are tested at least 5 days after the production of the conveyor belt. Before the test they must be con-ditioned for 3 days by temperature (23 2) C and relative air humidity (650 5)%, if the condition period is not speciedotherwise. The tests are realized by the temperature (23 2) C and relative air humidity (65 5)%. Each test specimen must

Fig. 1. The test device with the detail of the drop hammer with impactor.

control by metro-tomograph. From the measurement recording we determine in which size of the tension and shock force it

Fig. 2. Shaped of impactors.

G. Fedorko et al. / Engineering Failure Analysis 36 (2014) 3038 33reaches a breakdown, so the damage of the test specimen and it is determined a disruptive energy. In the case of mathemat-icalphysical model creation the measured data are next presented to the tables, graphs (Fig. 5), where the output createsClassication of the tests results from the impact stress is not currently dened by any standards or regulations. By theevaluation we use all available mathematical, physical and statistical procedures.

After the tension of the test specimen and drop hammer release it is measured the size of tension and shock forces by thehelp of two strain-gauge scanner. By the help of laser sensor it is measured the actual height of the drop hammer in regard ofthe top layer of the test specimen. An electronic device records the whole time slope of the measurement, actual height ofindividual time points, the size of the tension and shock force (Fig. 4). During one measurement cycle it was used the sam-pling frequency 1000 Hz. The duration of the measurement of one impact test (10 s) results from the request of solver torecord completely the full test course, also with grading margin. This time exactly sufces to impact of the drop hammerby maximum shatter height and maximum possible weight by free fall and at the same time it reects several times.

Evaluation of the test in the case of determination of the breakdown is based on a visual control of the test specimen andbe controlled after the test at the point of the testing. If it is determined a material defect, the result does not regard andthere is recorded a kind of error in the record [27,28].

2.5. Basis of the test

The test specimen (conveyor belt type P 1600/5 + 2) is inserted into the space among hydraulic jaws and it is rmly sta-pled. After pressing by hydraulic jaws the test specimen is strained by the force, which is equal to 1/10 strength of the beltwhat corresponding with allowable stress of the conveyor belt in operation. After the test specimen tension the drop ham-mer is freed on which it can be additional weight, but the drop hammer falls out on the test specimen. After the test it isvisually determined the rate of the test specimen damage and it is subtracted the size of the shock and tension force atthe moment of the impact. Subtracted data are subsequently appreciated with the aim to determine the dependence amongthe size of shock force and the test specimen damage.

2.6. Evaluation of the testdependence of all parameters affecting the test results.

Fig. 3. Preparation of the test specimen.

3. The

This m

data (

Fig. 4. Dependence of the shock force Fi at the time of the drop hammer impact on the conveyor belt.

34 G. Fedorko et al. / Engineering Failure Analysis 36 (2014) 3038Within the frame of the next procedure, the regression function (1) was replaced by selected regression function (2).

Y b0x0 b1x1 . . . bnxn: 2

Coefcients b0; b1; . . . ; bn are estimates of the unknown coefcients b0; b1; . . . ; bn. For the estimate of these coefcients itwas used the least squares method, by which we minimize the residual sum of squares for deviations (3).

SR Xn

i1yi Yi2 3Table 1), which were obtained by experimental measurements on the test device (Fig. 1).It is described by regression function (1), which can have various forms.y b0x0 b1x1 . . . bnxn: 1

Within the frame of our research for conveyor belts resistance we applied as a based for the regression analysis appliesodel is an expression of the assumption of dependence and it presents starting parameters for the regression analysis.3.1. Design of the regression function

A regression analysis helps to nd a mathematical model, which determines the correlation of two or more parameters.ory/calculationFig. 5. Compare of shock forces values obtained experimentally and by calculation using the model (10).

is the

the form (6), where d f a; x is the differential of the k- place value of the function f x.

Thy f

For

2.2 14.038 17.344 19.993 23.220 26.222 29.3022.4 15.382 18.080 21.386 24.103 28.115 31.0682.6

G. Fedorko et al. / Engineering Failure Analysis 36 (2014) 3038 35y f x f m;h after modications in form (8).

y b0 b1 h b2 m b3 h2 b4 m2 b5 h m: 8Th

4. Res

4.1. Pr

Weimumimpacpacts

Infrom t

Forrst pf x d0f a; x 11!d1f a; x 1

k!dkf a; x: 7

e nal model was created on the basis of related reections related to the possibility of approximation of the functionxby Taylors polynomial so thatwe regarded themembers of the series (7) ending by the differential of the second series.asmuch as our model is dependent on two variables, weight m and height h, it is possible to transform the function1! k!

If we neglect the function Rkx (approximation error) in the Eq. (6), we will get an approximate expression, i.e. theapproximation for this function in the form (7).f x d0f a; x 1 d1f a; x 1 dkf a; x Rkx; 6From the methods of mathematical analysis it is known that the function f x is possible to approximate on the basis ofTaylor expansion to the series by the polynomial of denite degree, that means to write this function using of differentials of

kSy i1

yi y2 5variable.

I2 1 SRSy

4

Xnsum of squares of deviations yi y, and y in the Eq. (5) is the arithmetic average of the measured values of the

The index of determination (4) was used as the characteristics of sententious for the regression function, where the Sy (5)16.010 19.502 22.592 29.195 32.383Table 1Maximum values of the shock force Fi [kN] depending on the weight m and shatter height h.

m [kg]/h [m] 50 60 70 80 90 100

0.2 3.100 3.806 4.542 5.268 5.582 6.5630.4 4.768 5.631 6.651 7.603 8.329 9.6330.6 6.082 6.995 8.260 9.761 10.762 12.0860.8 7.250 8.123 9.741 11.164 13.450 14.4311.0 8.456 9.516 11.301 13.322 15.461 16.7261.2 9.437 10.801 12.871 14.970 17.315 18.5021.4 10.575 12.292 14.519 16.657 19.316 21.7101.6 11.586 13.616 16.000 18.227 20.905 23.3871.8 12.606 14.999 17.462 20.130 22.857 25.1232.0 13.675 16.137 18.708 21.435 24.447 27.370e regression function (8) was supplanted by selective regression function (9).

Y b0 b1 h b2 m b3 h2 b4 m2 b5 h m: 9

ults

esentation of the results (practically)

can determine from each measurement recording for example the actual size of forces in dependence on time, max-values of the forces acting on the belt, duration of the shock. The maximum loading force is generated by the rstt of the drop hammer on the test specimen; next extreme forces (Fig. 4) are caused by reections and consequent im-of the drop hammer on the test specimen.the rst approximation we selected values of forces from the rst phase of drop hammer impact on the test specimenhe le of measured data.the determination of forced needed for research in all next measurements we used maximum values of forces of thehase of drop hammer impact on the test specimen. On this basis, in accordance with the statistical procedures related

Table 2

36 G. Fedorko et al. / Engineering Failure Analysis 36 (2014) 3038Maximum values of the tension force Fs [kN] depending on the weight m and shatter height h.

m[kg]/h[m] 50 60 70 80 90 100

0.2 36.670 38.220 40.064 41.712 42.232 44.116to theregresof the

Analo

Thshocksmall,and teof tenrmed

0.40.60.81.01.21.41.61.82.02.22.42.6Fig. 6. Compare of tension forces values obtained experimentally and by calculation using the model (11).suitability of the data selection from the basic set it was created the Table 1 with maximum values of shock force, thension function (9) for shock force is (10) and to this the calculated index of determination which presents the suitabilityfound model is I = 0.9988.

Fi 0:3859 2:0998 h 0:0081 m 0:8029 h2 3:3556 m2 0:1089 h m 10

gously it was also determined the regression function for tension force (11) with an index of determination I = 0.9959.

Fs 19:9463 11:1003 h 0:3122 m 2:2634 h2 0:0011 m2 0:1262 h m 11

e Fig. 5 graphically presents the tightness of shock forces values obtained experimentally (Table 1) and values offorces obtained by calculation using the model (10). Anomalies among experimental and calculated values are verysome as insignicant. The Fig. 6 presents the tightness of tension force values obtained experimentally (Table 2)nsion force values obtained by calculation using the model (11). Analogous to the case of shock force, also in the casesion force, anomalies among experimental and calculated values are very small. In both cases it is presented and con-the suitability of obtained models.

40.260 42.124 44.204 46.225 47.363 50.00243.615 45.391 47.677 50.286 52.582 54.05345.666 47.618 50.620 52.591 56.653 58.90948.167 50.463 53.710 57.644 59.635 62.51950.100 53.523 56.702 60.312 62.970 64.73652.317 56.192 60.214 63.510 67.179 70.53454.740 58.948 62.578 66.198 69.416 72.67256.221 61.234 64.942 69.945 71.937 73.08558.507 63.147 67.199 71.701 74.458 79.43258.379 65.609 68.405 74.850 76.871 82.33561.832 66.590 71.073 73.938 81.080 84.67063.353 69.474 74.115 83.120 84.739

Fig. 7. Analyzed test specimen with created depression on the bottom side.

[5] Hou Y, Meng Q. Dynamic characteristics of conveyor belts. J China University Mining Technol 2008;18:62933.[6] Cheng XW, Du HB. Resistance analysis of belt conveyor during horizontal turning section. Adv Mater Res 2011;201203:46770.

G. Fedorko et al. / Engineering Failure Analysis 36 (2014) 3038 37[7] Kumar D, Mandloi RK. Analysis & prospects of modication in belt conveyors a review. Int J Eng Res Appl 2013;3:5817.[8] Lodewijks G. On the application of beam elements in Finite Element models of belt conveyors. Part I. Bulk Solids Handling 1994;14:72937.[9] Lodewijks G. On the application of beam elements in Finite Element models of belt conveyors. Part II. Bulk Solids Handling 1995;15:57191.[10] Fedorko G, Ivanco V. Analysis of force ratios in conveyor belt of classic belt conveyor. Proc Eng 2012;48:1238.[11] Mazurkiewicz D. Problems of identication of strength properties of rubber materials for purposes of numerical analysis: a review. Arch Civil Mech Eng

2010;X:116.[12] Romani NK. Belt speed and tension control on long conveyors. Bulk Solids Handling; 2011:111.[13] Zimroz R, Krl R. Failure analysis of belt conveyor systems. Prace Naukowe Instytutu Grnictwa Politechniki Wrocawskiej; 2009:501.[14] Mazurkiewicz D. Analysis of the ageing impact on the strength of the adhesive sealed joints of conveyor belts. J Mater Process Technol

2008;208:47785.[15] Ballhaus H. A new conveyor belt wear test stand. Bulk Solids Handling 1982;2:5963.[16] Pavliskov A, Andrejiov M. Comparison of model life of conveyor belts with selected statistical methods. Trans Logist 2009;9:4215.[17] Pavliskov A, Jadronov M. Application of mathematical model of operating costs for calculating conveyor belts life. Manufact Eng 2006;5:612.[18] Andrejiov M. Attestation of assumptions about random error of the regression model. Creative Math Inform 2011;20:47.[19] Andrejiov M, Pavliskov A. Annalysis of regression model of functional dependency in impact force from height and weight of ram for conveyor belt.

Annal Faculty Eng Hunedoara: Int J Eng 2010;8:26770.4.2. Analysis of the conveyor belt sample

During the measurement it began to create an oval bulge (Fig. 7) on the bottom side of the test specimen (bottom coverlayer of the test specimen). We did not regard signicant changes in the substructure of the top cover layer after the rstthree impacts at the point of test impactor impact on the topside of the analyzed test specimen. We regarded small cracksduring next impacts of the impactor.

After completion of the experimental measurements the test specimen was submitted to nal visual control for damagesformation. Its result was the statement that the analyzed test specimen was not damaged by breakdown (it was not syn-chronic breakdown of the top and bottom cover layer of the test specimen). By visual control it was observed just cracksin the top cover layer and the above-mentioned formation of local oval bulge.

But the existence of this bulge predicted that there were some changes by the test in the internal structure of the testspecimen.

5. Conclusions

Damage of the conveyor belt by breakdown is often case that occurs in continuous transport of sharp-edged materials. Ifwe want to identify the conveyor belt as breakdown, it must be satisfy the conditions of simultaneous damage of all struc-tural layers of the conveyor belt, and the destructive line runs steady through its top and bottom cover layer.

Based on this denition, we can state that there was not damage of the test specimen by breakdown during realizedexperiments on the test device. From the measured data we created mathematicalstatistical models describing the depen-dence of the shock or tension force on the weight and shatter height (10) and (11). These models show very good results(Figs. 5 and 6).

These models enable to recalculate the sizes of the shock or tension force for arbitrary values of weights and heights byuse of interpolation. We can collect these values for the weight from the interval between minimal and maximal weight ofthe drop hammer and for the height from the interval between minimal and maximal shatter height used by experiments.

Determination of the boundary value of the impact energy, by which it comes to the breakdown, so damage of the con-veyor belt, has a practical relevance for users. A user can set the shatter height and maximum weight of the falling weight atthe point of shifting right in operation, so that it does not make, or exceed the size of end-point impact energy.

Despite of the statement that the test specimen was not damage by breakdowns during the test, it was identied a localanomaly on the test specimen which was at the point of simulated place of sharp-edged material impact.

Acknowledgements

This work is a part of research project VEGA 1/0922/12 Research of effect of material characteristics and technologicalparameters of conveyor belts on size of contact forces and resistance to motion in pipe conveyors with experimental andsimulation methods.

This work is a part of research project VEGA 1/0085/12 New strategy for effective measurements with coordinate mea-suring machines with multi sensor systems.

References

[1] Markusik S, Nowakowski P, Opasiak T, Pypno C, Sajdak C. Protection of environment in conveying of bulk powder materials. Australasian Institute ofMining and Metallurgy Publication Series, Faculty of Transport, Silesian University of Technology, Krasinskiego Str. 8, Katowice 40019, Poland: 2003,p. 3613.

[2] Lihua Z. Typical failure analysis and processing of belt conveyor. Proc Eng 2011;26:9426.[3] Cun GY, Gang C, Kun H, Fen WZ. Research on the idler spacing of belt conveyor. Appl Mech Mater 2012;127:2959.[4] He Q, Li H. Review of dynamic modeling and simulation of large scale belt conveyor system. In: Chen R, editor. Intelligent computing and information

science SE - 27, vol. 134. Berlin Heidelberg: Springer; 2011. p. 16772.

[20] Pascual R, Meruane V, Barrientos G. Analysis of transient loads on cable-reinforced conveyor belts with damping consideration. Proc XXVI IberianLatinAmerican Congress Comput Methods Eng CILAMCE 2005;2005:115.

[21] Aldrich C, Jemwa GT, van Dyk JC, Keyser MJ, van Heerden JHP. Online analysis of coal on a conveyor belt by use of machine vision and kernel methods.Int J Coal Preparation Utilization 2010;30:33148.

[22] Fiset M, Dussault D. Laboratory simulation of the wear process of belt conveyor rollers. Wear 1993;164:10125.[23] Ballhaus H. Die Impulskrfte beim Aufprall grobstckingen Gutes auf den Frdergut. Braumkohle 1980;32:1847.[24] Ballhaus H. Stress and wear of conveyor belts by loading point impact. Bulk Solids Handling 1983;3:34753.[25] Flebbe H. Untersuchung von Frdergurten auf ihre dynamische Festigkeit. Braumkohle 1982;34:18691.[26] Hardygora M, Golosinska G. Effect of belt and loading station design on impact resistance of steel cord conveyor belts. Bulk Solids Handling

1986;6:5616.[27] troffek E. Conveyor belts in industry. Kosice: Technical University of Kosice; 1995.[28] Standard STN EN 260370:1987, 1987. Conveyor belts with textile carcass. Method of test of main properties.

38 G. Fedorko et al. / Engineering Failure Analysis 36 (2014) 3038

Failure analysis of belt conveyor damage caused by the falling material. Part I: Experimental measurements and regression models1 Introduction2 Material and methods2.1 Problem formulation2.2 Description of the experiment2.3 Description of the test device2.4 The test specimens2.5 Basis of the test2.6 Evaluation of the test

3 Theory/calculation3.1 Design of the regression function

4 Results4.1 Presentation of the results (practically)4.2 Analysis of the conveyor belt sample

5 ConclusionsAcknowledgementsReferences