clanok Elsevier 2010

This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and education use, including for instruction at the authors institution and sharing with colleagues. Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party websites are prohibited. In most cases authors are permitted to post their version of the article (e.g. in Word or Tex form) to their personal website or institutional repository. Authors requiring further information regarding Elseviers archiving and manuscript policies are encouraged to visit: http://www.elsevier.com/copyright

Author's personal copyEngineering Failure Analysis 18 (2011) 244255

Contents lists available at ScienceDirect

Engineering Failure Analysisjournal homepage: www.elsevier.com/locate/engfailanal

Analysis of crack initiation in the press frame and innovation of the frame to ensure its further operation Frantiek Trebuna a,, Frantiek imck a, Jozef Bocko a, Peter Trebuna b, a a Miroslav Pstor , Patrik argaa b

Technical University of Koice, Faculty of Mechanical Engineering, Department of Applied Mechanics and Mechatronics, Letn 9, 042 00 Koice, Slovakia Technical University of Koice, Faculty of Mechanical Engineering, Department of Management and Economy, Letn 9, 042 00 Koice, Slovakia

a r t i c l e

i n f o

a b s t r a c tThe paper describes numerical and experimental analysis of the causes of press frame failure with the aim to propose an optimal variant of its strengthening in order to guarantee safe operation of the press in its original working regime. The analysis of stress states in the frame during the operation was performed using the nite element method. Experimental analysis was focused on determination of residual stresses in the locations of failures. 2010 Elsevier Ltd. All rights reserved.

Article history: Received 18 February 2010 Received in revised form 9 September 2010 Accepted 9 September 2010 Available online 30 October 2010 Keywords: Cracks Press frame Stress concentration Numerical and experimental analysis

1. Introduction Supporting structures of the technological equipment in mechanical engineering are exposed to intensive inuence of force during their operation. The increase in productivity and the volume of production leads to the increased loading of the carrying elements. It results in increasing stress levels in the carrying elements, especially in the locations of their concentration. These effects decrease the time of safe operation of the technological equipment (its lifetime) and in some cases they initiate damage. Numerical and experimental methods [1215] are most suitable for the stress and strain analysis of such carrying structures. These methods allow both prediction of the residual lifetime of supporting structures (on the basis of the known loading history) and identication of the causes of possible failures. After approximately 10 years of operation of the press (Fig. 1a) used for cutting and forming side panels of washing machines, there were detected transversal cracks in the vertical columns of the press frame (Fig. 1b). They reached 3040% of their cross-sections. The press with maximum projected force of a ram 3500 kN was equipped with an automatic limiter of the maximum loading force. With respect to operational conditions it was necessary to operationally analyze the causes of crack initiation in the columns of the press frame and on the basis of such analysis to propose measures to be taken to ensure its further failure-free operation. The aim of this paper is to present the results obtained in the solution of this problem.

Corresponding author. Tel.: +421 55 6022462; fax: +421 55 6334738. E-mail addresses: [email protected] (F. Trebuna), [email protected] (F. imck), [email protected] (J. Bocko), [email protected] (P. Trebuna), [email protected] (M. Pstor), [email protected] (P. arga). 1350-6307/$ - see front matter 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.engfailanal.2010.09.004

Author's personal copy F. Trebuna et al. / Engineering Failure Analysis 18 (2011) 244255 245

Fig. 1. (a) View of the press, (b) crack in the column of the press frame.

Fig. 2. Computational model of the press frame.

Fig. 3. (a) Cutting after the rst operation, (b) pressed piece after the second operation.

2. Basic parameters of the press frame and the nature of its failure The design of the press frame is apparent from the computational model for the application of nite element method as shown in Fig. 2. During press operation two operations are accomplished simultaneously in one working cycle (one ram stroke of the press): cutting pieces from a rectangular shape sheet, forming a pressed piece from the piece cut in the previous operation.

Author's personal copy246 F. Trebuna et al. / Engineering Failure Analysis 18 (2011) 244255

Fig. 4. Cracks in the bottom part of the press frame columns.

Fig. 5. Cracks in the columns (a) on the inner side of column no. 2, (b) on the frontal side of column no. 1, (c) on the upper wall of the bottom crossbeam between column nos. 2 and 4.

Fig. 6. Cross-section of column no. 2.

Fig. 7. Local cracks in the press pedestal.


Fig. 8. Model of the press frame (a) whole frame with boundary conditions, (b) one-quarter press model, (c) boundary conditions for the one-quarter press model.

Fig. 3a and b show the shapes of the cut and pressed pieces that were produced by one stroke of the press ram during the past period. Sheets made from S235 material with modied mechanical properties Remin = 200 MPa, Rm = 320470 MPa, Agmin = 20% were used for the production of the pressed pieces. From the analysis of loading in the individual operations it was found out according to [3] that maximum load on the press ram is Fs = 1800 kN during cutting and F0 = 1200 kN during forming, i.e. for simultaneous execution of both operations the maximum force acting on the press ram reached 3000 kN. According to the data obtained from the operator, the press force limiter was adjusted to the magnitude of force in the interval from 3200 kN to 3500 kN during the whole period of its previous operation. Despite the above-mentioned fact there were detected cracks in three columns of the press approximately after 10 years of operation (with approximately 8 106 working cycles). Fig. 4 shows in more details the location and shape of cracks in the bottom part of the column nos. 1, 2 and 4. The photos in Fig. 5 illustrate the cracks. In Fig. 5a the crack runs along the whole length of the inner side wall of the column no. 2, Fig. 5b shows a crack in the column no. 1 in the frontal wall of the press and in Fig. 5c the crack spreads from the back side of the column no. 2 to the upper wall of the bottom crossbeam between column nos. 2 and 4.


Fig. 9. (a) Meshed model, (b) the eld of equivalent stresses.

As the presented gures show, the cracks were initiated in three columns and they were located in the bottom part of the press frame in locations of transition of the columns to the bottom crossbeams of the press frame. The columns had a closed rectangular cross-section created from the sheets as shown in Fig. 6 and the cracks were spread through the whole thickness of the sheets that form the column walls. Inspection of crack lengths showed that as a result of these cracks the carrying cross-sections of the columns were decreased by 30 to 40%. Subsequent inspection detected further cracks in the press frame (Fig. 7). However, these were of local character and affected a small part of the supporting structure. They were probably initiated after formation of cracks in the press columns. Material properties of the press frame were determined using tensile test according to STN 100 002+AC1. As a result of the test the following values were obtained: ReH = 376 MPa, ReL = 343 MPa, Rm = 499 MPa, A5 = 31.7%, Z = 67.3%. On the basis of chemical analysis and experimentally determined data the material of the press frame can be considered as being equivalent to the material according to STN 41 1483, S355J2G3 (EN 10025-94), or Gr50 (ASTM A572) [2]. 3. Stress analysis of the press frame by the nite element method The computations were accomplished by program COSMOS DesignStar using parabolic tetrahedral elements in linear area. Fig. 8a shows the press frame model together with the boundary conditions applied for numerical solution. The hatching marks the areas on which technological forces act on the bottom part of the frame with the cutting force on the left and the forming force on the right side. Reaction forces are transmitted through a ram to the crank mechanism in the upper part of the press frame. Computation was carried out for the whole loading force in the ram of 3500 kN (maximum allowable force determined by the force limiter). This force was divided into two parts: 2100 kN for cutting and 1400 kN for forming. At the rst stage was analyzed the whole press frame. The mesh of elements was coarse and for the computation was considered different loading on both sides of technological operations. As the computation showed that there are no substantial differences between stress levels on both sides, in the next computations was used only one-quarter of the frame. In order to identify stress concentrators in the locations of the cracks computations were performed in one-quarter of the frame on the more loaded part of the press using geometrical symmetry of the frame (Fig. 8b). It increased the precision of computation by using smaller nite elements in the critical part and at the same time it decreased the time of computations. Boundary conditions


Fig. 10. Location of the junction between the column and the bottom crossbeam (a) meshed model, (b) the eld of equivalent stresses.

according to Fig. 8c assume symmetrical loading of the press, which represents an adverse condition of loading in comparison with the real state. Fig. 9a shows a mesh of nite elements in the one-quarter press model. Fig. 9b gives the eld of equivalent stresses in this model with a previously dened press loading. Fig. 10a shows a detail of the mesh in the area of cracks (junction of the column to the bottom crossbeam). Stress concentration in this location is apparent from the eld of equivalent stresses in Fig. 10b. In this computation a sharp corner (without rounding) was considered in the junction of the column to the crossbeam. In Fig. 11a and b are given elds of equivalent stresses in locations of stress concentrators. As the result of the stress analysis shows, the maximum equivalent stresses in the area of question marked by the capital letter A reach the value 273 MPa (Fig. 11b). 4. Determination of residual stresses in the frame columns by the hole-drilling method Residual stresses also occur in the structures without loading and they can be caused by machining, assembling, but mostly by operational overloading [6]. The residual stresses are superimposed to stresses from the operational load and they


Fig. 11. Details of equivalent stresses in locations of stress concentrators.

can, under certain circumstances (magnitude and character) cause failure of the structure. The levels of residual stresses in the material can be determined by several methods. One of them is a hole-drilling method. Hole-drilling equipment RS-200 was used for the measurement of residual strains [7,9]. The measurement was performed by self-compensating strain-gages 1-RY-21-3/120 with k-factor 2.06 on all grids. The strain-gages were applied by strain-gage glue X60 and protective silicon coat SG-250. The location and orientation of the strain-gages on the columns of the press frame is illustrated in Fig. 12. The strain-gages were located at distances approximately 20 mm from the cracks. This is sufcient distance for possibility to quantify residual stresses invoked by manufacturing. Arising cracks do not release signicantly residual stresses in the areas of measurements. The measurement was performed using strain-gage rosettes with the radius 5.15 mm and the diameter of the drilled hole 3.21 mm. The depth of the drilled hole was 5 mm and the drilling was accomplished in 10 steps, each with 0.5 mm of length. The magnitudes of residual stresses were determined according to standard ASTM E-837-01 [1] as well as using Integral method and the method of Power-Series. For the measurement of strains released during drilling the strain-gage apparatus P3 was used [11]. Determination of residual stresses according to ASTM E-837-01 supposes constant distribution of residual stresses along the depth of the hole. Despite this fact it is the most widely used method because it is the only method in the world that is standardized. In Table 1 are given magnitudes and directions of residual stresses determined according to ASTM E-837-01.


Fig. 12. Locations of the strain-gages on column no. 1 and 2 of the press frame.

Table 1 Values of residual stresses and their directions according to ASTM E-837-01. Column no 1 Location of measurement X.1 X.2 X.3 X.4 Column no 2

rmax [MPa]16.42 31.25 37.21

rmin [MPa]61.10 5.52 13.97

u []18.86 52.21 27.73

rmax [MPa]6.46 26.02 28.33 20.19

rmin [MPa]56.84 13.73 15.05 4.45

u []2.12 77.57 75.85 15.61

5. Discussion of possible causes of cracks initiation in the press frame and suggestions for press frame innovations The computed values of residual stresses determined from the measured strains released by drilling on the inner sides of the column walls did not exceed 40 MPa (see Table 1). All the values were found to be positive, i.e. they facilitated the growth of cracks. Their magnitudes on the sidewalls lay in the interval 2040 MPa that clearly documented the fact that they were not produced during manufacturing of the frame. It can be supposed that they resulted from loading during the operation of the press. With respect to their magnitudes they cannot be considered as direct causes of crack initiation, because during the operation of the press vanishing loading was invoked where the middle stress was equal to halve of the stress amplitude. The numerical computation of the press frame by the nite element method showed that in the junction of the frame column with the bottom crossbeam (location A in Fig. 11b) there was extreme stress concentration with maximum equivalent stress 273 MPa. Maximum stress was initiated on the inner 30 mm thick wall side in the location of the sharp (not rounded) corner. The critical value of the upper stress level in the vanishing cycle and fatigue of the material was 0.61 Rm for vanishing tension and 0.74 Rm for vanishing bending according to [8,10], where Rm is the strength of the material. With respect to the facts conrmed by numerical computation, the bending stress in the columns caused by different stiffness of columns and crossbeams as well as resulting from eccentric loading during the working cycle exhibited very low levels. The critical stress value was therefore largely inuenced by tensile loading, i.e. the critical stress for vanishing loading can be determined from the formula 0.61 Rm. For the determined strength of the material it corresponds to the value of the upper stress level of approximately 300 MPa. As we can see, even when considering stress concentration in the corner without rounding, the magnitude of maximum equivalent stress is smaller than the upper critical value of the material from which the column is made. The excess of the upper critical stress could only occur in the case of non-anticipated increase of force (e.g. tight pull of the cutting) and nonfunctional force limiter. This would result in forces exceeding 3500 kN. The second cause of crack formation can be proneness of the material to brittle fracture. In order to verify this possibility there were performed tests of the material toughness by Charpy hammer in accordance with STN EN 100 E45-1, STN EN 875. For the test, pendulum hammer PSW 300 was used.


Fig. 13. Fracture surfaces of the specimens for evaluation of toughness.

Fig. 14. Shapes and dimensions of the strengthening plates (a) inner side of the columns, (b) outer side of the columns.

The results of the test carried out showed that the magnitudes of fracture energy KV lay in the interval from 77 J to 115 J and the magnitudes of notch toughness KCV in the interval from 96 J/cm2 to 144 J/cm2. According to STN 73 1401, the recommended magnitudes of impact toughness are connected with yield stress and the factor that also inuences the desired magnitude is the type of loading, i.e. it is different for static and dynamic loading. For the steel materials with yield stress up to 275 MPa, the notch toughness KCV = 50 J/cm2 is required for the dynamic loading. For the materials with yield stress up to 355 MPa, the notch toughness 70 J/cm2 is required [4,5]. The notch toughness measured from test specimens exceeded the values given in the standard. Fig. 13 shows the shapes of the fracture surfaces gained from two impact tests. The rolled out inclusions were observed on the fracture surfaces of both specimens that substantially inuenced the test results. If these are found in two specimens, it can be stated that this fact signicantly inuences the behaviour of the press frame and it can initiate a crack even at lower levels of the stress amplitude. As it is evident from the fracture surface in Fig. 13, the percentage of brittle fracture is close to 50%. The measurement was taken at the temperature 20 C. It is evident that there exists a hazard of a brittle fracture, because the material is near to the transit area and with the temperatures under 20 C the danger of brittle fracture is real. From the analysis of the results it was found that the cracks in the structure of the press frame were initiated in the sharp corners of the junction of the press columns and the bottom crossbeams and were probably caused by inappropriate geometry of the press frame and low purity of the material. Lack of time prevented to provide metallurgical examination of material and take material from the areas of cracks for analysis of crack surfaces.


Fig. 15. Computational model with the strengthening plates.

Fig. 16. Field of equivalent stresses on the press frame after modications.

The shapes and positions of cracks unambiguously document locations of their initiations and they are in agreement with results of analysis provided by FEM. For the analysis of the press frame strengthening, in order to ensure further operation of the press without failures, the nite element method was used. The stress analysis by the FEM showed that rounding of the corner with a radius of 40 mm in locations of the junction of steel sheets of the columns with the bottom crossbeams (location A, Fig. 11b) the maximum value of the equivalent stresses decreased from 273 MPa to 180 MPa, i.e. by approximately 35%. On the basis of the above-mentioned facts it was proposed to provide (after welding cracks) strengthening of the press frame by welding eight plates made of material S355 to all columns (one both on the inner and outer side of each column). The shapes of plates are shown in Fig. 14a (the plates on the inner side of the columns 25 mm thick) and Fig. 14b (the plates on the outer side of the columns 16 mm thick).


Fig. 17. Details of equivalent stresses on the press frame after modications.

A computational model of the one-quarter press with welded plates is given in Fig. 15. Fig. 16 shows the eld of equivalent stresses in the frame after its strengthening by the plates. In Fig. 17a and b there are details of equivalent stresses in the area of the outer side wall of the column (Fig. 17a) and the inner side of the column (Fig. 17b). The stress analysis showed (Fig. 17) that maximum equivalent stresses did not exceed 190 MPa, i.e. strengthening decreased stresses by approximately 30%. This is sufcient for further safe operation of the press. 6. Conclusions On the basis of numerical and experimental analysis of the press frame it can be concluded that: Numerical analysis of the press frame showed that occurrence of sharp corners (without rounding) in the junction of the columns and bottom crossbeams caused (in the case of maximum allowable loading of ram 3500 kN) equivalent stresses of 273 MPa.


The levels of residual stresses in the press frame columns did not exceed, according to ASTM E 837-01, the absolute value 62 MPa. The stress levels on the sidewalls of the columns did not exceed 40 MPa and they lay in the interval from 20 to 40 MPa. Residual stresses on the frontal walls of the columns did not exceed their absolute value level 62 MPa and these stresses were invoked by the operation of the press. Crack initiation in the inner corners of the columns (location A, Fig. 11b), and also in other locations could be most probably caused by overloading during the operation which, however, was not caused by the technological process of pressing itself (cutting and forming), but as a result of inappropriate position of the semi-nished product and consequently arising of additional forces during the movement of the ram. However, in this case it appears to be due to malfunction of the force limiter. Another possible reason for crack initiation was also low purity of the press frame material that was documented by probably rolled out sulphides and lamellar splitting of material as a consequence of that. Strengthening of the bottom parts of the columns by welded plates will result in decreasing of maximum equivalent stresses by approximately 30%, which is sufcient for further safe operation of the press. In order to eliminate possible press frame overloading the authors proposed calibration of the press limiter and its regular accuracy testing.

Acknowledgments This work was supported by Project OPVaV-2009/2.1/03-SORO-772253. References[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] ASTM E 837-01 Standard test method for determining residual stresses by the hole drilling strain-gage method. New York; 2001. Frbacher I, Macek K, Steidl J. a kol. Lexikon technickych materilu. Verlag Dashfer; 2005. Hrivnk A. a kol. Technolgia plonho tvrnenia. Alfa Bratislava; 1989. Juhas P. a kol. Navrhovanie ocelovych kontrukci. STN 73 1401. Juhas P. Spolahlivost a pevnost materilov ocelovych kontrukci. ELFA, Koice; 2009. Kobayashi AS. Handbook on experimental mechanics. Cambridge: VCH Publishers; 1993. Redner S., Perry CC. Factors Affecting the Accuracy of Residual Stress Measurements Using the Blind Hole Drilling Method, In: Proceedings of 7th International Conference Experimental Stress Analysis. August 1982. p. 604-16. Trebuna F, Burk M. Medzn stavy lomy, Grafotlac Preov; 2002. Trebuna F, imck F. Kvantikcia zvykovych napt tenzometrickymi metdami. Preov, Grafotlac; 2005. Trebuna F, imck F. Odolnost prvkov mechanickych sstav. Emilena, TU SjF Koice; 2004. Trebuna F, imck F. Prrucka experimentlnej mechaniky. Koice, TypoPress; 2007. Trebuna F, imck F, Bocko J, Pstor M. Stress analysis of casting pedestal supporting structure. Acta Mechanica Slovaca 2009;13(1). Koice. Trebuna F, imck F, Bocko J, Gainec J. Methods for verication of safety of the sluice gates in water power plants. Acta Mechanica Slovaca 2009;13(4). Koice. Trebuna F, imck F, Bocko J, Trebuna P. Identication of causes of radial fan failure. Eng Fail Anal 2009;16(7):205465. ISSN 1350-6307. Trebuna F, imck F, Bocko J. Failure analysis of storage tank. Eng Fail Anal 2009;16(1):2638. ISSN 1350-6307.

Documents

clanok Elsevier 2010