MATERIALBEHAVIOUROFPOWDER-METALLURGICALLY PROCESSED TOOL ... · Keywords: tool steels, high speed steels, mechanical properties, tensiletest, bending test, ﬁnite element simulation

MATERIAL BEHAVIOUR OF POWDER-METALLURGICALLYPROCESSED TOOL STEELS IN TENSILE ANDBENDING TESTS

S. Marsoner, R. Ebner and R. MinichmayrMaterials Center Leoben

Franz Josef Strasse 13

8700 Leoben

Austria

F. JeglitschDepartment of Physical Metallurgy and Materials Testing

Franz Josef Strasse 18

8700 Leoben

Austria

Abstract The paper concentrates on the static mechanical propertiesof powder metal-lurgically processed (PM) tool steels tested in tensile andbending tests. Arecently developed tensile test based on a especially designed tensile speci-men is used to characterise the mechanical properties of a PM-tool steel indifferent tempering conditions. So far there is no standardtensile testing pro-cedure available for high strength tool steels. Main goal ofthe investigationsis to study the influence of heat treatment on the mechanical properties likeyield strength, ultimate tensile strength and strain to fracture. The resultsof the tensile tests are compared to the results of bending tests which arecommonly used for characterising the mechanical properties of high strengthtool steels. In these tests the bending rupture strength is predicted from thefracture load based on the assumption of a linear stress distribution withinthe bending specimen. This assumption of linear elastic material behaviourhas to be recognised as the major uncertainty in the prediction of the bendingstrength from the fracture load. Finite element (FE) simulations are per-formed to model the bending test based on material properties determined in

197

198 6TH INTERNATIONAL TOOLING CONFERENCE

the tensile test. Experimental load – displacement curves are used to validatethe model.

Keywords: tool steels, high speed steels, mechanical properties, tensile test, bending test,finite element simulation

INTRODUCTION

High speed steels were mainly developed for cutting applications liketurning, drilling or milling. However, they are increasingly important astool materials for cold work applications like cold forging, blanking, cuttingor shearing because of their outstanding property profile. Besides a highwear resistance, mechanical properties like toughness, strength and ductilityare most relevant for the application of high speed steels. Especially a highresistance against plastic deformation in combination with a high fractureresistance is important to realise highly loaded tools. Generally, the strengthof high speed steels is characterised by three or four point bending tests[1, 2, 3]. Only a few results on tensile properties are reported in literature(e.g. [4, 5]). The bending strength is calculated from the maximum load atthe onset of fracture assuming a linear stress distributionwithin the bendingspecimen. The assumption of a linear stress distribution isin principleinvalid if plastic deformation occurs. Plasticity is sometimes indicated bya curvature of the load deflection line, especially in case ofhigh speedsteels tempered at higher temperatures. The rupture strength of powdermetallurgically produced high speed steels is often found to be in the range of4000 to 5500 MPa. In many cases deviation from the linear elastic behaviouris found on exceeding a stress level of about 2500–3500 MPa atthe surfaceof the bending specimen (e.g. [2, 3]). It can be thus assumed that the bendingtest is suitable to get information on the material properties in case of a specialloading geometry, but it is not suitable to characterise material properties likeyield strength, ultimate tensile strength and ductility.

In the paper the mechanical behaviour of one selected powdermetallurgi-cally processed high speed steel is investigated using the recently developednew tensile test procedure. For comparison, bending tests were additionallyperformed. Finite element simulations are employed to simulate the bendingtest using material data determined in the tensile test.

Material Behaviour of Powder-Metallurgically Processed Tool Steels in Tensile and...199

Table 1. Chemical composition of HS 10-2-5-8 (Bohler S390PM) in weight percent

C Cr W Mo V Co

1,63 4,66 10,54 2,00 4,70 7,83

Table 2. Heat treatment program of HS 10-2-5-8 (Bohler S390PM)

austenising tempering, each with 3×2h

1130◦C/ 6 min 475◦C 500◦C 525◦C 550◦C 575◦C

MATERIALS, SPECIMENS AND EXPERIMENTAL METH-ODS

A powder metallurgically processed high speed steel of the type DIN HS10-2-5-8 (BOHLER S390PM) was investigated in this paper. The chemicalcomposition of this steel is summarised in Table 1. The heat treatment wasperformed by austenising of the specimens in a salt bath withsubsequentquenching and three times tempering for two hours. To investigate theinfluence of the heat treatment the tempering temperature was varied fromabout 475 to 575◦C. Details on the heat treatment program are shown inTable 2.

The microstructure of the fully heat treated material consisted mainlyprimary carbides embedded in a tempered martensite matrix.Fig 1 shows ascanning electron micrograph of the microstructure indicating micrometersized primary carbides of the types MC (grey) and M6C (white). The meansize of the globular primary carbides is about 1µm, the maximum size isabout 2,5µm. The volume fraction of the primary carbides is about 0,15.For details about the microstructure of high speed steels the reader is referredto Ebner et al. [6].

The tensile properties were determined with a new tensile specimen ([6, 7]), which was developed to characterise the tensile behaviour of highstrength tool steels. The shape of the new tensile specimen is shown in Fig 2.The specimen has a diameter of 8 mm and a measuring length of 40mm.The shape of this specimen was optimised by finite element simulation inorder to minimize stress concentrations. The specimens forthe four point


Figure 1. Scanning electron micrograph of the high speed steel HS 10-2-5-8PM

Figure 2. Shape of the new tensile specimen

bending test had a circular cross section of 5 mm and a length of 55 mm. The


specimens were finished by grinding to a mean surface roughness of about0,2µm after heat treatment.

All mechanical tests were performed in an universal tensiletesting ma-chine (ZWICK UPM 1485). In the tensile tests the strain measurement wasperformed by a video extensometer. The bending tests were carried out ina four point bending set-up. The load was applied via high speed steel rollswhich were heat treated to a hardness of about 66,5 HRC. Load displacementcurves were measured and the load to fractureFB was used to calculate thebending rupture strengthσb,f by assuming a linear elastic stress distributionin the bending specimen (1).

σb,f =16FBx

πD3, (1)

where D is the diameter of the bending specimen and x the minimum hori-zontal distance between upper and lower rolls.

Figure 3. FE–model of the bending specimen

3D-finite element simulations were carried out using the software packageABAQUS® to verify the stress and strain distributions in a bending specimen.A three dimensional model had to be used because of the symmetry of thebendingspecimen. The specimenwas definedwithanelastic-plastic material


behaviour, the load was applied via rolls with rigid surfaces. The used FE–model is shown in Fig 3. Main attention had to be laid on the simulationof the contact zone between rolls and specimen. The use of rolls instead ofsingle forces was necessary to avoid numerical problems dueto extremelyhigh local stresses, which would lead to very high plastic deformation in thecontact zone. Stress-strain data of the high speed steel in the various heattreatment conditions determined with the new tensile test were used in theFE simulation. Experimentally determined load displacement curves wereused to validate the FE model.

RESULTS AND DISCUSSION

Figure 4. Influence of tempering temperature on bending strength and hardness

The results of the mechanical tests are summarised in Fig 4 and Fig 5.Figure 4 shows the bending strength determined by means of Equation 1 andthe hardness (Rockwell C), Fig 5 shows the yield stress, the ultimate tensilestrength and the ductility as a function of the tempering temperature. Thebending rupture strengthσb,f increases from about 3600 MPa at a temperingtemperature of 475◦Cto a maximum of about 5400 MPa at a temperingtemperature of about 550◦Cfollowed by a slight decrease on further increaseof the tempering temperature. In contrast the maximum hardness occurs


Figure 5. Influence of tempering temperature on ultimate tensile strength, 0,2% proofstress and fracture strain

at a tempering temperature of about 500◦Cfollowed by a decrease withincreasing tempering temperature. The 30 to 50◦Cshift in the temperingtemperatures which lead to the maximum values of the hardness and thebending strength is in good accordance with results from literature [3]. Theresults of the tensile tests indicate that the maximum tensile strength isabout 3200 MPa for a tempering temperature of about 540◦C, whereas themaximum in the yield stress is achieved at a tempering temperature of about525◦C. The tensile tests furthermore indicate a low but remarkable ductilitywhich is ranging from about 0,2% for tempering at 500◦Cto about 1,3%for tempering at 575◦C. Scattering of the yield and tensile strength data ishigher at the lower tempering temperatures.

A comparison of bending and tensile strengths of Fig 4 and Fig5 revealsthat the tensile strength is about 500 to 2400 MPa lower than the bendingstrength. It is argued that a non-linear stress distribution within the bendingspecimen caused by local plastic deformation is responsible for these differ-ences. FE simulations are employed in order to verify this assumption. Forthe FE simulations the stress–strain behaviour of the material and a suitablefracture criterion are necessary.


Typical stress–strain curves of investigated material subjected to differenttempering temperatures are shown in Fig 6. The results indicate that allspecimens fracture prior reaching the point of plastic instability, which ischaracterised by attaining a maximum in the technical stress–strain curvefollowed by a subsequent stress reduction. The plastic instability followedby localised deformation (necking) is achieved in case of the HS 10-2-5-8for tempering temperatures above about 600◦C.

Figure 6. Effect of tempering on stress–strain behaviour

For the FE simulation of the bending test the experimental measuredstress–strain data from Fig 6 were used. It was necessary to extrapolatethe stress–strain curves to higher strain levels. The extrapolation was doneby assuming a linear stress–strain behaviour with the slopeat the onset offracture. Main reason for this extrapolation was that higher strains occur inthe contact zone between the bending specimen and the loading cylinders.Non-linear effects from this contact zone affect the load displacement curvesand have thus be considered in the simulation.

The validation of the FE model was performed by comparing calculatedload–deflection curves with experimentally determined ones. A comparisonof calculated and measured load-deflection curves is shown in Fig 7. Slightdifferences between these two curves only occur in the region of high loads


Figure 7. Experimental and simulated load displacement curves

which can be attributed to uncertainties resulting from thebending test, thetensile test and the FE simulation. Despite these slight discrepancies athigher loads it can be concluded that the results from the simulation arein suitable accordance with the experimental results. The FE simulationswere stopped on reaching the true strain to fracture ( = fracture criterion) asdetermined from the tensile tests in the outer fibre of the bending specimen.The calculated stress distribution in the bending specimenat reaching thefracture criterion is shown in Fig 8 (solid curved line). Assuming that thesame bending moment is applied in case of a linear elastic material behaviourthis would lead to the dashed straight line.

The following conclusions can be drawn from Fig 8 :

The FE simulations reveal that a significant fraction of the cross sectionis subjected to plastic deformation.

The non-linear stress distribution seems to be responsiblefor the sig-nificant over-estimation of the fracture stress.

Figure 9 shows a comparison between experimentally determined andsimulated fracture loads at fracture criterion. The results indicate that theexperimentally found increase of the fracture load with increasing temperingtemperature can be well reproduced by the FE simulations butthe predicted


Figure 8. Curved line: FE–simulated stress distribution; straight line: stress distributionby assuming linear elastic material behaviour

Figure 9. Experimental and simulated load to fracture

fracture loads are generally too low. It is argued that this effect is caused byan underestimation of the fracture strain which was taken from the tensile


tests. One possible reason is that the stressed volume is significantly smallerin the bending specimens than in the tensile specimens. The probabilityof finding a fracture initiating defect is thus lower in the bending specimenwhich causes a higher ductility. Choosing fracture strainswhich are about 50to 100% higher than those determined in the tensile tests lead to calculatedfracture loads which are comparable to the experimentally measured ones.

CONCLUSION

The aim of the paper was to study the mechanical behaviour of ledebu-ritic tool steels especially high speed steels. Standard bending tests andtensile tests based on a recently developed specimen and procedure wereperformed. The investigated material was a PM high speed steel DIN HS10-2-5-8 (BOHLER S390PM). Significant differences were found for thematerial strengths determined in the tensile and the bending tests. In orderto understand the reasons for these differences FE simulations of the bendingtests were carried out based on material data which were determined in thetensile test. The results of the study can be summarised as follows:

A developed tensile test (specimen and testing procedure) enables thedetermination of tensile properties of fully heat treated high speedsteels.

Variations of the heat treatment reveal its strong influenceon the tensileproperties.

The tensile strength values are significantly lower than thestrengthdetermined in the bending specimen.

The FE simulations of the bending test indicate that plasticdeforma-tion takes place over a significant fraction of the cross section.

The non-linear stress distribution due to the plastic deformation is themain reason for the significant overestimation of the strength levels inthe standard bending tests.

Good accordance between experimentally determined and calculatedfracture loads in the bending tests can be achieved by assuming thatfracture occurs at a strain which is about 50 to 100% higher than thetensile fracture strain.


The results of the present paper show that it is possible to determinetensile properties of high strength materials quite easily. This test has theadvantage to determine clearly defined material propertieswhich can beused in simulations. The bending test reveals the material behaviour in caseof a special loading situation namely in a bending bar. In thebending testnot only the strength of the material is considered but also the effect of theductility of the material.

ACKNOWLEDGMENT

The authors would like to thank Bohler Edelstahl GmbH & Co KG forsupplying with tool steels. Financial support for this workby the Tech-nologie Impulse G.m.b.H, the County of Styria, the Innofinanz – SteirischeForschungs- und Entwicklungsfurderungsges. m.b.H. and the Municipalityof Leoben in the frame of the Austrian Kplus Competence Center Programis highly acknowledged.

REFERENCES

[1] G. HOYLE, "High speed steels" (Butterworths, London, 1988) p. 123.

[2] S. WILMES, Stahl und Eisen 81 (1961) 676.

[3] W. SCHMIDT, Thyssen Edelst. Techn. Ber. 13 (1987) 141.

[4] J. A. RESCALVO and B. L. AVERBACH, Metall. Trans. 10A (1979) 1265.

[5] P. BRØNDSTED and P. SKOV-HANSEN, Int. J. Fatigue 20 (1998) 373.

[6] R. EBNER, H. LEITNER, D. CALISKANOGLU, S. MARSONER and F.JEGLITSCH,Z. Metallkde 92 (2001) 820.

[7] J. M. LACKNER, "Entwicklung eines Zugversuches fur hochstfeste Werkzeugstahle",Diploma thesis (University of Leoben, Leoben, 2001).

MATERIALBEHAVIOUROFPOWDER-METALLURGICALLY PROCESSED TOOL ... · Keywords: tool steels, high speed steels, mechanical properties, tensiletest, bending test, ﬁnite element simulation

Documents