Notch Tensile Testing Of High Strength Steel

2nd International Symposium on High Strength Steel, 23.-24. April, 2002, Verdal, NorwayNotch tensile testing of high strength steel weldmentsV. Olden*, Z. L. Zhang*, E. Østby*, B. Nyhus*, C. Thaulow*****SINTEF Materials Technology, Trondheim, NorwayNorwegian University of Science and Technology, NTNU, TrondheimAbstract: Notch tensile testing has proven to be a suitable method of testing the tensile properties ofwelded joints. In combination with a simple procedure of geometry correction of the stress level, truestress strain data for the transversal direction of the heat affected zone and weld metal can beretrieved. The paper presents the method of notch tensile testing and its application on welded joints of690 MPa yield strength structural steel.IntroductionTransversal stress strain curves for the heat affected zone (HAZ) and weld metal(WM) can not be obtained by the traditional cross weld tensile testing method.However, by using round notched tensile specimens, deformation will occur only inthe notched area and material data for limited material zones, as the HAZ and WMcan be established.By a simple geometry correction, material specific true stress strain curves for thezones can be retrieved from the tensile results. This implicates the possibility ofhaving correct input to material models calculating load bearing capacity and criticalfailures of structural welds.This paper will give a thorough description of the method of notch tensile testing andgeometry correction. Results from the method applied on welded joints in highstrength 690 Mpa structural steel will also be presented.Test specimen geometryLDd17.5 a)Figure 1Notch tensile specimen a) specimen geometryb)b) notch geometry 5In general there are no specific requirements related to the length (L) or diameter ofthe specimen (D) outside the notched area (Figure 1). The requirements are onlyrelated to the notch geometry and the width of the material zone of interest, ref. 11

2nd International Symposium on High Strength Steel, 23.-24. April, 2002, Verdal, NorwayIf the notch radius is less than the specimen radius in the notched area, the anglebetween the straight area of the notch surface and the perpendicular axis of thespecimen should be 17.5 , as specified in Figure 1b.dRHFigure 2 Notch area geometry of tensile specimen 1, 51)The diameter of the specimen in the notch (d) should at least be twice theradius of the notch (R). The d/R relationship should however be limited to a value of6 1, 3, 5.2)To ensure that plastic deformation is limited to the material zone of interest,the diameter (d) should not exceed the length of this zone (H) 1.To ensure the quality of the machined notch and an accurate monitoring of the notchdeformation during testing, the radius should however not be less than 0.4 mm. Thetest specimen geometry requirements are summarised in Table 1.Table 1Geometryd0/R0d0/HR0Specimen geometry requirements 5Requirements min20.4 mmRequirements max61H/2The diameter of the specimens in the notch area is measured at the position giving thesmallest cross-sectional area. The dimension measurements should be performed witha micrometer or other measuring equipment with the sufficient accuracy (Preferredresolution 1/100 mm). The radius control can be performed by randomly pickingsome specimens for R0 control by feeder gauge or by using a microscope measuringtable. 5Location of the notchThe notch should be located in the middle of the material zone of interest. In case ofwelds, specimens extracted in the transversal direction of the weld should apply to thefollowing general directions:Testing of WM:Notch located in the cross mid-weld position2

2nd International Symposium on High Strength Steel, 23.-24. April, 2002, Verdal, NorwayTesting of HAZ:Notch located in coarse grained HAZ.Positioning of the notch in HAZ should be verified by macroetching of the microstructure near the fusion line beforemachining.TestingTesting shall be performed in a calibrated tensile testing machine with sufficient loadcapacity. The cross head displacement rate should be in the range 0.002 - 0.011 mm/s.The sharpest notched specimen should be tested at the lower displacement ratesMeasuring of diameter reduction during deformationTo obtain a true stress-strain curve, it is required to measure the smallest cross-sectionin the notch area during deformation. The diameter must be measured in twoperpendicular directions either optically or with inductive displacement gauges. Ifusing displacement gauges with knife edges, the edges must be rounded to avoiddamaging the surface. (suggestion R 0.2 mm) It is vital however that the knife edgesare straight, located in the same plane and perpendicular to the tensile direction. Thedisplacement gauges must be carefully monitored to ensure that the smallest crosssection is measured at all times. 5Registration of test data and determination of material curves for notchedsamplesThe following data can be retrieved during testing:Load vs timeLoad vs cross head displacementLoad vs vertical displacement in notchFrom the load vs vertical displacement data, the following curves can be determinedLoad vs diameter reductionNominal stress vs nominal strainTrue stress vs true strainDiameter reduction is taken as the mean vertical displacement of the two measureddirections at every registration point.Nominal stress, s [MPa], is P/A0, where P is the registered load in kN and A0 is theinitial minimum cross-sectional notch area in mm2 calculated from the initial diameterof the notch.True stress, sT [MPa], is P/A, where A is the true cross-sectional notch area calculatedfrom the diameter reduction.True strain, e [mm/mm], is calculated by:æd öe 2 lnç 0 , where d0 is the initial diameter and d the actual diameter at the notch.è d ø3

2nd International Symposium on High Strength Steel, 23.-24. April, 2002, Verdal, NorwayNominal strain, e [mm/mm], is retrieved from the established relation betweennominal and true strain: e ln( 1 e )Typical material curves for different notch geometries compared to a smoothspecimen curve, all for a 500 MPa yield strength steel is given in Figure 3. Load andstress level rises with the sharpness of the notch.a)b)Figure 3 Material curves for a 500 MPa yield strength steel for different notchgeometries 2 a)Load vs diameter reduction b)True stress vs true strainEstablishing true stress strain curves for the material by a geometry correctionof the notch stress strain curvesThe geometry dependent stress strain curves, both nominal and true, can be convertedto stress strain curves for the material by dividing the notched sample stress, sN , by ageometry factor G (d0/R0, ePmax).1sM sNG1G is a function of the relation between initial diameter and notch radius, d0/R0, and thestrain at maximum load, ePmax, and is described by the following equation.12éæ d0 öæ d0 ö ùG ê1.007 0.18777çç - 0.01313çç ú (1.053 - 0.53e P max )êëè R0 øè R0 ø úû2Figure 4 presents a plot of the geometry factor. The equation is valid within thefollowing ranges:d0/R0 : 2 - 6ePmax : 0.05 - 0.204

2nd International Symposium on High Strength Steel, 23.-24. April, 2002, Verdal, NorwayDo/R 6Do/R 4Do/R 2Figure 4The geometry facture presented as a function of fracture strain atmaximum load and geometryNotch tensile testing of 690 MPa welded joints.The base metal used in the investigation was a 50 mm thick QT steel plate with aprescribed yield strength of minimum 690 MPa. The chemical composition is given inTable 2. Yield and tensile strength in the mid-thickness position were 655 MPa and745 MPa respectively.Chemical composition of the 690 MPa steel 4Table 2C.16Si.430Mn1.20PSAlNCu.013.001.039 .008 .035VNbTiB.043 .026 .003 .0002Mo.31Ni.34Cr.425The welding procedures were designed to give one yield strength evenmatch and oneyield strength overmatch welded joint. The weld configuration was a K- butt jointwith a root angle of 45 . The welding position was horizontal. Flux cored arc welding(evenmatch ) and submerged arc welding (overmatch) were used. The heat input wasin the range of 1.7-2.5kJ/mm. Preheating was minimum 100 C, and interpasstemperature was max 200 C. A typical macro section is presented in Figure 5.TopMiddleFigure 5: Typical macro section of the welded joints 45

2nd International Symposium on High Strength Steel, 23.-24. April, 2002, Verdal, NorwayTest specimensLongitudinal and transversal notch tensile specimens were machined from the top andmiddle section of HAZ and WM. Specimens were also extracted from the transversaldirection of the base metal. The notch centre in the HAZ transversal specimens werelocalised approximately 1mm from the straight fusion line. All the specimens had asample diameter of 6 mm in the notch area. Three different notch geometries weremachined: R 3.0 mm, R 1.5 mm and R 0.8 mm.Typical tensile stress strain curves for weld metal middle section are presented inFigure 6. In the evenmatch weld metal (Figure 5b), somewhat higher stress and lowerfracture strain could be observed in the transversal direction. This tendency was notpresent in the overmatch weld metal (Figure 5a) or in the HAZ. G-corrected truestress strain data for R 1.5 mm in HAZ and WM overmatch is presented in Figure 7.G-correction within the small strain rangeThe yielding pattern for smooth and notched specimens is different. Smoothspecimens usually have a sharp transition between the elastic and plastic part of thestress strain curve. This transition is not present in the notched specimens. Thisimplicates that the G -correction strictly not is appliccable within 1% strain 1. Bycomparing the G-corrected yield strength (Rp0.2) in this study with the smoothspecimen results, it is evident that the yield strength after G-correction always islower than in the smooth specimens, Figure 8. Hence, it is a conservative estimate ofthe stress level. 3a)b)Figure 6: Tensile tress strain curves from notch tensile testing ofa) overmatch WMb ) evenmatch WM 46

2nd International Symposium on High Strength Steel, 23.-24. April, 2002, Verdal, Norwaya)b)Figure 7: True stress strain curves derived from notch tensile resultsa) HAZb) overmatch WM 4The difference in yield strength increases with sharpness of the notch. For D0/R0 2,the yield strength is reprodused within an error of 5%. For the sharpest notchedspecimen the G-corrected yield strength is about 10-15% lower than in the smoothspecimen curves. This implicates that if a reproduction of the yield strength level is amain objective in the notch tensile testing, a D0/R0 ratio of 2 preferably should beused. 3Best overall fit with respect to the plastic range of the stress strain curve is obtainedwith a D0/R0 ratio of 4 ( Figure 7). This implies, that when stress strain data for theplastic range is the main objective in the notch tensile testing, the D0/R0 4 is thegeometry which should be preferred. 3G corrected true stress strainG corrected true stress strain690 MPa - Base metal - top1200.0690 MPa - HAZ- longitudinal - middle - 0.0True StressStress600.0Bridgeman (smooth)Do/Ro 2, R 3.0 mmDo/Ro 4, R 1.5Do/Ro 7.5, R 0.8 mmE-mod at 0.2% strain400.0BridgemanD0/R0 2, R 3.0 mmD0/R0 4, R 1.5 mmD0/R0 7.5, R 0.8 mmE-mod at 0.2% inFigure 80.030mm/mm0.0400.0500.00.0000.0100.0200.030Eq strain mm/mm0.0400.050Stress strain curves for BM and HAZ within 5% strain7