Orientation effects in pulsed magnetic field treatment

Materials Science and Engineering A398(2005)344–348Orientation effects in pulsed magneticfield treatmentCai Zhipeng∗,Lin Jian,Zhao Haiyan,Lu AnliDepartment of Mechanical Engineering,Tsinghua University,Beijing100084,ChinaReceived5January2005;received in revised form17March2005;accepted29March2005AbstractThis paper presents the orientation effects on residual stress release by pulsed magneticfield treatment.Specimens are produced by tungsten inert gas(TIG)welding and the initial residual stresses in them are measured by the hole-drilling method.With different orientations these specimens are treated in the same magneticfield and the residual stresses are measured for the second time.By comparing the stress values before and after treatment it is obvious that magneticfield treatment effects are related to orientations.Based on a referred ferromagnetics formula,a supposition is proposed and a further survey of the magnetostriction constant is carried out.All experimental results support the proposed model and a prescription is given for pulsed magneticfield treatments.©2005Elsevier B.V.All rights reserved.Keywords:Pulsed magneticfield treatment;Orientation effects;Magnetostriction constant1.IntroductionThis paper presents the orientation effects on residual stress release by pulsed magneticfield treatments.As a new materials processing technology pulsed magneticfield treat-ment has been proposed for several years.The benefits of this new method have been reported in some papers[1–10]. The magneticfield treatment can alter surface hardness,re-duce friction,extend wear and fatigue lifetime,alter frac-ture and corrosion resistance,and retard crack growth.And some equipments utilizing this technology have been created to improve the lifetime of cutting tools.Though almost all the previous investigators have speculated that the release of residual stresses caused by magneticfield treatment was sig-nificant for the various improvements,the mechanisms for the substantial boost in mechanical properties have not yet been clearly determined.Tang et al studied the effects on stress release by pulsed magneticfield treatment[11–15].In their experiments,the residual stress was reduced by nearly40%.They held that the dynamic magnetostriction,which provided energy for the ∗Corresponding author.Tel.:+861062773796x13;fax:+861062773637.E-mail addresses:czpdme@,caizhipeng97@(C.Zhipeng).inelastic strain movements,was the principal variable affect-ing the stress release.But in their publications,the effects of different treating orientations were not mentioned.So the role treating orientation plays in the pulsed magneticfield treatment needs to be studied;that is the objective of this paper.2.Experimental procedures and instrumentationExperiments in this paper include three parts.Thefirst is to fabricate specimens with certain residual stresses.The second part is to survey the residual stress values.And the last part is to treat specimens with different orientations in the same magneticfield.In thefirst part,we use welding to produce specimens with residual stresses.The sketch of specimens is shown in Fig.1.Each specimen contains two#20steel plates,400mm long,50mm wide and 3mm thick,jointed by tungsten inert gas(TIG)welding.The cold-rolled plates have a composition of0.18%C,0.67%Mn, 0.21%Si,0.11%Cr,0.03%S by weight.TIG welding is chosen for its even residual stresses.The welding parameters are:230A welding current,9V welding voltage and4.3mm/s welding velocity.0921-5093/$–see front matter©2005Elsevier B.V.All rights reserved. doi:10.1016/j.msea.2005.03.097C.Zhipeng et al./Materials Science and Engineering A398(2005)344–348345Fig.1.Sketch of specimen and defined directions:(a)orientation in case1, (b)orientation in case2and(c)orientation in case3.The welding direction is defined as the x vector.And vec-tors of y and z are defined as perpendicular to the weld-ing bead and vertical to the plate,respectively,as shown in Fig.1.In the second part,we use the hole-drilling method to survey the residual stresses.Before the magneticfield treat-ment,three points along the welding bead are measured as references.Then the specimen is treated with a certainfield orientation.After treatment,the residual stresses of four other points are measured in the same way.By comparing the stress values before and after treatment,the stress release in this ori-entation can be determined.The seven measured points are located in the welding bead center.The three reference points measured before magnetic field treatment are located180mm,200mm and220mm away from the bead origin,respectively.The other four points measured after magneticfield treatment are located170mm, 190mm,210mm and230mm,respectively,away from the same origin.In the last part of this study,to treat samples with different sample orientations in the same magneticfield,an apparatus with the capability of creating pulsed magneticfields in the central region of a pair of iron yokes is used in this work.The magneticfield strength,pulse pattern and pulse frequency are set as constants.Then three treatments with different orienta-tions are processed as‘case1’,‘case2’and‘case3’,respec-tively.In case1,the direction of magneticfield is parallel to the vector x of the specimen.In case2and case3,magnetic field directions are parallel to vectors y and z,respectively, as shown in Fig.2.To reduce accidental errors,two speci-mens were processed identically in each case.Thus,six total specimens are treated with three different orientations in this work.The characteristics of the magneticfield treatments ap-plied are proprietary information.In general terms,the mag-neticfield treatment consists of a series of magneticfield pulses with magnitude of the order of thousands of Gauss. The frequency of the pulses is of the order of several Hertz. Total treatment time is about5min.After treatments,the residual stresses are -paring the stress values before and after treatments wecanFig.2.Field orientations in cases1–3.determine the variations caused by the treatments.Then the effect of different orientation can be determined.3.ResultsWe define stresses along and vertical to the welding di-rection asσs andσy,respectively.Stresses before and after treatments are shown in Tables1–3for different cases.In all tables,figures in the same columns belong to the same Table1Stress variations in case1Distance fromorigin(mm)σx(MPa)σy(MPa)Before treatment180.0256.6219.231.3−17.3200.0238.1193.324.8−13.4220.0190.1187.25.78.5 Average228.3199.920.6−7.4 After treatment170.0237.4209.44.5−21.3190.0275.3194.131.0−14.0210.0210.5187.040.325.0230.0235.1175.07.9−38.6 Average239.6191.420.9−12.2346 C.Zhipeng et al./Materials Science and Engineering A398(2005)344–348Table2Stress variations in case2Distance fromorigin(mm)σx(MPa)σy(MPa)Before treatment180.0263.0260.243.45.4200.0245.8208.019.33.7220.0206.8213.3−4.946.0 Average238.5227.219.318.4 After treatment170.0217.6204.236.0−6.5190.0187.2189.142.7−4.6210.0210.7182.0−7.538.4230.0185.3170.0−8.6−24.8 Average200.2186.315.70.6 specimen and the data for two specimens is shown in each case.It can be seen that the welding beads are nearly in uniaxial stress states becauseσx values are much larger thanσy values. In that case,σy is neglected andσx is regarded as the only stress.We use a thermal–elastic–plastic FEM to simulate the welding residual stresses(Fig.3).A3D model with3900shell elements and4142nodes is established and MARC software is adopted for the numerical analysis.The materials properties depending on temperature are non-linear and a Gaussflux model is used.More details are listed in Ref.[16].The residual stresses along the welding bead are shown in Fig.4.It can be seen that the measurements are consistent with the simulation results.The stress along the welding beadσx is about225MPa,and the transverse stressσy is about zero in the measured area.Accordingly,the center region is indeed in a unaxial stress state.The summary of variations ofσx caused by different mag-neticfield orientations is shown in Table4.It is shown that in case1,the magneticfield treatment has nearly no effect on stress release.The stress variationsTable3Stress variations in case3Distance fromorigin(mm)σx(MPa)σy(MPa)Before treatment180.0293.7232.420.84.2200.0281.3268.728.325.2220.0312.7286.143.4−18.5 Average295.9262.430.83.6 After treatment170.0225.3211.460.1−17.4190.0200.6194.634.859.3210.0219.4180.8−1.9−26.3230.0232.2192.3−15.39.8 Average219.4194.819.46.4Fig.3.Mesh of thermal–elastic–plastic FEM.are within5%.In case3,the variation of stress is the most remarkable and26%residual stress is relieved.In case2, the effect is in the middle and the reduction degree is about 17%.From the above experiments,we conclude that magnetic field treatments are influenced byfield orientation.Treating along the welding direction,the residual stresses have nearly no change.Treating along the normal direction of the plate would produce the most release.And treating perpendicular both to welding direction and to plate normal has a median effect.4.DiscussionsFrom the above experiments,it is obvious that the effects of magneticfield treatments are related to differentfield ori-entations.The explanation in this paper is based on an earlier derived formula relating the stress energy in aferromagnetuni Fig.4.Residual stresses along the welding bead by FEM.C.Zhipeng et al./Materials Science and Engineering A398(2005)344–348347 Table4Summary ofσs variationsCase1Case2Case3Averageσx before treatment(MPa)228.3199.9238.1227.2295.9262.4 Averageσx after treatment(MPa)239.6191.4200.2186.3219.4194.2 Stress increment(MPa)11.3−8.5−38.1−40.9−76.5−68.2 Relative decrement(%)−5.04.216.017.625.826.0to the magnetostriction constant and magneticfield orienta-tion.4.1.Referred ferromagnetics formulaIn Ref.[17],if an uniaxial stress is applied to a piece of isotropic material in a constant magneticfield,the stress energy should be formulated as follows:Fσ=−32λsσcos2θ(1) In formula(1),cos2θ=α1γ1+α2γ2+α3γ3Fσis the energy increment caused by applied stress,λs the magnetostriction constant,σthe applied stress,α1,α2,α3 the direction cosines of the magnetic moment,γ1,γ2,γ3the direction cosines of the applied stress.It can be seen that the applied stress would change the direction of magnetic moment to make Fσhave the lowest value.In formula(1),if the magnetostriction constantλs>0,θ=0orπwould cause energy Fσto be the minimum.While ifλs=0,the direction of magnetic moment would be perpen-dicular to the direction of the applied stress to keep the lower energy state(θ→π/2or3π/2).In magneticfield treatments,if the applied magneticfield and magnetic moment are in different directions,additional energy has to be applied to arrange the magnetic moment along the magneticfield.If the applied magneticfield den-sity is constant,the most efficient treatment is that the applied magneticfield and magnetic moment are in the same direc-tion.In this paper,the welding beads are in a uniaxial tensile state along the x direction.Ifλs<0,the vector of magnetic moment would be perpendicular to the x direction,including either y or z directions.This conclusion implies cases2and 3would be more efficient than case1at reducing Fσ.More stresses would be reduced since the applied magneticfield is the same.This deduction is coincident with the experiment results.So,it is necessary to investigate the magnetostriction constant to make sureλs of#20steel is negative.4.2.Measurements ofλsThe magnetostriction constant of single grains of ferrum is anisotropic.The coefficients in the directions of[100] and[111],marked asλ100andλ111,are20.7×10−6and −21.2×10−6,respectively[17].For the multi-grainedmate-Fig.5.Magnetostriction constantλs of#20steel.rials,such as the steel used in this work,the magnetostriction constant is isotropic and the value is not really determined yet.In this paper,the magnetostriction constantλs of#20steel is surveyed.The results are shown in Fig.5.It can be seen that the material has a non-linear negative magnetostriction constant(λs<0).Thus,it can be predicted from formula(1)that the most efficient direction for mag-neticfield treatment in this work should be perpendicular to welding direction,including y and z directions.And treat-ing along x has the least effects.These predictions are fairly coincident with the experiment results.So,the release of the residual stresses is correlative with thefield orientations of the magneticfield treatments.For steels,the directions perpendicular to the maximum stress would be efficient directions for magneticfield treatments.5.SummaryEffects of pulsed magneticfield treatment are related with field orientations.Applied magneticfield perpendicular to the direction of maximum stress has potential to release more residual stresses.AcknowledgementThis work was supported in part by the Beijing Nature Foundation(Project no.3042010).348 C.Zhipeng et al./Materials Science and Engineering A398(2005)344–348References[1]D.Paulmier,Surf.Coat.Technol.76–77(1995)583–588.[2]ler,Tool.Prod.55(12)(1990)100–103.[3]R.F.Hochman,N.Tselesin,V.Drits,Adv.Mater.Process.8(1988)36–41.[4]O.Batainech,B.Klamechi,B.G.Koepke,J.Mater.Process.Technol.134(2)(2003)190–196.[5]M.S.C.Bose,Phys.Stat.Sol.A86(1984)649–654.[6]Y.Fahmy,T.Hare,R.Tooke,H.Conrad,Scripta Mater.38(9)(1998)1355–1358.[7]L.Baotong,Q.Shengru,S.Xiaoyan,Scripta Mater.40(7)(1999)767–771.[8]F.D.Snegovskij,V.A.Uvarov,Trenile I Iznos3(1991)535–539.[9]S.N.Prasad,P.N.Singh,V.Singh,Scripta Mater.34(12)(1996)1857–1860.[10]B.E.Klamecki,J.Mater.Process.Technol.141(3)(2003)385–394.[11]F.Tang,A.L.Lu,H.Z.Fang,J.F.Mei,Mater.Sci.Eng.A248(1998)98–100.[12]A.L.Lu,F.Tang,X.J.Luo,J.F.Mei,H.Z.Fang,J.Mater.Process.Technol.74(1998)259–262.[13]F.Tang,A.L.Lu,J.F.Mei,H.Z.Fang,X.J.Luo,J.Mater.Process.Technol.74(1998)255–258.[14]F.Tang,A.L.Lu,H.Z.Fang,X.J.Luo,Trans.China Weld.Inst.21(2)(2000)29–31.[15]F.Tang,Research on residual stress reduction in steel by pulsedmagnetic treatment,Ph.D.Dissertation,Department of Mechanical Engineering,Tsinghua University,1999.[16]Z.Cai,H.Zhao,A.Lu,Sci.Technol.Weld.Joining8(3)(2003)195–204.[17]Z.Wending,Ferromagnetics,vol.2,Science Press,Beijing,2000,pp.3–44.。