DEVELOPMENT OF MAGNETIC INSPECTION TECHNIQUES FOR EVALUATION OF FATIGUE DAMAGE AND STRESS IN LOW ALLOY STEELS M.K. Devine, S. Hariharan, t.j.h. Brasche, and D.C. Jiles Center fr NDE Iwa State University Ames, IA 50011 INTRODUCTION It is knwn that the magnetic prperties f ferrmagnetic materials change under fatigue and applied lads (1,2,3). These changes in the magnetic prperties culd be used as indicatrs f the stress state f the material, r pssibly fr predicting the remaining fatigue life. Previus reprts have shwn successful implementatin f this magnetic measurement technique fr NDE f steel samples in a labratry envirment. Hwever, fr this technique t be practical, a field-usable instrument must be develped. This paper will describe measurements using ne such instrument, the Magnescpe. It will als discuss the techniques used t evaluate the effects f applied lads, bth tensile and cmpressive, and lw cycle fatigue n a variety f materials. A brief descriptin f the Magnescpe will be given here. Further details have been given by Jiles, Hariharan, and Devine [4]. The Magnescpe is a prtable magnetic inspectin device cnsisting f a persnal cmputer, gaussmeter, fluxmeter, and a biplar prgrammable pwer supply. Bth the meters and the pwer supply are cntrlled by, and send input t, the cmputer. The sftware fr cntrlling the system was develped at the Center fr NDE. The Magnescpe is capable f perfrming an inspectin n a specimen and immediately reprting the magnetic hysteresis parameters. A variety f inspectin heads can be used with the Magnescpe. The type f inspectin head used depends n the specimen gemetry. An inspectin head cnsists f a C-shaped sft irn electrmagnet, with a flux cil t measure the magnetic flux density. The magnetic field strength is measured by a Hall sensr, permanently psitined between the ples f the electrmagnet s as t be n the surface f the specimen being measured. The Magnescpe is capable f demagnetizing smaller specimens. If the specimen is large, it will demagnetize the regin where the measurement is taken. It then magnetizes the sample thrugh a hysteresis lp. The field is swept at a frequency f typically 0.01 Hz. Experimental Prcedure All the magnetic measurements reprted in this paper were perfrmed by the Magnescpe with a specially cnfigured inspectin head. Each sample was demagnetized by the system and a hysteresis lp was measured. Frm this the remanence, cercivity and maximum differential permeability were recrded. Review f Prgress in Quantitative Nndestructive Evaluatin. Vl. lob Edited by D.O. Thmpsn and D.E. Chimenti. Plenum Press. New Yrk. 1991 2021
T determine the effect f applied stress n the hysteresis parameters, samples were taken frm the web sectins f varius railrad rails representing 9 different metallurgical cnditins. Flat tensile specimens were machined frm chrmium/vanadium, chrmium/mlybdenum allyed steel rails. There were als a number f standard carbn rails representing intermediate strength and intermediate hardness levels, a head hardened and a fully heat treated rail, and a stretch straightened rail. The stresses applied were within the elastic limits fr all the specimens. Five stress levels were applied, -200, -100, 0, +100, and +200 MFa. Magnetic measurements were taken while the stress was applied. Fur parameters frm the hysteresis lp were tabulated: remanence, maximum differential permeability, cercivity and hysteresis lss. Tw materials were studied fr the effects f fatigue. The first was a medium carbn steel taken frm a railrad bridge whse heat treatment histry was unknwn. The samples were cylindrical bars 10 cm. in length with a 0.6 cm. gage diameter. The secnd material studied was 4340 steel. This was austenitized at 8500 C fr 1 hr. fllwed by an il quench, tempered fr 1/2 hr. at 4000 C and il quenched. The samples were cylindrical bars, 7 cm. in length and a 0.6 cm. gage diameter. Fatiguing fr bth materials was at cnstant ttal axial strain amplitude. Strain cntrl was perfrmed by an extensmeter. The rail samples were fatigued t a predetermined number f cycles and remved, with n further fatiguing. As the apprximate number f cycles t failure, Nf, was knwn, the percentage f expended fatigue life was calculated. Nne f the samples were fatigued t failure, s the percentages calculated are nly an estimate f expended fatigue life. The magnetic measurements were made n the samples after fatiguing was cmpleted and stresses remved. The samples f 4340 were fatigued at tw strain amplitudes, 0.003 and 0.005. The frequency f the cycling was changed s that the strain rate was kept cnstant. All samples were fatigued t failure (defined as the cycle at which the applied lad drpped 50% frm its plateau value). Magnetic measurements were perfrmed generally every 100 cycles, while the sample was in the grips. Each measurement was taken with the specimen under zer stress. RESULTS AND DISCUSSION Appljed Stress Figure 1 is a plt shwing the remanence vs. applied stress fr the 9 specimens with different metallurgies. As can be seen, all the samples shw changes in respnse t cmpressive stresses. There are als changes as the materials underg tensile stresses but this is much diminished. Apparently the metallurgy f the specimen des nt play a significant rle in determining the dependence f remanence n cmpressive stresses. This is nt true fr tensile stresses hwever. Sme specimens have strnger respnses t tensile stresses than thers, ntably W6 (a standard carbn rail) and WlS (a standard carbn, stretch straightened rail). One specimen (WI, a Cr-V ally rail) even shwed a decrease in remanence under tensile stresses. The variatin in maximum differential permeability with applied stress is shwn in Figure 2. Here it can be seen that fr mst f the specimens there is a decrease in the maximum differential permeability under cmpressive stress, and an increase under tensile stress. This was nt seen in three f the specimens: WI, WS and WIS. All three f these samples shw decreases in maximum differential permeability in respnse t bth tensile and cmpressive stress. These three specimens are the Cr-V and Cr-M allys and it is believed that the allying elements diminish the changes in maximum differential permeability f these materials. 2022
Changes in the cercivity with applied stress are shwn in figure 3. As can be seen, all f the samples shw an increase in cercivity with cmpressive stress and a decrease in cercivity with tensile stress. This respnse appears t be independent f the metallurgy f the specimens. Hwever, the magnitude f the cercivity des change with cmpsitin. The effect f applied stress n the hysteresis lss can be seen in figure 4. All the samples shwed either minimal changes with stress r slight decreases in hysteresis lss with bth tensile and cmpressive stress. ~,---------------------------~ 5500..- Standard Carbn (WI).-".. (Cf+l) aeei SC fully H. Treat.(W2)... SC Inter. Hard._ (Wl) ~ SC Inter. Str. (W4)... SC Stretch Straight (W15) _ Cr IV ally (WI) +++ Cr/l.l ally (W5) Cr/l.l ally (W2l) ~i---~----~---r----r----r--~ -200-100 0 1 200 JOO Applied Stress (MP) Figure 1. Variatin f remanence with applied stress fr varius metallurgical cnditins.,..., ";;;-500 II) :::J (!) -400 ~,--------------------------'r----' E ~ Cl..JOO -i5 )( :IE 200 l00i----,----,-----r----r----~--_1-300 -200-100 100 200 JOO Applied Stress (MP)..- Standard Carbn (WI) _".. (Cf+1 aeei SC Fully H. Treat. )... SC Inter. Hard. 3) ~ SC Inter. SU. 4)... SC Stretch Straight (W15) ~g~~aal~'lry(~5) Cr/l.l ally ~23) Figure 2. Variatin f maximum differential permeability with stress fr varius metallurgical cnditins. 2023
45,--------------------------------, 40 G-- e e e 0 ~, +--.,.... 15..- Standard Carbn (W6)...".. (CF+I) EEEI SC Fully H. Tre~t. ) _ SC Inter. Hard. 3) ~ SC Inter. Str. 4)... SC Stretch Straight (W15) eeecr/y ally (WI) +++ CrZM ally (Vi5) - Cr/M ally (W2:') IO;-----r----,-----r----,-----r---~ -300-200 -100 0 100 2 300 Applied Stress (MPa) Figure 3. Variatin f cercivity with applied stress fr varius metallurgical cnditins. Fatigue Figure 5 illustrates the variatin in cercivity fr the railrad bridge steel specimens, measured at certain pints in their expended fatigue life. It can be seen that the samples fatigued at higher strain amplitudes shwed larger changes in the cercivity. This effect was als reprted by Bse (1). It may be expected that the change in cercivity caused by different strain amplitudes is due t increases in the dislcatin density. Higher strain amplitudes wuld increase the dislcatin density, prducing mre pinning sites fr the dmain walls, which in turn wuld lead t higher cercivities. The "e" sample, fatigued at strain amplitude 0.005, shwed a drp in cercivity near the end f its fatigue life. This can prbably be attributed t crack frmatin prir t failure. I~,-------------------------------,,... 01I0000 "- 0' ~ -140000 fii fii...j fiil20000 "iii ~ ~IOOOOO fii >- :I: Figure 4. t:::.v±== e ~ ~ It;::::: === ::i 1OOOOi----,----,----,----,----,r---~ -300-200 -100 0 100 300 Applied Stress CMPa)..- Standard Carbn (W6)..... (CF+I) EIBEI SC Fully H. Tret.(W2) _ SC Inter. Hard. (Vi3) ~ SC Inter. Str. (W4)... SC Stretch Straight (WI5) eeecr/y ally CV<!~ +++ CrZM ally 5)... Cr/M ally 2:') Variatin f hysteresis lss with applied stress fr varius metallurgical cnditins. 2024
7.5 7 6.5 a " 6.5.5 >- ~ 5.~ 4.5... 0 u 4 3.5.: 3 ci 0 2.5 0 10 20 30 40 50 60 70 BO 90 Applied cyclic fatigue (% f life) -EI- "A" samples, 0.002 --Ar "A" samples, 0.005 --- "e" samples, 0.005 3.B ~ 3.6 f3.4.:;.~ 3.2 " 8 3 2.B 4,---------------------------~----------~ 2.6+-----.------r-----,----~r_----._----._----~ 10 20 30 40 50 Applied cyclic fatigue (% f life) I-e-- "8" samples, 0.002...-. "8" samples, 0.005 I 60 70 Figure 5. Cercivity vs. expended fatigue life fr the railrad bridge steel specimens. Figure 6 is a plt shwing the remanence vs. expected fatigue life fr a number f railrad specimens. Again we see the large initial increase in remanence which may be fatigue sftening. Hwever, there was n reductin prir t failure. Previus wrk by Brasche et al. (2) has shwn the reductin t ccur after 90% expended fatigue life. As these samples were fatigued nly t 90% f expected fatigue life, a larger reductin may ccur upn further cycling. Figure 7 shws the cercivity vs. percentage expended fatigue life fr the 4340 samples at tw different strain amplitudes. As can be seen, bth shw a large reductin in cercivity early in the fatigue life. This reflects fatigue sftening which the samples underg in early stages f fatigue. There is a larger drp in the cercivity fr the materials fatigued at strain amplitude 0.005, similar again t the effect seen by Bse. The remanence vs. percentage fatigue life can be seen in figure 8. Here it can be seen that there is a steep increase early n and then a levelling ff, again a reflectin f the fatigue sftening the material has undergne. 2025
13 12 11 ";;;' 1 0 '0 C iii a 9 () II> C :l ~ 0 8 C..c I-... 7 E ~ 6 5..... 4~---r---,----~---r--~----r---'---~--~ 10 20 30 40 50 60 70 Applied cyclic fatigue (% f life) 80 90 -B- "A" samples, 0.002 --.- "A" samples. 0.005 ---- "C" samples, 0.005 Figure 6. Remanence vs. expended fatigue life fr the railrad bridge steel specimens. ".00.,--------------,.8.00,.-------------~ ---.. CI) 14.00 : ~'3.00 :?;-.';; " U 11.00... CI) U 11.00 SINI. Amplllud.: 0.003 10.00 -+---,-----,.----.---.,----1 20 40 &0 a Percentage fatigue life Figure 7. ".00 ---. CI) 14.00 ~ Z'13.00.;; 'u... CI) '2.00 U 11.00 Sltel. Amptlt.d.: 0.00' 10.00 +---.---------.----.---,.-----l 20 40 &0 110.00 Percentage fatigue life Cercivity vs. % fatigue life fr the 4340 specimens. 2026
~~------------------------~ ~~-------------------------, 'Ui'7000 UI :J '-'8000 () C ~5000 E 0::4000 StraIn Amplitude 0.003 'Ui'7000 UI :J... 8000 () C ~5000 E 0::4000 Stn.I" Amplitude: 0.005 5OOO+-----r----,-----r----~--_4 20 40 60 e Percentage Fatigue Life 5OOO+-----r----,-----r----~--_4 20 40 60 e Percentage fatigue life Figure 8. Remanence vs. % fatigue life fr the 4340 specimens. CONCLUSIONS The magnetic hysteresis parameters, such as remanence, maximum differential permeability, cercivity and hysteresis lss have been shwn t vary with applied stress fr a variety f metallurgical cnditins. The remanence was seen t decrease under cmpressin and increase under tensin. The maximum differential permeability was seen t decrease with cmpressin and increase with tensin. The cercivity was seen t increase under cmpressin and decrease under tensin. This hysteresis lss decreased under bth tensin and cmpressin. The metallurgy f the samples played a significant rle in the magnitude f changes in the parameters. In general, the specimens with fewer allying elements shwed larger changes. The specimens with Cr-V and Cr-M shwed smaller changes in prperties when cmpared with the ther specimens. The magnetic hysteresis parameters, such as remanence, cercivity and maximum differential permeability have been shwn t vary with fatigue in sme ferrmagnetic materials. Previus wrk (4) has shwn that the remanence decreases sharply near the end f fatigue life. This was nt bserved in the samples f 4340. The magnetic hysteresis parameters have been shwn t reflect either strain hardening r strain sftening during fatigue. The strain amplitude seems t affect the magnetic parameters althugh this is nt entirely understd. Larger strain amplitudes lead t decreases in cercivity at the end f fatigue life. Mre wrk needs t be dne t determine why sme materials shw larger changes in magnetic prperties than thers. The pssibility f using a prtable magnetic inspectin system fr detecting fatigue in the field appears feasible. It has been shwn t be capable f detecting applied stress in test specimens. The Magnescpe has als been shwn t be capable f detecting changes in materials due t fatigue. ACKNOWLEDGEMENT The authrs wish t acknwledge the cperatin f D. Utrata and the Assciatin f American Railrads fr prvisin f the rail specimens and fr the use f facilities fr fatiguing. M.K. Devine als acknwledges prvisin f a summer internship frm the Assciatin f American Railrads. 2027
This wrk was funded by the Center fr NDE at Iwa State University and was perfrmed at the Ames Labratry. Ames Labratry is perated fr the U.S. Department f Energy by Iwa State University under Cntract N. W-7405-ENG-82. REFERENCES 1. Bse, M.S.C. "A study f fatigue in ferrmagnetic materials using a magnetic hysteresis technique", NDT Internatinal, 19, 83, (1986). 2. Brasche, L.J.H., D.C. Jiles, O. Buck, and S. Hariharan "Nndestructive methds fr the determinatin f mechanical prperties f materials", in D.O. Thmpsn and D.E. Chimenti (eds.), Review f Prgress in Quantitative NDE, Vl 9B, Plenum, New Yrk, 1990, p. 1581. 3. Shah, M.B. and M.S.C. Bse "Magnetic NDT Technique t Evaluate Fatigue Damage", Phys. Stat. Sl. (a) 86, 275, (1974). 4. Jiles, D.C., S. Hariharan, and M.K. Devine "Magnescpe: A Prtable Magnetic Inspectin System fr Evaluatin f Steel Structures and Cmpnents", IEEE Transactins n Magnetics 26, 2577, (1990). 2028