1、+c)m#. - L./-1qNATIONAL ADVISORY COMMITTEE r-armzFOR AERONAUTICocopy *#AFWLiDOUlKiRtL but, despite precautions taken to maintainflatness, the unnotched hardened-steel specimenswere warped to a degeevarying between virtual flatness and 0.25 inch out of a plane. The bendingstress introducedby straiten
2、ing a spechmn assmned to have a circularcurvature of the specimen face with O.= inch as the rise of the arc is7.5 ksi.All the notched specimens tested at the Lsmgley Laboratory wereunpolished. Most of the unnotched specimenswere electropolishedaswere all the notched smd unnotched specimens tested at
3、 Battel.leMemorialInstitute. (Seerefs. 4, 5, and 6.)EQUTFMENTTwo types of fatigue testing machines were used in this series oftests. One was a subresonantmachine which operates at 1,800 cpm. (Seeref. 5.) The natural frequency of the system was adjusted to about1,90 by mg tie maSS of tie log it ch W=
4、S excited bya rotating eccentric.A photograph of the second type of testing machine, a double-actinghydraulic jack, is presented as figure 2. The principal parts of thismachine are: a constant-dischargepump, a rate-control valve, a four-wayvalve to direct the hydraulic pressure, a double-acting hydr
5、aulic rem,and a null-method air-operated weighing system. The machine operates ina manner similar to that of other hydraulic testing machines. This machtiewas modified by the addition of an electricweighing system end an airservo for operating the four-way valve. Contacts on the electric loadindicat
6、orwere adjusted to actuate the air servo whenever the load on thespecimen reached the desired value. The hydraulic pressure was thusProvided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-4 NACA TN 3866directed to the opposite side of the losd piston to reve
7、rse the direction *of load application. Special grips stiilar to those used in the subreso-nant machines were used to permit testing of sheet specimens. (Seeref. 5.) .Guide plates similar to those described h reference were used toprevent buckling of the specimens. A low-voltagecurrent was passed co
8、n-tinuously tlm?oughthe specimensto operate a relay which stopped thehydraulic pump when the specimenfailed.An electronic load-measuringdevice was used to monitor the appliedlosds in the automatically controlled tests. Monitoring was necessarybecause time delays in the automstie-controlmechanism mad
9、e it difficultto preset the limiting contacts on the electricweightig system with suffi-cient precision. The loads were measured with the electronicmonitoringequipment with a msxinnm error of approximately*1 percent.TFSTSAND TESTING F!ROCEDUREFinal load adjustmentswere necess during the initial stag
10、es ofeach fatigue test. Since the high-stress tests terminated after a small #nwber of cycles, a relatively slow acting machine (thehydraulic jack)was required in order to allow the adjustmentsto be made before a largepercentage of the total life had elapsed. A faster machine (the subreso- *nsnt typ
11、e) was required to perform the low-stress tests within a reason-able length of the.During those tests in the jack in which failure was eected to occurafter 30 cycles, the rate-control valve was fully opened to allow maximumtesttag speed. Ioads were controlled automaticallyby the electric con-trollin
12、g device described in the section -titled “Equipment”. Cyclingspeed was dependent on the load rage and varied from about 14 to 50 cpm;the higher load ranges correspondedto the lower frequencies.Tests in which failure was expected to occur in less than about30 cycles were manually controlled in the d
13、ouble-actinghydraulic jack.In these tests, the rate-controlvalve was used to decrease the loadingrate when approaching the maximum and minimum loads for more preciseload control. The frequency of manual cycling varied from 0.4 to 1.0 cpm.Load-time curves for the jack are illustrated in figure 3. The
14、 precipitousunloading was due to the sudden release of oil pressure which occurredwhile shifting betwea tension and compression. The curved portionsresulted from manipulation of the rate-control valve.Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-N
15、ACA TN 3866 3The fatigue behaviors of four materials with various combinationsof and were investigatedby covering the life range from 1 toapproximatelylC cycles for each combtiation shown in the followingtable:Mean stress, , ksi, for -Material = 1.0 = 2.0 Q = 4.02024-T3 aluminum alloy o 0 and 20 0 a
16、nd 207075-T6 aluminum alloy o 0 and 20 OandmNormalized SAE 4130 steel o 0 and 20 0 and 20Hardened SAE 4130 steel osnd Oando OandMost tests were run at stresseswhich caused failure in less than10,000 cycles. A few tests in each group were run at lower stresses toafford comparison of the results with
17、data obtained at BattelleMemorialInstitute on similar specimens. (Seerefs. 4, 5, and 6.)T!heeffect of cycling speed on the fatigue strength was investigatedin a limited way by testing identical specimens at the ssme stress condi-tions but at different cycling rates. For practical reasons these tests
18、were ltiited to stress levels which were expected to cause failure in theneighborhood of 10,000 cycles. High-speed tests at shorter lives werealmost impossible to perform and low-speed tests at longer lives wouldhave been extremely ttie consuming.The greatest errors in load applicationwere less thsn
19、 5 percent andoccurred during the first few cycles of the automatically controlled testswhile final adjustments were being made.RESULTS AND DISCUSSIONThe results of the fatigue tests are given in tables II to V and arepresented in figures 4 to 15 as maximum nominal stress plotted agahst thenmber of
20、cycles to failure(designatedherein as S-N curves). Thescatter in the results of the tests in the short-lifersnge was remarkablysmall, whereas the tests at long lifetimes indicated considerablymorescatter in the results.Of the unnotched harden-steel specimens, 19 were appreciably warped -after heat t
21、reatment. During these tests the guide plates, which wereemployed to prevent buckling, straightenedthe specimens and necesssril.yProvided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-6 NACA TN 3866introducedbending stresses,with the maximum stressesprobabl
22、y occurring at -the minimum cross section. The fatigue cracks tn 13 of the 19 warped spec-imens were initiatedon the concave face (theface that probably contained .-tensile bending stresses due to straightening). However, the scatter in theS-N curves Por the unnotchedhardened-steel specimens (fig. 1
23、3) was notextreme and indicatedthat these bending stressesplayed a minor role hdeterminingthe fatigue life.The minimum number of cycles to failure, greater than 1, for all theS-N curves regardless of the value of mean stress fell between 2 sad 58.Minimum lives for those groups subjectedto completely
24、 reversed loadingonly (R = -I) were less than 16 cycles. These mtnimur.ulives differedfrom those published in reference 3 which showed that, for R = O, fatiguefailures at stressesnear the ultimate tensile strength did not occur inless than roughly Ld, 1, and 102 cycles for specimenshaving values of
25、equal to 1.0, 2.0, and 4.0, respectively. The materials used fn thatinvestigationwere 6061-T6 aluminum alloy and 347 and 403 stainlesssteels.The present investigationresulted in S-N curves that sre concaveupwsrd at the long-life end and have a reversal of curvature at a life-time dependent on the st
26、ress-concentrationfactor and, to a Lesser extent, #on the mean stress. These inflectionpoints occur at roughly 105, 103,and 102 cycles for stress-concentrationfactors of 1.0, 2.0, and 4.0,respectively,for all four materials. The S-N curves for mean stresses .greater than O generallyhave the reversal
27、 at a somewhat greater numberof cycles than the curves for mean stresses of O.Of practical interest to the aircraft designer is the fatiguebehavior of specimens subjectedto repeated stresses in the victiity oftwo-thirds of the ultte tensile strength. This stress correspondstothe limit design stress
28、of a given aircraft part. Table VI gives thenumber of cycles to failure at this stress level for each material andtype of specimen. The specimenswith the highest stress-concentrationfactor had the shortest lives at this loading with the aluminum alloyshaving the lawest values. The results of the tes
29、ts on steels at R = -1compared on this basis show that the hardened steel has a longer fatiguelife than the normalized steel for unnotched spec3mens,whereas the reverseis true for the notched specimenswith = 2.0 and 4.0.If it is assumed that for R = -1 an unnotched spechen would failin the same numb
30、er of cycles as a notched specimen,provided the maximumlocal stresses are equal in both specimens, it follows that the effective *stress-concentrationfactor of the notch would be equal to the ratio ofthe maximum nominal stresses in the two specimens. !lhisratio KF ofthe nominal stresses at the same
31、nwiber of cycles is plotted agatist the vmaximum nominal stress of the notched specimens in figure 16.Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-NACA TN 3866 7221figure 16, the limits of-the scatter bands are the ratios of thecorresponding limit
32、s of the scatter of the S-N curves. The KF curvesextend to the ultimate tensile strengths of the notched specimens. Themaximum values of KF were generally smaller than because sizeeffect reduced the severity of the notch. (Seeref. 8.) b general,KF decreased with increasednominal stress because the m
33、aximum localstress entered the plastic range. The width of the scatter band for KFalso decreased with increasednomimal stress.It was found in previous investigations,such as those reported in ref-erences 3 and 9, that the tensile strengbh of notched specimens sometimesexceeded that of unnotched spec
34、imensmade of the same material. Zn thepresent investigation,the notched 7075- specimenshad somewhathighertensile strengths than the unnotched specimens; for q = 2.0 the increasewas 9 percent and for = 4.0 the increasewas 4 percent. For notched2024-T3 specimens,however, the reverse was true; that is,
35、 for = 2.0there was no static-strengthchange smd for KT =4.0 a reduction of8 percent was produced. The tensile strengths of the notched steel speci-mens, both normalized and hardened, were about 8 perc”enthigher than thoseof the unnotched steel specimens.No effect of polishing was found. Also, no de
36、finite difference intest results was found between specimenstested at x and 1,800 cpm; however,it should be noted that only a very small number of tests entered intothis ccuuparison.CONCLUSIONSFatigue tests were performed in the life range from 2 to 10,000 cycles,and previously published data have b
37、een included to extend the data toUfetiroes up to I 87.4.*.O 168.0e-xi59479m.78.ak).7.4.1-.Oltiuetstende !lbtal.elmgatbn, Z-ilwh we L?a8th,” “ -* (:) ” -“ M=”72.1 70.3 73.4 0.9 20.3 15.0 a.oEm 79.8 wk. 1.1 12.3 7.0 15.015.9 IJLk 1.24.61.8 15.2 Iz.o 18.oB 12 JA76m 7 10b C=tnilltiA?321s % 1232AU5M2 IA
38、* SI.6$: * MaIuaUyCantrOuedad tically cmt,rDlldA* 218 663%7 z Autitf. 15 Y cOntniUdA.103M2Alo5K2 1,em% Zz .aarenamntlrucms (M. 5fiM2 l,am: g % “1tic Cmtmlled. ,Elf: 1,%3JTz: % Smrwcmsllts (ref.5)N09K2 45.0l,U07H3 m.o 1*WIU03M1 l,eahktl z: l,acaN13m2 %$%30.0 l,eloI19,wo 1,520M,!35M3 30.0AUEM2 L,exl31
39、$%kg9Hl %: l,exlWM1 X.o Z#J 1,5X3UMl1,wNKVU2 l, 1,8C0179213 1,WUC5M1 - - 1A53 !5 . 21 1.2nhuh ccdmU.EdA33SI.3 62.5 39 14.AZJ2J.9 62.5 41 - kbnum.yOmltmwAXM2 Ss.o - 13 AntautA5+=2 3.0 1387cOnfrOu.eAAg+2ilo p.: 139 zA39l25% a.AZQ7 40.0 WQ7 23.A*SL6 1,049 21.A7922B % 3,km l,lm -tte (ref.6)A 35.0 3,%U l
40、,lw 12,%fJ 2sM3A3B Autaa.ta Oaltmllul$: 6,500 l,lmAa3 30 Autc$mtic25.0 YOcm U.o 20Jc2S 29 Autamat.iYcontrolled5,43049:0 vAj+322B 6,m I,loo Hteu.e (r=.6)22Z:kg.o 1,1.W45.0 2,E l,lmA92.22B 45.0 W;g p 29.5 V m,w% 1,1oo 1wtdkL(ref. 6)mgB2B w,ti,m 1,1ooA% 017 25:0 w,w, 1,8C$I 2utuema5mtDacldnes.4“Provide
41、d by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-13m.mE.n.-AxIAbIt2ADFAmous!msTBs9uL :2A30s15 6s.0 ; .9A30s12 62.0 1A32S 9 E 32 :A31s15 47.5 377 29A%s15 ko.o 57A5J3BII.O 37.4 t% 37 IA12S3B 3s.0 1,1.Wwlm 33.0 :%mttszk (ref.6)21 AYltmaticallywntrolka=9S3B 3s.5
42、 9,0XI lJCQ JAtiBlk 22,m mt (ref.6)A49S3B l,ttxl% ti,tw SubrasmantmcbineaAI.6S3B l,loa27.5mtt (ref.6)39,403 1,100 JA38=a3 hg,fxrJ I,mA37S3B Suhr-t D!achinenz: l,343,m I,m =e (ref.6)WS3B 22.5 lo,3a,m 1,100 JProvided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-
43、,-,-NACA TN 3866TAHLE III.-AXJAL-KIADFAIL(GUETESTUSUIilSKIS -!l% MLM13KM-AIU3Y SEUCC FFFCD4EKS(a) -1.O;C=OSpecimen Meximm stress, Fatiguellfe, Frequency, kei N,cycles F R.emerksB33Sl8 t-.- - -B39S1k :.: 46 Statictensileteet end 2attdJ.e(ref.4)-*O S16 Neweuy Controlled-twjs13244s17m:o 1% - AutOnlEtic
44、ellyCIm.tron.ea143 12B35s15 %: 228 lk- -(0.0 13B5 m 3 l,%2:;20E54KL3 1,688 L5?A2sll 50.0 y,182 19alsl50.0 8,L32m#ma.16 I50.0 18,0w 1,8R Subre80n0ntm.chines (ref.5)19,0KI 1,mB3TSI.L %: 27,0m 1,8XBIOlK 1 W.o 33*W l,em2232M2 50.0 36,W3 1,:; 1,800B37S1225.0 248,m 1,800Blu Ml 25.0 *,mo l,WJOMl 293,003W.i
45、n61,C% l, -m.o.32k9m 3B4821Ul %: J .-7ma! %; 10248217 u :3%! 62.5 15Cay .717B49M.9 .7 I25.0 lb Autmti.cdlycmtrOIJdB$9!31-2 55.0 -J%5? 47,5 1.047.5 = c-$ 17249=41.xdrolled: e-249217 WBt9215 1.9E4,9.218 52.5 329 2332.5 * .-B99816 2,622 e249211 % 2,228E47s17 24.5 lW247215 2Q.O 5,2:;.25583B l,lm2XM6 :;
46、2%$%B3683B l,ex BmraacQnllt9.25 A539,2X l,lsa mttdm (-. 6)BW23B 8.5 S69,200 JB51B19 8.0wSmB D,232,CC0 :% 21bres-t mOh.fJMsl,652,m l,lofl mtteua (w. 6)B3B;:? 4,722,CCJ3 l,lmW23B?:X2,11ql,ym 1,mB2923B X4,zw,m l,lul 1.$nm6215 e6.o + tanueUyWntrollea%; .%:B4721L63.0w m 10 23.0 2 :B4621.8 23.0B4.6213 R1.
47、o z t:26 l.o I24.623-14719B47219 A_ticmlly oontrolld18B47M4 55:0 $ 24WlmlR 1P 23B51218 6322532!.2 M.o m .-7% 30 IB21s3BBh62L6 35.0 Z,m L,m35.0 k (ref.6)3,LkJ4w Autaaaticdlycmtrol!dMB3B 52.5 5,503 l,KXI B98Bll mttel.la(m?.6)B9721330.0 2,639 20mAutitlu acdrollea9,W0 1,8XIBXL2L6 Wtmmlmut.Io,m 1,mBl183B
48、 %: lo,xu J.l,UXI Fattau.e(m?.6)B923B xl.0 lo,m l,mB98212 13!).0 U,alo 1,mg:;mbreml!antma-1,mZ.: 2% Mttam (rd.6)l,lcn 4$,% l,mB99BI.9 -Cmnt mohimc%25.02?mmB1 22.5M,!$g :? JB23B 22.3 wttau.e(ref.6),“.Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-JM.NACA TN 3866TABLIZIV.- AXIAL-LOADFATIGUEMSTRESULTSFORNORMALIZEDSAE4130- SHEE?rsPEcIMms(a) =1.O;=O17Specfmen Msximum stress, Fatigue life, Frequency, ksi I?,cycles RemarksCpmC204M2 120.5 - - Staticte
