1、tOZb-l_ but the exact reason for variationin life with block size cannot be explained at this time. Possible linesof investigation which might explain this variation are systematic studyof (1) fatigue crack propagation in multilevel tests, and (2) the rela-tive effect of single or multiple cycles at
2、 the high-load level.Effect of Truncating the Load ScheduleThe lowest stress level in load schedule 1 was well below thefatigue limitand theoretically contributed no damage (n/N = O, sinceN _ _) although this stress level contributes approximately one-thirdof the total cycles. A comparison of result
3、s for tests with and with-which wasout this load level (fig. 5) shows a slight increase in Nnnot found to be significant (table VI). The small increase in llfe mightbe explained by the actual damage contributed by the lowest load levelafter the fatigue crack had begun to propagate; thus 3 the effect
4、ivefatigue limit is lowered. Another possible explanation is that thisomitted load level probably contributes to the decay of the residual com-pressive stresses caused by higher loads.Omitting the highest load level in load schedule 1 gave a statisti-(table VI) when compared with testscally signific
5、ant decrease innwith the highest load level included (table III and fig. 5). Thisdecrease is probably due to the decrease in magnitude of the residualstresses; thus, the crack initiation and propagation phases of the testare shortened. Although the critical crack length for residual staticstrength a
6、t the highest load was increased, the increase in the crackinitiation and propagation rates overshadows this effect.L1462Effect of Reducing the Number of Load StepsIn the series of tests for which the number of load steps wasreduced from 8 to 4 (table III) the reduction in the geometric mean ofProvi
7、ded by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-13L1462n was significant (table VI and fig. 5). This reduction is probablydue to (1) a decreased magnitude of beneficial residual stresses becausethe highest load had been reduced, (2) a possible accumulatio
8、n of actualdamage from the lowest load level after the crack had been initiated,since the lowest load is now closer to the fatigue limit Se, and (5) anincrease in actual damage at the highest load because of the increasednumber of applications.In reference i there was not a significant difference fo
9、und between8 and 18 stress levels for gust-load histories. In the light of thepresent results it might be surmised that there is a minimum number ofstress levels above which the number of stress levels used would haveno effect.Effect of Adding Load Levels Above Design Limit LoadFo_ tests in which tw
10、o load levels above design limit load were_ nadded (load schedule 2), _ decreased slightly but not significantlywhen compared with results of tests without these load levels (table VI).The predominant effect of the higher loads appears to have been to causestatic failure when a shorter crack was pre
11、sent. Since the criticalcrack length was shorter for residual static strength considerations,but the difference in life was small, it would seem that the higher loadsdid increase the residual stress effect.Effect of Including Negative Load FactorsTwo methods of including the negative load factors we
12、re used, inone method the negative loads were applied in groups (load schedule 5).The other method combined the negative and positive cycles such thateach negative cycle followed a positive cycle (load schedule 4).For the tests in which negative load steps were applied in groups,(table VI) was prese
13、ntna statistically significant reduction inwhen compared with similar tests without negative loads (table V andZfig. 5). This reduction in _ is thought to be caused by the reduc-tion of beneficial residual stresses when compression loads were applied,thus crack initiation and propagation rates are i
14、ncreased.Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-14The amount the residual stresses are lowered appears to be dependenton the magnitude of the compressive load. The relative location of thehighest load and the negative loads within the random
15、ized blocks is alsoimportant because (i) damage from those loads applied between the highestload and the negative load will be reduced by the residual compressivestress, and (2) damage from those loads following the negative loads maybe greater than that computed because the residual stresses may be
16、tensile.Z nThe values of _ for tests with combined cycles were found tobe significantly lower than those for tests without negative loads butnot significantly different (table VI) from tests with loads applied in2groups. The reduction in _- is thought to be caused by a reductionin or elimination of
17、the beneficial residual stresses when compressiveloads are applied. This load schedule had a positive high load followedby a negative load and formed one load cycle. In this case the compres-sire one-half cycle decreased the effect of the residual stresses dueto the tensile one-half cycle. The exact
18、 value of the residual stressis not known and it is possible that a tensile residual stress existedin the specimen following the full cycle. This action was deleteriousto the fatigue llfe until a load was applied which was sufficient inmagnitude to reverse the nature of the residual stress. The cycl
19、e whichproduced the maximum compressive residual stress was lower than the max-imum load and therefore the increase in llfe was less than that assumedfrom similar tests without negative loads. A second reason for thedecrease in _ for tests with combined cycles is the reduced cracklength for which th
20、e specimen residual static strength becomes critical.Another factor which may have had a small effect on the value ofis the reliability of the assumed N values in _ for thecombined load steps. As has been stated earlier, these values wereinterpolated and some error is probable.Li462Other Observation
21、sLoad step at failure.- In 58 of the 64 tests the specimen failedduring the application of the highest load. This failure then illus-trates the importance of residual static-strength considerations. For,Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,
22、-15Li462although a high load applied periodically does influence the fatiguelife by introducing beneficial residual stresses, there is also thepossibility that the load will exceed the residual static strength ofthe specimen containing a fatigue crack. The higher this load theshorter the fatigue cra
23、_k need be for failure coaditions to be satisfied.Fracture surface.- In all tests the specimen fracture surface boretypical variable-amplitude fatigue markings_ thst is, dark, coarse bandsalternating with smooth shiny bands. The dark bands are thought to becaused by semibrittle _rack propagation dur
24、ing the application of oneor more high loads and th_ shiny, smooth bands by the application ofmany small loads.Method of comparison.- As has been previou_ ly stated_ linear cumula-tive damage has been asst_:ed in the analysis of these tests because ofits simplicity. However, airline operators are in
25、terested in flying time,not theoretical damage; therefore_ many fatigae tests are designed sothat a block represents ELgiven number of fligh!s or flight hours. Ifthe test results of this :-nvestigation had been compared by this methodit would be possible to o_tain different resulted. For example, lo
26、adwhile load schedule 4nschedule 3 produced the lowest value ofproduced the shortest lif, if number of blocks _,o failure was the crite-rion. (See table V. ) This latter comparison more forcefully demon-strates the deleterious effect of removing resilual stresses as soon asthey are formed and of th_
27、 tendency for high loads to cause completefailure when short creeks are present.CONCITJS IONSThe data obtaine_ from axial-load fatigue tests of 7075-T6 aluminum-alloy sheet specimens? programed to simulate mamuever-load experience_support the followkb: _onclusions:ni. All values of the summation of
28、cycle rstiosa_ N at failurewere greater than i whic_ is due to the formation of beneficial residualstresses during application of the highest loa_i. (In the ratio n/N,n is the number of cycles applied and N is the fatigue life at thesame stress level.)2. A standard load schedule, consisting o_ 8 loa
29、d levels andrepresenting positive load peaks from the i g level-flight stress toProvided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-16design limit load, was varied or modified in several ways. The followingeffects were present when compared with this sta
30、ndard load schedule.(a) Adding load levels to simulate negative load factors gaveI na significant reduction in _. Adding load levels to simulateI nloads above design limit did not give a significant change in H.(b) Truncating the load schedule at the high stress end_, ndecreased the value of _ signi
31、ficantly, whereas truncatingthe load schedule at the low stress end did not have a significanti neffect on _.size.Z n(C) increased and then decreased with increasing blockL1462I n(d) The value of _ was greater when using eight stepsto approximate the load spectrum than when using four steps tosimula
32、te the same load experience.Most of the variations in fatigue life observed in this investiga-tion may be explained qualitatively with the aid of residual stressesand residual static strength considerations.Langley Research Center,National Aeronautics and Space Administration,Langley Air Force Base,
33、 Va., February 13, 1962.Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-_L17REFERENCESi. Naumann, Eugene C., Hardrath, Herbert F., and Guthrie, David E.:Axial-Load Fatigue Tests of 2024-T3 and 7075-T6 Aluminum-AlloySheet Specimens Under Constant- and
34、 Variable-Amplitude Loads.NASA TN D-212, 1959.2. Neuber, Heinz: Theory of Notch Stresses: Principles for ExactStress Calculation. J. W. Edwards (Ann Arbor, Mich.), 1946.3. Spaulding, E. H.: Deslgn for Fatigue. SAE Trans., vol. 62, 1954,pp. 104-116.4. Grover, H. J., Bishop, S. M., and Jackson, L. R.:
35、 Fatigue Strengthsof Aircraft Materials. Axial-Load Fatigue Tests on UnnotchedSheet Specimens of 2hS-T3 and 75S-T6 Aluminum Alloys and of SAE4130 Steel. NACA TN 2324, 1951.5- Grover, H. J., Hyler, W. S., Kuhn, Paul, Landers, Charles B., andHowell, F. M.: Axial-Load Fatigue Properties of 24S-T and 75
36、S-TA_uminum Alloy as Determined in Several Laboratories. NACARep. 1190, 1954. (Supersedes NACA TN 2928.)6. Mayer, John P., Hamer, Harold A., and Huss, Carl R.: A Study ofthe Use of Controls and the Resulting Airplane Response DuringService Training Operations of Four Jet Fighter Airplanes. NACARM L5
37、3L28, 1954.7. Hamer, Harold A., Huss, Carl R., and Mayer, John P.: Comparison ofNormal Load Factors Experienced With Jet Flghter Airplanes DuringCombat Operations With Those of Flight Tests Conducted by theNACA During Operational Training. NACA RM L54EI8, 1954.8. Anon.: A Tentative Guide for Fatigue
38、 Testing and the StatisticalAnalysis of Fatigue Data. Special Tech Pub. No. 91-A, AS_M, 1958.9. Sigwart, H.: Influence of Residual Stresses on the Fatigue Limit.Proc. Int. Conf. on Fatigue of Metals (London and New York), Inst.Mech. Eng. and A.S.M.E., 1956, pp. 272-281.i0. Heywood, R. B.: The Effect
39、 of High Loads on Fatigue. Colloquiumon Fatigue. Waloddi Weibull and Folke K. G. Odquist, eds.,Springer-Verlag (Berlin), 1956, pp. 92-102.ii. SchiJve, J. : Fatigue Crack Propagation in idght Alloy Sheet Mate-rial and Structures. Rep. MP. 195, Nationaal Luchtvaartlaboratorium(Amsterdam), Aug. 19(10.P
40、rovided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-1812. Hudson, C. Michael, and Hardrath, Herbert F. : Effects of ChangingStress Amplitude on the Rate of Fatigue-Crack Propagation in TwoAluminum Alloys. NASA TN D-960, 1961.13. McEvily, Arthur J., Jr., l
41、llg, Walter, and Hardrath, Herbert F.:Static Strength of Alumlnum-Alloy Specimens Containing FatigueCracks. NACATN 3816, 1956.L1462Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-19TABLE ITENSILE FROPERTIES OF AI/;MIN94-ALLOY MATERIAL TESTED7075-T6 (
42、152 tests); data from ref.Ave rageYield stress (0.2-percent offset),ksi 75.50Ultimate tensile strength, ksi . 82.94Total elongation (2-inch gage length),percent 12.5Minimum71.5479.847.0Maximum79.7984.5415.0TABLE IIMANEUVER-LOAD STATISTICSEquivalent NumberAcceleration, stress,g ksi exceedingPositive
43、load distributioni2345677.3897.014.021.028.035.042.049.051. i56.063.0i0, 0005,6002,800i,22043011523133.70.53Negative load distribution0-i-20-7.0-14.0140121.3Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-2OI-.IHI,-,Ir.oiO1, 502.89C6 81, 0OO,OOC9 20,
44、 521,61Ceometric mean i,554Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-2304TABLE VRESULTS OF VARIABLE-AMFLITJDE AXIAL-LOAD FATIGUE q_STS OF7075-T6 ALUMINUM-ALLOY SPECIMENS USING A MANEUVER-LOAD SPECTRUMWIT_ Stain = 7 ksi and KT = 4.0, EDGE NOTCHS
45、pecimen MachineFailureBlock StepIAfe,cycles f nNLoad scaedule 1 ( 0.i); 295 cycles/blocKB94NI-6B96NI-7BSONi-4B128SI- i0BSON1-8B52Nl-717216715912h112io5_)o,301I_8,957_6,500%, 5_o52,67950,7571.721.661.591.25i.121.05Geometric mean 40,150 1.37Load schedule i (X 0.3); 879 cycles/blockB95_I-6B ONi- 5B eNi
46、-7B94NI-7B95NI-9B95NI-58O726968615oGeometric meanLoad schedule 1 (8 step); 2,954 cycles/blocki 69,5158 62,8088 59,9528 59,7788 55,1298 45, 11957,4202.572.162.052.031.811.471.96B52NI-4B95NI-2B5_i-2B50NI-9B56NI-1BSONi-5B52NI-2*2425212O19Z9Z369,91164,69499,81555,76654,08554,08235,25059,4_02.542.252.041
47、.911.85i.85i.22Geometric mean 2.02Load schedule i (X 5.0); 14,670 cycles fb:;ockB52NI-9B94NI- i0Boy2N1-4B92NI-5B94Nl-IB52NI-8Geometric meanLoad schedule i; load step 1 omitted70,56970,55056,52456,51356,25756,15360,6102.042.021.761.751.731.671.82B h_I-8B129SI- 1B IN1-7B128SI-2B52NI- i0B9_I-5BSINI-9*G
48、eometric mean242422222221ii_5,186L5,18250,03250,05150,05139,11o19,64941,5202.3_2.342.122.122.121.961.092.16*Not included in geometric mean.Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-24TABLE V - ConcludedRESULTS OF VARIAN_-_4PLITUDE AXIAL-I_AD FATIGUE TESTS OF7075-T6 AI/;MINUM-ALIEY SPECIMENS USING A MANEUVER-LOAD SPECTRUMWITH Stain = 7 ksi and KT = 4.0, EDGE NOTCHSpecimen MachineLoad scheduleFailureBlock StepLife_cycles f nNi_ load step 8 omittedBSINI-8B92N1-6B%NI-IOB28S1-9B96_1-2B92NI-3BglNI-5*Geometric mean22212121191815Load schedule i
