NASA-TN-D-3149-1966 Influence of programming techniques and of varying limit load factor on maneuver load fatigue test results《编程技术和不同限制荷载系数对机动荷载疲劳试验结果的影响》.pdf

《NASA-TN-D-3149-1966 Influence of programming techniques and of varying limit load factor on maneuver load fatigue test results《编程技术和不同限制荷载系数对机动荷载疲劳试验结果的影响》.pdf》由会员分享，可在线阅读，更多相关《NASA-TN-D-3149-1966 Influence of programming techniques and of varying limit load factor on maneuver load fatigue test results《编程技术和不同限制荷载系数对机动荷载疲劳试验结果的影响》.pdf（17页珍藏版）》请在麦多课文档分享上搜索。

1、GPO PRICE CFSTI PRICE(S) s s NASA TN D-3149 / fl Hard copy (HC) Microfiche (M F) ff 653 Juk 65 a 8 z P * - -% AulMTNuM ALLOY F52 tests; data from reference 5 1 Yield strength Total (offset 0.2 percent) elongation ksi I wa percent . in 2 in. (5 cm), Average . . . 82.94 572 75 * 50 521 32.3 79.84 71.5

2、4 7.0 Maximum. * . . 1 84.54 1 52,” 1 79-79 1 1 15.0 RESULTS Test Data The results of the variable-amplitude fatigue tests are presented in table N and in figures 4 and 5. establish whether the variations investigated have an effect on fatigue life. For completeness, table IV contains the load step

3、at failure and the specimen life (total cycles) in addition to life indices computed by Miners linear cumulative damage theory. excessive and is indicated by the ticks on the symbols in figures 4 and 5. Data taken from reference 4 have been used to The scatter in the test results is not considered A

4、utmatic and Seloiautomatic Tests A comparison of results from program l(a) , semiautomatic block and auto- matic block, showed no significant difference (table IV); therefore, it was concluded that any effects due to machine differences, load accuracy, speed differences, and so forth, were negligibl

5、e. Block and Random Tests The results of the three sets of tests in the block and random series are shown in figure 4. The random test lives were invariably shorter than the block test lives but this effect was most pronounced for the program which contained negative loads. were about 40 percent sho

6、rter than the block test lives. effect of negative loads was also noted for gust load tests in reference 1. Figure 4 also indicates that including negative loads in the test program has reduced specimen life by a factor of approximately 2 as compared with the same program without negative loads. Thi

7、s substantiates the findings of several investigations of this particular effect. The random test lives for this particular program This perturbing (See, for example, ref. 1.) 7 Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-TABLE IV VARIABLE-AMPLIT

8、LIDE TEST RESULTS MANEWER LOAD SPECTRUM 2.43 2.35 2.08 2.06 1.98 1.88 2.17 2.04 2.08 191 054 184 430 162 804 161904 156 295 147 705 171 1-17 160 191 - 163 200 Geometric mean. 1.95 54 260 B84N2-3 B85N2-7 B85N2-9 B85N2-5 B84N2-3 1051-2 B84N2 -6 2.28 1.74 1.62 1.44 1.44 1.44 1.21 1.47 60 666 46 300 42

9、997 38 221 38 221 38 221 32 174 42 200 - - , 1.88 Bl9N2-2 B19N2-3 1.79 1.70 2.03 102 741 100 347 - 109 500 Specimen I Load step at failure I n/N I Cycles Specimen I Load step at failure1 n/N I Cycles Program l(c); block 8 B104N1-2 8 B10 4N1- 10 8 B96N1-3 8 B104N1-6 8 B97N1-7 7 Geometric mean . Progr

10、am l(a); block; semiautomatis B52N1-4 B95N1-2 B50N1-9 51-2 561-1 Geometric mean . rl 2.34 2.04 1.85 - 2.23 1.91 1.85 2.02 - .3 69 911 64 694 59 815 55 766 54 083 54 083 59 440 Program 2; block; n = 6 Program l(a); random; automatic 8 8 8 8 2.32 2.17 1.91 1.89 1.78 1.69 - B2ON2-10 B2N2-2 m2-1 6N2-10

11、22-9 BbN2-2 B4N2-5 Bl9N2-9 Geometric mean 64 653 64 413 53 228 52 672 49 578 47 065 Bll2N2-1 B84N2-1 1122-3 B105N1-7 B84N2 -7 B84N2-4 8 8 . I I Program 2(a) (25 percent program l(a) plus program 2 W2-3 22-7 72-1 22-8 B3N2-2 6N2-4 Geometric mean 7 8 8 8 8 8 . 3.61 3.52 3.22 3.06 3.22 3.23 2.77 - 285

12、454 251 ooo 228 760 227 606 215 564 192 131 227 599 ?rogram l(b) (program l(a) + 2 levels NR); random 10 10 10 10 10 10 10 rogram 2(b) (50 percent program l(a) plus program 2) 102-7 32-7 B7N2-5 6N2-5 B6N2-9 BjN2-9 Geometric mean 8 8 5 8 7 6 . 3.54 3.36 3.34 2.74 2.66 2.57 3.01 - 228 124 214 201 211

13、978 164 878 158 152 592 298 185 900 . I I Program l(b); block* ?rogram 2(c) (75 percent program l(a) plus program 2) 2.80 2.19 2.19 2.00 1.67 1.67 1.67 B49N1-5 BgON1-2 961-1 901-1 901-5 91-6 B94N1-2 Geometric mear 10 10 10 10 10 10 10 . 79 069 60 5% 60,586 54 797 46 978 46 978 46 978 55 800 3.38 2.3

14、1 2.17 2.16 2.61 5-13 2.74 - - 191 360 168 979 140 207 105 557 95 579 94,148 127 600 Geometric mean I. Program (c) (program 1(a) + 2 levels Se Positive+2 Sl S8) Program l(b), block Program 1( c) , random (prog. l(a) + 2Si o Program l(c), block % -4 -P -0 P-+ Ld .A kf Mal or0 e2 % 0.92 NO 1.32 1.02 2

15、.60 Yes 1.28 - X 0 P- P - P P- v ! M Ll a - No Yes 2.52 - Yes Yes 1 1.57 5 P- P - U P- v 9 M h a - Yes - -Sample b/N geometric means are significantly different. I i.281 Top group Side group -atio Xn/N geometric means, 10 Provided by IHSNot for ResaleNo reproduction or networking permitted without l

16、icense from IHS-,-,- - -c om -0 Nh fa Mr- 0- kk aa Yes Yes Yes No Ah N- +- 74 d+ ?IN MM 00 k$ c,+ -c .da, -0 CUh ma k0 2a Mln 0- Yes Yes Yes Yes , # TABLE V.- Concluded RESULTS OF STATISTICAL ANALYSIS OF VARIA2JE-AMPIlTLmE FATIGUE TESTS hneuver load spectrum; 7075-T6 aluminum-alloy specimens; 1 g st

17、ress = 7 hi (48.3 /m2i .-. N + a rl - v ti k P4 t, -G dm -0 NFI Eg kln MN 0.- a - N + d d M k a t, -G Po) -0 Nh Ea MA 0- a v W /I F N 2 M h a - 1.75 Yes Yes Yes rogram l(a), block q = 7.3 rogram 2 7 = 6 Yes Yes rogram 2(a) (25 percent prog. l(a) + 2, 5-82 L.39 NO 5 0.69 1.5L - rogram 2(b) (50 percen

18、t prog. l(a) + 2 L.14 - 3.78 i.36 - 0.82 - 0.56 - 1.89 rogram 2( c) (75 percent prog. l(a) + 2 rogram 2(d) (50 percent prog. 2 + l(a) Yes -Sample simulated flights geometric means are significantly different. Itio si-ted flights geometric means, Top group Side group DISCUSSION OF ESULTS Damage and F

19、ailure Considerations Trends in fatigue life observed in the present tests are explained quali- tatively on the basis of residual stress and residual static strength considerations. 11 Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-Residual stresses

20、.- Residual stresses are obtained whenever a local stress, such as at the root of a notch, has exceeded the elastic limit of the materidq-. The plastically deformed material must be stressed to return to its original shape, and the necessary force is provided by the adjacent elastically strained mat

21、erial. Residual stresses cannot be computed accurately or determined by non- destructive testing; however, their effects can be determined through experi- mental methods and used to advantage. Compressive residual stresses delay fatigue crack initiation and propaga- tion, whereas tensile residual st

22、resses have an adverse effect. The beneficial effects of compressive residual stresses will decay under repeated cycling, the rate of decay being determined by the relative magnitude of the highest load level and successive load levels. Residual static strength.- Failure of the specimen occurs when

23、the applied load equals the residual static strength of the specimen. The residual static strength of a specimen first decreases sometimes precipitously as a crack is initiated and then deteriorates further with increasing crack length. ref. 7.) if any, effect on the residual static strength. High l

24、oads which may produce residual stresses that increase fatigue life by retarding crack initiation and propagation may also cause early failure of a specimen containing a short fatigue crack if the load exceeds the residual static strength of the specimen. Table IV indicates that almost every specime

25、n failed on the highest load in the program, which substantiates the above argument. (See In engineering materials, residual stresses probably have very little, Block and Random Tests In the block and random test series, program l(c) showed the largest vari- ation in life; this indicates that the pr

26、esence of negative load cycles is one of the most disruptive factors in comparisons of block and random tests. This variation was probably due to the fact that in the block form test, the negative loads,which reduce beneficial residual stresses,occurred in groups at widely spaced intervals and in th

27、is form had little more effect than would single neg- ative loads at like intervals. The same number of negative loads occurred in the random test, but in this case they were distributed throughout the test program and therefore, in effect, occurred at a much higher frequency. This multiplied their

28、residual stress destroying capability and a correspondingly shorter life was obtained for the random test. For test programs l(a) and l(b) the differences between lives of random and block tests were small. These differences were probably due to the fact that the random programs introduced more high

29、 load cycles in the interval of program used than was the case for the block tests. The random test schedules were programed to have the same statistics as the block tests for the total load history; however, the test life actually involved only a small interval of the complete history and the above

30、 situation was found to be true in the inter- val used. It was noted that summation of cycle ratios were approximately 2 for the tests with all positive load factors, but were close to 1 for the tests 12 Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-

31、,-containing negative loads. pumished in reference 4. These results are consistent Kith the results Varying Service Limit Load Factor Tests In test programs 2(a), 2(b), and 2(c) the lives were considerably longer than would be expected from linear damage accumulation theories. in life may be explain

32、ed on the basis of residual stresses; that is, the high residual stresses introduced by the large amplitude loads of the delayed crack initiation and/or growth at the subsequent lower stresses of the q = 6 level. This increase q = 7.3 level For program 2(d), in which the low stress levels preceded t

33、he high stress levels, the total life was approximately the sum of one-half the life at q = 6 and one-half the life at q = 7.3 of linear damage theories. the other tests in which the high stresses preceded the low stress levels. which would be expected on the basis As noted, however, this concept do

34、es not hold for CONCLUDING REMARKS Variable-amplitude axial-load fatigue tests of 7075-T6 aluminum-alloy sheet specimens were conducted according to loading schedules designed to approximate maneuver load histories. following observations: The results of these tests support the Maneuver load fatigue

35、 lives were shorter for random form tests than for block-form tests having the same load spectrum. when the loads were applied in random sequence and negative loads were included. The shortest life occurred Negative loads in a test program reduced fatigue lives by a factor of 2 as compared with the

36、same test without negative loads. tion of cycle ratios was found to be approxixrately 1 and 2, respectively. The corresponding summa- Fatigue lives up to 60 percent above the original test life were obtained by preloading with a portion of a test program having a higher limit load factor. . All of t

37、he trends noted herein may be explained qualitatively with the aid of residual stress and residual static strength considerations. Langley Research Center, National Aeronautics and Space Administration, Langley Station, Hampton, Va., August 5, 1965. Provided by IHSNot for ResaleNo reproduction or ne

38、tworking permitted without license from IHS-,-,-APPENDIX Specimens The material for specimens used in this investigation was taken from part of a stock of commercial grade 0.090-inch-thick (2.3 mm) sheets of 7075-T6 alu- minum alloy retained at the Langley Research Center for fatigue test purposes.

39、The material properties are given in table 111. given in figure 2 of reference 8. The material blank layout is Each specimen was stamped with a number identifying the specimen as to material, sheet number, and location within the sheet. For example, specimen 1151-7 is 7075-T6 (B), taken from sheet 1

40、15, blank N1, seventh position. The specimen dimensions are shown in figure 3. The specimen surface was left as received, and the longitudinal edges were machined and notched to give a theoretical elastic concentration factor of 4.0. chosen because it has been found to have fatigue characteristics r

41、epresentative of aircraft components (ref. 9). form the notch root and then slotting to the specimen edge with a 3/32 inch (2.4 mm) milling tool. an undersize hole was drilled first and enlarged to the proper radius by using progressively larger drills. nesses and then replaced. The last three drill

42、 increments were 0.003 inch (0.076 m) and a drill press with constant automatic feed was used. This configuration was The notch was formed by drilling a hole to In order to minimize residual stresses due to machining, Drills were used to drill four specimen thick- Burrs left on the specimens by the

43、machining process were removed by holding the specimen lightly against a rotating composition dowel impregnated with a fine grinding compound. This procedure was used to keep the present tests consistent with past tests conducted at the Langley Research Center. All specimens were inspected with a fi

44、ve power magnifying glass, and only those free of defects in and near the notches were used. 14 Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-REFERENCES I 1. Naumann, Eugene C.: Evaluation of the Influence of Load Randomization and of Ground-Air-Gr

45、ound Cycles on Fatigue Life. NASA TN D-1584, 1964. 2. Mechtly, E. A.: The International System of Units - Physical Constants and Conversion Factors. NASA SP-7012, 1964. 3. Mayer, John P.; Hamer, Harold A.; and Huss, Carl R.: Controls and the Resulting Airplane Response During Service Training Oper-

46、ations of Four Jet Fighter Airplanes. A Study of the Use of NACA RM 5328, 1954. 4. Naumann, Eugene C.; and Schott, Russell L.: Axial-Load Fatigue Tests Using Loading Schedules Based on Maneuver-Load Statistics. NASA TN D-1253, 1962. 5. Anon. : A Tentative Guide for Fatigue Testing and the Statistica

47、l Analysis of Fatigue Data. Spec. Tech. Fubl. No. 9l-A, Am. SOC. Testing Mater., 19%. 6. McEvily, Arthur, J., Jr.; Illg, Walter; and Hardrath, Herbert F.: Strength of Aluminum-Alloy Specimens Containing Fatigue Cracks. Static NACA TN 3816, 1956. 7. Grover, H. J.; Bishop, S. M.; and Jackson, L. R.: F

48、atigue Strengths of Aircraft Materials. of 24S-T3 and 75S-T6 Aluminum Alloys and of SAE 4130 Steel. Axial-Load Fatigue Tests on Unnotched Sheet Specimens NACA TN 2324, 1951 8. Spaulding, E. H.: Design for Fatigue. SA?3 Trans., vol. 62, 1954, pp. 104-116. 9. Grover, H. J.; Hyler, W. S.; Kuhn, Paul; Landers, Charles B.; and Howell, F. M.: Axial-Load Fatigue Properties of 24s-T and 75s-T Aluminum Alloy as Determined in Several Laboratories. NACA Rept. 1190, 1954. (Super- sedes NACA TN 2928. ) Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-