1、t a c a 2 NATIONAL ADVISORY COMMITTEE FOR AERONAUTICS / TECHNICAL NOTE 2324 FATIGUE STRENGTHS OF AIRCRAFT MATERIALS AXLAL-LOAD FATIGUE TESTS ON UNNOTCHED SHEET SPECIMENS OF 24S-T3 AND 75S-T6 ALUMINUM ALLOYS AND OF SAE 4130 STEEL By H. J. Grover, S. M. Bishop, and L. R. Jackson Battelle Memorial Inst
2、itute Washington March 1951 Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-NOTICE THIS DOCUM.ENT HAS BEEN REPRODUCED FROM THE BEST COPY FURNISHED US BY THE SPONSORING AGENCY. ALTHOUGH IT IS RECOGNIZED THAT CERTAIN PORTIONS ARE ILLEGIBLE, IT IS BEING
3、 RELEASED IN THE INTEREST OF MAKING AVAILABLE AS MUCH INFORMATION AS POSSIBLE. Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-1 . NATIONAL ADVISORY COMMITTEE FOR AERONAUTICS TECHNICAL NOTE 2324 FATIGUE STRENGTHS OF AIRCRAFT MATERIALS AXIAL-LOAD FATI
4、GUE TESTS ON UNNOTCEED SHEET SPECIMENS OF 24S-T3 AND 75S-T6 ALUMINUM ALLOYS AND OF SAE 4130 STEEL By H. J. Grover, S. M. Bishop, and L. R. Jackson SUMMARY This report presents information on the axial-load fatigue behavior of unnotched specimens of each of three sheet materials: 75S-T6 aluminum allo
5、ys and normalized SAE 4130 steel. 24S-T3 and The experimental investigation of these materials included the following items: 1. Determination of fatigue strengths, in tests at a speed of about 1100 cycles per minute, covering a range of mean loads from zero to a high tensile value and, for each load
6、ing condition, lifetimes from 10,000 to 10,000,000 cycles 2. Determination of fatigue strengths in tests at a slower speed of about 90 cycles per minute 3. Several measurements of damage or strengthening at one stress level due to previous loading at another stress level; these tests included interc
7、hanging the order of application of high stress level and low stress level In several respects, fatigue test data are extended beyond thos,e previously available. However, results are in general agreement with such previously reported data as are available for comparison, The main observation unpred
8、ictable from previous work is that fatigue strengths at 90 cycles per minute appear, in some ranges of loading, appreciably lower (up to 10 percent) than corresponding strengths at 1100 cycles per minute. Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,
9、-,-NACA TN 2324 INTRODUCTION A wartime survey (reference 1) showed a lack of complete information on the fatigue properties of sheet materials used in airframe construc- tion. Although a great deal of information was available, it appeared that no material had been investigated fully and that no str
10、ictly com- parative tests of large extent had been made on different materials under carefully controlled conditions. Therefore, it was planned to investi- gate rather fully the fatigue behavior of each of three metals commonly used in airframe construction: SAE 4130 steel. Each metal has been teste
11、d in one thickness (0.090 in. for the aluminum alloys and 0.075 in. for the steel), and all tests have been conducted under axial loading (of obvious importance in stressed- skin construction). 2bS-T3 and 75S-T6 aluminum alloys and The results, of interest in themselves, also furnish basic informa-
12、tion for further studies of the same materials. In view of this possi- bility, care has been taken to evaluate the experimental errors involved and to estimate, insofar as is possible, the residual “scatter“ of test points. This investigation was conducted at the Battelle Memorial Institute under th
13、e sponsorship and with the financial assistance of the National Advisory Committee for Aeronautics. Acknowledgment is due to Mr. David 0. Leeser, who, while on the staff of Battelle Memorial Institute, contributed a major part of the experimental work described in this report. The authors would also
14、 like to express appreciation to Mr. Paul Kuhn, of the Structures Research Division of the Langley Aeronautical Laboratory of the NACA at Langley Field, Virginia, for many helpful discussions during the course of the investigation. SHEET MATERIAL AND TEST SPECDENS Coupons cut from 135 sheets (0.090
15、in. thick) of each aluminum alloy and from 270 sheets (0.075 in. thick) of the steel were furnished by the NACA. Each sheet was laid out to provide four static tension blanks with the grain (rolling direction) and four static tension blanks across the grain, four static compression blanks with the g
16、rain and four static com- pression blanks across the grain, four fatigue test blanks with the grain, and a number of blanks for possible future use. As shown in figures 1 and 2, the layouts were such that a sample was taken from each section of each sheet for the various tests. Test pieces were mach
17、ined at Battelle from these coupons. . Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-NACA TN 2324 3 I- Static Test Specimens Static tension test coupons were machined to conform with the A.S.T.M. standard for sheet metals (reference 2). The static
18、compression test coupons were machined and ground 0.625 inch in width by 2.625 inches in length with ends parallel to each other and normal to the longitudinal or vertical axis of the specimen. Fatigue Test Specimens For fatigue test specimens, blanks of each material were cut approxi- mately 3 by 1
19、8 inches with the grain running the long dimension of the blank. These were protected on each face with a zinc chromate primer. With this coating still on, each blank was machined to the shape shown in figure 3. Previous experience had shown this to be a desirable speci- men for sheet fatigue tests
20、(reference 3). A reduction from a width of 1.000 to 0.800 inch in some of the steel specimens was necessitated by the load capacities of the available fatigue testing machines. Cross checks indicated that this decrease in width did not significantly affect test results. Specimens were polished elect
21、rolytically (after preliminary tests to justify this procedure for the materials concerned; see appendix A). TEST EQUIPMENT AND PROCEDURE Static Tests Tension tests were made in a Baldwin-Southwark universal testing machine with a Templin type recorder. Compression tests were made in the same machin
22、e with a Montgomery-Templin roller-type compression jig. Tests conducted at the Langley Aeronautical Laboratory showed that com- pression stress-strain curves obtained with the roller-type support were more accurate than curves obtained with other types of support available (reference 4). Loading ra
23、te for the static tests was 0.03 inch per minute. Fatigue Tests All fatigue tests were run on Krouse direct repeated-stress testing machines (reference 3) , one of which is shown in figure 4. have a nominal capacity of 10,000 pounds tension to 10,000 pounds com- pression. These machines When the mac
24、hines were operated at the normal speed of about I Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-4 NACA TN 2324 1100 cycles per minute, the determination of fatigue strengths covered a range of mean loads from zero to a high tensile value and, for
25、each loading condition, lifetimes from 10,000 to 10,000,000 cycles. A belt drive was arranged for the low-speed tests to give a speed of about 90 cycles per minute. The machines are of constant-deflection type; however, each is equipped with a sensitive means of detection of load decrease, so that t
26、ests can be generally considered as run at nearly constant load. Before this investigation was undertaken, the machines were recalibrated both statically and dynamically. The estimated pre- cision of setting and maintaining loads was about $3 percent for tension- tension tests and about k5 percent f
27、or tension-compression tests. Tension-tension tests were run with the self-alining type of grips used in previous investigations conducted at Battelle (reference 3). Measurements with bonded wire strain gages have shown that, with careful loading, the tension-tension grips have uniformity of stresse
28、s across the 1-inch gage length of a sheet specimen to about k500 psi. ment of the grips in the testing machine keeps bending stresses below about 500 psi. The aline- Tension-compression tests probably have somewhat less precision. Construction details of the tension-compression grips and guide plat
29、es are shown in figures 5 and 6. National Bureau of Standards (reference 5). There are two difficulties: This general method was developed at the (1) If the guide plates are too tight and specimens are not perfectly flat, an appreciable, unmeasured fraction of the applied load goes into friction (2)
30、 If the guide plates are too loose, the specimen buckles on the compression part of the cycle and bending stresses may become large Experiments with bonded wire strain gages were performed to determine optimum conditions for using the guide plates. These experiments are described in appendix B; it a
31、ppeared that errors did not exceed about 500 psi or 5 percent of maximum stress. Surface Finish Surface finish is known to be of major importance in determining fatigue strength. It appeared desirable to use a method of surface finishing which would: (1) Leave no transverse scratches (2) Slightly an
32、d reproducibly round edges to prevent development of “feather“ edges Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-NACA TN 2324 5 I- 1. - (3) Introduce negligible residual stresses (under 500 psi, if possible ) (4) Not cold-work the surface layers
33、(5) Be reasonably economical for use on the large number of speci- mens anticipated (6) Satisfactorily polish the roots of deep, narrow notches in anticipation of future tests to be run on notched specimens Several methods of mechanical polishj.ng were tried. was also investigated rather thoroughly
34、and was finally chosen as most nearly fulfilling the requirements listed above. Electropolishing While a considerable amount of work was done in selecting the sur- face finish, the results may be summarized briefly. Electropolishing gave as high (or higher) fatigue strengths on aluminum-alloy specim
35、ens as mechanical polishing, gave as little scatter in fatigue tests, presumably introduced negligible residual stresses, did not cold-work the surface, and was relatively simple and economical. Appendix A gives details of tests which were made to justify these conclusions. After polishing, fatigue
36、test specimens were coated with Vinylseal fcr protection against corrosion and against surface damage due to handling. This coating was removed, with acetone, only immediately before testing a given specimen. Testing Procedure Basic fatigue-strength values were obtained by testing specimens of each
37、type of sheet at constant-load ratios varying from R = 0.70 to R = -1.00 (R Min. stress/Max. stress). The range covered, as far as feasible, the values of fatigue strength for each material. Fatigue Damage Tests Fatigue damage tests were made for each material at a constant mean stress of one-fourth
38、 the ultimate tensile strength of that material. While this procedure has not been generally followed previously (refer- ences 6 to g), it seems useful for calculations with respect to gust loading (references 10 and 11). (one-fourth the ultimate tensile stress), while chosen arbitrarily, is one tha
39、t might be used in airframe design. The particular value of the mean stress Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-6 NACA TN 2324 s Tests were taken for each material at two levels of maximum stress. These levels were chosen with the followi
40、ng considerations: (1) The low level was above the relatively flat part of the S-N curve so that scatter in lifetime was not too large (2) The high level was below the yield stress (with some question in the case of 24S-T3) (3) The difference between stress levels was as great as possible in view of
41、 the above considerations A test was made in the following manner: One specimen was run at the higher stress for a predetermined fraction (say, one-half) of its average expected lifetime; it was then run to failure at the lower stress. A second specimen was run in the reverse order (lower stress for
42、 one-half of its expected life, then higher stress to failure). Each test was repeated on other specimens so as to obtain average results. The tests were then repeated with several different fractional lifetimes at the first stress level. A major purpose of these tests was to find out the effect of
43、order of occurrence of high and low stresses. EXPERIMENTAL FC3SULTS Static Strength Tests Table 1 gives the results of the static tension tests and the static compression tests on the three materials. The results of these static tests indicated that the sheet materials were up to standard in mechani
44、cal properties. The variations observed, from one sheet to another or from one specimen in a given sheet to another in the same sheet, were small in view of the precision possible in the fatigue tests. Fatigue Strength Tests Results of the fatigue tests for the 2kT3, 75S-T6, and SAE 4130 sheet speci
45、mens are given in tables 2, 3, and 4, respectively. (Some typical specimen failures are shown in fig. 7.) These results are shown plotted in the form of S-N curves in figures 8, 9, and 10. of scatter for the test data is illustrated in figure 11, which indi- cates that the scatter for the steel was
46、relatively slight. the S-N curves were extrapolated conservatively into the 1000- to 10,000-cycle range. The degree Some of Part of the difficulty in obtaining accurate values c Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-NACA TI1 2324 7 . in thi
47、s range, particularly at high load ratios, was attributed to the difficulty in maintaining loads well above the yield point and to the increase in strength due to the work-hardening effect. Some of the S-N curves represent intermediate test-ratio plots outlined with a few critical points and fitted
48、into the general pattern of the more com- pletely determined curves. Calculations indicated that, for a region f1/2 inch from the line of minimum cross section in each specimen, any variation in stress due to specimen shape was well within the loading precision (reference 3). The few specimens in wh
49、ich failures occurred beyond this region were not used in plotting the S-N curves. When the fatigue behavior of each material had been established for stresses up to the tensile yield point, some explorations at still higher stresses were pursued. It was anticipated that a specimen so loaded in the Krouse testing machines would elongate sufficiently to cause dif
