1、* 2-y Fe f NATIONAL ADVISORY COMMITTEE FOR AERONAiJTICS _ -. : -I-L.,. _ - . _ -_ Efa, 951 37 3lmer ?. 3ruhn Purdue Vniversit7 -.w - ._-, -i-m, RESTRICTED Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-F RESTRICTED -1 . l d NATIONAL ADVISORY COMMITT
2、EE FOR AERONAUTICS TECHNICAL NOTE NO; 951 T!ISTS ON THIN-WALLID CELLULOID CYLINDERS TO DETERMINE THE INTERACTION CURV3S UNDER GOMBINZD BElTDI?TG, TORSIOX, AND COMPRESSION OR TZNSION LOADS By Elmer F. Bruhn SUMMARY The report on this research project is divided into two parts. Part I presents the res
3、ults of preliminary tests to datermine the modulus of elasticity of celluloid sheet; the effect of tcmporaturo chanqo on the value of tho modulus of alasticity; the creep of celluloid sheot under stress as a function of time; and finally tho effect of ropeatcd buckling failures on the original buckl
4、ing strength of celluloid cyl- inders. Part II of this report givce the results of tests on a considcrablc numbor of thin-walled, circular celluloid cylin- ders of several lengths, diameters, and wall thicknesses when subjected to loads producing pure compression, pure bending, pure torsion acting s
5、eparately and in combination and of such magnitude as to cause failure of the cylinders. Ultimate strength interaction equations based on the test results are given for cfrcular cylinders subjected to combined compression and pure bending, combined compression .- and pur3 torsion, combfned puro Send
6、ing and puro torsion, and finally to . combined compression, bending, and torsion. I Limited results are given for the celluloid cylinders subjected to combined tension and pure bending, combined ten- sion and pure torsion, and finally to combined tension, bend- ing,and torsion. SlSYRICT3I.3 Provide
7、d by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-4 h d XACA PX Bo. 551 2 . - / The modern airDlane body is essentially a thin-walled shell. In some bod$ designs, all longitudinal skin stiffen- - ., ers are removed or only three longibUdj.nal Skin Stiffener8
8、- - are used, thus the thin Shell becomes the major structural unit in resisting the applied external loads. In general, the airplane body is subjected to external forces which will- produce c2omprossive, bending, and twisting stresses in the - body structure. very little test information 18 availab
9、le at prasont on. _. - the ulti%afe strength of thin-rralled metal cylinder8 under combined ICads, Tarticularlg so if compression, bending, and tW-lsti:lg loads are acti.ng simultaneously. on0 reason for -m-m. the lack of this informat3.n j.8 due, no doubt, to the greaf -_. A cost and amount of time
10、 to carry out a complete test program on full-size metal cylinders. The purpose. of this research. Project was to determine whether the tasting of inexpensive -: small cylinders fabrficntod from thin cellulcrid shoet would -. -. give valuable reliable data for determining the ultimate strength inter
11、action relationship8 for circular thiG-waliSt .-. cylinders when subjected to combination8 of compressive, - ; bending;, clnd twistinq LOadS. 4 It is thought. that the results obtained in this project . have shovn dofinftsly that valu:gle desiza information can . -c be obtained from tests of cellulo
12、id cylfndors with coapara- tivelg small expenditure of time and money. . - s =- . . . The funds for this research :Jzoj. A .ir actual test work was carried on in the departm.ent of soronantics of the Sehaol of Ma-hanical and Aeronautfoal - Zingink+aring under the direct suporvlsion of Professor E.-y
13、; Bruhn. The test apparatus was collstructedand the majorjty of the cylinder tests yere made by Kr. B. 1;. Dickinson, former instructor of Aeronautical Za-:- -; 1. Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-6 BACA able L D r t B, ab Bs TN No. 95
14、1 3. The writer is indebted to Professor K. D. Wood for valu- suggestions and for readfng the final repor-t-.- c SYMBOLS cylinder length for minimum internal frame spacing, inches cylinder diameter, inches cylinder radius, inches cylinder wall thickness- compression load ratio (non-dimensional) ( Rc
15、 = applied compressive load on cylinder ultimate compressive load for ) cylinder - .- -z.- pure bending ratio (non-dimensional) ( Rb = app lied p ure bending load on cylinder ultimate pure bending load for ) cylinder -.- pure torsion load ratio (non-dimensional) i . .: - -2 , Rs = applied pure torsi
16、on load on cylinder ultimate pure torsion load for 1 cylinder R,“ + RbY + 11,“ = 1.0 (an equation referred to as the ultimate strength interaction curve where the exponents x, y, and z define th8 general relationships of the load ratios when a cylinder is subjected to a combined loading which causes
17、 failure of the cylppdor) - I- PRELIEtINARY TESTS ATD STUDIBS Summary The primary object of the test project as a whole was - -./-_. to determine the ultimate allowable load interaction curvss- - - ; Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-MA
18、CA TB Fo. 951 for thin-walled cylinders when subjected to combined compres- sive, bending, and torsional loads. The material used for the test cylinders was the nitrocellulose compound commonly . referrod to as celluloid. It is common knowlodgo that for such plastic materials the stiffness of the na
19、terial.is in- fluenced by temperature and also that the material suffers creep under stress. Preliminary tests were run to detornine the extent of these factors. The Suckling of thin-walled cylinders falls in the gen- eral category of elastic buckling since tho stresses that produce buckling are rol
20、atively low. Tests were run on cyl- inders to determIne whether the ultimate buckling strength of cylinders was affected by repeated loading to buckling failure and also to determine the influance of time of load- ing upon the ultimate buckling strength. Mat er i al ,The material for all the test un
21、its in the test project was cut from standard 20- by 50-inch sheets of transparent celluloid, which were purchased from the Celluloid Corporation. Three nominal sheet thicknesses were used: namely, 0.0075 inch, 0.010 inch, and 0.015 inch. Stress-Strain Diagrams of Celluloid Shaat Figures 1 to 4 show
22、 the results of tests to determine a portion of the tension stress-strain curve for celluloid sheet material. The tast specimens consisted of 12inch. strips varying from 0.125 to 0.25 inch in width, which were cut from the 20- by 50-inch celluloid sheets in directions parallel to the length of the s
23、heet and also parallel to the width direction. A lo-inch gage length was carefully marked on the test strips. The test strips were hold in position at one end by a rigid clamp placed to coincide with one of the gage marks and which was fastened to the table top. Tension loads were applied by means o
24、f a small flexible wire thread attached,to the other end of the test strip, which in turn passed over a nearly frictionless pulley, with small weights being suspended on the end of the wire to load the test strip in tension. The elongation of the lo-inch gago length was obtainod by a “Carl Zeissi me
25、asuring microscope set up over the gage mark near the free end of the test strip. Photograph No. 1 Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-5 shows test setup for measuring the strain for the various tension loads. Strain readings were taken f
26、or two series of loadings. In the first series the strip load was added, the strain was read within 6 seconds and then the load was re- moved before adding the next larger load. Therefore, fn this series the stress on the specimen existed for only a short interval of time. Curve A of figures 1 to 4
27、shows the results of this series of test.s. In the second series of tests the loads on the test strips were applied continuously and not .re- moved, and the loading was carried both up and down. To run a complete test on a strip requfred several minutes which meant that the test strip carried tensil
28、e stress for a consid- erable length of time. CurvesfB of figures 1 to 4 show the results of these tests. Creep of Celluloid Sheet Table 1 shows the results of tests to obtain information on the effect of stress intensity and time of stress duration upon the creep action of celiuloid sheets. A test
29、strip sim- ilar to that used for obtaining stress-strain curves was used in this test. The test strip was loaded with increasing ten- sion loads and the strain in 10 inches was measured at vari- ous time intervals up to a maximum of 2 minutes. The results in table 1 are presented in graphical form i
30、n the curves be- low the table. Effect of Temperature on Modulus of 3lasticity Zigure 5 shows the affect of temperature change upon the stiffness in tension for celluloid sheet material. Thin strip specimens similar to those used for obtaining the test data for the stress-strain curves as previously
31、 discussed were cut from the 20- by 50-inch celluloid sheets. These strips were loaded in tension by a single load and the elongation in a lo- inch gage length was obtained by a measuring microscope. The modulus of elasticity was computed on the basis of this unit strain and the accompanying stress,
32、 Tests were run ai room temperature varying between 65 and 95 F. Tht curves in figure 5 should not be considered as giving the cor- rect value of 3, since only one point on the stress-strain curve was obtained. The purpose of the tests was only to ob- tain a measure of the effect of temperature chan
33、ge upon the stiffness of the celluloid sheet In tension. Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-UACA TN No. 961 k - II General Oolnclusions from Stress-Strain, Creep, and Temperature Tests 6 l 1 The following genaral conclusfons can be drawn
34、 from a. - study of the results (figs. 1 to 5 and table 1): 1. The modulus of elasticity in tension for the 20- by 50-inch celluloid. sheets is not the same for stresses paral- lel to the length and wiidth of the sheets. The difference, however, is usually less than 10 percent. (See fig. 5.1 3 There
35、 is considerable variation in the value of the modulci of elasticity with sheet thickness. The thinner the sheet the higher the stiffness. The variation for sheets of 0.0075-inch and 0,015-inch thickness is around 20 percent. 3. Over the range of stress used in the stress-strafn tests, the resultfng
36、 curve is practically strafght. Refer- ence to the buckling stress for test cylinders in the latter portion of this report will show that the maximum stress at buckling was around 300 to 400 psi, which stress range falls on the lower portion of the stress-strain data of figures 1 to 4, I.uhere the r
37、elationship between stress and strain is definitely a straight line. It is assumed that the comprss- sive stress-strain relation at low compressive stresses ia the same as for tension stresses. The buckling of the cyl- inders should therefore fall in ths elastic category if the influence of creep is
38、 eliminated. 4. An increase in temperature decreases the stiffness of the celluloid sheet materfal. For example, for a sheet G.010 inch in thickness, the stiffness changed from approx- imately 450,000 psi to 400,000 psi when the temperature was changed from 65 to 95 F. (See fig. 5.) 5. celluloid she
39、et under stress suffers the character- istic generally referred to as creep. Table 1 shows, however, if the stresses are kept under 500 psi and the time interval T of stress applfcation withiln 0.25 of a minute, the unit strain due to creep is negligible. These preliminary tests on stress-strain pro
40、perties l therefore definitely indicated that if reliable and consist- ent comparatZve results were to be obtained in testing cellu- loid cylinders under combined stresses, the tests on any cyl- inder should be run at the same temperature and that the time Interval used in applying the loads should
41、be short and that it should be kept nearly the same in order to eliminate the Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-: t ;-; NACA TX No. 951 7 4 . influence of temperature and creep on the buckling strength of cylinders fabricated from thin
42、celluloid sheet. . Preliminary Tests to Determine Effect of Repeated Failure and Effect of Time of Loadir;g on Cylinder Strength Two preliminary test cylinders were fabricated. They will be referred to as cylinders P-l and P-2. Cylinder P-l was 6.87 inches in diameter, 7 Inches long, and had a wall
43、thickness of 0.0075 inch. Cylinder P-2 was the same as cyl- inder P-l except the length was 28 inches. The method of fabrication of these two cylinders was similar to that of the cylinders in part II of this test program, which is de- scribed in detail in part II. These two test cylinders were loade
44、d in pure bending _. and pure torsion by means of a system of levers, wire thread, and nearly frictionless pulleys. (See fig. 6 for schematic diagram of the loading system.) (See photos. Nos. 2 to 9 . : for pictures of these-cylinders.) Table 2 gfves the results of the pure bending and tor- sion tes
45、ts on test cylinder P-l, Bine tests each were made in bending and torsion and the load which caused buckling failure each time was read on the platform scale. As shown in table 2, the time interval required to produce buckling faflure was varied between 0.10 to 0.93 minute for the pure bending tests
46、. The scale loads to cause failure for the first and ninth tests were 3 pounds 8.3 ounces and 3 pounds 8.4 ounces, reepectively, with time of loading 0.10 and 0.13 minute, respectively. Thus, after the cylinder had failed by buckling eight times, the ninth time gave a failing strength practically th
47、e same as the first loading when the time of loading was kept practically the same. When the time of loading was increased to 0.93 minute (see table 2), the scale load which caused failure dropped to 3 pounds 6.5 ounces, or a decrease of 3.3 percent from the strength in test No. 1 where time of load
48、ing was 0.13 minute. Table 2 also gives the results of nine tests in pure 7 torsion. The percent difference between the strength of the cylinder in the first and ninth tests was 0.66 and the max- imum varilatfon from the average strength of nine tests was 2.2 percent. Provided by IHSNot for ResaleNo
49、 reproduction or networking permitted without license from IHS-,-,-BACA TN Ro. 951 . 8 4 c Tables 3, 4, 5, and 6 give the results of repeated fail- ure tests on test cylinder P-2 in pure torsion and pure bend- ing. The cylinder in table 3 was 28 inches long. For the tests in tables 4, 5, and 6, intermediate bulkheads were a
