1、NATIONAL ADVISORY COMMITTEE A FOR AERONAUTICS - .- - TECHNICAL NOTE - - . NO. 1051 PRELIMINARY INVESTIGATION OF THE LOADS CARRIEDS .- - -1 -. BY INDIVIDUAL BOLTS IN BOLTED JOINTS; By Manford B. Tate and $amuel J. Rosenfel-d Uhgley Memorial Aeronautical. Laboratory Langley Field, Va. . -_ - *-$ - _ -
2、 - _ - . _i_- . .- - i. . n - a- Washington May 1946 ,- NACA LIBRARY 0NAmCAL LABORATORY a - - - .- kley field, Vr . -. - . . . _ Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-4 . L D - c 8 .a - NATIOKAL ADVISORY COlsLne . - - line in line wltk t+e
3、qolied load. Test rF-sults are - Ten in the form of curves s5ovilng bolt-load histo:=.ies- through the elastic Ed yt6ld ranges to joint f.ailxTCT- Lractical upper limit of elastic acti-on of a joint sub-. Seck to static loading and to obkab c*vos-.reresal;r%-lc of early investigations is given in re
4、ierence 2, ?Po 47b484. The uresent paper deals with tile problem of load distribution a-3ong the bolts of sgmetrical butt joints. Tests were conducted to deternine ex:)orLnentally, both within and above the elastic range, the manner in which load was distributed =.ow the verious bolts. The test sl?e
5、cirdens were doubly symmetrical two- and three-bolt - joint8 made of 3or! tbe followi-q3 con- dFtiona; (1) The ratio of stress to strain is constant, . - . . . Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-5 -(2) The stress is unifcrxly distributsd
6、 over the cross eectlons of main plates and buttrstraps (3) The effect of frtztion is negligible e slastic rmae On t;bB basfs of these ass-tiiipiiorfi, it 2s faufid far SF.- metrical hxct joints that the rslhtionshlp between the loads on any tim SUCCQSSVF bclts in a single line cf bolts is (5 ) The
7、ralationvhip bet*een ?tilt deflection end zi . . _- I . _ - .- SquatLon (1) is used in the computation of bolt; loads In appen6ix B. i3a3ed on assumptions 1 and 2, the plate ccnstant rL 2- may bcj stated as -. . . - .- ._ The llnear relation between bolt load and deflection (assunintion 5) may be ex
8、pressed in terms of the bolt cGnstslnt (3 as . . -. In the determinstion of C it is assumed that-the bolt acts as a fixed-end beam with tha holt.1oad H dls- tribute3 uniformly along a length equal ta the main-plate loa3 is tL?l.formly distrihuto? aloq two lengtta, each equsl to the butt=stran thickn
9、ess. Bearing stress is computed in the convent-ianal manner as bolt load divided c by an .area that is deterxinod bf projacting the bolt diameter on the plate thickness; arid bearing dsformations thickness, Acting in the opposite direction, the bolt * are ex7riasaod in terms of the compressive modul
10、i of the - - - - - P - Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-. . . . NACA TN NO. 1051 I 7 materials and dimensions of the bolt and plates. From these consideretions, C may be stated asro-ilow: Tor joints made of 24s-T plates with a butt-str
11、ap thickness of one-half the main-plate thickness ( = 2) and f a8 tened with elloy- steel bolts : For other symmetrical butt- joint arrangements, expressions for C are given in appendix A. (See equatiors 16 to A22. ) The third assumptbn (thbt the effect of friction is negligible) gives rise to a hig
12、hly cantroversi to act elastically and the other inelastically, but the yielding of any component constitutes the beginning of inelcastic action of the joint as a whole. Interpretation of this dewpoint, however, should be practical and should not include yielding of small reglona where there are rrt
13、ress concentrations when such yielding has no appreciable effect on the over-all elastlc behavior of the Joint. A part of the behavior of a joint may be pre- dictcid from an elastic theory, and empirical method8 may be employed in the determination of the upper llrnit of elastic action and of joint
14、behavior above thls elastlc limit,. This upper limit of elastic action is termed the “crit;ical bolt load“ R, in this paper. main plate and butt straps are of the same width and 4s and are joined by bolts that are materfal with t, = all of the 8ame Bize and material, the following procedure mag be u
15、sed to predict the Joint-load against the bolt- load (P-R) relationshimi throughout the elastic and yield ranges to joint failure. The Drocedure, -however, ha8 not been extended to include other joint arrsngements because only two- and three-bolt joints were teated. Far the elastic range, the P-R re
16、lationships may be established by means of equation (1). These relat3onehips may be plotted and the experimentally determined value of Rcr plottad on the P-R curve for the most heavily loa thus the P-R curve for either end bolt is completed. The end-bol-load R1 may be corn- puted for any joint load
17、from equations of the atraight lines obtained a8 outlined. Load R2 on the middle bolt may be found at any given joint load as R2 = P - 2Al. The vcilues of RZ, however, are of less importance man the greater loads R1 on the end bolts and in many cases it is unnecessary to computs-values or R2. - In t
18、he analysis of three-bolt joints in whlch the 2 - rrr 3 .:. . i e“ .- - -L . . .- Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-L A NACA TN BO. 1051 TEST SPrCIMENS AND PROCSNJRES . - Specimens and Apparatus . . Specimens. - Materials Cornon to airc
19、raft construction were used for the test specimens, which were seloctehto provide a plate-thickness range sufficient to check the applicabiiitg of the theory. six spe6rical butt-joint specimens - three of. tte two-bolt joints and three of three-bolt joints - .vere fabricated and ested. Each specimen
20、 was made symetrical about it3 transverse center line. Such a condition provides duplicate test specimena, as the theoretical behavior of the payt of the joint on one side of this center line is identical with that of the part on the other slde of the line. The specimens were classified in t_wo gror
21、zps, A and B. It wag decided ta test two-bolt joints (group A) in order to procure infor- determining bolt loads from strain measurements and to secure additLana1 information whlch might ser_ve_ -in the interpretation of data obtained from tests of tmeebol_$- joints, The three-bolt joints (group B)
22、were chosen to - furnish an experimental chsck of the theory and to dxpe- matlon in-regard to the reliability of the method for _. -2 dit6 testing and the analysis of data. - -_ In all cases, the material ofitbe plates was 24s-T aluminum alloy and the bolts were -inch aircraft bolts, equivalent to t
23、hose specified in reference 8 of heat- treated. alloy steel with minimum ultimate tensile and shearing strengtle of 125 ksi and 75 ksi, respectively. 4 - . Specimens- A-I and B-1 were of balanced deslgn -based The design stresses were 62 anq-90 ksi on the usual assumption that load is divided equall
24、y among the bolts. 75 ksi for shearing of the bolts. permits a greater allowable tensile stress, it was con- - shear atrerqth gf the bolts. ,Secimcn* A-2 and 372 vers- tension. In every case, the butt-strap thickness wag one-half that of the main plate. A width of 1 inchss 2nd a pitch of 2 inches we
25、re used for all specimens in order to acccjmmodate the strain gages. A wring fit was used to fit the bolts in all specimens. In preference to washers, collars made of and specbens -A-3 and B-3, in -t A1thG-h reference 8 t 2 Provided by IHSNot for ResaleNo reproduction or networking permitted without
26、 license from IHS-,-,-10 NACA TN EO. 1051 under the nuts to eliminate bearing of the plates on bolt threads. When the specimens wore assembled, the nuts were first Hghtened to bring the plates together and then, in order to eliminate friction farces insofar as practicable, were loosened to cause fir
27、m contact-between the bolt tieads or collars and plates, The specimens are shown in figure 1 and Cheir d1mension.s am given in table 1. a hycbau -%ET- e-sting mmhint. having 100-kip capacity and an ac:curacg within 0.5 percent. Wedge grips were used tu- apply load to specimen A-1, and the remaining
28、specimens were gripped wi tk Templin grips of 5O-kip capacl ty. A aratus.- Load was applied in tension by mean8 of 1 Ctrain was -measured by -Inch Sn-4 electrical gages. 2 With these gages, the error in strain measurement did not exceed 2 percent. An attempt WLLS made to measure bolt deflections by
29、means of micmmcter microscopes but wag abandoned because th3 inskrurnents wore not sufficiently precise to measure the small deflections that occurrcd in the elastic range. As IQsd was applied, cnlargoment of tk0 gap bstween main plates was measurGd with 1-inch Tuckerman optical strain gages. The ar
30、rangement of elec- trical strain gages is shown in figure 1 cind the general test arrangement for a typical specimen Is shown in figure 2. Te s t lng Pro ce dum The width and thickness of each plate wece measured at several points along the length of the plam with a micrometer caliper of 0.0001-inch
31、 precision, and the bolt diameters were checked as a precaution against the use of appreciably Irregular bolts. YLoadinp, t6sks in the elastic range.- In the loadine teats-in the elastic range, load was applied In six or seven (usually equal) increments to a ioad approxlmatuly equal to 45 percent of
32、 the estimated ultlmate load. The specimen was than unloaded with repetition of the incro- ments used in the application of increasing load. This prococrs was repuated twice; with the specimen thus sub- jected to threa complete cycles of loading. Strain readings were made at each incrament of loud.
33、This pro- c;dure was followed in testing all joints with the oxcep- tlon cf specimen A-1, which wa8 loaded directly to failuro. . .- Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-KACA TN No, 1051 11 Loading tests to failure.- Xter the first phase d
34、f testing, the spechiens were loaded to failures - Load ivas asplied in 12 to 15 incremen-ts until failure occuzred. Strain and tile increase in wl.ath of the ga;! between _ mah olates were observea at eecli load increment.- rkotopa hs of? the fractured secimens are snown as figures 3 s PmplOY6d Ln
35、joinicg the syeciinens of $roxp A and 3, For thsse tests, the Dlates were of S.A.E. bl5a heat-treated steel and a w-r.lng fit mas used to fit the bolts. Fne dfnrrellulons of the syecbens are shown in table 1. As ir; was desired to coq3are sspzratolg 3etermiiied bolt deflections with avera?e values c
36、omputed 2rom the novenent of the gap, three sets of cieflsction neasuraments were obtained in eacl, of two tests. Iove- ment of the gap vas aetermined in tke manner used to secur6 sbilar data for specilllens cf groups A and 3. DeflectLon maasurecentn for both bolts were obtslned - smerately bg placi
37、ng 2-inch Tuckerman gages ori o3pcaite faces of 6 specimen IsitSthe fixed knife edge of a yge on the butt stra;? and with the lozenge -on the main p ate. Precautjons were taken to snsure aTrprox5mats pardlelimc between the gages and ?late surfaces, Auxiliary tests.- stress-strzFn data were secured -
38、_ -_ - PF(ESEI7TATIOE AI.? DISCUSSIOE 03 T3ESULTS Determination of Solt Loads and DeflectLon? from Yest Data . -_ _ T9s exaxrlmental bolt loads -tiere obtained 57 findf-ag the loads in the butt strms at sections n:lclwa between bolts; the difference batween loads at two adjacsnt asc- tions was consf
39、derad to be equal to tha load on the intar- vening bolt. Butt-strq loads wera comsuted f-rorn strain 3.sta; And for specimens -A-I, A-2, 3-1, and 5-2 tae ibaas w9re corrected for the e2fect of lateral bending acment in the butt stra?s, whisr_ acted in a plane normal to tne nlane of the stra3s. The l
40、ateral mozent was induced by eccentricity of the resultant of the part CJ? a bolt load Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-12 NACA TN NO. 1051 that. was transmitted to one butt strap. The curves of the butt-strap lotids Ps are plotted In
41、=Cures 5(a) tu- l5(a) to illustrate the effect of lateral bending and are shown in conjunction with the curve8 of joint load against bolt load because of the interrelationship of these curve8 owing to the use of P8 in the determination of correc- tion factors. Calculation of butt-strhp loads, the pr
42、es- ence of lateral bending monient, and the correction pro- cedure are explained in appendix D. Tho methods used in the determination of bolt deflec- tions are based on relative movements of the main plates and butt straps. Elongation of either the main plates or butt atraps ; depending upon the lo
43、 cation of instruments used to measure Joint momenta, was included in measure- ment obtained during the tests. JMflections were corn- puted by subtracting Such elonGation (considered to be PL,/AZ) from the mst measurements. A more detailed expla- nation of the methods ernployeil 28 given in appendix
44、 D. Curves of J3int Load Against Bolt Laad The main results of this investigation are presented in the form of curves of joint- load against bolt load shown in figures 5 trr 15. These curves show the load history of each bolt-md indicate the behavior thatmay be expected of bolts loaded under conditi
45、ons similar to thoe of the tests. Elastic behaVLQr of t-est 6peclmens.- The curves accompanying th plotted points represent experimental curves for the same bolts obtained from the tests to failure. During the tostFng of specimen A-3, the strain gagas at the centsr of the joint on one butt strap bec
46、ame loosened. As a result, the values of fS, R2, and R3 could not be determined; hence, only the curves for R1 and R4 are shown in figury! 6. prior to the testing of specimen A-3 to fttilure. Inspec- tion of figures 5 to 9 shows that there is 8ood agrecment among the.repeated bolt loads and between
47、the8e loads knd the csrves from tests tu-failure. Figures 10 to 12 give the results of testing the Replacement of the inoperative cages was made opeciinens of group A to failure. Both theory- (appendix C) . . Provided by IHSNot for ResaleNo reproduction or networking permitted without license from I
48、HS-,-,-NACA TN 170. 1051 1-3 and ths conventional method of analysis (that is, the assumption of equal loads carried by the bolts) indicate that the P-R curves should be represented by the equation R =0.5OOP. This curve is not shown, however, as It was considered more informative to give the experim
49、ental curves and their equations. In every case, tho equation given for a cwve applies to the initial straight-line Dortion of the curve. In general, it may be seen that deviatfons of 3 to 11 percent from an equal distribution of load to the bolts occurred in the two-bolt. joints. The maximum deviation from equ
