am (Wal, I, No. 3 we obtain [Aste de | dn data f fon eee at 82 “The quantities es, Aay can be easily computed from Eqs. (19), (20) and (16) with the help of the result* fre where t(n) is the Riemann Zeta-function, tabulated for integral » in J. Edwards, “The integral calculus,” vol. 2, Macmillan, London, 1922, p. 144 rhe solution for K-=1 differs from the well-known exact solution for this case by Jess than 0.2 per cent, when only the first three of b, are retained, provided that a4 For larger values of K and a it may be necessary to retain more terms to achieve the desired accuracy but for practical values the amount of computation required is not excessive, 7 When n=O, the result reduces to foe tanh idt=log 24 a Niger gure table i given in E. Jahnke and F, Emde, Tales of functions, Dover Publicatios New York, 1943, p- 273. LARGE DEFLECTION OF CANTILEVER BEAMS* By K. E, BISSHOPP ano D. C, DRUCKER (Armour Research Foundation) ‘The solution for large deflection of a cantilever beam cannot be obtained fr elementary beam theory since the basic assumptions are no longer valid. Speciicall fhe slementary theory neglects the square of the first derivative in the curvatug formula and provides no correction for the shortening of the moment arm as sonded end of the beam deflects. For large finite loads, it gives deflections greater t the length of the beam! The square of the first derivative and correetion factors the chertening of the moment arm become the major contribution to the solution @ ~ Reosived Apel 6, 1945 1 -Rasprobion was cnadered by H. J, Barten, *On the Deflection of Cantilever Beam,” Qu of Apt Siete 168-17 1940), Previously an approximate solution had been obtained by Grom Leh a Die Feder, Belin VDI Neri, 1938.(Vol. II, No. 3 (19), (20) and (16) with n-3, tegral m in J. Edwards, 1449 solution for this case by ned, provided that a}. ore terms to achieve the sputation required is not junctions, Dover Publications, R BEAMS* srch Foundation) -afinot be obtained from longer valid. Specifically, ivative in the curvature the moment arm as the deflections greater than ind correction factors for ibution to the solution of 4 Cantilever Beam,” Quarterly ad been obtained by Gross und 1945) K. E, BISSHOPP AND D. C. DRUCKER 23 large deflection problems. The following theory which utilizes these corrections is in agreement with experimental observations. ‘The derivation is based on the fundamental Bernoulli-Euler theorem which states that the curvature is proportional to the bending moment. It is assumed also that bending does not alter the length of the beam. Considering a long, thin cantilever leaf spring, let Z denote the length of beam, 4 the horizontal component of the displacement of the loaded end of the beam, 8 the corresponding vertical displacement, P the concentrated vertical load at the free end, B the flexural rigidity, that is B= EJ, when cross-sectional dimensions are of the , TN A TV a FP f—t-a Fie. same order of magnitude, and ‘=EI/(1—?) for “wide” beams, where » is the Pois- son ratio. The exact expression for the curvature of the elastic line may be stated ‘convenient in terms of arc length and slope angle denoted by s and ¢, respectively, ‘9 that if x is the horizontal coordinate measured from the fixed end of the beam, the Product of B and the curvature of the beam equals the bending moment M. 4 PUL ~ 2-4) = w or Paz P a gat ee ® iam Notes {Vol IT, No, L(Q)- Emote ° ‘The constant C can be evaluated directly by observing that the curvature at the loaded end is zero. Then if ds is the corresponding angle of slope whence % Bo” (ings — sin. Sm 4/F Gin ty sn) ® The value of de cannot be found directly from this equation but itis implied by the requirement that the beam be inextensible, so that VE [firm fone snore = A(R)" In order to evaluate this elliptic integral, denote PL*/B by a? and let 1+ sin Ge) sin* int 9)-"7a8, sin @ = 2/28 The next step is to represent the deflection 6 in terms of a and an elliptic integral Since f sin dé c (ingo — sing)" With the aid of Eq. (6) we obtain sin 6 r+ Gannon ah, “This equation can be split up into complete and incdmplete elliptic integrals of th fret and second kinds. In the notation of Jabnke and Exide, 6 1 / Sod rw - many - 260) +226) = F(R) — F(R, 1), so that fai? few - 24,9) Fotos Lew ~ Bt, 9)[Vol IT, No. 3 @) tthe curvature at the lope @) but it is implied by the (5) 6, © VI/2b. a and an elliptic integral. 7 sin? 6 — 1)d0 sin ¢ elliptic integrals of the (h, 0), ® 1945) H. J. BARTEN ams ‘The horizontal displacement of the loaded end is calculated from Eqs. (1) and (4) with ==0 when ¢=0. Thus rea) ~0(2) = 24/Z ons =F in 80nn, (40) From Eq, (6) we have sin y= 2k*—1 Numerical results can be obtained by: (1) selecting values of & corresponding to tabulated values of the modular angle in the elliptic function tables and (2) determin- ing és and a from Eq. (7). After this has been done, 8/L and (L—A)/L can be calculated from Eqs. (9) and (10) and plotted against a?=PZ#/B. The results of these calcula- tions are shown in Fig. 1. CORRECTIONS TO MY PAPER ON THE DEFLECTION OF A CANTILEVER BEAM* QuaRTERLy OF APPLIED MATHEMATICS, 2, 168-171 (1944) By H. J. BARTEN This paper is correct up to the equation fos conoas The next step Cheney i 08 8 is incorrect since @ is not only a function of L, but is also a function of s. This error makes Eqs. (9), (11), and (12) incorrect. Using the relation qe and the various steps used in the original paper, we find that eou~ a(42) He. By using 5 as an independent variable we can calculate corresponding values of # and Received June 25, 14.