NASA Technical Paper 3192

Stress Concentrations for Straight-Shank and Countersunk Holes in Plates Subjected to Tension, Bending, and Pin Loading
K. N. Shivakumar and J. C. Newman, Jr.

June 1992

Three-dimensional .A three-dimensional stress-concentration analysis was conducted on straight-shank and countersunk (rivet) holes in a large plate subjected to various loading conditions.

Based on the . remote bending. The plate material was assumed to be linear elastic and isotropic. three types of loading were considered: remote tension. Results for remote tension and wedge loading were used to estimate stress concentrations for simulated rivet or pin loading.nite-element analyses were performed with 20-node isoparametric elements.1 to 2.3. but the angle was held constant at 100 for most cases. with a Poisson's ratio of 0. For straight-shank holes. Stress concentrations along the bore of the hole were computed for several ratios of hole radius to plate thickness (0. The countersink angle was varied from 80 to 100 in some typical cases. only remote tension and bending were considered. and wedge loading in the hole. For countersunk holes.5) and ratios of countersink depth to plate thickness (0.25 to 1).

A riveted joint is a commonly used method of joining structural components. Exhaustive studies on stress-concentration factors (SCF's) for holes and notches in two-dimensional (2-D) bodies subjected to a wide variety of loadings have been reported in the literature (refs. The stress concentration at a hole in a plate subjected to bending was . changes in the load path due to lapping. local stresses are elevated in the structural component. A recent paper by Folias and Wang (ref. Whenever possible. Accurate estimations of these local stresses are needed to predict joint strength and fatigue life. 6) provides a review of these previous solutions and presents a new series solution. stress-concentration equations were developed. and additional loads such as rivet bearing and bending moments. Joining introduces discontinuities (stress risers) in the form of holes. the present results were compared with other numerical solutions and experimental results from the literature. The Folias and Wang solution covers a wide range of ratios of hole radius to plate thickness. Studies have also been made on three-dimensional (3-D) stress concentrations at circular holes in plates subjected to remote tension loads (refs.nite-element results. Because of these changes at the joint. 3 to 6). 1 and 2).

Cheng's results (ref. remote tension and remote bending. 10) conducted 3-D photoelastic measurements to verify Reissner's circular hole and Naghdi's elliptical hole solutions. Two types of loading. Rubayi and Sosropartono (ref. Furthermore. 16 and 17). 4) using the Love-Kirchho thin-plate theory (ref. and wedge loading in the hole (simulated pin loading). 17) include stressconcentration factors for both tension and bending loads.g. Only two papers in the literature report results on stress concentration at countersunk holes (refs. 13 to 15) results are reported in the literature for remote loading. Three types of loading. 7). but none consider 3-D e ects for rivet loading in the hole. Countersunk-rivet construction is commonly used in aircraft industries to achieve aerodynamically smooth surfaces. and stress concentrations for countersunk holes have not been reported in handbooks. refs. Reissner's SCF solution for bending loads is presented in Abstract terms of Bessel functions. 11 and 12) and experimental (e. remote tension. 9) extended Reissner's analysis to elliptical holes using Mathieu's functions. Naghdi (ref.. 8) rederived the plate solution including the e ect of shear deformation and showed that Neuber's solution was unconservative. Three-dimensional . remote bending. refs. The objective of the present study is to conduct a comprehensive analysis of three-dimensional stress concentrations for circular straight-shank and countersunk (rivet) holes in a large plate subjected to various loads encountered in structural joints. Both papers use the 3-D photoelastic slice method to obtain stress concentrations for holes in thick plates.rst presented by Neuber (ref. 3-D stress concentrations are needed to predict the strength and life of joints. are considered for the straight-shank hole. Reissner (ref. Three-dimensional stress concentrations at riveted joints are not fully understood. Knowledge of 3-D stress concentrations is needed to verify the adequacy of 2-D solutions in the handbooks. are considered for countersunk holes. These joints are not amenable to 2-D approximation. Many other analytical (e..g.

3. based on the 20-node isoparametric element. With the F-E results. series-type stress concentration Introduction . The in uence of the countersink angle on stress concentrations is also examined. with a Poisson's ratio of 0. simple. A wide range of hole sizes (ratio of hole radius to plate thickness) and countersink depths (ratio of countersink depth to plate thickness) are considered in the analysis. The plate material is assumed to be linear elastic and isotropic. The FRAC3D code is an elastic and elastic-plastic fracture mechanics code developed at NASA Langley Research Center for the analysis of cracked isotropic or anisotropic solids.nite-element (F-E) stress analyses of large plates with straight-shank and countersunk circular holes were conducted with the FRAC3D F-E code.

y. Whenever possible.equations are developed for a wide range of holeradius-to-plate-thickness ratios and for any location along the bore of the hole. z ij . Nomenclature b FEM h Kb Kbmax Kp Kt Ktmax Kw M P r S t w x. the present results are compared with results from the literature.

ij c yy  depth of straight-shank portion of hole .

nite-element method one-half height of plate stress-concentration factor along bore of hole under bending maximum stress-concentration factor along bore of hole under bending stress-concentration factor along bore of hole under pin loading stress-concentration factor along bore of hole under tension maximum stress-concentration factor along bore of hole under tension stress-concentration factor along bore of hole under wedge loading applied remote bending moment applied pin or wedge loading radius of straight-shank portion of hole applied remote tension stress plate thickness one-half width of plate Cartesian coordinate system coecients in stress-concentration equations coecients in stress-concentration equations countersink angle hoop stress at  = 90 angle de.

ning applied stress distribution in hole countersunk hole. the thickness was divided into two sections: the cylindrical section. The stress concentrations depend on the length of the straight-shank or countersink depth. The two sections meet to form an edge referred to as the countersink edge. (See . referred to as the straight-shank depth b. and the conical section. referred to as the countersink depth t-b. The two extreme cases of countersunk holes are when b = 0 (knife edge) and b = t (straight-shank hole). The countersink angle was  c .

the plate width and height were selected large enough so that the stressconcentration solutions were not greatly a ected by the remote boundaries.) In the present study. 1(b).g. Stress concentrations for .

For the straight-shank hole. A wide range of values for hole-radius-to-plate-thickness ratio r=t and straight-shank-depth-to-plate thickness ratio b=t were considered in generating the data base on stressconcentration solutions. six values of r=t (encompassing the range of structural con.nite-size plates have to be generated with the use of either analytical or numerical methods.

5.25. with w=r = 5 and h=r = 5. 0. 0.gurations used in industry) were selected: 2.5. and 0.0. 1. 1.5.1. For the countersunk-hole con.

0. and 0. 0. and 0.25 and b/t values were 0.75 with w=r = 7:5 and h=r = 7:5.50.25. 0.0. r=t values selected were 2. 0.guration. 1. Although the straight-shank hole con.5.

guration is a special case of the countersunk hole (b=t = 1). for convenience the two con.

remote bending moment per unit width M . Loading Conditions Figure 2(a) shows the three types of loadings that were applied to the plate with the straightshank hole: remote tension stress S .gurations are considered separately. The wedge loading was imposed on the hole boundary as a normal pressure loading that has a cosine distribution and is assumed to be constant through the plate thickness. Appendix A explains how the stress concentrations for a pin-loaded hole were approximated from the remote tension and wedge loading solutions. The surface pressure is de. and wedge loading P .

The angle  is measured from the y-axis. (See . 18 and 19) and was applied over the angle  = 690 .ned as (2P=rt) cos  (refs.

remote tension and remote bending. two loading types. were considered. 2(a).) The bending moment M was applied as an equivalent remote stress that varies linearly through the plate thickness.g. (See . For countersunk holes.

2(b). rivet (pin) loading was not considered in the current study. Rivet Hole Con.g.) Because of the lack of understanding of 3-D load transfer between the rivet and the countersunk hole.

gurations Two types of rivet hole con.

gurations. were considered. straightshank and countersunk holes in a large plate. The two types of holes and nomenclature used are shown in .

In the plate with a 2 .gure 1.

De.

nition of Stress-Concentration Factor Although the de.

many of these solutions are associated with 2-D con.nition of the stress-concentration factor is given in many classical books on theory of elasticity and in stress-concentration handbooks.

For 3-D con.gurations.

Herein. the stress-concentration factor is de. however. such as along the bore of the hole.gurations. the stress concentration varies along the structural discontinuity.

For con.ned as the stress at any location along the bore of the hole normalized by a characteristic stress (related to applied loading).

) The stress-concentration factors for the three loading conditions are de.gurations and loading conditions considered in this study. the peak stresses occurred at  = 90 because the pin contact angle was assumed to be 90. (See appendix A for details. the highest stresses occurred along the bore of the hole at the intersection of the hole surface and the y = 0 plane. Even for the case of pin loading.

ned as follows. Remote tension. The stress-concentration factor for tension Kt is the hoop stress yy at  = 90 along the bore of the hole normalized by the applied remote tension stress S and is given by  (z) Kt (z ) = yy S (1) Finite-Element Modeling A three-dimensional .

20) for solving matrix equations of equilibrium. The F-E model includes the full thickness of the plate.nite-element code FRAC3D developed at NASA Langley Research Center for analyzing cracked isotropic and anisotropic solids was used in this study. The plate with the countersunk hole was symmetric about the x = 0 and y = 0 planes. and z = 0 planes. The program uses a vector skyline Choleski decomposition algorithm (ref. The sti ness matrix and the consistent load vectors were generated with the 2 by 2 by 2 Gaussian quadrature formula. The remote bending was symmetric about the x = 0 and y = 0 planes and antisymmetric about the z = 0 plane. The FRAC3D code has an option to impose symmetry and antisymmetry boundary conditions. Because many con. hence. only one-eighth of the straight-shank hole plate was modeled. The plates with the straightshank hole and remote tension and wedge loading were symmetric about the x = 0. The code is based on the 20-node isoparametric element formulation. Because of these conditions. one-fourth of the plate was modeled. y = 0.

a simple 3-D modeling procedure was developed to generate the .gurations were to be analyzed.

nite-element meshes. a 2-D F-E mesh in the x-y plane was generated with re. In this procedure.

ned elements near the hole boundary. Then the 2-D mesh was translated in the z-direction (with appropriate x-y transformation to account for the countersunk hole). Typical 3-D F-E meshes for one-eighth of a straight-shank hole in a plate and for one-quarter of a countersunk hole in a plate are shown in .

10. the hole radius was kept constant and the plate thickness was scaled by t=r . and 2 percent of the total plate thickness. The layer thicknesses (starting from the z = 0 midplane) were 15.gure 3. 6. The small thickness layers were used in the high-stress-gradient regions (near the free surface). 13. the halfthickness of the plate was divided into six layers of unequal thickness. The F-E mesh for r=t = 1:0 is shown in . The F-E model had 936 elements and 4725 nodes (14 175 degrees of freedom). For all straight-shank hole models. For di erent values of r=t. 4.

Table 1 gives the details of the F-E idealizations used for countersunk holes with b=t = 0. 3 Remote bending . In the countersunk hole. di erent through-the-thickness idealizations were used for di erent countersink edge locations b=t . The stress-concentration factor for bending Kb is the hoop stress yy at  = 90 along the bore of the hole normalized by the remote outer-.gure 3(a). there are three regions where the stress gradient is high: near the two free surfaces of the plate and at the countersink edge. Therefore.

The stress-concentration factor for wedge loading Kw is the hoop stress yy at  = 90 along the bore of the hole normalized by the average bearing stress P =2rt and is given by  (z ) Kw (z ) = yy P=2rt (3) Pin loading.ber bending stress 6M=t2 and is given by  (z ) Kb (z) = yy 2 6M=t (2) Wedge loading. The stress-concentration factor for pin loading Kp is obtained from a superposition of remote tension and wedge loading.) The factor Kp is de. (See appendix A.

ned as the hoop stress yy at  = 90 along the bore of the hole normalized by the average bearing stress P =2rt and is given by  (z ) Kp(z ) = yy P=2rt (4) .

17) for thick plates subjected to tension and bending.75. The distribution of the stressconcentration factor Kt along the bore of the hole for remote tension is shown in .02. and 0. Comparison With Other Solutions Remote tension. even for r=t as small as 0. The present 3-D stress-concentration factors (SCF's) for the straight-shank hole are compared with Folias and Wang's solution (ref. Straight-S ha nk Hole reported in reference 6.25. 8) for remote bending. Three-dimensional stressconcentration solutions for wedge loading or simulated pin loading have not been reported in the literature.50. For countersunk holes.0. the present solutions are compared with Cheng's photoelastic measurements (ref. 6) for remote tension and with Reissner's solution (ref. 0. The reason for not capturing this expected trend in reference 6 is unknown. Figure 3(b) shows a typical F-E model (r=t = 0:25 and b=t = 0:50) for one-quarter of a plate with a countersunk hole.

gure 4 for various values of r=t.) In all these cases. the plate width and height were selected large enough (w=r = h=r = 5) so that Kt values are not greatly a ected by the . (Note the expanded scale on the ordinate axis. The stress concentrations are symmetric about the midplane (z=t = 0).

nite plate. 6) in. The SCF's for r=t  0:5 are about 2 percent larger than Folias and Wang's (ref.

6 are not shown in . results from ref.nite-plate solutions for all z-values. (For clarity.

4.g.) Part of this di erence may have been caused by the .

and from Neuber's thin-plate theory (ref. (These results are shown subsequently. For thicker plates (r=t < 0:25). the present F-E solutions show that the location of Kbmax is at z=t = 60:5 (free surface) for thin plates (r=t  0:5).nite-size plate used in the present study. the Kt value appears to peak at a value of 3. but the magnitude of the present results is 3 percent lower than the magnitude of their solution. For r=t = 0:1. It is expected that the classical value of Kt = 3 would be obtained for much thicker or thinner plates. from Reissner's shear deformation plate theory (ref. z=t Remote bending. Figure 5 shows a comparison of maximum bending stress-concentration factor Kbmax calculated from the present analysis. 4). the present results show the same trend as that of the Folias and Wang solution (maximum SCF near the free surface). the maximum Kt location shifts toward the free surface (z=t = 60:5). For r=t  0:5.) In . At z=t = 0. the maximum Kt occurs at = 0 (midplane). From the plots of Kt for various values of r=t. 2. Kbmax always occurs at z=t = 60:5. but the maximum SCF is slightly interior from the free surface (jz=tj < 0:5) for thick plates (r=t < 0: 25). Because both Reissner and Neuber assumed that the stress distribution was linear through the thickness.22 for r=t = 0: 5. However. 8). two observations are made: 1.

ref.) However. The SCF's for the two con. The geometric parameters of models 7 and 8 are given in table 2. as observed by many investigators. b=t = 0:6 and  c = 90 . (See. 17) measured 3-D stress-concentration factors for countersunk holes in thick plates using a photoelastic slice technique for both tension and bending. for example. the maximum SCF values from the F-E analysis are plotted for various values of r=t extrapolated to r=t = 0. (See the dashed curve. Neuber's thin-plate theory. is inadequate even for r=t = 2:5 and produces values about 6 percent lower than those for Reissner's solution and the present results. Cou nters un k Hole These two trends are consistent with crack-tip stressintensity factors for cracks in thick plates. The di erence between results for Reissner's solution and the present results for r=t less than unity is about 4 to 8 percent. Cheng's photoelastic models for tension (model 7) and for bending (model 8) were analyzed through the generation of separate F-E meshes. Kb = (5 +  )=(3 + ). 21. such a drop in Kt near the midplane (z=t = 0) for very thick plates was not 4 Cheng (ref.gure 5.) Results for Reissner's shear deformation theory and the present results agree well with each other for r=t > 1:5. For both models.

The maximum SCF calculated from the F-E analysis is within 3 percent of Cheng's measured value. whereas the countersink edge is at z=t = 0:1). The F-E results show that the maximum SCF for remote tension occurs slightly away from the countersink edge and in the straight-shank portion of the hole (at z=t = 0:08. (See table 2. note that percent error is de.gurations at the critical locations are presented in table 2.

ned as the di erence between solutions divided by the largest stress-concentration value. 0.5 percent at z=t = 00: 5. three locations on the hole (z=t = 00:5.1. and 0. but the di erence .) For bending. The di erence between Cheng's measurements and the present solution is about 2.5) were considered for comparison.

As previously observed for straight-shank holes in thick plates.is about 8 percent at the countersink edge (z=t = 0:1) and along the countersink ank (z=t = 0:5). 21 and 22). The two parameters that can in uence the SCF for countersunk rivet holes are the countersink angle c and the countersink depth t 0 b. (See . the maximum SCF is not at the free surface (z=t = 00:5) but is slightly interior to the free surface (z=t = 00:48). The drop in SCF at the free surface is attributed to the well-known free-boundary-layer e ect (refs.

g. Count ers in k Angle E ect of Countersink Parameters on SCF . 1(b).) The e ects of these two parameters on SCF at countersunk holes in plates subjected to tension and bending were analyzed.

SCF's for countersunk holes are only about 30 percent higher because the .nite-thickness plates.

nite thickness elevates the SCF. as shown in .

The variation of maximum Kb (at z=t = 6 0: 5) is less than 1 percent for 0:25  b=t  0: 75. The maximum bending SCF at z=t = 00:5 is 02: 41 for b=t = 0.gure 4. except for b=t = 0 (near the knife-edge location).) In this study. (Note that the SCF will be positive at z=t = 00:5 if the moment is reversed. this Kb is about 24 percent lower than that for b=t = 0:50. Figure 7(b) shows the bending SCF distribution for various values of b=t. the maximum bending SCF is almost una ected by b=t. In contrast to tension loading. 3-D stress-concentration factors for a wide range of hole con.

gurations and loadings were generated with the .

These solutions may be used in structural design as they are or they may be interpolated to calculate stress concentration at any other location along the hole boundary or for other hole (r=t and b=t) con.nite-element method.

SCF equations are developed by .gurations. In this section. The F-E solutions will be easier to use if equations are developed.

Multiparameter least-squares equation . Separate equations are developed for straight-shank and countersunk holes subjected to di erent loading conditions.tting the F-E results to double-series polynomial equations.

The weight factor for each of the SCF values along the bore of the hole is selected such that the weight is proportional to the length between the two neighboring points on either side of the point under consideration. This procedure of selecting the weight factor minimizes the area under the SCF curve on the z -axis. Also. For example. a weight factor for the ith point is (zi+1 0 zi ) =2t. which uses the Householder transformation for matrix factorization (ref.ts were performed with the International Mathematical and Statistical Library routine QRASOS. this procedure gives good . 23).

ts to even unequally spaced data points without higher order oscillations. which are generally present in a high-order polynomial .

Sha nk H ole Stress-Concentration Factor Equations The e ect of small variations in the countersink angle c on tension and bending SCF was analyzed with Cheng's model 7 con. Straight.t.

Figure 6 shows the distribution of Kt and Kb along the bore of the hole for c = 80. and 100 . However. the variation in Kt is much smaller at all other locations on the hole boundary. (See . Kb varies less than 1 percent. 90. A change in  c of 610 from the reference angle of 90 changes maximum Kt by about 3.5 percent at the countersink edge. 17).guration (ref. For the 610 variation in  c.

For thin plates. the e ect of  c variation on SCF is of the same order as that shown for the thick plates. where r=t = 0:24.g. the countersink-depthto-plate-thickness ratio is de. used in aircraft applications (r=t of about 2). 6(b). Count ers in k Dept h As previously mentioned.) These results are for a thick plate.

50. and 0.75. r=t = 2:0. For convenience.25. 0.ned as 1 0 (b=t). and  c = 100 (a typical value for the aircraft industry). Figure 7 shows the distribution of Kt and Kb along the bore of the hole (00:5  z=t  0:5) for b=t = 0. b=t is used as a depth parameter. In . The plate w=r = h=r = 7:5. 0. where b=t represents the ratio of the straight-shank depth to the plate thickness.

3. 0. on the straight-shank portion of the hole. 4. and 0. and 3. for The con. the maximum tension SCF's occur at the countersink edge for all values of b=t. In contrast.50.82. However. the maximum SCF's occur slightly away from the countersink edge. The highest SCF for the countersunk hole under remote tension is about 37 percent higher than the classical 2-D value (Kt = 3). 0. respectively.25.10.gure 7(a).39 for b=t = 0. in thick plates.75. The maximum tension SCF's are 4.06.

(See .guration for a plate with a straightshank hole is symmetric about the z = 0 plane.

Therefore. an even-power polynomial in z and a general polynomial in r=t were used to . whereas the bending is antisymmetric about the z = 0 plane.g. 2(a).) The tension and the wedge loading are symmetric about the z = 0 plane.

An 5 .t SCF results for tension (Kt ) and for wedge loading (Kw ).

odd-power polynomial in z and a general polynomial in r=t were used to .

The forms of the SCF equations are Km = X X (r=t) (z=t) 4 4 i=0 j =0 ij i 4 ij i 2j (5) development of equation (7) is given in appendix A.t the SCF results for bending (Kb ). The results from equation (7) are shown in .

Of course. these results show the same trends as those shown in .gure 11.

respectively. Countersunk Holes The con.gures 8 and 9 for tension and wedge loading.

gurations of the countersunk hole dictate that two separate SCF equations be .

separate equations were developed for each value of b=t.t: one equation for the straight-shank part (00:5  z=t  (b=t 0 0: 5)) and the other equation for the countersunk portion ((b=t 0 0:5)  z=t  0: 5). Furthermore. A general polynomial series equation in terms of r=t and z=t was .

(Note that an enlarged scale is used on the ordinate axis in . and 5 for remote tension. The coecients ij are given in tables 3. Figures 8 to 10 show a comparison between the F-E results and the equation results for various values of r=t for remote tension. The orders of the polynomials for r=t and z=t are selected by trial and error such that the maximum di erence between the F-E results and the equation results is minimal and the sum of residuals is a minimum. respectively. Equations (5) and (6) apply over the range 0: 1  r=t  2:5. respectively. and remote bending. and remote bending. The SCF equations are given by Km = where m = t for remote tension and m = w for wedge loading and Kb = X X (r=t) (z=t) 0 4 i=0 j =1 2j 1 (6) for remote bending.t to the F-E results with the leastsquares procedure previously discussed. wedge loading. 4. wedge loading.

The maximum di erence between the F-E solutions and the equation results is about 1 percent. For both remote tension (. 8 to 10 to magnify the di erence between the equation and the F-E results.gs.) The equation results agree well with the F-E results for all values of r=t and for all three loading conditions.

g. 8) and wedge loading (.

9). The drop is larger for thicker plates (smaller r=t ). The bending SCF (. the SCF drops near the free surface.g.

the SCF equation for simulated pin loading is written as Kp = Kw + (r=w )Kt 2 (7) X X (r=t) (z=t) 3 4 i =0 j=0 ij i i j (8) for 00:5  z=t  (b=t 0 0: 5) and Km = 3 4 i=0 j=0 ij X X .g. 8) and Neuber (ref. 9) is almost linear for r=t  1:5 and becomes nonlinear for thick plates (r=t  1:0). Now that the SCF equations for remote tension and wedge loading have been established. For r=t = 0:5 and 0. it is located in the interior of the plate (jz=t j < 0:5). 4) analyses is valid only for r=t  1:0. the assumption of linear stress (or strain) distribution through the thickness made in the Reissner (ref. Thus.25. the maximum SCF is not at the free surface (z=t = 60:5). particularly near the free surface.

(r=t) f[z 0 b + (t=2)] =(t 0 b)g j Equation (7) is restricted to r=w = 0: 2 because Kt and Kw are generated for a plate with r=w = 0:2. Equations (8) and (9) apply over 0:25  r=t  2:5. Coecients ij and . The 6 (9) for (b=t) 0 0:5  z=t  0:5.

the maximum SCF is within 2 percent of the F-E results for all b=t values. Even for thick plates. see . The equation results and F-E results agree well. except near the free surface for thick plates.ij for various values of b=t are given in tables 6 and 7 for remote tension and remote bending. Note that the bending SCF at z=t = 0: 5 for the straight-shank hole (b=t = 1:00. respectively. Figures 12 and 13 show comparisons between results for equations (8) and (9) with the F-E results for remote tension and remote bending. respectively.

g. 10) is slightly less than that at z=t = 00:5 for the countersunk hole with b=t = 0:50 (see .

pin loading. To generate the SCF's for values of b=t other than those used in this study. remote bending.g. and wedge loading. . 13(c)). This program is based on equations (5) to (9) with the coecients presented in tables 3 to 7. a FORTRAN program is given to evaluate the SCF's for straight-shank and countersunk holes subjected to remote tension. In appendix B. This program may be used to generate three-dimensional SCF's for any value of b=t and r=t and at any location along the bore of the hole. an interpolation scheme between the available solutions has been implemented in the program.

The plate material was assumed to be isotropic. Three-dimensional . with a Poisson's ratio of 0.3.A comprehensive three-dimensional stressconcentration analysis of straight-shan k and countersunk (rivet) holes in a large plate subjected to various loading conditions encountered in service was conducted.

remote tension. Stressconcentration factors for wide ranges of hole-radiusto-plate thickness and countersink-depth-to-plate thickness ratio were generated. Series-type equations were . for the countersunk hole only remote tension and remote bending were considered.nite-element analyses were performed with 20-node isoparametric elements. The countersink angle was varied from 80 to 100 in some typical cases. but the angle was held constant at 100 for most cases. were considered. For straight-shank holes. three types of loading. remote bending. and wedge loading.

t to the .

nite-element re- Concluding Remarks sults. These equations generally agreed within 1 percent of the .

However.5 percent) for both remote tension and remote bending. 1992 7 .nite-element results. the bending SCF was almost una ected by countersinking the hole. VA 23665-5225 April 15. Variation in the countersink angle (80 to 100) had little e ect on the peak SCF (a change of less than 3. the SCF was about 30 percent higher than the 3-D SCF for a straight-shank hole with the same hole-radius-to-plate-thickness ratio. except for the knife-edge case (no straight shank). NASA Langley Research Center Hampton. Tension stress-concentration factor (SCF) for a countersunk hole was about 37 percent higher than the classical (2-D) solution for a circular hole (SCF = 3).

the pin and the plate surface are assumed to be smooth and the pin . In this analysis. as already mentioned in the text.Appendix A Computation of Stress-Concentration Factors for a Pin-Loaded Hole in a Large Plate This appendix shows how the SCF solutions for the wedge loading and for the remote uniform stresses can be used to predict the SCF for a pin-loaded (or rivet-loaded ) plate.

ts snugly into the hole (no clearance). 24 to 27) have clearly demonstrated that the maximum tensile stress concentration due to pin loading occurs at the end of the contact between the pin and the hole boundary. Two-dimensional studies (refs. For a snugly .

t pin joint and modeling the pin. Because the present analysis assumes that the contact angle is 90 . see . the maximum stress concentration occurs also at 90. the superposition of the wedge loading and remote loading solutions (both give the maximum stresses at 90) give the highest stress concentration. If the contact angle is assumed to be 83. the pin-load reaction is approximated by a cosine distribution over the contact angle  of 690 (refs. Crews et al. Therefore. these references show that the contact angle is about 83. 18 and 19). However. 25) show that a contact angle of 83 or 90 (between the pin and the hole) has very little e ect on the maximum stress concentration (about a 3-percent di erence). (ref. In the present analysis. (The angle is measured from the y-axis. then the maximum stress at 90 is within 3 percent of the maximum value at 83.

) Consider a pin-loaded hole as in the plate shown in .g. 2(a).

shown in .gure 14(a). where the plate height h is large compared with the hole radius. Consider another problem. This condition results in a uniform stress of P =2wt at y = 0h. The SCF for this problem is measured by the yy stress at y = 0 and x = r.

where the pin load acts on the lower half of the hole and the corresponding remote stress is P=2wt at y = h. The two problems in .gure 14(b).

the yy stress for these two problems is identical. The sum of these two loading conditions in . On the x-axis.gures 14(a) and 14(b) are identical except that the stress is 180 out of phase.

as shown in .gures 14(a) and 14(b) can be represented as wedge loading and remote tension.

Alternatively.gures 14(c) and 14(d). Therefore. yy stress on the x-axis for the pin-load case is half the sum of the yy stress due to wedge loading and remote tension. the stress concentration for the pin load Kp is de.

ned in terms of the stress-concentration factor for the wedge load Kw and the remote tension load Kt as follows: Kp = Kw + (r=w)Kt 2 (A1) A plane strain analysis of r=w = 0:2 was performed with the 3-D .

26). such as Crews et al.985) reported by Chang et al. and Eshwar et al.058. ref. 27). Some di erences are observed between the present results and those of other analyses.87. 24). ref.nite-element analysis described in the text.922. De Jong (1. The computed pin-load SCF from equation (A1) is Kp = 0:994. 8 . (ref. ref. (0. 25). (0. This solution agrees reasonably well with the experimental results (0.

LCASE FORMAT(A1) PRINT *. .' r/w = '.RT. remote bending.ZT.LOADING CASE AS DEFINED BELOW C 9 . PIN LOADING AND WEDGE LOADING CHARACTER*1 LCASE PRINT *. Jr.BT. REMOTE BENDING. C. C C PARAMETERS IN THE CALL STATEMENT: C (A) INPUT C C BT = b/t.Appendix B Computer Code Used To Ca lculate Stress-Concentration Factors This appendix p resents a FORTRAN program and su broutine (SCF3D) used to calculate the stressconcentration factor at any location along a straight-shank or countersunk hole sub jected to various loadings. Kt Kb Kp Kp Kw Kt Kw C C C C C PROGRAM MAIN COMPUTES THREE-DIMENSIONA L STRESS-CONCENTRATION FACTORS FOR STRAIGHT-SHAN K OR COUNTERSUNK HOLES SUBJECTED TO REMOTE TENSION.LCASE. z/t.ZT. Shivakumar and J.'INPUT LOAD CASE (TENSION. BENDING. C C THREE-DIMENSIONAL STRESS-CONCENTRATION IN COUNTERSUNK AND STRAIGHT C SHANK RIVET HOLES. and wedge codes. LOCATION WHERE STRESS-CONCENTRATION FACTOR IS REQUIRED C AS A RATIO OF PLATE THICKNESS C NOTE: 'z' IS MEASURED FROM THE MID-PLANE OF THE PLATE C RW = r/w. and for remote tension. HOLE RADIUS TO PLATE WIDTH C LCASE . RT. pin loading.RW. . Newman. The simulated pin-load value appendix A.ZT.' b/t = '. b/t.RW. WEDGE): READ 1. STRAIGHT SHANK LENGTH TO PLATE THICKNESS RATIO C RT = r/t.VERSION CREATED APRIL 1991 C C DEVELOPED BY: Kunigal N. 'Stress-Concentration Factor = '. SCF STOP END SUBROUTINE SCF3D(RT. P or W' 1 CALL SCF3D(RT. and procedures that are presented in loading. PIN.BT.' z/t = '.RW PRINT *.RW T. r/w ?' READ *. The program returns was calculated from .'r/t = '.SCF) C---------------------------------------------------------------------------C C SCF3D .ZT.SCF) PRINT *. B. . This program was developed such that it may be readily incorporated into other stress-analysis or life-prediction . 'INPUT: r/t.LCASE. HOLE RADIUS TO THICKNESS RATIO C ZT = z/t.BT.BT.

WEDGE LOADING C C COUNTER SUNK HOLE C LCASE = T. 2) PRINT *.GT.NE. STRESS/(P/(2rt)) C P = WEDGE LOAD C C NOTE: SCF FOR SIMULATED RIVET LOADING IS OBTAINED BY C ADDING ONE-HALF OF THE SCF FOR REMOTE TENSION C S = P/(2wt) AND ONE-HALF OF THE SCF FOR WEDGE C LOADING (2w IS TOTAL WIDTH OF PLATE). BT .GT. 3) PRINT *. STRESS/(6M/(t*t)) C M = REMOTE APPLIED MOMENT PER C UNIT WIDTH C C PIN LOAD: SCF = MAX. 0. 0. PIN LOADING C LCASE = W.5) IERR = 2 IF(RT . -0.25) IERR = 4 10 . REMOTE BENDING C LCASE = P. 1) PRINT *. 0 (PARAMETER OUT OF RANGE) IERR = 0 IF(BT . ZT .25) IERR = 2 IF(IERR .LT.LT.LT.NCASE/1HT. REMOTE BENDING C C (B) OUTPUT C C SCF .'INPUT PARAMETER z/t OUT OF RANGE' IF(RW .C STRAIGHT SHANK C LCAS E = T.GT.'INPUT PARAMETER r/t OUT OF RANGE' IF(ZT .501) IERR = 3 IF(IERR . C---------------------------------------------------------------------------CHARACTER *1. STRESS/S C S = REMOTE APPLIED STRESS C C REMOTE BENDING: SCF = MAX. RE MOTE TENSION C LCASE = B. LCASE. 2.0) IERR = 1 IF(IERR .'INPUT PARAMETER b/t OUT OF RANGE' IF(RT .THREE-DIMENSIONAL STRESS-CONCENTRATION FACTOR C C REMOTE TENSION: SCF = MAX.0 .501 . 1.EQ.EQ. REMOTE TENSION C LCASE = B.GT.OR.1HW/ C C BEGIN ANALYSIS C C INPUT ERROR WHEN IERR .OR.EQ. 0.NCASE DATA MCASE. STRESS/(P/( 2rt)) C P = PIN LOAD C C WEDGE LOAD: SCF = MAX. .MCASE.

4) PRI NT *.ZT1.25) THEN CALL NZT(BT.ZT.0 .ZT1.SCF1) CALL CSHANK(RT. BT .25. 0.ZT2) CALL CSHANK(RT.SCF1) CALL SSHANK(RT.GT.ZT2) CALL CSHANK(RT.LCASE.ZT.AND.0.SCF) RETURN ENDIF IF(BT .LCASE.25 * BT RETURN ENDIF IF(BT .0.75) RETURN ENDIF 11 .LT.0.SCF1) CALL CSHANK(RT.0. 1.ZT.EQ. 0.BT. 0.0.ZT.NCASE.5 20 RETURN ELSE IF (BT . 0.75.EQ. 0.0. 0.ZT.ZT1.LCASE.OR.75. 0.LCASE. BT .0.ZT. 1.'INPUT PARA METER r/w OUT OF RANGE' IF(IERR .ZT2. 'P' .ZT2. 0) STOP C IF (BT .75.LCASE. LCASE .LE.ZT2.SCF2) SCF = SCF1 + (SCF2-SCF1)/0.IF(IERR . BT .50.ZT.NE.0.GT.GT.ZT2) CALL CSHANK(RT.GT.ZT1.SCF) GOTO 20 10 CONTINUE CALL SSHANK(RT.AND.0) THEN CALL NZT(BT.MCASE.LCASE.ZT1.1.AND.5.25 .0) THEN CALL CSHANK(RT.25 * (BT-0.25.0.75) THEN CALL NZT(BT.25 * (BT-0. 'p') GOTO 10 CALL SSHANK(RT.0.LE.SCF2) SCF = SCF1 + (SCF2-SCF1)/0.5.50 .AND.LE.ZT.0.75.LCASE.ZT1.LCASE.ZT1.ZT.SCF1) CALL CSHANK(RT.SCF2) SCF = SCF1 + (SCF2-SCF1)/0.0.SCF2) SCF = SCF1 + (SCF2-SCF1)/0.0.50. BT .0.0.EQ.0) THEN IF(LCASE .0.75 .5) THEN CALL NZT(BT.LCASE.SCF W) SCF = (SCFW + RW * SCFT) * 0.25) RETURN ENDIF IF(BT .25 * (BT-0.EQ. 0.25.50) RETURN ENDIF IF(BT .LCASE.SCFT) CALL SSHANK(RT.25.EQ.ZT1.0.ZT2) CALL CSHANK(RT.

1679. . -.1503. -2.9450. (BT-0.5684. -2.8632. 1. .8610.2539.0317.0518.6981.25) C (4) WEDGE LOADING IN HOLE C C RANGE OF PARAMETERS: -0. .9801.5) * (1.AND.1527. .5) * BT1/BT ZT2 = BT2-0. 4 2. -3.5 < z/ t < 0.7469.1875. -.6614/ -.2088. LCASE DATA ALP/ 1 3.2063. 9 3. 1. . 2 .2427.0075.2831.0148. -1.5) THEN ZT1 = BT1-0.8331.ZT.0000. 8 -1. .LE.-BT1)/(1.1644. 4. 5 -2.3626.ZT. -. . ZT .7130. -. 3. -.ENDIF END SUBROUTINE NZT(BT.0395. . DATA ALPB/ 1 3.5 + (ZT-BT+0.GT.5 + (ZT-BT+0. C -2.BT2.5) .4169. 4 -1. .5.1825.5) * BT2/BT ENDIF RETURN END SUBROUTINE SSHANK(RT. 4.5 + (ZT-BT+0. -.1773.5555. 2 -. 0.1356. 2. 4.8337.9983.-BT2)/(1. 5. -.2924.BT 1.0242.7202.4239. .5125. -3. 3 .25 < r/t < 2.ZT1.5 C C---------------------------------------------------------------------DIMENSION ALP(4.4847.1651. 3 -1.0040. ALPB(4. -. -1. -2.0712.ZT2) C C C EVALUATE APPROPRIATE Z-LOCATION FOR COUNTER-SUN K HOLE IF(ZT . 2.5.4096.8804. 6 1. -1.3957/ 12 .5357.3673.7242.1650.SCF) C--------------------------------------------------------------------C C THREE-DIMENSIONAL STRESS-CONCENTRATION EQUATION FOR STRAIGHT C SHANK RIVET HOLE SUBJECTED TO: C (1) REMOTE TENSION C (2) REMOTE BENDING C (3) PIN LOADING IN HOLE (r/ w < 0. 0. -6.2).5767.3370. .5 + (ZT-BT+0. -1.4) CHARACTER *1.2778. . 7 .0206.1390.LCASE. -6.1870.5) * (1. .-BT) ELSE ZT1 = BT1-0.-BT) ZT2 = BT2-0.

4. 'w') L = 2 LCASE . L .0 Z2T = 2 * ZT IF(L. 5 J1 = (J-1)*2 SCF = SCF + ALP(I.L) * RT**I1 ELSE DO 10 J = 1.J) * RT**I1 * Z2T**J1 20 CONTINUE ENDIF RETURN END SUBROUTINE CSHANK(RT.5.OR. 3) STOP SCF = 0.5 < z/t < 0.ZT.1 SCF = SCF + ALPB(I.EQ.EQ.BT.25. 't') L = 1 LCASE .1 DO 20 J = 1. LCASE 13 .EQ.LCASE. 4 I1 = I-1 IF (Z2T .L) * RT**I1 * Z2T**J1 10 CONTINUE ENDIF 11 CONTINUE ELSE DO 20 I = 1. C RESULTS FOR ANY OTHER b/t VALUE ARE COMPUTED BY LINEAR C INTERPOLATION BETWEEN THE TWO LIMITING b/t VALUES.GT. BET(3.EQ. L .EQ.5.OR.5. . 0.25 < r/t < 2. . 'b') L = 3 3) PRI NT *.GT.2) CHARACTER *1.OR.LE. 'B' 1 . .75).50 AND 0.2) THEN DO 11 I = 1.LT. 4 I1 = I .OR. 1 .LT. 4 J1 = 2*J . .'LOAD TYPE NOT DEFINED' IF(L . LCASE . 'T' .0) THEN SCF = SCF + ALP(I. 0.EQ.2). 'W' .SCF) C----------------------------------------------------------------C C THREE-DIMENSIONAL STRESS CONCENTRATION FACTOR FOR COUNTER-SUNK C RIVET HOLE SUBJECTED TO: C (1) REMOTE TENSION C (2) REMOTE BENDING C SOLUTION IS FOR THE COUNTER-SUNK ANGLE OF 100 DEGREES AND A C SELECTED VALUE OF b/t (RATIOS ARE 0.4. .EQ.IF (LCASE IF (LCASE IF (LCASE IF(L . . C C RANGE OF PARAMETERS: -0.J.OR.5 C C----------------------------------------------------------------C DIMENSION ALP(3.1.

-2.6817. -2.7503.6036.9384.9311.2163. 1.6204. -. -2. . C 6. 1. C -2.4904.3049. -.8632. -3. C -5.4454.1330. C 6. -2.0493.8987. -.6774. 2.4494. 10. . -2. .2723.0127. 2.1675.2240. -2. L .8507. . -.-30.-15. -7. .6559.4740. -. 3.1933. 9. 1. -4. 2.-12.6186. -.GT. -. -1.LT.2839.2021. 3.8911. 22.3341.7056.8948.4221. -3.5507.3341.0844.8632. 4.7146.OR. 2) PRINT *.1981.8402.4221.4878. C 3.4568.6866.7327. -. C 23. 2. -4. 3.4071. -.6364. -.2565.3649.9136. C -. -2. . C -3. .0883.0428.9018.7330.EQ. 8.0611. 1.9691. .7382.7803. 1. 3. -7. -1.3481. 3.8236.-30.5229.1620. 1.4773. .4453. 3. .7708. 3. C .1557. C -1.2419.3670.1089.3460. LCASE .9993. 2.8492.1806.7327. -3.4052.3866/ C C *** BEGIN ANALYSIS C IF (LCASE . -.9081. . . 7. C -1.9691. C .0363.0720. 2. . .7198. -2.0574. 9.1049.3757. -.8286.2863.2472. -1. -1.OR.5507.9708.6470.1185. 1 . 't') L = 1 IF (LCASE .1064. -2.6870. 2.2145.-10. 4. 6. 3. 9.8507.9445.5264.4746.0077. C 23.0053.0777.0711.DATA ALP/ C 3. -1.3538.7805.3731.4071. 5. . -4.3438. C -2.2715.8774.1888. 2) STOP 14 . 10.6093.1386. 3.7184. . C -2. -4. 3.0843. -4. 1.5102. -2.3737. -2. C 3.1620.1485.1933.0259.7056.6002.2565. -.-19.3213.8222.0428.8738. -. 1. . -8.1465. -4.OR.0730. C 1. LCASE .1424.2035.-11.0730.-12. C -10.7731. 3. 3.4835.0777. 2562. 22.4482.4629. 2.5510. 2.2232.7898. C . 6.1234. -. . -8. .0259. C -2. -1.1888.4322. -6. C 1. 2. 1. DATA BET/ C 3.1465.5303.6958.0883. -1. -.7985. -.1287.8939. 1. 3.EQ. 5. -. -.6093.4920. -8. -1.1604.-13.8201. 4.5036. 12. -3.0031.2356.2768.1237.8862. 3. C .7192.7192.4322.5453.4773.6911.3670.1424. 2. . -3. C -10. -.4036.4706.1992. . -1. -6.4454. -. 'T' . 1. . 2. 1.6655.OR. . -2. C . 1. 1.6036.0805. C . -.7898.2713. C -.4422.LT. -6.6784.1384.2380.9410. -9. 1. -.8887.0229. 8.1981. 1.1675.4058. -. 14.2150. -1. 3.2552.5498.EQ. 6. 15.2623. C -5.8097. -.5698. -1. .4144. 1 . 3. -1.7503.0135. .GT. 8.-11. .2128. 15.2616. -.0135.8037.4835.1888. 3. . . C .5453. -.1486.5768. C .7146.9384. -.7525. -. 5. -. C -9.7420. -. .3056.4552/ 1.-15.6850.8885.2768. -1.1935.1485. 5. -2. . C 8.-10.5880.5784.6156.2839. -.0875. C -1. -.EQ.7198.4052. . 2.8178.2021. -.5876.1969. 13.3109. . 'b') L = 2 IF(L .1935. 2. 'B' . 3.7318.2356. . C -4.'LOAD TYPE NOT DEFINED' IF(L .2562.2232.2713.1265.7020.7020. C -4. . . -7.8948.5101. L . -1.

1 SCF = SCF + ALP(I. 0.25) K = 2 IF(BT . 0. 0.L) * RT**I1 * Z**J1 CONTINUE ENDIF CONTINUE ELSE T2 = CCOR/(1.-BT) Z = T2 + ZT/(1. 'W') PRINT *. 0. 5 J1 = J .1 SCF = SCF + BET(I.L) * RT**I1 ELSE DO 21 J = 1.1. ZT .0) THEN SCF = SCF + BET(I. 0. 'W') STOP IF(BT .-BT) DO 30 I = 1. 3 I1 = I-1 IF(Z .0) THEN SCF = SCF + BET(I.0) K = 1 IF(BT . SCF = 0.EQ. . 3 I1 = I-1 IF( Z .IF(LCASE .0) THEN SCF = SCF + ALP(I. 0.EQ. ZT .75) K = 4 CCOR = (1.-BT) Z = T2 + ZT/(1.0) THEN IF (-0.K.NE.5 .EQ.L) * RT**I1 ELSE 10 11 21 20 15 .K.EQ.1. -CCOR) THEN T1 = CCOR/BT Z = T1 + ZT/BT DO 11 I = 1.K. 0. 3 I1 = I-1 IF(Z .L) * RT**I1 ELSE DO 10 J = 1.'WEDGE LOAD SOLUTION NOT AVAILABLE' IF(LCASE .J.LE.50) K = 3 IF(BT .LE. 5 J1 = J .1.EQ.EQ.-BT) DO 20 I = 1.EQ.L) * RT**I1 * Z**J1 CONTINUE ENDIF CONTINUE ENDIF ELSE T2 = CCOR/(1.EQ.K.AND.K.0 IF(BT . 0.EQ.J.2*BT)/2.

K. L) * RT**I1 * Z**J1 CONTINUE ENDIF CONTINUE ENDIF RETURN END 16 . 5 J1 = J .31 30 DO 31 J = 1.1 SCF = SCF + BET(I.J.

pp. J . A. 7. 26.References 1. N. Drucker.. May 1974... J . Dattaguru. C. 69{80. J. A. Heft 3. D.: St ress Concentration Around an Open Circu lar Hole in a Plate Subject ed t o Bending Nor mal to the Plane of t he Plat e. Rep. Pergamon Press. Fourth ed. Jr. vol. NASA TP-1862.: Treatment of Singularities in Cracked Bodies. 1965. Sept. 1961. Dep. 929{932. Neuber.. no. N.: Stress-Concentration Factors for Countersunk Holes. June 1943. John H. vol. NACA TN 740. and Raju. Mech . Businger. 16. J . Mar. Virginia Polytechnic Inst. Mag. Mat er. A. Royal Soc. Gene H. Savin. 379{391. J. of Science. pp. ser. 1955.: An Extension of the Photoelastic Method of Stress Measurement to Plates in 17 . 8. no. A. pp. 2. 3. J. and Springer. George S. ser. no. P. 1. Peterson. 22. Eshwar. 45. Mar. J. Trans. 240. vol. pp. 21. 269{276. Aerona ut. Dec. Bd. 39{61. 257{261. 11. Bickley. 1928.. Nov. G.. 27. Mech.: The Distribution of Stress Round a Circular Hole in a Plate. 1944. 6. E. Theo: Stresses Around Pin-Loaded Holes in Elastically Orthotropic or Isotropic Plates. A-69{A-77. Dec. 23. 12. 383{415. no. Whaley.: The Photoelastic Analysis of Transverse Bending of Plates in the Standard Transmission Polariscope.. July 1936.: Stress-ConcentrationFactors for a Countersunk Hole in a Flat Bar in Tension and Transverse Bending.. Aug. and Rao. no. Ph.. Mater. 11. I. Hong. vol.: Three-Dimensional Solution for the Stress Concentration Around a Circular Hole in a Plate of Arbitrary Thickness. C. A. J. 470{494. Mar.. vol. 20. E. vol. Dec. Exp.. Oct. H.. 7. N. 6. Peter. Compu t. 89{94. 8.: A Treat ise on the Mathematical Theory of Elasticity.: A Review of the Three-Dimensional Stress Problem for a Cracked Plate. A. 227. no. 1971. 16. pp. Chang.D. ser. W. Heinz: T heory of Not ch Stresses: Principles for Exact St ress Calcu lat ion. 15. June 1945. c.. C. S.. 3. Aug. vol. Philos . 10. pp. 1941. Nu mer .. pp. E. B. Fra ct.: The In uence of Circular and Elliptical Holes on the Transverse Flexure of Elastic Plates. 45.: With a Pin-Loaded Hole.: Three-Dimensional Stress Systems in Isotropic Plates. and Golub. vol. A. 1948. Sternberg. vol. no. Crews. pp. 4. 19. no. no. 14. 561{597. 25.1974. Green... C. 145. and State Univ. Cheng. Inc. W. 1. J . G. London. Sih. R. vol. 9. N. StressConcentration F actor s for Fi nit e Ort hotropic Laminates Michigan). 2. 1981. A. Int.: Strength of Mechanically Fastened Composite Joints. and Raju. vol. Trans.. G. vol.: Photoelastic Analysis of a Thick Plate With an Elliptical Hole Subjected to Simple Out-of-Plane Bending.. 1. 1942. Richard E. S. 1. 1946. Richard A.. St ress Concent ration Factor s|Chart s and Relations U sefu l in Making Strength C alculations for Machine Parts and St ructu ral Element s.. K. vol.-J. Dover Publ. Mech. 1939. Ro yal Soc. Jr. Goland. and Lee. Compos. G. Mech. no. no. J . Appl. 24. 1. 287{303. no.: Linear Least Squares Solutions by Householder Transformations. Int . J. I. 4. pp. 159{178.. no. No. V.. Eric: The E ect of Transverse Shear Deformation on the Bending of Elastic Plates.. pp. J. John Wiley & 2. N. and Sosropartono. 13.: On the Three-Dimensional Stress Field Around a Circular Hole in a Plate of Arbitrary Thickness. 10. J . Love. J. pp. pp. W. Rubayi. 9. Mech..Indian Inst. E. H. Edwards (Ann Arbor. Mech. Mech. F.. 6. A ppl. Folias. I. Naghdi. Mech. A-69{A-75. 4. 7. Mech . ARDBSTR-5010. Thesis. Philo s. Mech. S.. Jr. C. 1978. 1978. M. 22. 17. 18. 16.. Martin: The In uence of the Shape and Rigidity of an Elastic Inclusion on the Transverse Flexure of Thin Plates. Transverse Bending. no. 1990. J. and Wang. 3. J. July 1977. (Eugene Gros. Newman.. vol. Shivakumar. Goodier. 7.. M. A-27{A-29. 5. 4. E. Goodier.. 12. Fu-Kuo. Math . transl. pp.: A Survey of Strength Analysis Methods for Laminates With Holes. 9. 8.: Finite-Element Analysis of Fatigue Crack Propagation|Including the E ects of Crack Closure. Exp. 36. J. DeJong.. pp. Mech. 1984. 1982. pp. of AeronauticalEngineering. 22. Soc.: Sons. A ppl. 1990. and Sadowsky. 1977. Appl. 313{331. 1965.. A-161{A-164. 5.. John H. Appl. Reissner. vol. pp. 1949. pp. Dumont. pp. Nov. Inc.: Partial Contact and F r ict ion in Pin Joi nts . 18. Mar. Crews. A ppl. Scott. vol. vol. no. Com pos. 27{38. 335{343.: The E ect of Elliptic Holes on the Bending of Thick Plates. Fract .): Stress C oncentration Aro und H oles. India . Appl. Y. Inc. G. pp. vol. London. S. K.

including sugg estions for r educing this burden. Send comments r egar ding this burden estima te or any o ther a spect of this co llectio n of inf ormatio n. Newman: Langley Research Center. SUPPLEMENTARY N OTES Shivakumar: Analytical Services & Materials. Paperwork Reductio n Pr oject (07 04 -0 18 8). AUTHO R( S) K. 12 a. SPON SO RING/ MO NITORING AGEN CY REPO RT NUMBER NASA TP-3192 11 . including the time for reviewing instr uctions. Shivakumar and J. Wa shing to n. Suite 12 04 . 1 21 5 J e erso n Dav is Highway. DC 2 05 03 . AGENCY USE ONLY(Leave b lank) 2. VA 2 22 02 -4 30 2. VA. g ather ing a nd maintaining the data needed.REPOR T D OCU MENTATIO N PAGE Form Approved OMB No. FUND IN G N UMBERS WU 505-63-50-04 6. 8 . DISTRIBUTIO N/AVAILABILITY STATEMENT 1 2b . 1. Hampton. DC 20546-0001 1 0. Newman. Bending. Arlington. to Washington Headquar ter s Se rvices.. REPORT TYPE AND D ATES COVERED Technical Paper 4. DISTRIBUTIO N CO DE Unclassi. C. 0704-0188 P ublic r eporting burden for this co llectio n of informatio n is estimate d to a vera ge 1 hour per response. Hampton. TITLE AN D SUBTITLE Stress Concentrations for Straight-Shank and Countersunk Holes in Plates Subjected to Tension. PERFO RMING ORGAN IZ ATIO N NAME(S) AND ADD RESS(ES) NASA Langley Research Center Hampton. a nd co mpleting and reviewing the co llectio n o f inf ormatio n. Dir ectora te fo r Information Operations a nd Repo rts. N. Inc. sear ching ex isting data source s. VA. and Pin Loading 5 . PERFORMIN G ORGANIZATION REPORT NUMBER 7. Jr. a nd to the Oce o f Ma na gement a nd Budg et. SPON SO RING/ MO NITORING AGENCY N AME(S) AND AD DRESS(ES) National Aeronautics and Space Administration Washington. REPOR T DATE June 1992 3. VA 23665-5225 L-17027 9.

ABSTRACT (Maximum 200 words) A three-dimensional stress-concentration analysis was conducted on straight-shank and countersunk (rivet) holes in a large plate subjected to various loading conditions.ed{Unlimited Subject Category 39 13 . Three-dimensional .

For straight-shank holes.3.5) and ratios of countersink depth to plate thickness (0. three types of loading were considered: remote tension. with a Poisson's ratio of 0. Based on the . but the angle was held constant at 100  for most cases. Results for remote tension and wedge loading were used to estimate stress concentrationsfor simulated rivet or pin loading. and wedge loading in the hole. only remote tension and bending were considered.25 to 1).1 to 2.nite-element analyses were performed with 20-node isoparametric elements. The countersink angle was varied from 80 to 100 in some typical cases. For countersunk holes. remote bending. The plate material was assumed to be linear elastic and isotropic. Stress concentrationsalong the bore of the hole were computed for several ratios of hole radius to plate thickness (0.

Countersink angle. 3-D stress-concentration analysis. Finite-element analysis. PRICE CODE A03 17 . LIMITATION Unclassi. Whenever possible. SUBJECT TERMS Countersunk holes. 14 .nite-element results. SECU RITY CLASSIFICATION OF REPO RT 18. stress-concentrationequations were developed. SECURITY CLASSIFICATION 19 . the present results were compared with other numerical solutions and experimental results from the literature. SECURITY CLASSIFICATION 2 0. Isoparametric elements OF THIS PAGE 1 5. N UMBER O F PAGES 34 1 6.

ed Unclassi.

Z3 9-1 8 2 98 -1 02 .ed O F ABSTRACT OF ABSTRACT NSN 754 0-01 -280 -550 0 Sta ndard Fo rm 2 98 (Rev. 2-8 9) Pre scr ibed by ANSI Std.