SP 41-8 Stress Concentrations Caused By Reinforcing Bar Deformations By 1.C. Jhamb and J.G. Mac Gregor Synopsis: Because research on metal fatigue has shown that fatigue failures tend to occur in regions of high stress concentrations, finite element stress analyses were used to determine stress concentration factow (K,) for deformed reinforcing bars. The influences of various geometrical paraneters of the bar deformations on K, were investigated The test findings suggest that the ratio of the lug base radius (r) to the lug height (h) was the most pronounced effect on K,. The value of Ky decreased with an increase in r/h value. For r/h ratios greater than 1,25, no further appreciable reduction in the K, was found. The analysed results were checked by strain measurements. These measure- ments also suggested that the stress concentrations at the intersec- tions of the longitudinal and transverse lugs were not as critical as Previously suggested. Fatigue tests of reinforcing bars have shown a reduction in fatigue strength for r/n ratios less than 1.25. Keywonts: bere; deformations (retnforcing steels); deformed reln- forcenent; fatigue (Materials); fetime tests; Finite elenent method; reinforcing steels; research; stress analysis; stress-strain rela- tionships, 169170 fatigue of concrete ‘AGI member Istwar C. Jhanb is a Materials Engineer with Grace Construc— tion Materials Ltd., Edmonton, Canada. Me cbtained a B.E. degree from the University of Jodhpur, India, An M.E. (Honors) degree from the University of Roorkee, India and a Ph.D. degree from the University of Alberta, Eduonton, Canada. Currently, he is engaged in the develop~ ment and marketing of construction materials. ACT member James G. MacGregor is a Professor of Civil Engineering, University of Alberta, Edmonton, Canada. In 1973 Dr. MacGregor was co-recipient of the Wason and Reese medals. Currently, he 48 Chairman of ACI-ASCE Conmittee 426, Shear and Diagonal Tension, and a mexber of the ACT Board of Direction. INTRODUCTION Research on metal fatigue has shown that fatigue failures tend to occur in regions of Hgh stress concentrations. The severity of a stress concentration or "notch" measured by the stress concentration factor (K,), which 4s defined as the ratio of the maximum elastic stress in the region of the notch to the nominal stress in the member. A typical hot rolled deformed reinforcing bar contains two longi tudinal ribs and a regular pattern of equally spaced transverse lugs. The ribs and lugs cause stress concentrations in the reinforcing bar. Although technical Literature (1,2) contains K, values for numerous types of notches, none of the published cases truly represent the deformations on reinforcing bars, For this reason an investigation was undertaken to determine K, for the deformations. A theoretical stress analysis was performed for circular bars having axi-synmetrical projecting lugs. The K, values for a variety of projecting lug shapes similar to deformations in reinforcing bars were determined. To check these results for one type of bar, electric resistance strain gages were employed to determine the locations of high stress concentration and the maximum etresses for one type of deformed reinforcing bars. DETERMINATION OF K, USING FINITE ELEMENT STRESS ANALYSIS A series of finite element analyses were used to study the stress concentrations in circular bars with protruding notches resembling reinforcing bar deformations. These were carried ait using a computer program entitled "Finite Element Analysis of Solids with Non-Linear Material Properties" (3). Because the program was limited to axi~ symetric solids or planar structures, it vas not possible to include the longitudinal deformations or to consider a lug inclination other than perpendicular to the bar axis. A section through a diameter of the simplified model used to compute K, for the bar deformations is shown in Figure 1.stress concentrations = 171 In the analysis of the projecting lugs the material was assumed to be elastic, homogeneous and isotropic. A uniform axi-synmetrical tensile loading was applied on the face c-d of the idealised model shown in Figure 1, The boundary conditions on the face a-b were assumed symmetrical, Radial deformations were not permitted along the center line c-b. A typical finite element mesh {8 shown in Figure 2. This model consists of 150 elenents. Similar models with 100 and 200 elements resulted in stress concentration factors of 88 and 100.9 percent, respectively, of those obtained using 150 elements, and 150 elements were. felt to give a reasonable estimate of K,. Sim{larly, models of 3 consecutive ribs led to K, values of about 99.7 percent of those for the one rib models and it was concluded that the K, values obtained for a single lug were representative of a de- formed bar consisting of many equally spaced lugs. The strain measure— ments presented later in this paper suggest that a bar with 70° lugs and a longitudinal rib could be represented by axi-synmetrical lugs having the same lug base radius in the plane of the bar axis. Referring to Figure 1, it may be wen that the following parameters adequately define the geometry of the deformation lugs: r= lug base radius (radius of the circular transition curve) h = height of lug, measured from the base of lug away from transition curve w= one-half width of the lug, measured from centerline of lug to the intersection of the transition curve tangents © = lug face angle D = bar diameter The selection of the ranges for these parameters was based pri- marily on the comparable ranges specified in ASTM A615~68 except, that a wider range of radif at the base of lugs vas analysed than ex- pected in North American reinforcing bars. For each set of these five parameters defining a lug it was possible to develop a finite element model. In this model attention was paid to accurately sinulating the geometry close to the lug base. Since the smooth circular transition curve at the lug base was sinulated with straight lines, these straight lines were kept very small. On the average the triangles close to the surface defining the lug were .004 times the diameter of the bar. The stresses in these triangular elements were calculated at a distance equal to .0013 times the diameter below the surface. From the computer output of each finite element model the stress state in each element was available. Figure 3 presents this data in the form of contours of equal ratio of calculated stress to nominal172 fatigue of concrete stress. The contours of high value are concentrated very close to the start of the curvature at the base of the lug. The type of stress state shown in Figure 3 is similar to that observed in photo-elastic models. The maximum value of this ratio is the stress concentration factor (K,) for the notch. Initially, the effect on Ky of the radius at the lug base (r) and the bar diameter (d), were investigated for constant values of lug height (h), Lug width (28) and lug face angle (@). A total of ten finite element models were analysed for this purpose. The value of Ky, obtained in each case ts reported in columns (5) and (8) in Table 1. It may be noted that the values of Ky are almost the same for the two bar diameters. From these results it was concluded that for the practical range of diameters and lug parameters, the K, of the lug is independent of bar diameter. Although this conclusion may not hold for cases outside the ranges considered, such cases would be beyond the current ASTM Standard (A615~68) for deformations. In the subsequent analysis, it was assumed that the results for h/D equal to 0.05 would be applicable to all practical h/D ratios. Table la and Figure 4 sunnarize the stress concentration factors for all the cases studied. EXPERIMENTAL STUDY OF Ky To check the validity of the finite element analyses strain concentration measurements were made on @ 30 inch (76 cm) specimen of No. 8 (25 mn) reinforcing bar using electrical resistance -strain gages, The gage locations are shown in Figure 5. Gages 1 and 2 were of 1/4 inch (6 mm) gage lengths and were mounted on diametrically opposite sides of the bar midway between two transverse lugs at a point adjacent to the maufacturer's markings, where the lugs were further apart than normally. The rest of the gages wereof 0.3 mm (0.01 in.) gage length and were located at possible zones of high stress concentrations. Stresses were calculated assuming the modulus of elasticity of steel to be 29 x 10° psi (2-04 x 10° kgf/en?). Each stress concentra~ tion value in Figure 5 is the average of 15 readings. The ratio of measured stress to the nominal stress corresponding to each strain gage is shown in a box adjacent to the gage in Figure 5. Gages 1 and 2, due to their location and longer gage lengths than the rest of the gages, were intended to show the nominal stress in the bar. The center of these gages was 10 mm from the nearest lug. The measured stresses in these two gages were within 2 percent of the nominal stress. Gages 4 and 9 were mounted parallel to the bar axis and perpen- Aiculor to the transverse lug, respectively. The centerline of these gages vas located at 0.15 mm from the transverse lug. Gage 4 always showed 142 higher stresses than gage 9. Gages 10 and 11 were mounted in an identical fashion to gages 4 and 9, except that gages 10 and 11stress concentrations ©9173 were 0.7 mm away from the lug base. Gage 10 measured 5% higher stress~ es than gage 11. From these observations, it may be noted that though the radius at the base of the rib 1s smaller in the plane perpendicular to the transverse rib than in the plane of the bar axis, lower stresses were observed in the former case. Higher stresses were obtained for gages parallel to the bar axis because the plane of load application and that of stress measurement was the same. A Mohr's Circle for strain would suggest that the strains in gages 4 and 10 should be 9 percent higher than those in gages 9 and 11 due to the angle between the latter gages and the axis of the bar. ALL the gages located near the intersection of the transverse and longitudinal ribs showed measured stresses less than 1.10 times the nominal stress. This suggeststhat for these bars the junction of the longitudinal and transverse ribs was not critical. In fatigue tests of the sane bars 81 percent of the failures initiated at the base of the lugs avay from the intersection of the longitudinal and transverse lugs. The stress concentration measurements at gages 12, 13, 4, 14, 10 and 1 are plotted in Figure 6. These gages were located at 2.5, 1. 0.15, 0.3, 0.7 and 10 mm (0.1, 0.4, 0.006, 0.01, 0.027 and 0.39 4n) respectively, from the case of a transverse lug in the direction parallel to the bar axis. In Figute 6 the theoretical results obtained form the analysis of an axi-symmetrical projecting lug having identical dimensions to the deformation lugs of the bar specimen used for stress concentration measurements are also plotted for comparison. The horizontal scale has been non-dimensionalized. It may be seen in this figure that the experimental results He very close to the theoretical curve. It was not possible to place a gage at the point of maximum stress concentration, however. In order to find whether the stresses were higher on the steeper side of the transverse lug than on the other side, gages 10 and 3 were located 0.7 um (0.027 in.) away from the lug base on the steeper and the other sides respectively. The stresses in the steeper side of the lug were found 3% higher than the other side. From the experimental investigation of stress concentrations in this deformed bar, the foloowing conclusions may be derived: (1) The stresses at the base of the lugs were 142 higher in the direction of the bar than at the same location but perpendicular to the transverse lug. (2) Stresses at the intersection of the longitudinal and transverse ribs vere not as critical as those at the bases of transverse lugs for the bars tested.174 fatigue of concrete EFFECT OF THE GEOMETRICAL PARAMETERS OF DEFORMATIONS ON K, The stress concentration factors (K,) for projecting lugs reported /in Table 1 are shown in Figure 4. In this figure r, w, h and D are grouped into three non-dimensional variable: The stress concentration factors are plotted on a linear ordinate scale and the r/h ratios are plotted on a logarithmic absci: + The plots of K, versus r/h are shown for two values of w/h and three values of lug face angle (0) in Figure 4, These are assumed to be independent of the h/D ratios in the practical range of values. The following conclusions may be made regarding the effect of projecting lug parameter on stress concentration factors: qa) (2) @) @) (3) The Ky of the projecting lugs increases with s decrease in the lug base radius. This conclusion is in agreement with the theory of notch stresses and las been suggested in other studies of reinforcing bars (4,5). For all the projecting lugs investigated, the Kp vs r/h plot could be approximated by a family of straight lines on a semtlogarithnatic plot. For r/h ratios greater than 1.25, there seems to be no appreciable further reduction in the stress concentration factors. The stress concentration factor increased with an increase in the w/h ratio. The stress concentration factor increased with an increase in the lug face angle (©). This increase in K, is not as significant as those due to r/h and w/h. The stress concentration factor was independent of the h/D ratio in the range studied. The reduction in the stress concentration factors due to multiple lugs was insignificant in the cases studies and could be ignored. ‘These five conclusions on stress concentration factors for pro- Jecting lugs and the two in the previous section could be used as a guideline for designing the deformations on reinforcing bars. The first major attempt in proportioning the deformation pattern parameters should be to keep the r/h ratio equal to or greater than 1.25. Second, it wo 1d be desirable to keep w/h closer to 1 than 2. The lug face inclination will not have much significance if r/h and w/h are kept to the desired ratios. However, lug face angles (0) less than 45° are probably not desirable from the standpoint of bond splitting. Figure 7 examines the influence of r/h ratios on fatigue strength of bars. deformed Data from several fatigue projects (4 to 7) on hot rolled bars 1s used in thie figure. To put this data on a comparablestress concentrations 175 basis, all the fatigue strengths were corrected to correspond to 0.1 £, minimum stress level using Modified Goodman Diagrams. The test data in Fig. 8 indicated that the fatigue strength of the deformed bars decreases when the ratio of radius to lug height is less than about 1.25 and is almost constant for r/h ratios greater than 1.25. These results tend to confirm stress concentration calculations. The average r/h ratio for North American bare is in the order of 0.25. According to the line plotted in Figure 8 the fatigue strength of reinforcing bars could be increased by about 30 percent by increasing the r/h ratio for the deformations to 1.25 or greater. The above observations suggest that by adding additional clauses specifying a minimum lug base radius to lug height (r/h) equal to 1.25 and a maximum width to height ratio for lugs, a specification for higher fatigue strength deformed bars could be developed. ACKNOWLEDGEMENTS. The assistance of Professor D.W. Murray in selecting the computer model 1s acknowledged with thanks. Financial assistance was obtained from the National Research Council of Canada, REFERENCES 1. Peterson, R-E., Stress Concentration Design Factors, John Wiley ‘and Sons, New York, 1956. 2. Derecho, A.T., Munse, W.H., "Stress Concentration at External Notches in Members Subjected to Axial Bending", Engineering Experimental Station Bulletin No. 494, University of Illinois, Urbana, Illinois, 1968. 3. Wilson, E.L., "A Digital Computer Program for the Finite’ Element Analysis of Solids with Non-Linear Material Properties” University of California, Berkeley, 1965. 4, Hanson, J.M., Burton, K.T. and Hognestad, E., "Fatigue Tests of Reinforcing Bars - Effect of Deformation Pattern", Journal ‘of the PCA Research and Development Laboratories, Vol. 10, No. 3, September 1968, pp. 2-13, PCA Development Bulletin D145. 5. Kokubu, M., and Okamura, H., "Fatigue Behavior of High Strength Deforned Bars in Reinforced Concrete Bridges", First Inter- national Symposium-Concrete Bridge Design, ACI Publication SP-23, American Concrete Institute, Detroit, 1969. pp.301-316. 6. MacGregor, J.G., Jhamb, I.C. and Nuttall, N., "Fatigue Strength of Hot Rolled Deformed Reinforcing Bars", ACI Journal, Proceed- ings Vol. 68, No. 3, March 1971, pp. 169-179. 7, Sham, I.C. nd MacGregor, J.G., "Effect of Surface Characteristics on Fatigue Strength of Reinforcing Bars", ACI Symposium on Fatigue, 1973.fatigue of concrete 176 dest | ezort — sosrt zest | eset tenet ovy-t | ost vert | otc — ga9t gegrt | ces*t tenet gyrt | 00°T eet | ozorz = Lo" neG T | tLtT ~— OBa"T gso°t | oso zeerz | core = vearz sezz | vost t98*T ose't | sz-o €esz | y9ore — g9stz uestz | tet'z S60" 02 | 0 SP=0 | 09=0 S%S=0 Sy=Olo9=OS%S=0 Sy=0 @) wm om (~ | @ 2) @ o z= ya z= ua T= wa TOs Pt 500 = TT 8m Suy3o0fozg 103 T aTavsstress concentrations = 177 FIGURE 1 AX! - SYMMETRICAL PROJECTING LUG MODEL eae (a) STANDARD MODEL FIGURE 2 TYPICAL FINITE-ELEMENT MODELfatigue of concrete 178 ONT ONILIIFONd JO ISWE FHL LY SNOILVYLNIINOD SSIYLS 40 dvWYNOLNOD € INNDI4 31v2S eH ee : ost ty 009=8 : ot 4 G31NdWOD $a3ss3yls 4 3a3HM SINIOd 9 St0r> $ivi3d 9N1 ONILD3F08d BYNLINALS SSIYS INVIdstress concentrations 179 <1 ah D ec i+ - D ae h cla FOR RANGE 0.05< 03 04 05 02 FIGURE 4 EFFECT OF PROJECTING LUG PARAMETERS ON K « a 2 (44) YOLWS NOILVYLN3DNOD SS3¥LS180 fatigue UVd G3WYOIIG NO SLNIWIYNSVIW NOLLVYINIONOD SSIMIS $ INNDI4181 stress concentrations ONT ONILIIOYd JO ISVS AV INFIGVYD SSTULS 9 FUNIIS LHOIZH 9N1/3SV8 ON1 WOYI JDNVISIG << -—- ot ON1 ONILD3fO8d os so BAYND WIW3IOIHL SLINSaY IONVS NIVUIS W Sllv13d 9N1 ONILDIFOYd NOILVYLNZINOD SS3aLls182 fatigue of concrete (242/54) JONVY SS3uls ooo 0002 000e 4 vo = 4s ‘wonauis 3nolLvs NO 4/4 JO 1994592 JUNI 4HOI3H_9N7 sniava 3sv@ s z o1 sO zo “W613 NOSNVH © VaNWyAO GNV NENIOX V wigadiy Oo ° ° Ss ° o ‘JONVYE SS3uLs