20 FOUNDATION ANALYSIS AND DESION ‘The Terzaghi Bearing-Capacity Equation One of the early sets of bearing-capacity equations was proposed by Terzaghi (1943) as shown in Table 4-1. These equations are similar to Eq. (k) derived in the previous section, but Terzaghi used shape factors noted when the limitations of the equation were discussed. Terzaghi’s equations were produced from a slightly modified bearing-capacity theory devel- TABLE 41 Bearing-capacity equations by the several authors indicated ‘Terzaghi (1943), See Table 42 for typical values and for Ky, values TearEst an = ONeSe + Ny + OSYBN Sy, Ny = (N= Neots For: strip ound square s=10 13 13 5510 06 08 Meyethot (1963)* See Table 43 for shape, depth, and inclination factors. Vertical load: gay = eNeSde + IMS + OSYB'Ny Syd, Tncined oad: guy = CNdeie + TNydig + 05YB'Nydyiy ra tan? (45 4 2 ™ (+9) Ne = (Ny Deotd Percent Hansen (1970).* See Table 45 for shape, depth, and other factors, Generlst guy = eNeSedelesebe + INeSeig feb + OSYBNy dha wen ono se an = SMMC 4s ds E-bay ame as Meyerhaf above same as Meyethof above Drang Vesié (1973, 1975).* See Table 4-5 for shape, depth, and other factors Use Hansen's equations above 1 = same as Meyerhof above N Ny same as Meyethof above iN + Ian “These mains ee wl proves wo obi design bat dimensions since width and length Lar need compute shape, depth, ad inftence fcr, 1800 Se. 46 when i <1[REARING CAPACITY OF FOUNDATIONS 221 = face or ahd” For Hansen, Meyerhot Y= kad ot Fahd’ Bo ai’ = log spiral for ¢ > 0 12 =H © Figure 43 _(@) Shallow foundation wih ough base defined. Terzaghi and Hansen equations of Table 4-1 neglect shear along ed; (b) general footing soil interaction for bearing-capacity equation for stip footing —Ieft sie for Terzaghi (1943), Hansen (1970), and right side Meyerho (1951), ‘oped by Prandtl (ca. 1920) from using the theory of plasticity to analyze the punching of a rigid base into a softer (soil) material. Similar to Eq. (4), the basic equation was for the case in which a unit width from a long strip produced a plane strain case, all shape factors 51 = 1.00, but the N; factors were computed differently. Terzaghi used a = ¢ in Figs. 4-2 and 4-3 whereas most other theories use the a = 45 + /2 shown. We see in Table 4-1 that ‘Terzaghi only used shape factors with the cohesion (5.) and base (s,) terms. The Terzaghi bearing-capacity equation is developed, as was Eq. (k), by summing vertical forces on the wedge bac of Fig. 4-3. The difference in N factors results from the assumption of the log spiral arc ad and exit wedge cde of Fig 4-3. This makes a very substantial difference in how P, is computed, which in turn gives the different N; values. The shear slip lines shown on Fig. 4-3 qualitatively illustrate stress trajectories in the plastic zone beneath the footing as the ultimate bearing pressure is developed ‘Terzaghi’s bearing-capacity equations were intended for “shallow” foundations where D=B222. FOUNDATION ANALYSIS AND DESIGN TABLE 42 TABLE 43 Bearing-capacity factors for the Shape, depth, and inclination factors for Terzaghi equations the Meyerhof bearing-capacity equations Vales of Ny foe @ of 0, 34, sad 48° are orginal of Table 4-1 “Teraaghi valves and used to back-compute Ky et ______— ————nrre te actors Value For dee Ne NK Shape se= 1+ 02K; ‘Any oe in L 5 33 te Osa seseteaug® 6510 0) ee no erie eral ici) L a ig 1208 kale 2s| in sonal oe 2» 1774 50250 ee) : 3 812797 sg Ha tsO2lhyy Ans x 20s 9752.0 oo pe eed a= 1401/2 6>10 3578 1424820 dyn dyat 0 495713000 5 4517237832975 298.0 Inclination -£ Any 42583287970 Reavy 5o 475 415.1 11532 _ 9000 a wy ES 7 i 3) o>0 Sho = 15m + 1 (Sex Teraph 9p 127 © wee a # or Where Ky = WES + #2) a ig 42 1 = angle of rsuan measured fom veri witout aSigai'@ = Oa = £0. 8.1, = previously defined so that the shear resistance along ed of Fig. 4-3a could be neglected. Table 4-1 lists the ‘Terzaghi equation and the method for computing the several N; factors and the two shape factors s). Table 4-2 is a short table of NV factors produced from a computer program and edited for illustration and for rapid use by the reader. Terzaghi never explained very well hhow he obtained the Kp used to compute the bearing-capacity factor N,. He did, however, sive a small-scale curve of ¢ versus V, and three specific values of Ny at @ = 0, 34, and 48° as shown on Table 4-2. The author took additional points from this curve and used a computer to back-compute Kp to obtain a table of best-fit values from which the tabulated values of Ny shown in Table 4-2 could be computed from the equation for Ny shown in Table 4-1. Inspection of Table 4-4 indicates that the Meyerhof Nw) Values are fairly close except for angles of ¢ > 40°. Other approximations for N include the following Ny = 2N, + Dtang Vesié (1973) Ny = 1.MNy~ tan 136 Spangler and Handy (1982) The Ny value has the widest suggested range of values of any of the bearing-capacity N factors. A literature search reveals 38= Ny, = 192 ford = 40" In this textbook values from Tables 4-2 and 4-4 give a range from about 79 to 109, Recently Kumbhojkar (1993) presented a series of values of NV with the claim they are better representations of the Terzaghi values than those of Table 4-2. An inspection of these[BEARING CAPACITY OF FOUNDATIONS 223 [Ny values shows the following: ‘Termght” Bolion and Lau Kumbhojkar Table #2 098) (1993) (1993) (this text) af 6 35 32 36 a 780 638 650.7 780.1 “Sex Tezahi (10), Fg 38 ad peat 128 Fortunately the Ny term does not make a significant contribution to the computed bearing ‘capacity, so any of the values from Tables 4-2 or 4-4 can be used (or pethaps an average). Bolton and Lau (1993) produced new Nj and N, values for strip and circular footings for both smooth and rough ground interfacings. Their N, values for either smooth or rough strips are little different from the Hansen values for rough strips. The N values for circular footings range to more than two times the strip values. The Ny values for rough footings ‘compare well with the Vesié values in Table 4-4. Since the Table 4-4 values have shape 5; and depth d; factors to be applied, it appears that these “new” values offer little advantage and are certainly more difficult to compute (see comparison with Terzaghi values in preceding table). ‘Meyerhof’s Bearing-Capacity Equation Meyerhof (1951, 1963) proposed a bearing-capacity equation similar to that of Terzaghi ‘but included a shape factor sy with the depth term N,. He also included depth factors d and ‘TABLE 44 Bearing-capacity factors for the Meyerhof, Hansen, and Vesié bearing- ‘capacity equations Note that Ne and Ny are the sume for thor foe Ny a ee ee Bian bit ~ sind oS 10 00D 8s ‘000 see eee og 0 ate aa oz 0.136 838 ss 20205, 2s ise e057) 0 ok ee ee Ts 0294 oe sda ae 20 29h sa, oe 0315 2 m7 07,68 GR 19ST o3it % nds is 79 02S 833 0308 % om 470912167 0239 5013S AST 187 2240610 0289 338472 wR 039204683 0276 362d 88TH 0282 365058377 400562785, 0247 pas eellstes alot | sei Gunh © 71) 10787 0231 % 7525 Gal 794981093 0RS? ons 48153731347 200 262373007 on 30 2665031853674 _—BTLT_— 7613195 oni Sas 2avlimicwnen = 0 Stig dillerences in above table can be obtained esng program BEARING EXE on dikete depending oncom ter wed and whether or a it hae Rosting pont,224 FOUNDATION ANALYSIS AND DESIGN inclination factors i; (both noted in discussion of Eq. ()] for cases where the footing load is inclined from the vertical. These additions produce equations of the general form shown in Table 4-1, with select N factors computed in Table 4-4, Program BEARING is provided on disk for other Nj values. “Meyerhof obtained his N factors by making trials of the zone abd’ with arcad’ of Fig. 4-3b, which include an approximation for shear along line cd of Fig. 4-3a. The shape, depth, and inclination factors in Table 4-3 are from Meyerhof (1963) and are somewhat different from his 1951 values. The shape factors do not greatly differ from those given by Terzaghi except for the addition of s,. Observing that the shear effect along line ed of Fig. 4-3a was still being somewhat ignored, Meyerhof proposed depth factors di He also proposed using the inctination factors of Table 4-3 to reduce the bearing capacity when the load resultant was inclined from the vertical by the angle @. When the iy factor is used, it should be self-evident that it does not apply when ¢ = 0°, since a base slip would ‘occur with this term—even if there is base cohesion forthe ie term. Also, the i factors all = 10 if the angle @ = 0. Up to a depth of D ~ B in Fig. 4-3a, the Meyerhof qu is not greatly different from the ‘Terzaghi value, The difference becomes more pronounced at larger D/B ratios. Hansen’s Bearing-Capacity Method. Hansen (1970) proposed the general bearing-capacity case and N factor equations shown in Table 4-1. This equation is readily seen to be a further extension of the earlier Meyer- hof (1951) work. Hansen’s shape, depth, and other factors making up the general bearing- capacity equation are given in Table 4-5. These represent revisions and extensions from ear- lier proposals in 1957 and 1961. The extensions include base factors for situations in which the footing is tilted from the horizontal b, and for the possibility of a slope 8 of the ground supporting the footing to give ground factors gi. Table 4-4 gives selected NV values for the ‘Hansen equations together with computation aids for the more difficult shape and depth fac~ tor terms. Use program BEARING for intermediate Nj factors, because interpolation is not recommended, especially for = 35°. Any of the equations given in Table 4-5 not subscripted with a V may be used as appropriate (limitations and restrictions are noted in the table). ‘The equations shown in this table for inclination factors j will be considered in additional detail in Sec. 4-6. Note that when the base is tilted, V and Hare perpendicular and parallel, respectively, to the base, compared with when itis horizontal as shown in the sketch with Table 4-5. For a footing on a slope both the Hansen and Vesié gi factors can be used to reduce (or increase, depending on the direction of H;) the bearing capacity using N factors as given in Table 4-4. Section 4-9 considers an alternative method for obtaining the bearing capacity of footings on a slope. ‘The Hansen equation implicitly allows any D/B and thus can be used for both shallow (footings) and deep (piles, drilled caissons) bases. Inspection of the GN, term suggests a great increase in guy with great depth. To place modest limits on this, Hansen used D 4, = 14045 et W = singe 2 1+ 2tangul ~ sing?[BARING CAPACITY OF FOUNDATIONS 225 D de = 1+ 04tan-! 2 1+ 2ean g(t ~ sing)? tan"! ‘These expressions give a discontinuity at D/B ~ 1; however, note the use of = and >. For = O (giving d’) we have bes oo 1 ise 2 5 0 2 10 “ 0 040 082 044 055 059 061 __ 082 *Aewally computes 039 We can see that use of tan“! D/B for D/B > | controls the increase in de and dy that are in line with observations that qu appears to approach a limiting value at some depth ratio D/B, where this value of D is often termed the critical depth. This limitation on uy will be further considered in Chap. 16 on piles. Vesié’s Bearing-Capacity Equations ‘The Vesié (1973, 1975b) procedure is essentially the same as the method of Hansen (1961) with select changes. The N- and N, terms are those of Hansen but N, is slightly different (See Table 4-4). There are also differences in the i, by, and g, terms as in Tuble 4-5e. The ‘equation is somewhat easier to use than Hansen's because Hansen uses the i terms in computing shape factors s; whereas Vesié does not (refer to Examples 4-6 and 4-7 following). Which Equations to Use ‘There are few full-scale footing tests reported in the literature (here one usually goes to find substantiating data). The reason is that, as previously noted, they are very expensive to do and the cost is difficult to justify except as pure research (using a government grant) or for a precise determination for an important project—usually on the basis of settlement control, Few clients are willing to underwrite the costs of a full-scale footing load test when the bearing capacity can be obtained—often using empirical SPT or CPT data ditectly—to a sufficient precision for most projects. Table 4-6 is a summary of eight load tests where the footings are somewhat larger than ‘models and the soil data are determined as accurately as possible. The soil parameters and gay (in kg/em®) are from Milovié (1965). The several methods used in this text and the Balla (1961) method used in the first edition, which is a subroutine in supplemental computer pro- gram B-31 noted on your diskette] have been recomputed using plane strain adjustments where L/B > 1. Comparing the computed qux to the measured values indicates none of the several theories/methods has a significant advantage over any other in terms of a best predic- tion. The use of py instead of gy, when L/B > 1 did improve the computed quu forall except the Balla method. ‘Since the soil wedge beneath round and square bases is much closer to a triaxial than plane strain state, the adjustment of dy to dy is recommended only when L/B > 2.pe 4 Bugera 90 2G HOY 21} = soepes oe (oon z-y oe (oumtiz—pho = "y cae owo!y +A O<) Apnre ol Hos L0 - itl x pmol ta oy o-9 Glan | ‘ese pon) SO EE seer Ore eeeivea |. pmso—D = f= peor tA y) a ddunso- eta ol oy a fo'N yay dey Bit af -so-y (ios wo 3589) 01 pO “Soe woneuIPT “suoprenbe 91594 auapeaznba 209 25-p argu, 29g ‘suonrenbo (QL61) Hoste, 4p 405 suo Dey seq pul ‘punoss ‘uoRIeUTU Jo 2qK, os sav, ‘oon poy san #30 py PAN Re ME WALSHE ovo SNES SALE "Sq was gf naa 258,899, 9380 NT = dyg-o1 2 so= Tyo 902 Syo-o1~ ms 1m) Gol a Puen oe 2noge pauyep Pung rou a ydous— ppumz +1 = "p pus 7 +01 sumper ur 1 < a/a 595 (a/q),-0a = ¥ wo? soe oa “ipapadsax 9994 “UasueH] 40) AH SiduDsGNg ‘suoneNbs UWasueH] J0y Aqu0 Q = — UDA 2p“ A801 “TH+ QUI, 10 suopenba H9ys 07 s9]—4 ue Mojaq sayOU 29g “suopenba Ayouduo-3uprwaq (75161 “CL61) 315A 2 410} s40y9e aseq pu “punoss “woeUEOUY Jo a1GeI, os TEV canoe (4) sus ney Up woe) .06 +g SON228 FOUNDATION ANALYSIS AND DESION ‘The Terzaghi equations, being the first proposed, have been very widely used. Because of their greater ease of use (one does not need to compute all the extra shape, depth, and other factors) they are still used—probably more than they should be. They are only suitable for a concentrically loaded footing on horizontal ground. They are not applicable for footings carrying a horizontal shear and/or a moment or for tilted bases (see Example 4-7 following). Both the Meyerhof and Hansen methods are widely used. The Vesi¢ method has not been ‘much used {but is the suggested procedure in the API RP2A (1984) manual]. As previously noted there is very litte difference between the Hansen and Vesié methods, as illustrated by the computed 4, values shown in Table 4-6 (see also Example 4-7). From these observations one may suggest the following equation use: Use Best for Tervaghi ‘Very cohesive soils where 2/B 1 or fora quick estimate of gu to compare with other methods. Do ‘no use for footings with moments andlor horizontal forces of for tilted bases andlor sloping ground, Hansen, Meyethof,Vesié Any situation that applies, depending on user preference or familiarity with a partcular method, Hansen, Vesié ‘When base is tilted: when footing is om a slope or when D/B > 1 Iis good practice to use at least two methods and compare the computed values of gu. If the two values do not compare well, use a third method, a trivial exercise where the equations have been programmed for computer use. Use either an arithmetic or weighted average? value for the qo provided for design (unless settlement is controlling). 4-4 ADDITIONAL CONSIDERATIONS WHEN USING THE BEARING-CAPACITY EQUATIONS. One should avoid using tables of N factors that requite interpolation over about 2°, It is a trivial exercise to program the equations (or use program BEARING) to obtain N values for any angle, For angles larger than 35° the factors change rapidly and by large amounts. Interpolation can have a sizable error (or difference), so someone checking the work may not be able to verify gun ‘The methods used to develop the bearing-capacity equations do not satisfy moment equi- librium but do satisfy 3: Fy = S Fy = 0. This errors not serious since statics is obviously 2 Weighting i a follows: Ifyou have values of 2 and 4 but boieve 2 is 15 times as good asthe 4, 9 weighted average gives med 7 T+is "33 ‘An arithmetic average is simply (2 + 4V2 val = =28