Está en la página 1de 13
DETERMINATION OF SHEAR FAILURE ENVELOPE IN Rock Masses By Roberto Ucar,! A. M. ASCE, Avsrnacr: In this work, an analytic expresion has been developed that en- ables the shear strength to be determined using the empirical failure criterion developed by Hoek and Brown. Applying this new hypothesis, along with elas- ticity theory and basic mathematical concepts for obtaining the failure envelope: oF of a family of failure circles, the shear strength has been determined to be such a that 1. = f(¢..0.m,s), where a, = the normal stress on the discontinuity, «, |! + + the uniaxial Compressive strength of the intact rock and the parameters mand : are constants dependent upon the properties of the rock. Art exp this type will help us develop new methods of analysis of slope st support of galleries, design of anchor support systems in rock masses, est mation of the skin friction of poor quality rock in driven pile foundations, as well as in other innumerable related areas of geotechnology’ IntRODUCTION In the 1970's, in the Department of Mineral Resources Engineering of the Imperial College of Science and Technology in London, Hoek-and Brown (4) performed a great many experimental tests on rocks with dif- ferent geomechanical properties in order to interpret in detail the failure mechanism in rock masses, At the same time they used as a fundamental basis the postulate of Griffith (2), which established that in order for a crack to develop, the relationship between the major ptincipal stress 0, and the minor prnci- pal stress 03 should be equal to (a; — 03)? = 8,(0; + 03), as well as (1 ‘+ 303) < 0. As a; is the uniaxial tensile strength, the uniaxial compres- sive strength o, = 89; taking into consideration that 03 = 0. Thus, with this hypothesis and its modifications and extensions ré- alized by other investigators, Hoek and Brown were able to obtain a new empirical failure criterion using a quadratic equation that relates o and 03 a8 a function of a, and the parameters m and s, which in turn depend upon the properties of the rock mass. Also, through this formula, a great advance in the area of rock me- chanics would be achieved if true analytical solution to determine the shear strength, 7,,, for a certain range of stresses in rocks with different planes of weakness could be obtained. This was achieved only in an approximate and empirical form by Hoek and Brown (4) due to the com- } 1 plexities of finding a precise mathematical result. 1 ‘The work described herein is a detailed analysis of an exact solution using the aforementioned failure criteria to calculate the shear failure envelope. This calculation provides an adequate base from which to es- timate the shear strength in intact rock as well as rock masses with di ferent discontinuities, thus enabling the resolution of a great variety of "Assoc. Prof., Dept. of Civ. Engrg., Univ. de Los Andes, Mérida, Venezuela. Note.—Discussion open until August 1, 1986. To extend the closing date one month, a writlen request must be filed with the ASCE Manager of Journals. The manuscript for this paper was submitted for review and possible publication on March 27, 1985. This paper is part of the Journal of Geotechnical Engineerin Vol. 112, No. 3, March, 1986. ©ASCE, ISSN 0733-9410/86/0003-0303/$01.00. Pa- per No. 20431 303 problems that are habitually encountered in rock mechanics involved with mining and civil engineering projects. DETERMINATION OF SHEAR FAILURE ENVELOPE IN Rock Masses Through laboratoty tests, in conjunction with the fundamental theo- ries existing of fractures and propagation of cracks in rock, Hoek and Brown (4) developed a new strength hypothesis establishing the follow- ing relationship between the principal stresses 0 and 03: 3 oe (ay in which o, = the major principal stress at failure; ¢3 = the minor prin- cipal stress at failure; ¢, = uniaxial compressive strength of the intact rock material; and m,s = constants dependent on the properties of the, rock. In Table 1, the different values of m and s can be observed, depending on the extent of jointing and weathering of the mass. Referring to Figs. 1-2, the uniaxial compressive strength of the intact rock ¢; is obtained by taking into account that no lateral confinement exists («3 = 0) and that s = 1. This results in Eq. 1 being simplified to a = O. When the mass presents planes of weakness, s < 1, and the com- pressive strength is on the dominant discontinuity then oa = os). Squaring both sides of Eq. 1 and solving for 95 results in 1 o3= (« + Zo,) + pire + Amaya, + 4802)!” Taking the negative root of (mo? + 4maa, + 4s02)'?, since o3 corre- sponds to the principal minor stress, we have - Q) m 1 o> («. + i «| a (ne? + dmojya, + 4502)? 0. eee 7 -@) The tensile strength, o,, is determined by considering 0; = 0. Thus the former equation takes the form of oy = = a it — (WP + ASA] cece bec terete eee eeeeeeeeees 4) From Eqs. 1 and 4, the limits of s can be seen, i.e., s = 1 and o = 9 for intact rock; and s = 0 and a3 = @, = 0 for highly jointed rock. From the previous result, for the intermediate states of the rock mass, s, is to be found between the interval of 0 as7 |a= oon |p B= 07s -o007 |e $= 000 ery good eaty | 100 73, | ns as rock me cr [2 oa SS ‘on Tightyiteecting oss [A= osm fas come nde’ ck orm {o> ons |a= on sith srmethred -o0 | g= nor |e 0008 joints at +3 m. ‘Good quaty | 10 me 1s |me rock mas s= om [r= Slightly weathered A= oso fac te istabed with a= 06s |p inte at 1-3 «= -00m | Fairquatiy | 1 m= 0% Ime rock mass, == oom |S Several ets of A= om |A moderately a= ase |e weathered joints f= ~00003 | ¢ = Spaced 903-1 oor quality | to“ m= 008 om w= ox rock mass = o.o00ni|"s = ooot|'s = o.coot Numerous weathered A= 01 oi [a> oa joints at 30-509 a= 062 owe |a= 68 om &= -oom | ¢= -cer | ¢= -vooo1 Very poor quality | 10-2 m= ons ao {m= os tock mase = 0 ° ° Numerous heey A= 006 cuss |A= 0078 weathered joins Bo Ose ose |5= 556 spaced 50 mm to [ee a fen o Quality index ebinined fom the NGI (Nonwegian- Geotechnical sui) dstifcaon r 12 ‘a0,\"" m tana = (t 1+ See (5) a0, . > TA afm 2+ :) L « Taking into account that s = 1 for intact rock, and 3 = 0 for uniaxial compressive strength test, then tan? a = (: + 4) 2 Atan?a-1)....... © m + Balmer (1) developed an analytical solution to determine the failure envelope for any distribution of pressures o, and 0, in the mass. Applying the equations of equilibrium and observing Fig. 1, it can be seen that 305 Tar Lig-d, sin 20 Gar FG $0) + (G,Gy) cov 1S THE ANGLE BETWEEN THE FAILURE SumFAce AND THE DIRECTION OF THE MINOR PRINCIPAL STRESS Oy swear strewom & oy G G = Slow + er Gy DED |a% waa fg Gay | per Fic. Specimens oft} FIG, 2—Re Graphical Representation of Stress Conditions at Fallure of Intact Rock | t 4 MAXIMUM PRINCIPAL STRESS B+ fremsasnr| von _Fancirae svaess G Cs pELATONSHIP BETWEEN PRINCRAL STRESSES AF FAILURE lationship between Maximum and Minor Principal Stresses at Fallure 306 1 : 0)" a [>.-de | tis (252) : in which a, and 1, represent the respective normal and tangential stresses on the failure plane. Taking the derivative of o; with respect to 03 on both sides of Eq. 8 and rearranging leaves: 1 1 (a0, = 5(% 293) 1 (ae. _ >. govt on][-3 (52 +4)] -2f 2 i= :) ‘ao, ‘81 — 93) ‘ao, 1 (-1)( i ) + (+1) [o.-de. +00] dos _j) (m= 23), (am, ,) 2 oo (@ 1)( ; ) + 06(2 +1) Fort oa(Zo4 (1+) ~6,-0(#2) <0... ae vee (12) aos, -® a (1 =) 803, anes a ( * =) 003 Substituting Eq. 13 into Eq. 8 and solving for 7, results in (04 = 03) = a-95 faa, Dividing Eq. 15 by Eq. 14, we obtain ns in S57 (E) enrol) (.~ 03) \doy oN Gos On the other hand, upon observing the triangle ABC of Fig. 1, the angle of failure is calculated according to the following expression: 12 tana = (2) ee ee eee - (17) 8053 (16) Using Eq. 1, the corresponding value of 804/d05 is 80, Sela 803 3 a 2m 45 oe 307 Transforming Eqs. 13, 15, and 17 as follows: . 20,(m 2+ :) a 3 - 19) a am % +.) oe Likewise, as may be observed in Fig. 1, the angle of failure (a) and the inclination (B) of the envelope to Mohr’s circle are related by the expres- sion ra= (E+8) beceeeeeeteeees . seeeeeess 22) - (26) Eq. 26 may be transformed into of — Ano’? -1=0 (27) in which (oj)? = 1 + (1 + (A)? (28) Solving for o in Eq. 11, we have 1 aint nadie 09-09 5] eer -. 29) +0304 eel ae re ee ee ee 2... 30) lta) On the other hand, the principal minor stress from Eq. 15 results in ta(1 + 91) osm ot Substituting this value into Eq. 30 leaves , a t(1 + of) o,(1 + oj) = 0, + oe - a inwhich 9, = 0, + 1,(0})'?... Substituing Eq. 28 into Eq. 33 we obtain a= 0, +r, + (147 That is to say: G1 = 04 + Tete + ToL + 1)'? «2. (35) Equally 03 =o, + tt, ~ T(1 + 14)? + 36) Subtracting the two equations yields (01 — 95) = 2ra(1 + 14)'7 (37) Squaring Eq. 28 gives 1 = (ra)? + ral + (RYN? + [1 + (ey) + (38) - 89) ac Theses (2) = 1+ 2¢02)8 + Deed + PI? 90, Replacing (mo3/o, +s)" = 0; ~ 03/0, from Eq. 1 in Eq. 21 we obtain placing ‘0,\'? mo, _]'” tana=(—") =]1+—"_| |. 803, 2(0; ~ 93). Thus Eq. 39 is converted into mae 1+—_ 2(e; — 63) Finally, knowing ffom Eq. 37 that (0; ~ 03) = 2z,(1 + 14)'?, the previous equation can be simplified to Mee os aartles VI Fae + (1 + 22). 8 = 1+ 2m)? + Arif + PYF... (41) - (42) SOLUTION OF DiFFERENTIAL EQuaTiOn For convenience, we express the differential equation indicated in Eq. 42 in the following manner: ca = yyy Vary) + 1+ y" where logically y = t., y' = dy/dx = x, - (43) 309 Upon studying the equation, it can be seen that it corresponds to form Q): Flyy')=0 Introducing a new parameter and replacing in Eq. 44 the equations y = (8) and y’ = 6(8), we have F[W(B), 6(B)] = 0 In addition to this pees peer et (45) 4) yee nay=y' ( ) YOR ody = Y'(B)AB . . y a) On the other hand 4 _ ¥ On yo) Integrating the previous equation results in ve) exes | ape ° J 4a“? in which C = a constant to be defined later. Considering that y’ = 6(B) = tan B (slope of the envelope), Eq. 43 is transformed into cds) =m (ape 9° =e [tan B(tan B sec B + sec’ B). Simplifying trigonometrically we are left with _ mo tant] 7° —e [tanp J ‘That is to say _ mo, {2 — sin B)(1 — sin? B)'” 8 sin 8 Equally, it is known that dy/dB = '(@); whereupon, according to Eq. 50 we have dy __ mo; [cos 6 tan B + (1 ~ sin B)tan B a OB tan’ B Simplifying we are left with Ta see GL) vgs Me YOq-F [ss sin 6] eee (63) Thus, according to Eq. 48, it can be seen that 1 _ ing | I[ea-*] Bre eoeoeeoaasaRabpaaoooo 4) 310 Integrating Eq. 54 we obtain Oe mee ( Lisi Jee 55 & (yantp t 8B) HCo rere beset eeeeeesaneees (55) Boundary Conditions.—Using Eqs. 26 and 19, and at the same time taking into account the particular case where o, > ~a;/m s [since, at this point, the line tangent to the curve (envelope) is perpendicular to the o,-axis] we have Likewise 0, = 05+ TA 9% m+4(m—+s % ; Ge limo; -—s ... m cee (586) Thus, if we substitute o, the arbitrary constant, C, og m(i+1) +c ~% = mae (1 m 8 \2 ~o,/m s and B = 17/2 into Eq. 55, we obtain a "8 |2sin®B The next step is to replace the value of sin B in Eq. 51 in order to obtain the shear strength 1, = {(o.,a-,m,s). This indicates that it will be necessary to transform Eq. 61 as follows: sin? B ~ k sin? B + in which k=— The solution to the cubic equation is . ak 1 27\] , 4e sinp= 4 (2cos [arecs [(: -2)| +4 +1) peeeae6Ha00666 (64) For the particular case where 27/4k° -> 0, of course, sin B > 0, which indicates that great precision is required in the calculator. This problem can be resolved taking into consideration that in Eq. 62, sin’ B is an infinitisimal of higher order. Since the value of kis very large, then k sin? B >>> sin’ B. That is to say 57 sin? pHocaEooq00099neGe=00 (65) 1\"" sinp~(3) eee ee = (66) Therefore, Eq. 51 may be transformed into the reduced form: PRACTICAL APPLICATIONS Example 1.—Granite, which has the following characteristics: Rock mass of good quality, slightly weathered with joints spaced at +2m. Accord- ing to these characteristics, from Table 1 it can be seen that: m = 2.5; 5 = 0.004; o,/, = —0.002; A = 0.603; and B = 0.707. Likewise, we will consider that o,/o, = 2. Exact Solution: from Eqs. 63 and 64 we have 27 1 — ————— _} + 240°] + 1} = 0.2557 + (69) arecos ( ax am) | } bed Therefore, according to Eq. 51 we obtain = (70) Applying the approximate equation proposed by Hoek and Brown and using the values of Table 1 results in Ta . ac: - 2) = 0.603(2 + 0.002)°7” = 0.985 ..... 2.0.2 eee eee (71) % Se Fe Example 2.—Amphibolite, which has the following characteristics: Rock 312 mass of very poor quality, highly jointed and weathered, joints < 50 mm. According to Table 1, we have m = 0.025; s = 0; o,/0, = 0; A = 0.078; and B = 0.556, Taking into account that o,/s, = 10 results in 80 3 Dope 15 | = 3/2010 cece eect eeeeeeeeeeeeeeeeeeee » (72) 0.025 2. Since upon substituting the value of k in Eq. 64 sin B > 0, the.expres- sion in Eq. 67 is therefore used m 25, any? = 90? (6.403)! = 0.2 3) g (6-403)' = 0.25 « According to Hoek and Brown, we obtain 7 = 0,078(10)° = 0.281 0.0... cceceveevseees . (74) Example 3.—Gneiss, which has the following characteristics: Rock mass of poor quality, with numerous weathered joint at 50-100 mm. From Table 1, the material constants are m = 0.13; s = 10 = 10; A = 0.203; and B = 0.686. The assumption that o,/a, = 5 results in K = 309.20............00005 1\"" en p= (2) = 0.04 . TABLE 2.—Comparative Results of ./c, between Hoek and Brown Equation and Analytical Solution al aloe 1/0, m=35 s=01 m=05 s= 10 m=013 s= 10% Hoek and Analytical Hoek and | Analytical | Hoek and Analytical e,/0, | Brown | solution | Brown | solution | Brown | —soltion w ) ® 4) @ @ ” 05 0.42 |) 0.43, 0.21 0.20 0.13, 0.1L 1 0.66 0.66 0.35 0.30 0.20 017 2 1.05 1.01 056 oat 0.33 0.24 3 1.38 129 075 0.36 043 0.30 4 1.68 153 os 065 053, 0.35 5 195 174 1.07 073 ost 039 6 2.20 193 121 oat 0.69 0.43 7 2.45 2n 1.35 0.88 077 0.46 8 268, 228 148 0.94 0.85 0.49 9 290 244 Let 1.00 os 02 10 3 238 173 1.06 0.99 0.55 n 3.32 273 1.85 1 1.05 0.58 2 352 2.86 197 117 12 ost B 37 259 2.08 1m 118 0.63 4 391 312 219 1.26 124 0.66 15 410 3.24 230 131 1:30 0.68 313 and Fi 2k)" = 0.404... ca According to Hoek and Brown, for a poor quality gneiss rock mass, we obtain: 0.203(5 + 10°*)°* = 0,612 Finally, in order to more clearly show the difference between the em- pirical and analytical equation, certain values of m and s chosen from Table 1 are shown in Table 2, even though we could have chosen other values. ConcLusions: The excellent contribution of Hoek and Brown involves the new hy- pothesis of failure, which relates the principal stresses at the instant of failure through a quadratic expression, and at the same time takes into account the analytical method described in this paper to evaluate the shear strength of rocks. It is therefore logical to conclude that a new area of investigation has been opened through the shear failure enve- lope. The applications include slope stability, gallery support, and de- sign of anchors in unstable rock masses, as well as new methods of calculation for foundations in poor quality rock and other problems that often occur in the area for geotechnology. This new formula is an improvement on the former linear approxi- mation to the Mohr-Coulomb envelope since it is known that the true equation that governs the resistance to failure in rocks is nonlinear. On the other hand, comparing the results obtained by the analytical formula with the Hoek and Brown's empirical equations we can appreciate (as shown in Table 2) that for particular determined values of m and s, when the relation o,/a, increase, the empirical solution for obtaining 1,/, in- creasingly differs from the exact one. Finally, possibly the most difficult problem that faces the geological engineer occurs at the moment that he chooses the input values of m and s that depend upon the properties of the rock mass. ACKNOWLEDGMENTS I would like to express my most sincere gratitude to Gerard Defives, Florencio Plachco, and Luca Rojas of the Mathematics Department of the Engineering College of the Univ. of the Andes for their valuable collab- oration during the elaboration of this work. Equally, I would like to thank Miguel De Luca of EPICA. APPENDIX |.—REFERENCES 1. Balmer, G., “A General Analytical Solution for Mohr’s Envelope,” American Society of Testing Materials, June, 1952, pp. 1260-1271 2. Griffith, A., “The Phenomena of Rupture and Flow in Solids,” Philosophical Transactions of the Royal Society, Series A, Vol. 221, 1921, pp. 163-198. 314 3. Elgosts, L., Differential Equations and the Calculus of Variations, Mir Publishers, Moscow, USSR, 1980. 4. Hoek, E., and Brown, T., “Empirical Strength Criterion for Rock Masses,”” Journal of the Geotechnical Engineering Division, ASCE, Vol. 106, No. GT9, Sept., 1980, pp. 1013-1036. AppeNoix II.—NOTATION The following symbols are used in this paper: k = constant, m = constant that depends on properties of rock and controls cur- vature between major and minor principal stresses at failure, s = constant for given rock type, which controls location of curve between principal stresses; @ = angle between failure surface and direction of minimum prin- cipal stress, B inclination of envelope to Mohr's circle, ¢, = uniaxial compressive strength of intant rock, % tensile strength of rock mass, 7 major principal stress at failure, 3; = minor principal stress at failure, oi = d0,/d03 = partial derivative of a; with respect to 03, slope of (01,05) curve; %, = normal stress at failure, t. = shear strength, and ta = tan B = 6(8) = slope of envelope. 315

También podría gustarte