Modification of Slope Mass Rating (SMR) by Continuous Functions
Modification of Slope Mass Rating (SMR) by Continuous Functions
Technical note
1365-1609/$ - see front matter r 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ijrmms.2007.02.004
ARTICLE IN PRESS
R. Tomás et al. / International Journal of Rock Mechanics & Mining Sciences 44 (2007) 1062–1069 1063
Table 1
Correction parameters for SMR (modified from Romana [1] by Anbalagan et al. [20])
These expressions allow us to calculate basic RMR for Romana [19] proposed a function for the computation of
several kinds of data (point load strength or uniaxial the F1 parameter (Fig. 1a):
compressive strength, water flow or pore pressure relation-
F 1 ¼ ð1 sin jAjÞ2 , (14)
ship). As hydrogeological conditions are usually defined
using subjective descriptions (dry, damp, wet, dripping or where A is the parallelism between discontinuities and
flowing), the last two expressions allow us to calculate slope dip direction for planar and toppling failures. For
basic RMR. wedge failure, A is the angle formed between the
When no borehole data are available and we use scanline intersection of the two discontinuities (the plunge direc-
data or volumetric measures to estimate RQD, the Priest tion) and the slope dip direction [24]. This function is valid
and Hudson [22] expression can be used: for all possible values of A and provides more conservative
values for F1 than the original discrete function.
RQD ¼ 100e0:1l ð0:1l þ 1Þ (10) The same author established the following continuous
or the Palmstrom [23] expression: function for F2 computation (Fig. 1b):
where l is the discontinuity frequency and Jv is the where B corresponds to the discontinuity dip (bj) in
volumetric joint count, respectively. Consequently, the degrees, for planar failure and toppling and to the plunge
basic RMR of (9) can be expressed for RQD obtained from (bi) of the wedge failure intersection line [24]. Its valid
expression (10) and taking into account that l ¼ 1=X̄ range is for values of B lower than 451. For higher values,
newly as F2 is set to 1. Consequently, this function is not valid for all
8 values of B constituting a discrete function. Moreover, it is
>
> 20ð1 þ 0:1lÞe0:1l 15:1 log l þ 1:670sP 2:9 log G þ 35:67 þ rJ ; unsafe because it provides values of F2 lower than those
>
>
>
> 20ð1 þ 0:1lÞe0:1l 15:1 log l þ 0:075sC 2:9 log G þ 34:00 þ rJ ;
>
>
> proposed in the original function.
< 20ð1 þ 0:1lÞe0:1l 15:1 log l þ 1:670sP 10 log ru þ 24:17 þ rJ ;
RMRb ¼
> 20ð1 þ 0:1lÞe0:1l 15:1 log l þ 0:075sC 10 log ru þ 22:50 þ rJ ;
>
>
>
> 4.2. Alternative functions
>
> 20ð1 þ 0:1lÞe0:1l 15:1 log l þ 1:670sP þ 24:67 þ rG þ rJ ;
>
>
: 20ð1 þ 0:1lÞe0:1l 15:1 log l þ 0:075s þ 24:00 þ r þ r ;
C G J
In this section, alternative functions for the F1 and F2
(12) correction parameters are proposed. Moreover, new
and for RQD obtained from expression (11), expression continuous functions are proposed for the F3 parameter.
(9) is expressed as The proposed F1 continuous function that best fits
8 discrete values is expressed as (Fig. 1a)
>
>
>
3:3J V þ 15:1 log X̄ þ 1:670sP 2:9 log G þ 150:67 þ rJ ;
>
> 16 3 1
>
> 3:3J V þ 15:1 log X̄ þ 0:075sC 2:9 log ðGÞ þ 149:00 þ rJ ; F1 ¼ arctan ðjAj 17Þ , (16)
>
>
< 3:3J V þ 15:1 log X̄ þ 1:670sP 10 log ru þ 139:17 þ rJ ; 25 500 10
RMRb ¼
>
> 3:3J V þ 15:1 log X̄ þ 0:075sC 10 log ru þ 137:50 þ rJ ; where parameter A and arctangent function are expressed
>
>
>
> 3:3J V þ 15:1 log X̄ þ 1:670sP þ 139:67 þ rG þ rJ ;
>
>
>
in degrees.
: 3:3J V þ 15:1 log X̄ þ 0:075sC þ 139:00 þ rG þ rJ : We propose an alternative continuous function for F2
(13) estimation, which is valid for all possible values of B (Fig.
1b):
As a result, basic RMR can be expressed in 18 different
forms (Eqs. (9), (12) and (13)), depending on the 9 1 17
F2 ¼ þ arctan B5 , (17)
parameters available for determining rock mass quality. 16 195 100
Note that the discontinuity conditions term (rJ), which where parameter B and arctangent function are also
depends on roughness, continuity, alteration, aperture and expressed in degrees.
fill, has a descriptive character and consequently a Tomás et al. [25] proposed continuous functions for F3
continuous function cannot be used to calculate it. calculus from dip relationship (C) using sigmoidal and
Furthermore, it should be pointed out that Eq. (12) has asymptotical arctangent function. These functions have
only four independent parameters while Eqs. (9) and (13) been slightly readjusted to the following new formula (Figs.
have five different independent parameters. 1c and d), which is more conservative and shows a better fit
to the F3 discrete values:
4. Continuous slope mass rating (SMR) 1
F 3 ¼ 30 þ arctan C, (18)
3
4.1. Available functions
1
F 3 ¼ 13 arctan ðC 120Þ. (19)
As previously stated, SMR is calculated from Eq. (1) by 7
adding four correction factors (F1, F2, F3 and F4) to the Function (18) is used for slopes with planar or wedge
basic RMR (Table 1). failure and expression (19) is used for toppling failure
ARTICLE IN PRESS
R. Tomás et al. / International Journal of Rock Mechanics & Mining Sciences 44 (2007) 1062–1069 1065
Fig. 1. Proposed functions for SMR correction factors estimation: (a) F1; (b) F2; (c) F3 for planar and wedge failures; (d) F3 for toppling failures.
cases. The C variable expresses dip relationship and is ence values approach interval borders. This observation
equivalent to bjbs for planar failure, bibs for wedge can be performed equally on the remaining parameters.
failure and bj+bs for toppling failure. Finally, F4 factor depends on the excavation system
Note that the use of arctangent functions has the employed for slope construction. It has an unavoidably
advantage of being asymptotic to the extreme score values descriptive character and is not suitable for continuous
and, consequently, for values higher than those corre- functions.
sponding to the properties near the extreme borders, they
do not cause significant deviations from the discrete values. 5. Comparison between original and modified by continuous
Functions (18) and (19) are novel for Romana’s functions SMR
classification application because they enable the F3
parameter to be calculated using the original values of In this section, we applied the proposed functions to 61
Bieniawski’s classification (varying from 0 to 60), which slopes in order to establish their confidence with the
very much condition the final SMR value. For instance, a original, discrete functions for SMR. Most of the cases
slope with planar failure and C ¼ 01, F3 is equal to 25 correspond to Spanish slopes measured by the authors,
points. However, for values of 01oCp101 F3, is equal to although several of them have been taken from historical
6 points and for values of 101pCo01, F3 is equal to and bibliographical cases.
50 points, which, in the worst case scenario, implies The results obtained from the comparison between
maximum SMR indexes with differences of 44 points continuous and discrete SMR correction parameters in
(more than 2 classes). Consequently, 711 can cause the cases considered are shown in Fig. 2 and Table 2.
important changes in the chosen score and, for the slope The continuous functions provide a unique correction
under consideration, we will select a value of F3 equal to value for every case (Fig. 2) in comparison with discrete
50, to be on the safe side. functions. It can be observed that Eqs. (14) and (15)
As a result, when this correction index is continuous, we generally provide lower values than those provided by
avoid the ambiguity that sometimes appears when refer- original SMR discrete functions. Moreover, the correction
ARTICLE IN PRESS
1066 R. Tomás et al. / International Journal of Rock Mechanics & Mining Sciences 44 (2007) 1062–1069
Fig. 2. Comparison between discrete and continuous F1, F2 and F3 parameters using different functions (61 case studies considered).
Fig. 3. Comparison between continuous and discrete indexes: (a) Basic rock mass rating; (b) slope mass rating for different continuous functions; (c) slope
mass rating for different failure types and Eqs. (14) and (15) from Romana [19] and Eqs. (18) and (19). (c) slope mass rating for different failure types and
Eqs. (16), (17), (18) and (19).
Table 3
Statistical differences among continuous and lump geomechanical indexes
Table 4
Absolute differences in discrete SMR and continuous SMR according to failure type and slope rock mass class (No. of considered cases ¼ 61)
W T P I II III IV V
Cases (n) 13 29 19 0 16 15 25 5
Average difference Eqs. (16)–(19) 574 674 573 — 473 574 674 472
Eqs. (14) and (15) from Romana [19] and Eqs. (18) and (19)) 12710 775 1179 — 674 775 1479 373
Maximum difference Eqs. (16)–(19) 11 14 10 — 10 10 14 7
Eqs. (14) and (15) from Romana [19] and Eqs. (18) and (19) 28 16 27 — 15 18 28 6
Minimum difference Eqs. (16)–(19) 0 0 0 — 0 0 0 1
Eqs. (14) and (15) from Romana [19] and Eqs. (18) and (19) 2 0 0 — 0 1 2 0
are, in the worst-case scenario, 14 points, which is less than tion System but in contrast they are difficult to apply in the
three quarters of a geomechanical class. field, where chart classification is more useful.
6. Conclusions Acknowledgments
Rock mass classifications are a very useful tool for the This study was partially funded by the Spanish Ministry
characterization of rock masses. Among all available rock of Science and Technology and FEDER (Project TEC-
mass classifications, SMR is the most extended and is 2005-06863), by the Valencia Regional Government
applied for rocky slopes due to its ease and exhaustive, well (Project GV06/179 and GRUPOS03/085) and by the
established, quantitative definition of correction factors. University of Alicante (Project VIGROB-157).
These factors are: F1, which depends on parallelism
between discontinuities and the slope face; F2, which
References
depends on discontinuity dip; F3, which depends on the
relationship between the slope face and the discontinuity [1] Romana M. El papel de las clasificaciones geomecánicas en el
dip and F4, which depends on the excavation method. estudio de la estabilidad de taludes. In: Proceedings of IV Simposio
In this study, new arctangent and asymptotical contin- Nacional sobre taludes y laderas inestables, Granada, Spain, 1997.
uous functions for F1, F2 and F3 SMR correction factors p. 955–1011.
[2] Bieniawski ZT. Rock mass classification in rock engineering. In:
are proposed. These functions show maximum absolute
Proceedings of the symposium on explor for rock eng, Johannesburg,
differences with discrete functions lower than 7 points and 1976. p. 97–106.
significantly reduce subjective interpretations. Moreover, [3] Bieniawski ZT. Engineering rock mass classification. Chichester:
the proposed functions for SMR correction factors calculus Wiley; 1989.
reduce doubts about what score to assign to values near the [4] Selby MJ. A rock mass strength classification for geomorphic
border of the discrete classification. Only the F4 parameter purposes: with tests from Antarctica and New Zeland. Z Geomorph
1980;24:31–51.
preserves its subjective and descriptive character. [5] Romana M. New adjustment ratings for application of Bieniawski
These functions have been combined with Sen and classification to slopes. In: Proceedings of Int Symp Role of Rock
Sadagah’s proposal [18] to also reduce basic RMR Mech, ISRM, Zacatecas, Mexico, 1985, p. 49–53.
subjective interpretations. These continuous basic RMR [6] Robertson AM. Estimating weak rock strength. In: Proceedings of
SME ann mtg, Phoenix, 1988. p. 1–5.
functions have been slightly modified to increase the
[7] Laubscher DH. A geomechanical classification system for the rating
possibilities of expressing different parameters. of rock mass in mine design. J South Afr Inst Min Metall
The resulting functions obtained by combining basic 1990;90:257–73.
RMR and SMR correction parameter continuous func- [8] Haines A, Terbrugge PJ. Preliminary estimation of rock slope
tions allow a unique SMR value to be assigned for every stability using rock mass classification system. In: Proceedings of 7th
slope and allow discrimination among slopes that have the cong ISRM, Aachen, Germany, 1991. p. 887–92.
[9] Chen Z. Recent developments in slope stability analysis. In:
same discrete SMR index. The differences observed among Proceedings of 8th international congress ISRM, Tokyo, 1995. p.
lump and continuous functions are less than 14 points for 1041–8.
the SMR with an average value of 574 (a coincidence of [10] Shuk T. Key elements and applications of the natural slope
8978%). These differences are higher if we use that methodology (NSM) with some emphasis on slope stability aspects.
proposed by Romana [19] for F1 and F2 estimation. In: Proceedings of 4th South Amer cong rock mech, Santiago, 1994.
p. 255–66.
The continuous functions make it easy to program data [11] Ünal E. Modified rock mass classification: M-RMR system. In:
processing routines for SMR calculus and facilitate the Bieniawski ZT, editor. Milestones in rock engineering, the bieniawski
implementation of SMR using the Geographical Informa- jubilee collection. Rotterdam: Balkema; 1996. p. 203–23.
ARTICLE IN PRESS
R. Tomás et al. / International Journal of Rock Mechanics & Mining Sciences 44 (2007) 1062–1069 1069
[12] Hack HRGK. Slope stability probability classification. Netherlands: [19] Romana M. A geomechanical classification for slopes: slope mass
ITC Delf Publication; 1998. rating. In: Hudson JA, editor. Comprehensive rock engineering.
[13] Lindsay P, Campbell RN, Fergusson DA, Gillard GR, Moore TA. Oxford: Pergamon; 1993. p. 575–600.
Slope stability probability classification, Waikato coal measures, New [20] Anbalagan R, Sharma S, Raghuvanshi TK. Rock mass stability
Zealand. Int J Coal Geol 2001;45:127–45. evaluation using modified SMR approach. In: Proceedings of 6th nat
[14] Romana M, Serón JB, Montalar E. La clasificación geomecánica symp rock mech, Bangalore, India, 1992. p. 258–68.
SMR: Aplicación experiencias y validación. In: Proceedings of V [21] Romana M. Nota sobre la valoración del 51 parámetro (condiciones
Simposio Nacional sobre taludes y laderas inestables, Madrid, 2001. hidrogeológicas) del RMR básico en función de ru (razón de
p. 393–404. presiones intersticiales). In: Proceedings of VI Simposio de taludes
[15] Romana M, Serón JB, Montalar E. SMR Geomechanics classifica- y laderas inestables, Valencia, 2005. p. 221–26.
tion: application, experience and validation. In: Proceedings of the [22] Priest SD, Hudson JA. Estimation of discontinuity spacing and trace
international symposium on the role of rock mechanics, South Afr length using scanline survey. J Rock Mech Min Sci Geomech Abstr
Inst Min Metall, 2003. p. 1–4. 1981;18:183–97.
[16] Romana M, Serón JB, Jordá L, Vélez MI. La clasificación [23] Palmstrom A. The volumetric joint count—a useful and simple
geomecánica SMR para taludes: Estado actual, aplicación y measure of the degree of rock mass jointing. In: Proceedings of 4th
experiencia internacional. In: Proceedings of VI Simposio Nacional conf int ass eng geol, New Delhi, 1982. p. 221–28.
sobre taludes y laderas inestables, Valencia, 2005. p. 239–50. [24] Singh B, Göel RK. Rock mass classification. a practical approach in
[17] Irigaray C, Fernández T, Chacón J. Preliminary rock-slope-suscept- civil engineering. Ámsterdam: Elsevier; 1999.
ibility assessment using GIS and the SMR classification. Nat Hazards [25] Tomás R, Cano M, Cuenca A, Cañaveras JC, Delgado J, Estévez A, et
2003;30:309–24. al. Nuevas funciones continuas para el cálculo del slope mass rating
[18] Sen Z, Sadagah H. Modified rock mass classification system by (SMR): aplicación mediante un sistema de información geográfica a los
continuous rating. Eng Geol 2003;67:269–80. taludes rocosos de una cantera. Rev Soc Geol España 2006;19:87–97.