Professional Documents
Culture Documents
Osgoui 2010
Osgoui 2010
Osgoui 2010
net/publication/240362845
An assistant tool for the Geological Strength Index to better characterize poor
and very poor rock masses
CITATIONS READS
13 1,098
3 authors, including:
Reza Osgoui
GEODATA Engineering SPA
16 PUBLICATIONS 121 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
All content following this page was uploaded by Reza Osgoui on 29 October 2018.
International Journal of
Rock Mechanics & Mining Sciences
journal homepage: www.elsevier.com/locate/ijrmms
Technical Note
An assistant tool for the Geological Strength Index to better characterize poor
and very poor rock masses
Reza R. Osgoui a, Resat Ulusay b,, Erdal Unal c
a
GEODATA SpA. Corso Duca degli Abruzzi, 48/E, 10129, Torino, Italy
b
Department of Geological Engineering, Hacettepe University, 06800 Beytepe, Ankara, Turkey
c
Department of Mining Engineering, Middle East Technical University, 06531 Ankara, Turkey
a r t i c l e in f o
Article history:
Received 17 June 2009
Received in revised form
5 February 2010
Accepted 6 April 2010
1. Introduction (W), more sensible and measurable in highly jointed poor rock mass,
is recommended. The proposed approach results in the modified-GSI,
Rock mass classification systems such as Bieniawski’s Geome- which mainly relies on measurable quantitative parameters. By
chanics Classification (RMR) [1,2] and Barton’s Q-system [3,4], in adding measurable quantitative input in N/A parts of the existing GSI
which RQD plays a dominant role are, to a certain extent, charts, they will be enhanced in characterizing poor rock mass while
subjective to apply in poor rock masses due to limitations and maintaining its overall simplicity. The modified-GSI is recommended
difficulty in the estimation of RQD (in most cases the value of RQD as an assistant tool for GSI to better characterize the poor and very
for such rock masses is lower than 10 and even zero) as discussed poor rock masses without disregarding its original concept.
in detail by some investigators [5–9]. Even though the introduc-
tion of Palmström’s Rock Mass index (RMi) [10,] and Hoek’s
Geological Strength Index (GSI) [11–13], independent of RQD; 2. Critical review
instead based solely on rock structural fabric (i.e., degree of
jointing and joint condition) [14] have alleviated such constrai- After several years of use, it became obvious that the RMR and
ners and difficulties, there are still difficulties in measuring the Q-system were difficult to apply to rock masses that are of very poor
degree of jointing of such a poor rock mass as denoted by N/A (not quality since they are dependent on RQD, which in most of the poor
applicable) parts at the bottom of the existing GSI charts. rock masses is essentially zero [5–9]. With the advent of the GSI,
In the GSI system, very poor rock masses are generally defined by based on an assessment of the lithology, structure and condition of
a GSI value below 27 [15,16] or 25 [12]. Due to its widespread joint surfaces in the rock mass, such a problem has been overcome.
acceptance in the field of rock engineering over the past 15 years, an This characterization index (Fig. 1) is directly linked to engineering
attempt has been made to provide an alternative practical approach parameters such as the Mohr–Coulomb or Hoek–Brown strength
to assist GSI in characterizing such poor rock masses where GSIo27. parameters, or rock mass modulus. It has been a universal rock mass
To improve the applicability of the GSI in poor rock masses and to characterization approach capable of characterizing wide spectrum
effectively quantify the joint properties, in this study a linkage of rock masses and has attracted keen interest in rock engineering
between descriptive geological terms and measurable field para- community. However, the application of the existing GSI system is
meters such as broken structure domain (BSTR), structure rating (SR), hindered by the facts that its use is to some extent subjective and
and some modifications are suggested. BSTR and SR are considered to requires long-term experience [17].
define the degree of jointing of poor rock structure and to overcome The recent form of the GSI charts is on account of tunnelling in
the limitations of RQD. To characterize joint condition for poor rock difficult ground conditions in Greece [18–21]. These charts
masses, the combination of intact core recovery (ICR) and weathering specially oriented to determination of the GSI for foliated/
laminated/sheared, heterogeneous (such as flysch) and for very
weak (such as molasses) rock masses. Most recently, Marinos
Corresponding author. Tel.: + 90 312 297 7767; fax: + 90 312 299 2034. et al. [6,7] have published papers putting forward some
E-mail address: resat@hacettepe.edu.tr (R. Ulusay). significant suggestions related with the appropriate selection of
1365-1609/$ - see front matter & 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ijrmms.2010.04.001
ARTICLE IN PRESS
R.R. Osgoui et al. / International Journal of Rock Mechanics & Mining Sciences 47 (2010) 690–697 691
Fig. 1. General chart for GSI estimates from the geological observations [18].
the GSI index for a range of rock masses under various conditions. Although in the available GSI charts, careful consideration has
In these papers, recommendations on the use of GSI and cases been given to the precise wording for each category and to the
where the GSI is not applicable have been discussed. relative weights assigned to each combination of degree of jointing
However, the definition of each component of the original GSI and joint condition, the use of the GSI involves some subjectivity
charts is quite subjective, as it is based on non-scaled sketch and [17,22–24]. Even though Marinos et al. [6,7] have pointed out that it
linguistic non-quantified terms [22]. In an effort to provide a is not meaningful to attempt to assign a precise number to the GSI
quantitative means for characterizing jointed rock masses, the value for a typical rock mass, and GSI is best described by assigning
original GSI was modified by Sonmez and Ulusay [23,24] and Cai it a range of values, it seems that assigning a range of values for poor
et al. [17]. The former uses the volumetric joint count (Jv) for rock mass causes misleading characterization of such rock mass.
determining the SR and Bieniawski’s RMR system for rating the joint There could be ranging of GSI values for a given poor rock mass in
surface condition in terms of roughness, weathering and infilling of which the differences in quality might not be well distinguished. For
joint (Fig. 2). The latter one, on the other hand, utilizes a new instance, if the range of GSI is considered lower than 20 to define a
category of weak rock structure namely; foliated/laminated/sheared poor rock mass, to what extent this range is able to differentiate the
introduced first by Hoek et al. [18]. Cai et al. [17] also suggested a quality between GSI¼20 and GSI¼10. Moreover, some engineers
3-D method for estimating block volume and used the RMi system tend to be uncomfortable with the system if some parameters
for rating the joint condition. Marinos et al. [6,7] have recently cannot be measured.
acknowledged that the quantifications do not work well in The lower part of the original GSI chart related to poor rock
tectonically disturbed rock masses in which the structural fabric masses is believed to be fairly improved to the extent of assisting
has been destroyed. In such rock masses, they recommend the use of the sensibility of such a quality index. The less sensible
the original qualitative approach based on careful visual characterization of the available GSI charts are attributed to
observations. Furthermore, they have denied the correlation of examples of a fair joint condition is assigned for a very poor rock
adjusted RMR and Q values with GSI, as it becomes meaningless in mass like a fully sheared zone or the line indicating GSI of 50 (fair
the range of weak, very weak and heterogeneous rock masses [6,7]. rock), signifying a fair and good joint condition, is embedded in
ARTICLE IN PRESS
692 R.R. Osgoui et al. / International Journal of Rock Mechanics & Mining Sciences 47 (2010) 690–697
Fig. 2. The quantitative GSI chart proposed by Sonmez and Ulusay [24].
the disintegrated category of rock structure or a sheared rock be noted that the boundary which defines the threshold of poor
mass structure as shown in Fig. 1. Such confusion could be, in rock mass is GSI ¼27 [15,16].
turn, overcome using the modified-GSI that acts as a supplement In order to aid in using the modified-GSI chart, it is necessary
for the original GSI charts without any incompatibility. that its input parameters suitably be known. Those parameters
are readily obtained from the geological visual impression,
judgments and the core-box logs. Five distinguishable stages in
3. Modified-GSI for poor and very poor rock masses estimating the modified-GSI is as follows: (a) Geological survey-
ing: overall field observation and core-box logging. (b) Determin-
3.1. The parameters used and the modified-GSI chart ing the type of broken structural domain (BSTR) to which poor
rock mass is mostly governed and corresponded. (c) Estimating
The approach recommended in this study is built on the important indicators of poor rock mass: ICR, volumetric joint
linkage between descriptive geological terms and measurable count (Jv), SR. (d) Determining the weathering condition of joint
field parameters; namely, SR and its relation with Jv, BSTR, intact surface correlated with filling. (e) Estimating GSI value for poor
core recovery (ICR) and joint surface properties. All required and very poor rock masses using the modified-GSI chart (Fig. 3).
parameters can readily be obtained from overall field observation
and core-box surveying.
For the purpose, a block in the matrix of 2 2 of GSI chart is
selected in terms of two axes showing the rock mass blockiness 3.2. The parameters associated with degree of jointing (blockiness
(interlocking/degree of joining) and joint surface conditions. As and interlocking)
shown in Fig. 3, the vertical axis of the matrix demonstrates the
degree of disintegration of rock mass. One quantitative parameter 3.2.1. Broken structural domain (BSTR)
SR and one qualitative parameter BSTR are embedded in this axis Broadly speaking, broken drill-core zones recovered from a
to assist the user to characterize the blockiness of the rock mass. very weak rock mass having a length greater than 25 cm are
The horizontal axis, on the other side, is assigned for the joint defined as the BSTR [25]. Various types of BSTR domains can be
condition based on ICR and Joint Condition Index (IJC). It should categorized into five groups, based on their degree of size and
ARTICLE IN PRESS
R.R. Osgoui et al. / International Journal of Rock Mechanics & Mining Sciences 47 (2010) 690–697 693
Fig. 3. The section of the GSI chart modified for poor and very poor rock masses.
composition. BSTR types included in the modified-GSI to define or intact, blocky, very blocky, blocky/disturbed/seamy, disinte-
the rock-mass structure are demonstrated in Figs. 4a–c. grated/decomposed and extremely crushed/sheared. In the exist-
Having recognized the BSTR type, one can distinguish the poor ing GSI charts, for example Sonmez and Ulusay [24], the intervals
rock mass from very poor rock mass. As can be indicated from between the structural categories were equally divided and
Fig. 4, the poor and very poor rock masses are best characterized selected as 100, 80, 60, 40, 20 and 0 (Fig. 2).
by BSTR types. It should be noted that the assignment of only one It is felt necessary here to adjust the previous SR intervals in
type of BSTR for a rock medium is not logical; rather, it is order to include the poor and very poor rock mass categories or
suggested to know the types of BSTR governing whole poor rock BSTR types. Therefore, the SR intervals are modified between 100,
mass. Otherwise, the assignment of only BSTR 1/2 for a domain 80, 60, 40, 20, 10 and 0 to adopt also the BSTR types. Also, the
cannot signify very poor rock mass; rather it stands for a soil corresponding values of the Jv (0.1, 0.3, 1, 3, 10, 30, 60, 100) are
material. For the very poor rock mass, it is supposed that a large taken into account in accordance with its definition by ISRM [27]
amount of medium is governed by the BSTR 1/2 while poor rock and Sönmez and Ulusay [24]. By doing this, a plausible way to
mass is dominantly governed by BSTR 3/4/5. quantitatively define the BSTR types is attained as illustrated in
This approach can also be regarded as one of the overall Fig. 5. The correlation between SR limits and corresponding Jv
observation methods for description of rock mass quality, but it is values, encompassing the wide range of the rock mass quality is
appropriate that BSTR be described through core-box logging given in Fig. 5.
rather than overall rock mass observation. BSTR 3/4/5 are The BSTR values assigned for poor and very poor rock masses
assigned for disintegrated–decomposed rock masses while BSTR lie between 0 and 20 as presented in Fig. 5. A very helpful
1/2 signifies the extremely crushed/sheared rock masses. It is distinction to notice is that the BSTR can be categorized in range
worth noting that the disc types of broken rock cores (Fig. 4d), of the Jv between 30 and 100; i.e. in the regions of disintegrated–
denoted to high horizontal stress field, can be categorized into the decomposed and extremely crushed/sheared rock mass to
BSTR type 4. calibrate the modified-GSI chart.
Fig. 4. (a–c) Various types of BSTR and (d) disc types of broken rock cores.
Disintegrated Extremely Crushed/ where Dn is the estimated number of joint sets as mentioned
Decomposed (BSTR 3 / 4/5 ) Sheared (BSTR 1/2)
above, and Sb is the average size of the block or rock pieces, which
represents average spacing of joints and estimated from the
Blocky
Massive / Intact Blocky Very Blocky selected pieces of the BSTR 3/4/5. The BSTR types, very important
Disturbed
100 indicator of poor and very poor rock masses used in the modified-
90 GSI chart (Fig. 3), are pertained to the Jv intervals.
80
Structure Rating (SR)
BSTR 1/2) is assumed to be less than 25 to satisfy the modified-GSI Aksoy [31] and based on the ratio between the Schmidt
requirements. hammer rebound values obtained from fresh and weathered
The upper part of modified-GSI chart (Fig. 3) is divided into rock surfaces as given in Table 1.
two categories; namely, poor and very poor. For ICR o25%, the
total rating of IJC varies between 0 and 16. A simple way of
determining IJC is simply given in Table 1. With reference to 4. Application of the modified-GSI: A case study from Turkey
Table 1, there are three alternatives to assign a rating for
weathering condition. The first and second alternatives use the 4.1. General
weathering conditions described and classified by Unal [25] and
ISRM [27]. The third alternative is to use the Weathering The 357-m-long Malatya no. 7 railroad tunnel being 5 m wide
Condition Index (W), which is suggested by Gokceoglu and and situated in the southeastern part of Turkey, was excavated in
Table 1
Joint Condition Index (IJC) ratings for using in the modified-GSI.
ARTICLE IN PRESS
696 R.R. Osgoui et al. / International Journal of Rock Mechanics & Mining Sciences 47 (2010) 690–697
1930 through a toe of a paleo-landslide slope. This tunnel with complex lithological units surrounding the tunnel where very low
overburden of 60–70 m passes close to the creek side of the slope permeability schist and metavolcanics taking part as blocks in the
descended to a river. The water table varied between 8 and 14 m matrix material indicate a medium permeability. However, in the
below the tunnel discharged to the river level. Since its existence of limestone blocks the permeability level increases due
construction, this horseshoe shape tunnel has struggled with to the existence of dense cracks and fractures. The landslide
severe stability problems due to the existence of very poor rock morphology and poor quality rocks are illustrated in Fig. 6.
mass, presence of voids, underground water seepage and high
level of tunnel deformation. A large amount of time dependent 4.3. Rock mass characterization by the modified-GSI
deformation developing through many years resulted in mis-
alignment of the tunnel for about 160 m of the tunnel length. This Core box logging of 50 exploratory boreholes drilled at
amount of large deformation (1.67 m) shifted the tunnel towards different locations of the landslide slope proved very poor rock
the creek side and formed an undulating tunnel shape; ultimately mass quality around the tunnel. The value of RQD in most cores
resulting in transit blockage. In 2003, during the rehabilitation of was 0 while the ICR was lower than 25% as illustrated in Fig. 4b.
the tunnel two collapses occurred inside the tunnel followed by a The total core recovery (TCR) was very low and varied between
failure in tunnel supports. As a result of this, quite a big sinkhole 17–28% in the deformed tunnel section. During drilling boreholes,
was developed at slope surface following a flow of 500 m3 washed soil, presence of voids and water loss were recorded. In
material into the tunnel. order to characterize such a poor rock mass, the modified-GSI was
realized to be more practical due to limitations and constrains of
4.2. Geology comparable rock mass classification systems such as RMR and Q-
system as outlined in Table 2. The characterization parameters
The landslide medium around the tunnel is a mixture type which were either difficult or in many cases impossible to be
having a complex geology. The rock mass around the tunnel quantified are indicated with N/A in Table 2.
consists of metavolcanics, schist, fractured limestone blocks, Rock cores could have been well defined in terms of BSTR
antigorite and radyolarite in patches. The matrix material consists types. In addition to the presence of BSTR types 1, 4 and 5, most of
of clay and schist having low swelling potential. Limestone blocks the rock cores were dominantly governed by BSTR 2 and 3. Taking
are generally closely jointed, highly fractured and weathered. into consideration the ICR lower than 25% for completely
There are also voids within such a poor quality rock mass. Such weathered rock (class 5, W5) and very poor joint condition
Fig. 6. Views from the rock mass to which the modified-GSI was applied: (a) landslide slope at mountain side and (b) highly weathered rock mass with very poor quality.
Table 2
Applicability of three rock mass characterization systems used for the Malatya railroad tunnel.
without filling for most rock cores, the value of IJC ranged between [8] Hack R. An evaluation of slope stability classification. In: Proceedings of
7 and 9. Accordingly, the dominant modified-GSI for such a very EUROCK 2002, Madeira, Portugal, 2002, pp. 3–32.
[9] Pantelidis L. Rock slope stability assessments through rock mass classification
poor rock mass of landslide slope varied between 13 and 18. systems. Int J Rock Mech Min Sci 2009;46:315–25.
[10] Palmström A. RMi—A rock mass characterization system for rock engineering
purposes. PhD. thesis, University of Oslo, 1995.
[11] Hoek E. Strength of rock and rock masses. ISRM News J 1994;2(2):4–16.
5. Conclusion [12] Hoek E, Brown ET. Practical estimates of rock mass strength. Int J Rock Mech
Min Sci 1997;34(8):1165–86.
The use of rock mass classification systems, in which RQD plays [13] Hoek E. Putting numbers to geology—an engineer’s viewpoint. Q J Eng Geol
1999;32:1–19.
a dominant role, is not suggested for poor and very poor rock [14] Tzamos S, Sofianos AI. A correlation of four rock mass classification systems
masses due to difficulty in its estimation (in most cases the value of through their fabric indices. Int J Rock Mech Min Sci 2007;44:475–95.
RQD for such rock masses is zero). Due to its considerable potential [15] Morales T, Uribe-Etxebarria G, Uriarte JA, de Valderamma F. Geomechanical
characterisation of rock masses in Alpine regions: the Basque Arc (Basque–
in characterizing a wide spectrum of rock mass qualities, a Cantabrian basin, North Spain). Eng Geol 2004;71:343–62.
supplementary approach was developed to aid at the GSI, [16] Osgoui R, Unal E. Rock reinforcement design for unstable tunnel originally
particularly in better characterizing poor rock masses. In this excavated in poor rock mass. In: Proceedings of the World Tunnel Congress,
Turkey, 2005, pp. 291–296.
study, efforts have been made to improve the GSI parameters to
[17] Cai M, Kaiser PK, Uno H, Tasaka Y, Minami M. Estimation of rock mass
better classify poor and very poor rock masses and to develop a deformation modulus and strength of jointed hard rock masses using the GSI
quantitative supplementary approach (modified-GSI) which is easy System. Int J Rock Mech Min Sci 2004;41(1):3–19.
[18] Hoek E, Marinos P, Benissi M. Applicability of the Geological Strength Index
to use. The modified-GSI, based on both geological visual
(GSI) classification for very weak and sheared rock masses: the case of the
impression and quantitative measures, may be integrated with Athens Schist Formation. Bull Eng Geol Env 1998;57(2):151–60.
the original GSI. The drawback of RQD in evaluating rock cores with [19] Marinos PG, Hoek E. GSI: a geological friendly tool for rock mass strength
the length of less than 10 cm (commonly in core box logging of estimation. In: Proceedings of the International Conference on Geotechnical
and Geological Engineering (GeoEng 2000), Melbourne, 2000, pp. 1422–1440.
poor rock masses) can be compensated with introduction of BSTR [20] Marinos P, Hoek E. Estimating the geotechnical properties of heterogeneous
and SR in the modified-GSI, which are qualitative and quantitative rock masses such as flysch. Bull Eng Geol Env 2001;60(2):85–92.
means, respectively, for the definition of the degree of jointing. The [21] Hoek E, Marinos P, Marinos V. Characterization and engineering properties of
tectonically undisturbed but lithologically varied sedimentary rock masses.
joint condition in the modified-GSI is well defined by means of ICR Int J Rock Mech Min Sci 2005;42(2):277–85.
and weathering condition since the evaluation of joint roughness is [22] Tzamos S, Sofianos AI. A correlation of four rock mass classification systems
meaningless for such a very highly fractured poor rock masses. through their fabric indices. Int J Rock Mech Min Sci 2007;44:477–95.
[23] Sonmez H, Ulusay R. Modifications to the geological strength index (GSI) and
The modified-GSI is valid for poor rock mass that behaves as a their applicability to stability of slopes. Int J Rock Mech Min Sci
homogenous isotropic medium. These charts can be used for poor 1999;36:743–60.
and very poor rock masses with GSI ranging between 6 and 27. [24] Sonmez H, Ulusay R. A discussion on the Hoek–Brown failure criterion and
suggested modification to the criterion verified by slope stability case
During characterization of poor rock masses, the modified-GSI chart
studies. Yerbilimleri (Earth Sciences) 2002; 26: 77–99, /www.yerbilimleri.
which could be integrated with existing GSI charts, i.e. in the case of hacettepe.edu.trS.
GSI427, the existing GSI charts mentioned earlier should be used. [25] Unal E. Modified rock mass classification: M-RMR system. In: Milestone in
Rock Engineering: A Jubilee Collection, ZT Bieniawski, ed. Rotterdam:
Balkema, 1996, pp. 203–223.
References [26] Hoek E, Kaiser PK, Bawden WF. Support of underground excavation in hard
rock. Rotterdam: Balkema; 1995.
[27] ISRM. The Complete ISRM Suggested Methods for Rock Characterization,
[1] Bieniawski ZT. Engineering classification of jointed rock masses. Trans S Afr Testing and Monitoring: 1974–2006. In: Ulusay R, Hudson JA, editors.
Inst Civ Eng 1973;15:335–44. Suggested methods prepared by the ISRM commission on testing methods,
[2] Bieniawski ZT. Engineering rock mass classification. New York: Wiley; 1989. International Society for Rock Mechanics, Ankara, 2007.
[3] Barton N, Lien R, Lunde J. Engineering classification of rock masses for the [28] Palmström A. The volumetric joint count—A useful and simple measure of
design of tunnel support. Rock Mech Rock Eng 1974;6(4):189–239. the degree of rock mass jointing. In: Proceedings of IAEG Congress, New
[4] Barton N. Some new Q value correlations to assist in site characterization and Delhi, 1982, pp. 221–228.
tunnel design. Int J Rock Mech Min Sci 2002;39:185–216. [29] Ozkan I. Modified rock mass rating (M-RMR) system and roof behaviour
[5] Palmström A. Measurements of and correlations between block size and rock model. PhD thesis, Middle East Technical University, Ankara, 1995.
quality designation (RQD). Tunnell Underge Space Tech 2005;20:362–77. [30] Ulusay R, Ozkan I, Ünal E. Characterization of weak, stratified and clay
[6] Marinos V, Marinos P, Hoek E. The geological strength index: applications and bearing rock masses for engineering applications. In: Myer LR, Cook NGW,
limitations. Bull Eng Geol Env 2005;64(1):55–65. Goodman RE, Tsang CF, editors. Proceedings of Fractured and Jointed Rock
[7] Marinos P, Marinos V, Hoek E. Geological Strength Index (GSI). A Masses Conference, Lake Tahoe, California. Rotterdam: Balkema; 1995, pp.
characterization tool for assessing engineering properties for rock masses. 233–240.
In: Romana M, Perucho A, Olalla C, editors. Underground Works under Special [31] Gokceoglu C, Aksoy H. New approaches to the characterization of clay-
Conditions, 2007, pp. 13–21. bearing, densely joints and weak rock masses. Eng Geol 2000;58:1–23.