SLOPE STABILITY ANALYSIS (REVISED, APRIL 06)

TABLE OF CONTENTS 1.0 2.0 3.0 4.0 5.0 6.0 INTRODUCTION ............................................................................................. METHOD OF SLICES .................................................................................... FACTOR OF SAFETY ................................................................................. DRAINED/UNDRAINED STABILITY ANALYSIS ......................................... UNCERTAINTIES AND ASSUMPTIONS ...................................................... SOIL SHEAR STRENGTH PARAMETERS .................................................. 6.1 Consolidated Drained Triaxial Test (CD TX) ....................................... 6.2 Consolidated Undrained Triaxial Test (CD TX)................................... 6.3 Unconsolidated Undrained Triaxial Test (UU TX) ............................... APPLICATION ................................................................................................ 7.1 Fills On Competent Ground ............................................................... 7.2 Fills on Soft Clay ................................................................................. 7.3 Natural Slopes .................................................................................... 7.4 Cut Slopes .......................................................................................... 7.5 Reinforced slopes ................................................................................ SUBMERGED SLOPES .............................................................................. 8.1 Effective Stress Analysis .................................................................... 8.2 Total Stress Analysis .......................................................................... EXAMPLES ................................................................................................. 9.1 Embankment on Strong Foundation ................................................... 9.1.2 Long Term Slope Analysis ....................................................... 9.1.3 Observations ............................................................................. 9.2 Embankment on Weak Foundation .................................................... 9.2.1 Short Term Slope Analysis ...................................................... 9.2.2 Long Term Slope Analysis ....................................................... 9.2.3 Observations ............................................................................. 9.3 Embankment on Foundation with Weak Stratum ............................... 9.3.1 Short Term Slope Analysis ...................................................... 9.3.2 Observations ............................................................................. 9.4 Excavated Slope .................................................................................. 9.4.1 Short Term Slope Analysis ...................................................... 9.4.2 Long Term Slope Analysis ........................................................ 9.4.3 Observations ............................................................................. 9.5 Back Analysis ....................................................................................... HOW RELIABLE ARE SLOPE STABILITY ANALYSIS ............................... FINITE ELEMENT METHOD FURTHER READINGS

7.0

8.0

9.0

10.0 11.0 12.0

SLOPE STABILITY ANALYSIS Ir Simon Tan Siow Meng, Director, SSP Geotechnics Sdn Bhd 1.0 INTRODUCTION There are many forms of slope failures. Slope stability analysis is an analytical tool for assessing the stability of a slope by using a simple failure model in analysis. This note is limited to slopes with homogenous material, soils or highly weathered, fractured rocks behaving like soils and for plain strain (2-D) problems only. Suction and reinforced slopes and rock slopes are not addressed. A few methods can be used in the slope stability analysis: a. Assumed failure surface methods. Via method of slices. Most common methods. Many computer codes available such as STABL, UTEXAS, SLOPE/W b. Simple closed-form solutions. Limited to simple geometry such as vertical cuts and infinite slopes. c. Advanced numerical methods. Such as FEM. For complex cases where instability includes combined failures structural members and soils. They enable deformation analysis as well. The first two methods satisfy either force or moment equilibrium or both. They do not stress equilibrium and strain compatibility. Advance numerical methods satisfy both of these conditions globally. 2.0 SIMPLE EQUATIONS AND METHOD OF SLICES 2.1 SIMPLE EQUATIONS (a) For a cohesionless (c=0) soil, the factor of safety (FS) can be estimated by: FS=tan /tan (b) For a vertical cut in =0 soil, FS=3.83c/H (c) For embankment on =0 soil FS=5.53c/H Note that the strength of the embankment is not considered therefore conservative. 2.2 METHOD OF SLICES Table 1 shows equilibrium of force or moment achieved in the various assume failure surface methods using method of slices in calculation of the factor of safety (FS). Note that tensor crack is ignored thus overestimate FS. being the slope angle.

Table 1 : Methods of slope stability analyses Factor of Safety (FS) Method (1) Ordinary (Swedish or USBR) (2) Bishops Simplified (3) Janbus Simplified (4) Janbus Generalised (5) Spencer (6) Morgestern-Price (7) Lowe-Karafiath (8) Corps of Engineers Force Equilibrium Yes Yes Yes Yes Yes Yes Moment Equilibrium Yes Yes Yes Yes Interslice Force Assumption (H=horizontal, V=vertical) Ignore both H, V V ignored, H considered V ignored, H considered Both H, V considered Both H, V considered Both H, V considered Both H, V considered Both H, V considered

The differences in the various methods are the assumptions on inter slice forces. For example, the Ordinary Method ignores inter slice forces (V=H=0), Simplified Bishop Method assumes inter slice forces are horizontal (V=0, H>0), Spencers Method assumes all inter slice forces are parallel (V>0, H>0) with an unknown inclination which is computed through iterations, Morgenstern and Price method relates the shear force, V to the normal force, H where V= f(x) H. The first three methods are not applicable to non-circular slips.

H V

Note that both Janbus methods satisfy only force equilibrium. FS is quite dependent on the side force assumption and can vary up to 20% for c, soils! A comparison of methods (1) to (6) was attempted by Fredlund & Krahn 1977. It shows that the FS calculated by Simplified Bishop, Morgenstern-Price and Spencers methods are very similar (difference <0.1%) while the Ordinary Method tends to underestimates the FS by a few percents. Janbus Simplified and Generalised methods tend to overestimate or underestimate the FS compared to Morgenstern-Price and Spencers methods by up to few percents. Duncan & Wright (2005) found that Janbus Simplified method underestimates the FS. 3.0 FACTOR OF SAFETY The definition of the Factor of Safety (FS) is expressed as: a) Fr / Fd b) Mr / Md c) Hc / H d) s / Resisting Force / Driving Force Resisting Moment / Driving Moment Critical Height / Slope Height Available shear stress/shear stress at equilibrium

The last definition is the most widely used. The required FS depends on the consequences of losses in terms of property, lives and cost of repair in the event of slope failure. FS is also dependent on the reliability of design parameters. The minimum FS recommended by various codes or guidelines vary from 1.2 to 1.5 as follows: a) BS6031 : 1981 FS = 1.3 to 1.4 for first time slide. FS = 1.2 for a slide with pre-existing slip surface. b) NAVFAC DM 7.1 FS = 1.5 for permanent loading condition. FS = 1.15 to 1.2 for transient load such as earthquake. c) JKR Road Work FS = 1.2 for unreinforced slope and embankment on soft ground. FS = 1.5 for reinforced slope. d) Geoguide, Hong Kong FS = 1.0 to 1.4 for new slope depending on risk. FS = 1.0 to 1.2 for existing slope depending on risk. e) Britain, National Coal Board, 1970 FS = 1.5 if risk of danger to person/property, peak strength. FS = 1.35 if risk of danger to person/property, residual strength. FS = 1.25 if no risk of danger to person/property, peak strength. FS = 1.15 if no risk of danger to person/property, residual strength. f) Canada, Mines Branch, 1972

FS = 1.5 if risk of danger to person/property, peak strength. FS = 1.30 if risk of danger to person/property, residual strength. FS = 1.30 if no risk of danger to person/property, peak strength. FS = 1.20 if no risk of danger to person/property, residual strength. Table 2, 3 and 4 are extracted from Hong Kong Geotechnical Manual for Slopes. They give a good guideline on the FS required for different cases. Risks are categorized into negligible, low and high (Table 2). The FS for new slopes (Table 3) and for existing slopes (Table 4) differs as the economic risk is taken into consideration for the design of new slopes. Natural slopes need not meet the FS recommended for New Slopes provided that the slope is undisturbed and there is no evidence of instability or severe surface erosion. Table 2 : Risk Categories (Geotechnical Manual for Slopes, 1994)

Table 3 : Safety factors for New Slopes (Geotechnical Manual for Slopes, 1994)

Table 4 : Safety factors for existing slopes (Geotechnical Manual for Slopes, 1994)

Figure 1 : Safety factors for natural slope with minor earthworks

Comparatively minor cut/fill

Localised slope FS>1.4

Natural slope Overall slope FS say around 1.2 and should be slightly improved after earthwork.

4.0 DRAINED/UNDRAINED STABILITY ANALYSIS Drained or undrained stability analysis is based on the soil mechanics principles. Typically, undrained analyses are critical in loading of soft clays, while drained analyses are critical in unloading of stiff clays (e.g. deep excavation) or cases involving cohesiveless (sandy) soils. Undrained analyses are usually performed using total stress approach particularly where piezometric data are sparse or unreliable. Effective stress approach may be used if reliable piezometric data are available. Drained analyses should always be performed with effective stress approach and steady state piezometric levels. The shear strength of soils is expressed as: - Total shear stress: = c+ tan - Effective shear stress, = c+ tan , where = -u Note that the effective stress cannot be measured, it can only be calculated whereas is the body stress generated by the mass acted by gravity. The pore water pressure u is a neutral stress. It cannot support shear stress and it acts equally in all direction. 5.0 UNCERTAINTIES AND ASSUMPTIONS Some of the basic assumptions and uncertainties are: a) In total stress: it is a very big assumption that u and the stress level is the same for the test specimen and in the field. b) In effective stress: It is difficult to determine the field u. Effective shear strength is difficult to determine in-situ. It is usually determined in the laboratory.

c) Anisotropy of soil material (Fig. 3) d) Principal stress axis (Fig.4) e) Peak vs residual strength (Fig. 5) f) Change of strength with time

Figure 3: Anisotropy

Weak planes

(a)

(b)

Figure 4: Principal stress axis

1 = Major principle stress 3 = Minor principle stress

1 A 1 3 1 3 3

Figure 5: Peak vs residual strength, Undisturbed vs remolded

Shear stress, (kPa)

Peak Shear Strength Residual Shear Strength

Axial strain,

6.0 SOIL SHEAR STRENGTH PARAMETERS Shear strength can be obtained from the laboratory tests, field tests, vane tests and correlation with SPTN or qc. Back analysis can also be conducted to back calculate the strength parameters. The soil in the analysis may be assumed as residual soil only because the inclusion of boulders and geological structure in the analysis may complicate the model. Some of the common lab tests carried out to obtain the following parameters are: a) Total stress parameters i) Unconsolidated Undrained Triaxial Test (UU TX), also referred to as the Q-test by Casagrande since the sample was loaded to failure much more quickly than the S-Test. ii) Unconfined Compression Triaxial Test (UCS) it is similar to the UU TX but the confining pressure, c = 0 b) Effective stress parameters i) Consolidated Drained Triaxial Test (CD TX), also referred to as the STest. c) Total and effective stress parameters i) Consolidated Undrained Triaxial Test (CU TX) with measurement of u value, also referred to as the R-Test. 6.1 Consolidated Drained Triaxial Test (CD TX) The CD TX is also known as the slow test or S-Test because it may take a few days for the completion of the test. It is not practical to carry out this test on clay due to cell leakage problem and the completion of the test would be too slow. There is less concern on the sample disturbance because soil structure is almost remolded near the failure plane. Figure 6 shows the stress conditions in the CD TX test. Note that at all times during the CD test, the pore water pressure, u is essentially zero. This means that the total stresses in the drained test are always equal to the effective stress.

The above behavior can be explained from the void ratio, e versus effective stress, curve in Figure 7a. When the vertical stress level is increased to reach point C on the virgin compression curve, rebounded to D, reloaded back past E and C and on to F, it would have the strength as shown at point F in Figure 7b. Note that it is now back on the virgin compression curve. The effects of the rebounding and reconsolidation have been in effect erased by the increased loading to point F. Once the soil has been loaded well past the preconsolidation pressure, it no longer remembers its stress history. Figure 7b also shows that E has greater strength than B although they are tested at exactly the same effective consolidation stress, which is due to the fact that E is at lower water content and void ratio, thus is denser than B as shown in Figure 7a.

Figure 6: Stress conditions in the consolidated-drained triaxial test (Holtz and Kovacs, 1981)
TOTAL,

NEUTRAL, u

+
vc =

vc At the end consolidation, hc = vc in CiD hc 0

vc

hc =

vc

~0

+ hc = hc

compression

f = ( 3 At failure
hc

Stress applied very slowly so that no u vf = vc + f = 1 f hc = hc = 3 f

Figure 7: (a) Compression curve; (b) Mohr failure envelope (DEC) for overconsolidated clay (Holtz and Kovacs, 1981)

(a)

(b)

6.2 Consolidated Undrained Triaxial Test (CU TX) The CU TX is also known as the rapid test or R-Test. In practice, to ensure 100% saturation, which is necessary for good measurements of the pore water pressure, a back pressure is applied to the pore water. Total, neutral and effective stress conditions in the specimen during the several phases of the CU test are shown in Figure 8. For a clay specimen, rate of loading should be slow enough so that the u is same in the whole specimen. This may not be practical for field testing. Figure 8: Conditions in specimen during a consolidated undrained triaxial test (Holtz and Kovacs, 1981)
TOTAL,

NEUTRAL, u

EFFECTIVE,

vc At the end consolidation, hc = vc ,in CiU, usually back pressure applied vc During axial compression hc +
u

vc = vc hc 0 hc = hc

vc = vc + + u = 1 hc = hc + u = 3

f = ( 1 3 ) f vc At failure hc
+ u

v f = vc + f + u = 1 f hc = hc + u = 3 f

10

The pore pressure vs strain curves in Figure 9 illustrate what happens to the pore pressure during shearing. The normally consolidated specimen develops positive pore pressure. In the overconsolidated specimen, after a slight initial increase, the pore pressure goes negative with respect to the back pressure uo because the specimen tend to expand during shear. Another distinctive behavior is the principal effective stress ratio, 1/ 3. Note that the ratio peaks early for the overconsolidated specimen, to normalize the stress behavior with respect to the effective minor principal stress during the test. Figure 9: Typical -, u, and 1/3 curves for normally and overconsolidated clays in CU TX test (Holtz and Kovacs, 1981)

For normally consolidated clay (Figure 10), because the specimen develop positive pore pressure during shearing, the Mohr circle diameter for effective and total stress are the same. The envelope for normally consolidated clay passes essentially through the origin, and thus, for practical purposes c can be taken to be zero which is also true for the total stress c parameter. T is less than and often is about one-half of . However, it is different for overconsolidated clay (Figure 11). The c and c parameter is greater than zero and the value for effective stresses are greater than that of the total stresses. The effective stress circle at failure is shifted to the right of the total stress circle. Figure 10 : Mohr circle for normally consolidated clay (Holtz and Kovacs, 1981) Figure 11: Mohr circle for overconsolidated clay (Holtz and Kovacs, 1981)

11

6.3. Unconsolidated Undrained Triaxial Test (UU TX) The unconsolidated undrained triaxial test is also known as the quick test or Q-Test because the sample is loaded to failure in about 10 to 20 minutes. Total, neutral and effective stress conditions in the specimen during the several phases of the UU test are shown in Figure 12. The sample must be 100% saturated so that even when confining pressure is applied, no consolidation can occur. Figure 12: Conditions in the specimen during unconsolidated undrained triaxial test (Holtz and Kovacs, 1981)
TOTAL, 0 Immediate after sampling, before applying c 0 -u r

NEUTRAL, u

EFFECTIVE, vc = u r

hc = u r

c After applying c ,(S=100%) c

-u r + uc u c= c

vc = c + u r - c = ur hc = c + u r - c = ur

c During axial compression c

-u r + c + u

v = c + + u r- c+ u h = c + ur
c+ u

f = ( 1 3 ) f c At failure c
-u r + c +uf

v f = c + f + u r- c + u f = 1 f hc = c + u r
-

= 3 f

c+ u f

The Mohr failure envelopes for UU tests are shown in Figure 13a for 100% saturated clays. All specimens are presumably at the same water content, hence they will have the same shear strength since there is no consolidation allowed. Therefore, all Mohr circles at failure will have the same diameter and the Mohr failure envelope will be a horizontal straight line and the slope T is equals to zero. The intercept defines the total stress strength parameter c. A high confining pressure, c is advisable in order to ensure 100% saturation However, for partially saturated soils, a series of UU test will define an initially curved failure envelope (Figure 13b) until the clay becomes 100% saturated due simply to the confining pressure alone. Then, the Mohr failure envelope becomes horizontal as in the case of 100% saturated clay. A high confining pressure is advisable in order to ensure 100% saturation is achieved.

12

Unconfined Compression Test (UCT) (Figure 14) is done similarly to UU TX except that there is no cell pressure applied. The results from UCT is similar to that of the UU TX when the specimen is 100% saturated, there is no weak plane in the specimen, specimen used is clay and the specimen must be sheared rapidly to failure. The relationship between the total shear strength, c and the Unconfined Compression Strength (UCS) can be defined as below: cu = 45 + /2

The value of cu is usually half of the value of UCS Figure 13: Mohr failure envelopes for unconsolidated unconfined triaxial test (Holtz and Kovacs, 1981) Figure 14: Unconfined compression test (Holtz and Kovacs, 1981)

7.0 APPLICATION Table 5 summarizes the application of the total stress analysis and the effective stress analysis for various cases of slope analysis. Table 5 : Undrained versus drained analyses Soil Type Cohesiveless Soils Cohesive Soils Composite Material Undrained Analysis Earthquake Dynamic Receiving Fills, Short Term Short Term Drained Analysis Normal Cases Excavation, Long Term Long Term

   

(1) (2)

(a)

(b)

13

7.1 Fills On Competent Ground (A) Cohesiveless Fills on Competent Ground For cohesiveless fills on competent ground, drained analysis is used. The slope failure curve in Figure 15 indicates a shallow sliding or surface unraveling. The slope is also susceptible to surface erosion. The analysis uses the infinite slope method: FS = tan / tan (3) When loose saturated sands are subjected to strains and shocks, there is a tendency for the sand to decrease in volume. This tendency causes a positive increase in pore pressure which results in a decrease in effective stress within the soil mass. Once the pore pressure becomes equal to the effective stress, the sand loses all its strength, and is said to be in a state of liquefaction. Liquefaction of saturated sands (Figure 16) may occur when a slope is subjected to large strain (steepening of riverbank), dynamic loading (pile driving), cyclic loads (train traffic), piping and earthquake. Figure 15: Slope failure for cohesiveless fills on competent ground

Figure 16: Liquefaction of loose sand adjacent to a waterfront (after Casagrande, 1975)

14

(B) Cohesive Fills on Competent Ground for Short Term / End of Construction Undrained analysis is used for short term cohesive fills on competent ground (Figure 17). Unconsolidated Undrained Triaxial Test (UU TX) or Q-Test shall be carried out on compacted samples. The pore water pressure, u is not considered in the analysis.

Figure 17: Slope failure for cohesive fills on competent ground (Short term)

(C) Cohesive Fills on Competent Ground for Long Term/Steady State Flow State Drained analysis is used to analyze the slope failure for long term cohesive fills on competent ground (Figure 18). The lab tests suitable to be carried out on the compacted sample to obtain the effective stress parameters are the Consolidated Drained Test (S-Test) or Consolidated Undrained (CU TX). Pore water pressure, u has to be established for the analysis. Figure 18: Slope failure for cohesive fills on competent ground (Long term)

(D) Cohesive Fills on Competent Ground under Rapid Drawdown For cohesive fills on competent ground under rapid drawdown (Figure 19), total stress analysis can be used. Hence the pore water pressure within the soils is ignored. Consolidated Undrained Test (CU TX) is carried out on the soil samples to obtain the total stress shear strength. Other than the soil shear strength, the FS depends on the drop in water level. Note that the free water contributes to slope stability by acting as a counter-weight. If effective stress analysis is used, the challenge is to accurately establish the piezometric line within the soil. Comprehensive transient seepage analysis involving water drawdown time will be required. Assuming the piezometric line stick to the slope surface is a conservative approach, and is not practical sometimes.

15

Water level before draw

Figure 19: Slope failure for cohesive fills on competent ground under rapid drawdown

7.2 Fills on Soft Clay Short term stability is more critical for fills on soft clay (Figure 20). Effective parameters, i.e, c and , should be used if the fills are cohesiveless material, whereas total parameters, c and , are used for cohesive fills. Unconsolidated Undrained Test (UU TX or Q-Test) is carried out on the foundation clay to obtain the value for su = cu and = 0. Alternatively, field or lab vane tests can be carried out to obtain the total stress parameters. However, for staged construction, Consolidated Undrained test (CU TX) is the preferred laboratory test for the foundation clay. The main concern for fills on soft clay is the progressive failure. Chirapuntu & Duncan 1975 suggested the application of reduction factors, Rf for the foundation soil and Re for the embankment soil (Figure 21). The required FS for the slope should be more than 1 after the strength reduction. A simple method to estimate the critical height of fills on soft clay was given by Taylor, 1937 (Figure 22) Hc = Ns su / Ns = 5.52 when slope angle <53 and depth of clay is infinite slope angle! Figure 20: Slope failure for fills on soft clay Figure 22: Critical height for fill on soft clay (4) (5)

Note that equation implies that when the critical height is reached, the slope will fail regardless of the

Soft clay

16

Figure 21: Application of strength reduction factor for fill on soft clay

Re Rf

7.3 Natural Slopes Natural slopes are slopes that have been in existence for a long time (Figure 22). Hence drained analysis is used. Usually Consolidated Drained (CD TX) and Consolidated Undrained (CU TX) laboratory tests are carried out on the soil sample. Standpipe and seepage analysis can be used to obtain the groundwater table (GWT). However it is difficult to obtain reliable soil strength parameters via laboratory or in particular via field test. It is often not practical to measure GWT on field due to time and budget constraint. Back analyse should always be carried out, by assuming FS or slightly above 1 and under sensible GWT, to obtain the soil strength parameters and compared with those obtained from testing. 7.4 Cut Slopes The long term stability of a cut slope (Figure 22) is more critical. Drained analysis is carried out. Consolidated Drained (CD TX) and Consolidated Undrained (CU TX) tests are used to obtain strength parameters and compared with parameters obtained from back analysis as mentioned above. During design stage, standpipe can only measure the existing groundwater table (GWT). The GWT after cut can only be predicted with seepage analysis and verified by standpipes after the cut slope has been completed.

17

Figure 22 : Slope before cut

Figure 23 : Slope after cut

7.5 Reinforced Slopes Reinforcement can be introduced to improve the stability of slopes. Two methods have been used for limit equilibrium analyses of reinforced slopes: Method A: The reinforcement forces used in the analysis are allowable forces, Ta, and are not divided by the factor of safety calculated during the slope stability analysis, only the soil strengths are. FS = soil resisting moment / (overturning moment reinforcement moment), or FS = shear strength / (shear stress required for equilibrium reinforcement resistance) Method B: The reinforcement forces used in the analysis are limit forces, Tlim, and together with the soil strengths, are divided by the factor of safety calculated during the slope stability analysis. FS = (soil resisting moment + reinforcement moment) / overturning moment, or FS = (shear strength + reinforcement resistance) / shear stress required for equilibrium Where Ta = Tlim/FS. FS varies from 1.5 to 2 depending on uncertainties in Tlim and load in reinforcement, and consequences of failure. for reinforcement. Tlim = Tult/(Fcr Fid Fd), short term ultimate load divide by the strength reduction factor, installation damage factor times the deterioration factor. These factors are material dependent. The product of the three factors ranges from 2 for Polyester under an ideal case to more than 15 for Polypropylene or Polyethylene under a less favourable situation! For steel, Tlim = Acfy, cross section of the steel after allowance for corrosion time the yield strength of the steel. The example given in the figure below can be used to determine which method a computer programme uses. It is found that STABL 6 uses Method B. SLOPE/W 2004 uses Method A. Method A is preferable as it allows different safety factors on soil strength and reinforcement which have different sources of uncertainties. Figure 23a : FS are 2.19 and 1.72 respectively for a computer programme using Method A and Method B respectively.

18

8.0 SUBMERGED SLOPES 8.1 Effective Stress Analysis In effective stress analysis, a submerged slope can be modelled by specifying a piezometric surface and applying an equivalent load on the slope to reflect the free water condition (Figure 24a). Alternately, the free water is modelled as a material with unit weight w, but has no strength whatsoever (Figure 24b). The third method is by using the submerged unit weight, for soil below the water table and for all soil above the water table (Figure 24c). 8.2 Total Stress Analysis In total stress analysis, the pore water pressure within the soils is ignored, only the free water body is considered. Free water can be modelled as a material with unit weight, w but has no strength (Figure 25a) or can also be modelled as an equivalent loading on the slope (Figure 25b). Since the free water body acts as counter weight, a submerged slope is always more stable than when it is dry in total stress analysis. Figure 24: Effective stress analysis of submerged slopes
(a)

Figure 25: Total stress analysis of submerged slopes

w,c=0,=0 (a)

w,c=0,=0

(b)
(b)

(c)

19

9.0 EXAMPLES 9.1 Embankment on Strong Foundation 9.1.1 Short Term Slope Analysis During the end of the embankment construction, a short term analysis is done on the slope. The soil parameters are assumed to be undrained if the soil is clayey. Shear strength can be obtained from the Unconsolidated Undrained Triaxial test or known as the Q-Test. Referring to Figure 26a, using soil parameter, c=50kPa and =10, the FS yielded by the analysis is 3.96. With the introduction of a tension crack, Dc, on the slope (Figure 26b), the analysis carried out is similar to the above but the termination point of the slip circle should end at the end of the crack. The length of the tension crack can be calculated using: Dc = 2cm/[ tan (45-m/2)] cm = c/FS m = tan (tan /FS) By introducing a tension crack, the FS is slightly lowered to 3.87 compared to FS from the above case. Figure 26a: Slip surface for embankment on strong foundation for short term analysis
FS=3.96
-1

(6) (7) (8)

Figure 26b: Slip surface for embankment on strong foundation with tension crack for short term analysis
FS=3.87

Tension crack

9.1.2 Long Term Slope Analysis For a long term analysis of an embankment on strong foundation, the effective stress analysis is used. The parameters can be obtained from the Consolidated Undrained (R-Test) or the Consolidated Drained (S-Test). Referring to Figure 27 and using soil parameters, c=5 kPa and =20 kPa, the FS calculated from the analysis is 1.43. However, if we assume that c=0 and =20, the FS calculated would be a mere 0.91. Observe that when c=0, the equation for the FS value (FS = tan / tan ) is similar to that of Eq 3 for the infinite slope.
FS=1.43 FS=3.96

Figure 27: Slip surface for embankment on strong foundation for long term analysis

20

9.1.3 Observations The following observations can be made from the previous example: a) Both short term and long term analyses are required. b) For the present example, long term condition govern but for higher slopes (such as dam embankment) the short term stability is more critical. c) The introduction of tension cracks (when c>0) lowers the FS by a few %. d) c=0 results in infinite slope 9.2 Embankment on Weak Foundation 9.2.1 Short Term Slope Analysis In the short term slope analysis, the properties for the foundation clay in Figure 28 is assumed to be 15 kPa for the first half and it increases with depth thereof (Figure 28b). The strength of the sand fill is c=0, =35 and the FS of the slope is 1.16. Now, let us assume that the sand fill does not contribute to the FOS, hence, the parameters are assumed to be c=0 and =0 (Figure 29). The parameters for the foundation clay remains unchanged and the FS of the slope from the analysis is 0.83. Alternatively, we can also assume the sand fill as an equivalent surcharge and the foundation clay remains unchanged (Figure 30). After carrying out the analysis, the FS is found to be 1.37! Figure 28: Slip surface for embankment on weak foundation (short term analysis)

(b)
Figure 29: Slip surface for embankment on weak foundation (short term analysis, no strength for sand) Figure 30: Slip surface for embankment on weak foundation (sand as surcharge)

21

9.2.2 Long Term Slope Analysis In the long term analysis, the foundation clay has been drained and the new parameters are c=0 and =23. The unit weight, of the clay increases as well due to consolidation. The FS of the slope is now 1.44 under drained condition for the clay (Figure 31). Figure 31: Slip surface for embankment on weak foundation (Long term analysis)

9.2.3 Observations The following observations can be made from the example of an embankment on weak foundation: a) Normally, FS increases with time as foundation clay gain in strength when it consolidates. b) Unless embankment fills govern stability, then long term condition may be critical as explained in the example for an embankment on strong foundation. c) It is not valid to simulate embankment fills as equivalent surcharges! 9.3 Embankment on Foundation with Weak Stratum 9.3.1 Short Term Slope Analysis Figure 32 illustrates an embankment founded on a weak clay stratum sandwiched between two layers of foundation sand. Assuming the strengths for the sand are c=0, =30 and the strength for the clay is c u=12kPa. The FS obtained from the slope analysis using a circular shear surface is 1.15. Alternatively, a non-circular shear surface (Figure 33) can also be used when carrying out the analysis. Applying the same soil parameters as above, the FS is reduced to 0.93.

22

Figure 32: Circular slip surface for embankment on foundation with weak stratum

Figure 33: Non-circular slip surface for embankment on foundation with weak stratum

c= 0, =30 o

cu = 12kPa

c= 0, =30 o

cu = 12kPa

9.3.2 Observation When analyzing the slope stability for an embankment on foundation with weak stratum, non-circular shear surface should be used when appropriate. 9.4 Excavated Slope 9.4.1 Short Term Slope Analysis For an excavated slope (Figure 34a), the short term condition of the slope is usually less critical. However, if the short term condition is to be analysed, use c and value from the Consolidated Undrained Triaxial Test but the pore pressure, u, is ignored. For a case whereby the clay is of a constant value, cu and =0 (Figure 34b), the slip circle will extend to a full depth if <53 (Taylor 1948). It means that the deeper the clay, the lower the FS. Figure 34: Slip surface for excavated slope
(a) (b)

CLAY Cu=CONS

9.4.2 Long Term Slope Analysis Refer to Figure 35 for a long term slope analysis with soil strengths of c=5 kPa and =25. No groundwater is introduced in this case. The FS obtained from the analysis is 2.05. However, when a piezometric line is introduced (Figure 36) for the same soil properties, the FS was reduced by 24% compared with no water to a value of 1.56.

23

Figure 35: Slip surface for excavated slope without ground water

Figure 36: Slip surface for excavated slope with groundwater

9.4.3 Observations a) Normally, long term stability condition is more critical for cut slopes. That is why some cut slopes failed years after construction. b) FS is significantly influenced by the presence of groundwater. 9.5 Back Analysis The example is taken from the back analysis of a landslide incident that toppled a power transmission pylon in the end of 2004. The photograph was taken during the progress of repair works by means of soil nails and shotcrete as can be seen near the crest of the slide on the right hand. The steps of back analysis were as follows: Estimate the subsoil condition using soil investigation as shown in Fig. 36a. Determine mode of failure based on the subsoil condition. Failure mode can be circular, noncircular, deep or superficial. Here the slip should be shallow and failure was assumed occurring within the granite residual soils. The difference between FS computed by noncircular and circular was not significant. Hence circular mode was assumed for ease of analysis as shown in Fig. 36b. Plot shear strength envelopes based on various ground water conditions: the worst case being full groundwater condition while the best case was when the slope was fully dry. The possible range of shear strength for the soils are superimposed on the plot as shown in Fig. 36c. Select the most like range of shear strength for the design of repair works.

24

25

Fig. 36b

Fig 36c
50

SEGALI PYLO N, SLO PE FAILURE

40

C o h e s io n C , k P a

30

20

P ossible R ange of S oil S trength F ully S aturated

10

P artially S aturated

D ry 0 0

1 0

20

30 Frictio n A n g le , (d e g )

40

50

60

26

10.0 HOW COMPETENT ARE YOU IN SLOPE STABILITY ANALYSIS?

The factor of safety predicted by various consultants, academic institutions and government departments in Sydney based on Mostyn & Small 1987 are summarised in Table 6. The slope is given in Figure 37. The following cases were analysed a) Case 1 Fixed circle, 30m radius, centred at (29.5, 20), soil parameters in Figure 37 b) Case 2 As Case 1 but Soil C: c=15kPa, = 18o c) Case 3 As Case 1 but Soil C: c=300kPa, = 0 d) Case 4 - Automatic search, soil parameter as in Case 1 e) Case 5 - Non-circular failure surface defined by (59.5,20), (49.5, -2.5), (9.5, -2.5) & (7, 0). Soil parameters as in Case 1. Figure 37 : Slope Section
25 20 15

Soil parameters for Case 1: Soil, dry, sat, c , A, 16, 20, 5kPa, 25o B, 18, 21.5, 0kPa, 40o C, 16, 20, 10kPa, 15o Unit wt. in kN/m3

H e ig h t(m )

10 5 0 -5 -1 0 -1 5 0 10 20

S oil B

S oil A

S oil C

D ista n ce (m )

30

40

50

60

70

Table 6 : Calculated Factors of Safety (data from Mostyn & Small, 1987)

Case 1, fixed circle 2, As 1 but c & of Soil C increased 3, As 1 but c=300kPa, =0 for Soil C 4, auto. search, soils as in 1 5, Fixed non circular, soils as in 1

Mean 0.92 1.12 4.41 0.85 1.10

Max. 1.04 1.25 5.14 1.00 1.71

Min. 0.77 0.91 3.32 0.75 0.84

No. of replies 16 16 16 14 11

Staff of SSPG 0.85* 1.05* 4.16* 0.78*/0.74** 0.84*

* by a trainee, using STABL6, simplified Bishop method for cases 1 to 4, Janbus method for case 5. **value in bold is by PLAXIS. Performed by a 2-yr experience engineer with MS degree. Note the run cannot be completed when mesh collapses with FS <1. The strengths of materials were then increased by 1.5 times (1.5c, tan-1(1.5tan)) to complete the run. Actual FS= FS calculated/1.5.

27

11.0 FINITE ELEMENT METHOD

Slope stability analysis using FEM is affordable nowadays thanks to powerful personal computer and affordable FEM software such as PLAXIS and Sigma/W. The safety factor is defined as: FS=max. available / needed for equilibrium The power of FEM method is illustrated by an example problem given in the PLAXIS manual for the construction of an embankment on saturated soft clay as shown in Figure 38. The material properties are tabulated in Table 7. As embankment fills are introduced, it leads to increase in pore pressure, u in the soft clay. A slip will be triggered if u becomes excessive. Therefore a rest period of 200days is allowed after the first 2m fills. This enables the dissipation of the excess u in the soft clay such that the clay gains in strength to receive further fills of another 2m. Using FEM analysis, wide range of information can be obtained such as: Pore pressure variation a any point within the soft clay through out the construction, Time settlement curves of the embankment fills FS against slope stability at the end of the first 2m fill, gain in FS during the rest period, FS at the end of the total fills, and the final FS in the long term. Shapes of slips

Figure 38 Situation of a road embankment on soft soil

28

Table 7 Material properties of the road embankment and subsoil

Parameter Name Clay MC Undrained 15 18 110-4 110-4 1000 0.33 2.0 24 0.0 Peat MC undrained 8 11 210-3 110-3 350 0.35 5.0 20 0.0 Sand MC drained 16 20 1.0 1.0 3000 0.3 1.0 30 0.0 Unit kN/m3 kN/m3 m/day m/day kN/m2 kN/m2

Material model Model Type Type of behaviour dry Dry soil weight wet Wet soil weight kx Horizontal permeability ky Vertical permeability Young's modulus Eref Poisson's ratio Cohesion cref Friction angle Dilatancy angle

Figure 39 Development of excess pore pressure under the embankment. From the curve it can be seen that more than 900days are need to reach full consolidation

Figure 40 Excess pore pressure contours after consolidation

29

Figure 41 Safety factor variation for various stages of the construction process

Figure 42 Shape of slip in the final stage. Note that the slip is log-spiral shape

30

12.0 FURTHER READINGS D. G. Fredlund and J. Krahn, 1977. Comparison of slope stability methods of analysis, Can. Geotech. J., 14, 429 S. G. Wright and J. D. Roecker, 1984. Example problems for slope stability computations with the computer program UTEXAS, Research Report 353-2, UT Austin R. D. Holtz and W. D. Kovacs, 1981. An introduction to geotechnical engineering, Chap. 11, p490, Prentice-Hall

31