Residual Shear Strength of Landslide Soils by Presheared Flush Testing for Stability Quantification
PhD Candidate, The United Graduate School of Agricultural Science,
Professor, Department of Biological Production System,
Associate Professor, Department of Biological Production System,
In landslide slope stability evaluation, the determination of strength parameters (c and f) is very important. For soil subjected to repetitive sliding, these parameters reduced to their residual values (cr and fr) are used in the calculation of the factor of safety. Currently, among the two types of testing methods that permits to assess these strength parameters, the torsion shear test by means of a ring shear apparatus is widely employed under numerous testing procedures. However, even though engineers have made efforts to minimize errors due to wall friction during the test progression, a considerable number of tests required to determine the shear parameters still extend in time. In this paper, we propose a short and effective procedure to obtain the residual shear parameters in which error due to metal friction is minimized and the shearing time is shortened: the Presheared Flush Test procedure (PFT). The strength parameters resulting from the new Presheared Flush Test Procedure (PFT) on samples from tertiary landslide slope of the Niigata landslide slope drainage tunnel project site are compared to the results obtained by the conventional Flush Test procedure. We also present the results of the quantification of the stability state of the landslide slopes at various groundwater conditions
Keywords: Landslide, PFT, Ring shear test, Testing procedure, Slope stability
For years landslides resulting in slope failures induced by the joined actions of rise in ground water table and long or heavy rain and snow melt during summer which cause the saturation of slope have been common in Niigata prefecture, Japan. The Niigata Experimental Laboratory (2000) estimated the average monthly generation of landslides to 200 with a peak of 737 in April. These phenomena trouble the serenity of the population in the neighborhood of hilly or mountainous areas. Consequently slope stabilization projects are under implementation to prevent disasters caused annually by landslides. Investigations have shown that the slopes move approximately 4cm every year in Takada village, Jouetsu region, where a drainage tunnel is under construction to stop the movement. Numerous cracks in the tunnel under construction indicate the continuing pattern of the movement. Samples taken from the drainage project site have been tested using the ordinary Flush Test procedure. However this procedure is time consuming. So we propose a new testing methodology which shortens the testing time and minimizes the errors originating from metal friction.
DESCRIPTION OF THE PROJECT
The large scale landslide of our interest is located in the southern portion of a hilly region (see Figure 1) within a highly-concentrated landslide zone. It is underlain by the Upper Tertiary Miocene marine mudstone. The region is situated within a NNE to SSW-trending structure exhibiting a fold axis (anticline and syncline axis) and faults. Due to deformation from folding and faulting, the strength of the bedrock is low. The mountain is composed of an andesitic intrusive body situated upslope of the slide, and functions as the source area for ground water. The other contributing cause of landslide is infiltration of prolonged rain water prior to the addition of snowfall and snowmelt which contribute to increase the groundwater. The past history of slide activity is unknown; however, there has been some erosion during every snowmelt season. It is apparent that these factors contribute to sliding. Based on the recent data the slope movement is about 4 cm per year and this continuous movement explains the fissures during the construction of the tunnel as shown in Photo 1. Furthermore, ground water moving near the slip surface is suspected to be the contributing factor for secondary sliding at the toe area. Due to the above reasons, mitigation measures are being implemented. Serious effort for mitigation work commenced in 1988. The primary concern for this large scale landslide is mainly ground water supplied from the andesitic body upslope of the slide which increases the possibility of movement of the slope. Hence subsurface drainage control works (see Figure 2) by tunnels is found to be the main thrust of mitigation work.
Figure 1. Location of the drainage project
Figure 2. Schematic view of the tunnel and its function as a drain
DESCRIPTION OF THE APPARATUS
For the determination of the shear strength parameters, among the multiplicity of ring shear apparatus reported by Hvorslev (1939), Lagatta (1970) Bishop et al. (1971), the Bromhead Ring Shear Apparatus (Figure 3) developed by Bromhead (1979) is becoming widely used due to its simplicity in operation compared to the previous models. A full description can be found in the technical literature WF25850 by WF Engineering. In this apparatus, the ring shaped specimen has an internal diameter of 7 cm and an external diameter of 10 cm. Drainage is provided by two porous bronze stones fixed to the upper platen and to the bottom of the container. At the present time, four testing procedures have been proposed for the use of the Bromhead Ring Shear Apparatus. Stark and Vetell (1992) have shown that the Single Stage Test procedure provides a good estimation of the residual strength at effective normal stress less than 200 kPa. When the effective normal stress is greater than 200 kPa, consolidation of the specimen during the test causes settlement of the upper platen into the lower platen giving higher residual strength values. Anayi et al. (1988) have pointed out that in the Preshearing Test procedure, the preshearing facilitates the creation of a shear plane and reduces the amount of length of the horizontal displacement required to reach the residual condition. This procedure causes extrusions of a substantial amount of soil during the shear process and therefore, as in the case of the Single Stage Test procedure, gives higher measured residual strength values. Stark and Vetell (1992) also concluded that in the Multistage Test procedure an additional strength, probably due to wall friction as the top platen settles into the specimen container, develops during consolidation and shear process; hence they proposed the Flush Test procedure in which, increasing the thickness of the specimen prior to shear reduces the wall friction and gives more trustworthy measured values. This procedure takes substantial time to reach the residual condition when it is conducted at low rate of displacement.
Figure 3. The Bromhead ring shear apparatus
Figure 4. Schematic of the PFT procedure
MATERIALS AND METHOD
As error due to settlement of the upper platen into the specimen container is minimized in the Flush Test procedure, we propose the Presheared Flush Test (PFT) procedure in which the specimen is presheared prior to the use of the Flush Test procedure. This technique combines the merits of the Flush Test and the Presheared Test procedures. In the PFT procedure, after the thickness was increased byƒ¢H (see Figure 4 ) equal 2 to 3 mm, the ring was removed, the specimen was consolidated and presheared by hand on a horizontal displacement of approximately 80 degrees which correspond to 6 cm of horizontal displacement and to 12 revolutions of the crank, then the system was left for about 2 to 3 hours to allow the dissipation of the pore water and sheared as opposed to Flush Test procedure in which the specimen was immediately sheared after consolidation. In this experiment, a rate of 0.03degree per minute was used following the hand shearing as suggested by Kakou et al. (2001). The natural specimens involved in this study were obtained from the interior of the tunnel of the Niigata Drainage Project (Photo2). The consolidation of the specimen followed the procedure described by Stark and Vetell (1992) in the Flush Test procedure. The procedure of testing is illustrated in Figure 4. Tests were carried out on previously remolded samples as it had been demonstrated by Bishop et al. (1971) that the ultimate residual strength is not affected by the initial structure of the soil.
TEST RESULTS AND DISCUSSION
The tests involved change in the normal effective stress (50, 100, 200, 300 and 400 kPa) to define the failure envelope obtained form the proposed PFT procedure and the ordinary Flush Test procedure for comparison. The results shown in Table 1 and in the typical stress displacement curve (Figure 5) appeal the following observations:
Reduction of error induced by wall friction
Stark and Vetell (1992) concluded that settlement of the upper platen into the lower platen induced wall friction and therefore increased the measured value of the shear stress. Consequently they proposed that the total settlement should not exceed 0.75 mm.
Table 1. Summary of tests results
|Normal stress (kPa)||Time to reach the residual state in
|Total settlement of the upper platen in the container
1 Value discarded in the calculation of fr (PFT procedure)
Table 1 shows that in the PFT procedure this settlement was lessened compared to the Flush Test procedure. This can be explained by the fact that the combination of precaution to avoid soil extrusion from the container during preshearing and the limited time to reach the residual state prevented the upper platen from settling into the lower platen and the error due to metal friction was minimized. Moreover in the Flush Test procedure the residual stress was achieved after approximately 4 to 5 cm of horizontal displacement and in the PFT procedure it was reduced to 2.5 to 3 cm, which correspond to respectively 33 to 38 hours and 13to 23 hours. Table 1 also shows that the average reduction in the shearing time is in the order of 50 per cent.
The residual strength parameters
The results of the tests summarized in Figure 6 and Table 1 show that there is no significant difference in the regression curves obtained from the Flush and the PFT procedures from the view point of internal friction angle when the consolidation load is lower than 200 kPa. As has already been demonstrated by Stark and Vetell (1992), when the normal load is greater than 200 kPa the difference between the shear stress becomes noteworthy and the Flush Testing procedure gives higher results. This can be explained by the incidence of the wall friction between the lower and the upper platens. The settlement of the upper platen into the specimen container (Table 1) increases as the normal load increases.
SLOPE STABILITY ANALYSIS
The strength parameters obtained were used for the quantification of the stability improvement after the completion of the drainage work. For comparison, three methods, based on two dimensional stability analysis, were employed. The factor of safety (Fs) was calculated by assuming a full water level and totally drained conditions. The result of the calculation is shown in Table 2. These results indicate that the factor of safety is in the order of one and even below one showing the instability state of the slope before drainage works. After completion of the works it is noticed that Fs is greatly improved to more than 2, indicating the effectiveness of the drainage works.
Table 2. Factor of safety (Fs) by different methods
|Water level condition||Simplified Method of Slices||Simplified Bishop||Janbu|
|Full water level||1.06||1.03||1.11||1.09||0.93||0.91|
|Low water level||2.40||2.38||2.50||2.5||2.21||2.19|
The Presheared Flush Test procedure (PFT) in the use of the Ring Shear Apparatus is proposed as a short and effective procedure to obtain the residual shear parameters with a reduced amount of error due to metal friction. Tests conducted on clay samples using the proposed PFT procedure showed that in this procedure the total settlement of the upper platen into the specimen container was minimized. Furthermore, the elapsed time to reach the residual state was reduced by 50%. The residual internal friction angle fr obtained was close to the one measured using the Flush Test procedure. Due to reconsolidation effect during the time allowed for the dissipation of the water pressure when the test is conducted under normal stress larger than 300kpa, the result obtained deviated from the regression line. The strength parameters obtained and used in the stability calculation show important improvements of the factor safety Fs indicating the effectiveness of the construction works.
We thank, through the Niigata Experimental Laboratory, the Publics Works Research Institute of the Ministry of Construction of Japan for the great assistance during this study.
We gratefully acknowledge Professor R. Nakano, formerly professor emeritus and dean of the United Graduate School of Agricultural Science, Gifu University.
|© 2002 ejge|