4th International Conference on Problematic Soils

21-23 September 2012, Wuhan, China

______________________________ †

P.E., MSCE, M. ASCE, Member of the Board AC Consulting Engineers, Lima, Peru

1

SLIDE STABILIZATION IN PERUVIAN RESIDUAL SOILS

Arnaldo Carrillo-Gil*, Arnaldo Carrillo-Acevedo†

*P.E., Dr. Eng., F. ASCE

Principal Professor Ricardo Palma University, Lima, Perú

e-mail: acsac01@speedy.com.pe, webpage: www.acingenieros.com

Keywords: Residual soils, Slide stabilization

Abstract. This paper describes the developing of a different method of slide

stabilization carried out in the riverbanks of the Amazon River and in the

Peruvian jungle. This method uses sand-cement bags as another form of

reinforcement by confinement, erosion protection and emergency repair. Many

cases of this slide stabilization method have been successfully implemented, and

the performance monitoring of one of these repairs is discussed, along with, a

history of the slide, description of the tropical soils, and the mechanism of failure.

The application of this method of slide stabilization is also evaluated.

1 INTRODUCTION

From the practical point of view, empirical correlations are established between the plasticity characteristics of the tropical soils, and probably classification that can be used for preliminary designs with the Peruvian Amazon soils. In Peru, the majority of tropical soils are saprolitics. This means, they are clay soils mixed mainly with lime and sand. Several results from shear stress tests include, from the conservative side, saturated lab-tests, and from the other side, field test that provides more confident and reliable results. Average values between the lab and field data are taken with the help of statistical methods that allow us to have better values closer to realistic scenarios, and to consider influences of suction. In the Amazon rain forest, there are some cases of slope stability problems due to rain precipitation. This is because slopes failures take place during rain season, leaving the soil in a saturated condition. The suggested procedure consists in verify the stability of the slope in its extreme conditions: at natural humidity, and at saturated condition. An intermediate factor of safety value has to be used for a specific suction level.

2 SITE CONDITIONS

The general description of the geomorphology of the Amazon region indicates that the low jungle is substantially flat and its height varies between 80 to 400 meters above the mean sea level. Due to this small difference of elevation, rivers flow slowly, getting in dry season the appearance of lakes. This region of the Amazon plain can be indicated as advance erosion type. The Amazon plain is characterize by its great humidity and soil covered by a dense tropical vegetation (Carrillo-Gil, 1997)

1.

Studies establish that in the high jungle and in the limits of the low jungle are found so much igneous rocks as sedimentary, while in the low jungle prevail saprolitic soils originated by the

sedimentary rocks of the tertiary, and quaternary and they are formed mainly by sandstones, shales, and clays (Carrillo-Gil, Dominguez, 1996)

2.

The general geology considers that a large part of the Amazon region has stayed covered during the interglacial periods of the quaternary by an interior sea of shallow water when the level of the oceans had 100 meters above of the existing now (330,000 years ago) it also to fluctuate during several glacial and interglacial periods forming terraces throughout the water courses, dropping to 100 meters below of the original level during the last Glacial Era (17,000 years ago) and remaining in these deep channels the large rivers, between them the Amazon river, raising afterwards to the current level (6,000 years ago).

3 SLOPE FAILURE MECHANISM

The most common failure in the Peruvian jungle happens in the riverbanks of the main jungle rivers. The failure of slopes in subjacent tropical soils is because of ascend and descend movement action of the water during the stage of water level decrease in the river. These rivers form meanders over a soft soil of variable thickness. The meanders flow downstream producing erosion and sedimentation. The continuo deformation of the upstream meanders modifies the riverbed and originates effects in the riverbanks during the stages of rapid draw down that results in dangerous landslides. The statistical analysis of the movement of the Amazon River performed between the years 1991 and 1996 clearly established that the landslides have occurred during the stage of water level decrease in the river. We consider that as a phenomenon of rapid drawdown that affects the bank, because of the water level decreases to an average of 12 meters in a very short time (Carrillo-Carrillo, 1999)

3.

1st LANDSLIDE 1992

NORTHERN ZONE P.T. IQUITOS

3rd LANDSLIDE 1994

NORTHERN ZONE P.T. IQUITOS

2nd LANDSLIDE 1993

NORTHERN ZONE P.T.

IQUITOS

MONTHS

JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DIC

HE

IGH

T (

m.a

.s.l.)

LEVELS OF THE AMAZONAS RIVER

Figure 1: Landslide ocurrence at the Amazon River drawdown season This rapid drawdown is interpreted as a process that increases the undrained deformation of the saturated zone in the affected banks. In other words, the reaction of the stability of the banks to the rapid movement when the water level decreases is similar to the response occurred in an open cut in which is produced a forced alleviation, due to material that previously was offered as lateral support and that was suddenly removed. In this case, as a consequence of the imbalance produced by the

rapid drawdown of the river, there is water that remains within the porous structure of the soil, since its level does not decrease to the same speed that the water level. This phenomenon causes an increase in the weight of the bank body, as in the pore pressure with the soil. This effect reduces the shearing strength of the soil, which, together with the effects of the river, causes the landslides (if it has not been possible to evacuate the water tricked within the soil of the bank).

SECTION

COLLAPSED

Figure 2: Typical landslide at the Amazon River Banks

4 INVESTIGATIONS POST LANDSLIDES

In the last years, a geotechnical instrumentation was installed in the most critical areas where dangerous landslides occurred. The results that we have obtained from the geotechnical instrumentation have permitted to study the phenomena showed previously. The analysis of the instrumental measures was performed by the years 1996 and 1997, 1998-1999, and between 2008-2009. It can be appreciated that the current average movements are not very significant. They are compared with the river movement to establish that the critical stage for the drawdown of water level is produce by the forward movement towards the river while at the rise of the river happens the opposite. This phenomena shows that any defense construction has to be designed in the most flexible way possible and not in a rigid way because it will be very expensive. The results of the final piezometrics measures indicate that, as a rule, the dissipation of the pore pressures in almost all cases has been effected in correspondence with the decrease and increase of the water level of the Amazon River. We found a good behavior in the drainage system installed at the critical zones. The slope stability analysis performed in high pore pressures gives factor of safety less than one. However, when it is installed an adequate sub-drainage system where pore pressures are neutral or even lower, high factor of safety are obtained. This indicates that the best solution is not necessary retention structures; a good drainage system needs to be included (Carrillo-Gil, 2009)

4.

PORE PRESSURE

PO

RE

PR

ES

SU

RE

(K

pa)

MONTHS

RIV

ER

LE

VE

L (

m.a

.s.l.)

RIVER

LEVEL

Figure 3: Variation of pore pressures and water levels at the Amazon River

5 METHODOLOGY FOR RIVERBANKS SLOPE PROTECTION

Taking in account all previous studies, it has been solved some critical landslides problems using a variety of methods from structures of rigid retention to systems of geo-textiles reinforced soil wall. But the method that we have applied is the most economical and reliable solution for this type of project. The use of this kind of protection and flexible retaining structure in the Peruvian tropical region is quite simple. It does not need any special equipment and it provides jobs for local population. The sand cement bags are placed one over the other in the flexible walls. The design of this kind of structure follows the same procedures used for the design of conventional retaining walls. This project includes 20 sub-horizontal drains placed in two rows with spacing of 5 m with each other. The minimum penetration is 30 m. this setting is performed in order to dissipate pore pressures that affect the mass soil of the slope during the drawdown season of the river. This setting ensures better stability and less weight to the slope (Figure4). In addition, it was proposed a synthetic drainage system to be placed at the back zone of the flexible walls. The special system is a plain drain that contains a geo-red between two geo-textile phases. These geo-textiles, at the lower part, wrap a perfored tube where the captured water flows to the installed drainage canals at each extreme of the protected area. For the slope erosion protection due to the Amazon River, it is adopted an economical solution of polypropylene bags filled with cement and sand very similar to the ones proposed in the retaining walls. These bags are placed over the slope in a special disposition where they are lock to each other. They are anchored with steel bars of 3/8 inches. This setting does not allow the movement of any of these elements. This type of protection slope avoids also the framing and the use of geo-synthetic materials that have higher costs.

UNIVERSIDAD RICARDO PALMA 4646

Figure 4: Protection method for slopes at the Amazon Riverbanks

6 INSTRUMENTAL MONITORING INTERPRETATION

The analysis from the gathered instrumental data from the south and north area, as well as contain available data, has allow to establish slopes behavior. It was considered the detected accumulated landslides from the inclinometers compared with river levels and pore pressures measures with respect the time in between measurements periods. The slope general behavior indicates that during the season of high water levels, the movement oscillates with a slight tendency to back up. While the season of drawdown, the movement of water go straight to the river. However, these tendencies could be influenced by many factors such as the occurrence of rainfall, drainage facilities, and other unusual events like the increase of water levels during the cold waves season.

SOUTHERN ZONE

Figure 5: Actual conditions of Southern Zone research area

NORTHERN ZONE

Figure 6: Actual conditions of Northern Zone research area High pore pressures can be seen at the southern zone because the instrumentation was installed more deeply, among other reasons. Lower variations can be observed at the northern zone (range between 25 KPa to 40 KPa). Comparing these pressures with the available database and with their water level, it can be observed that for similar water levels, pore pressures are higher during the drawdown period. In the superficially installed instrumentation was observed that pore pressures increased proportionally with the river water levels up to certain high where correlation increases exponentially. This is very important to considerate because once the maximum water levels are reached, a rapid drawdown happens resulting slopes with no support and extremely high pore pressures at upper soil layers. According to the last results, values are found above average from the previous seasonal readings which means that drainage systems capacity has decreased since its installation (Figure 7). The worst case scenario could be considerate as a main event where the river shows a maximum high water levels and a maximum low drawdown levels in a very rapid way. Secondary factors are normal delays and poor performance of superficial and deep drainage systems during medium to high precipitation events that do not consider high water levels but high pore pressures due to infiltration.

Figure 7: Maximum and mínimum water levels at the Amazon River for the last 50 years

7 FINAL COMMENTS

Finally, all the acquired experience from past projects, either the ones with good performance or the ones that fail, a solution of protection and stability has been implemented in slopes of tropical soils using the described system of sand-cement bags displaced conveniently to form retention walls and to offer protection against slope erosion. All the projects, using this system, showed very good results and their behavior were adequate for the local weather, soil, and special conditions of construction at the Peruvian jungle.

REFERENCES

[1] Carrillo-Gil, A., Peculiarities of tropical saprolitic soils of Peru, XIV International Conference on Soil Mechanics and Foundation Engineering, Hamburg, Germany (1997).

[2] Carrillo-Gil, A & Dominguez, E., Failures in Amazon riverbanks, Iquitos, Peru, Seventh International Symposium on Landslides, Trondheim, Norway (1996).

[3] Carrillo-Gil, A & Carrillo-Acevedo, A., Slope Stabilization in residual soils of Peru, International Symposium on Slope Stability Engineering, Matsuyama, Japan (1999).

[4] Carrillo-Gil, A., Monitoring and performance in landslides control, II International Conference on Environmental Management, (ICEM2) University of Wollongong, Australia (2009).