Analysis of landslide effects along a road located in the Carpathian flysch

3.1 Determining of the minimum slope stability factors

The Regulation of the Polish Ministry of Transportation of March 2, 1999 (Polish Journal of Laws No. 43, item 430), regarding technical conditions that should be followed by public roads and their location, also defines the concept of earth structure [6]. In paragraph 144 (2), it requires that “slope stability factors determined individually by the methods given in Polish Standards should not be less than 1.5”. This is generally accepted in road design. In section 2.2 of Annex No. 4 to the above-mentioned Regulation, the requirement to check the stability of slopes and slopes in accordance with Polish Standard PN-B-03010: 1983 [7] is repeated. While considering the possible forms of destruction of free-free and supported slopes, one should accept:

1. for homogeneous soil in terms of compliance – circular–cylindrical slip surfaces,

2. for slopes made of several layers of soil, differing in shear strength – non–circular slip surfaces,

3. on fractured rocks – the shape of the slip surface depends on the occurring discontinuities; the balance of the solid limited by each of the possible slip surfaces is checked, for the calculation values of forces and geotechnical parameters, by the strip method; the permissible values of stability factors should not be less than 1.5,

4. loading from vehicles evenly distributed, of a value of 25 kPa [6].

The other aspect of checking the stability of landslides is the uncertainty related to the soil parameters. The soil parameters used to make the necessary calculations may come from in situ tests, laboratory tests of the soil samples, and from writings related to similar cases. They may differ significantly, on a case-to-case basis. In order to verify the correctness of the determination of the soil parameters, it was decided to drill two holes in the analysed cross-section of the road discussed. One of them is 17 m deep and the other one is 1 m deep. The safety and stability of the road excavation slopes are determined by local factors, such as the discontinuities of layers. In particular, layers of rock, surfaces of fuzzing caused by, for example, earlier landslides, and many other possible geological disturbances (e.g., soil lenses) can contribute to the safety and stability. Sometimes, these factors are difficult to detect in substrate studies, especially if substrate tests are made pointwise (e.g., in the form of boreholes or soundings). Soil samples for macroscopic and laboratory tests are collected and then analysed, as well. Care should be taken when estimating soil cohesion values. The PNB-03010:1983 standard [7], for example, required that the cohesion value should to be reduced by as much as 50%, in some cases. The use of shallow geophysical methods is very desirable in the case of roads on landslides. They allow us to determine local discontinuities and loosening of some layers. The appropriate methods are various types of seismic methods, electro-resistance or ground penetrating radar (GPR) methods.

The results of the slope stability analysis are also influenced by the calculation method. It is possible to apply computational techniques based on analytical and numerical methods. The analytical methods include the methods of Felleniuss [8], Bishop [9], Morgenstern-Price [10], Sarma [11] and others. The numerical methods are based on approximation algorithms, where the space is divided into a finite number of discrete elements. The Finite Element Method (FEM) is a method based on the discretization of a physical model [12, 13]. The discretization of the model replaces the medium (i.e., the continuum), which is actually a physical model in a discrete model. Physical quantities are not distributed continuously, but they are concentrated in elements resulting from discretization. Discretization dividing of the continuum into a finite number of elements is a simplification of the physical model that facilitates much of the computations. The computations check the convergence of the process and estimate the errors of the mesh. FEM is used worldwide in well-known FEA packages such as ADINA, ABAQUS, DIANA, MIDAS and others.

The calculation method used in this study is called Strength Reduction Method (SRM). This is non-linear method based on the limit equilibrium theory [14]. The method analyses the minimum safety factor and failure behaviour using various shapes, loads, and boundary conditions. The SRM method may be used to simulate the failure process without any previous assumptions and may be applied both to 2–D and 3–D problems. The method gradually decreases the shear strength and internal friction angle until the calculation does not converge. Hence, the obtained point is considered to be the failure point of the slope. The iteration starts with the factor of safety F_os = 1 unless otherwise specified by the user and subsequently F_os is incremented. With the updates of F_os the cohesion and the tangent value of internal friction angle are scaled down. The maximum shear strength reduction ratio at that point is used to calculate the minimum safety factor of the slope. Stability by SRM is achieved by weakening the soil in an elastic-plastic discrete analysis until the slope fails. The Factor of Safety (F_os) is considered as the factor by which the soil strength is to be reduced to reach failure [15].

3.2 Road stabilization project

The research area is located in the active part of the landslide. Therefore, the section of the provincial road located along this section is at risk. The landslide card, developed in 2009, classifies the entire landslide as an old slumbering landslide, which activates during periods of adequate hydro-meteorological conditions. The activation before 2009 damaged the local road and increased the risk of rupturing the road surface over a hundred meters. Therefore, the embankment of the road was protected in this section. The landslide was reported to the Verification and Evaluation Commission as caused by natural disasters. According to the landslide card [2], it is sub–compact–complex, rock–like wasteland, the colluvium material is detritally blocky. The types of material are identified as Magura sandstones with Quaternary colluvium forms over them.

Figure 1 shows the range of the landslide and comes from the landslide registration card. The landslide consists of a main slope of about 4 m high. In the upper part of the colluvium, there is a local road exposed to soil mass movements. On the asphalt surface, despite the renovation works, one may see the contour of a landslide niche. Above the local road, under the secondary escarpment (about 943 m above sea level), there is a zone of large wetlands. It is additionally supplied with a stream flowing periodically. The road surface is within the range of this zone, where the active part of the landslide is located. The landslide descends to the stream and it has a height of about 5 m there. The landslide has formed on the lower slope of the Sylec stream valley, and the direction of sliding off the soil masses runs from NE to SW. The colluvium area is regularly drained by a surface stream flowing through the culvert under the road, and at the foot of the slope.

There was no regulated drainage system along the road. The roadside drainage ditch did not fulfil its task, especially during periods of heavy rainfall and spring thaws. The culvert existing in the landslide was damaged, which, in turn, led to the collapse of the colluvium and the formation of wetness at its outflow.

As part of construction works, a project to stabilize the landslide and the road was carried out within the 100.0 m long road project. The reinforcement of the existing pavement structure at a width of 7.00 m, with both sides of the road paved with width of 1.60 m and 1.85 m, was made. The rainwater was drained, in a controlled manner, to two wells and further by the road to the roadside ditches.

Within the project framework, geological engineering documentation was prepared. On the basis of this documentation, it was found that the existing pavement structure in the upper 20 cm layer are bituminous layers. Below, to the depth of 2.00 m to 2.50 m, there is a road embankment composed of clayey sandstone particles. Under the road structure, there is natural soil of semi stiff and plastic texture. At the depths of 3.0 m and 6.0 m, groundwater filtration occurred. On the basis of the traffic measurements and geological engineering documentation, the required thickness of the pavement structure was assumed, taking into account the frost resistance condition. Taking the above conditions into consideration, the following pavement structure was used for the road system:

1. wearing layer made of asphalt concrete AC 11 with 5 cm of modified bitumen,

2. basic foundation made of asphalt concrete caliber 0/20 mm, 13 cm thick,

3. auxiliary foundation made of broken aggregate mechanically stabilized caliber 0/31.5 mm, 20 cm thick, layer of crushed aggregate 0/63 mm 35 cm thick,

4. geotextile with the following parameters: tensile strength (length/width) min. 80/80 kN/m, elongation (length/width) max. 12.5/12.5 %, and water permeability (perpendicular/in the plane) 50/30 l/m²s.

The total thickness of the renovated road was equal to 73 cm. The ordinates of the area covered by surface mass movements range from 965 m above sea level, in the upper edge of the main embankment, up to 880.0 m above sea level, in the area of the front of the landslide. There is no continuous aquifer in the quaternary area formations; however, the occurrence of water flow within the sandstone clayey deposits was identified. The water is coming from surface water infiltration. The intensity of the filtration is largely dependent on the amount of precipitation.

Several steps were taken in order to limit the inflow of water from the area above the road:

1. modernization and reinforcement with a reinforced concrete slab of the existing culvert of 1000,

2. strengthening the inlet to the road culvert,

3. strengthening and modernization of the outlet from the culvert.

The regulation of the water management of the entire landslide area by capturing the surface waters was not possible due to the terrain and its afforestation. After the drainage was installed, the current water flow protection is fulfilling its task. The retaining structure is located along the road below its edge. It was built in the form of a 20 cm thick reinforced concrete slab. The slab is fixed to the subsoil with ground anchors 40/16. For the purpose of calculations, it was assumed that the probable slip surface is located at a depth of about 7.0 m, but the anchoring was assumed to be in the layer of fine sandstone under the deepest slip surface specified by geological documentation. The length of ground anchors was 15.0 m and 17.0 m. The anchors were spaced 2.0 m in both directions and finished with 4.0 m long concrete bond. The tendons were fixed in a reinforced concrete slab with steel plates 20 × 20 cm. The retaining structure is divided by 5 cm wide dilatations into separate segments. Along the road, due to the need to widen the shoulder, the upper edge of each section was finished with a vertical wall thickness of 20 cm. The wall height of 0.60-0.74 m was adjusted to the slope of the longitudinal road. In a reinforced concrete slab, horizontally mounted filters 10 cm were placed at a distance of 2.0 m in both directions. Figure 2 presents the land development within the landslide and road, while Figure 3 shows the cross-section through the reinforced concrete slab with the reinforcement and schematically marked ground anchors.

Figure 2

Plan of development within the landslide and the local road.

Figure 3

Design of road protection against the landslide. Cross-section through a reinforced concrete slab with reinforcement and schematically marked ground anchors.

After the work to stabilize the landslide, in accordance with the design that was approved for implementation, a new road surface was made in this section and the road was re-opened. After a couple of years, considerable displacements were observed from the third, counting from the pass side, a section of the retaining structure. The as-built inventory shows that the sections of the retaining structure has been visibly dislocated. The displacement is clearly visible in the upper part of the plate and measures 8 cm horizontally and 3 cm vertically. There are several reasons for this:

1. water accumulation in the road body,

2. incorrectly made anchoring of a part of the plate,

3. soil quality locally poorer than shown in the design.

3.3 Field and laboratory tests results

The necessary field and laboratory steps were carried out to identify the cause for the excessive displacements of the reinforced concrete slab section. Among others, two boreholes were drilled together with the collection of the samples for laboratory tests. One borehole, marked in Figure 4 as OT-A, was made using the mechanical and rotary method, on the water scrubber, a double core machine with 116.95 millimetres drill. The drilling depth was 17 m in cross-section through the road corresponding to the third section of the retaining structure.

Figure 4

Cross-section with the drilling boreholes marked.

Another borehole was made by manual drilling to a depth of 1 m in order to take undisturbed soil samples (OT-B). The resulting cross-section is, according to the design, a road structure layer with a thickness of approximately 20 cm, followed by a 20 cm foundation and an additional 35 cm substructure. At a depth of about 75 cm below the ground level, a geotextile that separates the main road structure from the modified substrate was found. This subsoil consists of successive layers of up to 2.00 m below the ground level, a strongly compacted layer of building embankment, and then sandstone clay with varying degrees of plasticity, mainly with a stiff and firm consistency. At a depth of about 7-7.20 m below the ground level, a strongly fractured layer of sandstone appears at the drilling site, which is identified as a potential slip surface. Then, up to a depth of 17 m below the ground level, there is a sandstone with a different degree of cracking. Nowhere in the cross-section is there free or suspended water. Figure 5 presents the samples taken from the OT-A borehole in one–metre sections.

Figure 5

Overall view of the samples taken from the OT-A borehole.

During the fieldwork, a determination of the overall soil parameters from each geological layer was performed macroscopically [16]. Later on, laboratory tests for the water content [17], as well as the grain size distribution curve [18], were performed. The tests were carried out for eight representative samples in the OT-A borehole and one from OT-B. The grain size distribution (i.e., the sieve and hydrometer analysis) was carried out for four representative soil samples. Moreover, six determinations of soil strength parameters were carried out using a triaxial stress apparatus [19] to test the disturbed samples of unconsolidated soil in undrained conditions (UU). Figure 6 presents the Mohr circles obtained during the triaxial test with a confining pressure, namely 100 kPa, 150 kPa and 200 kPa, of a disturbed sample in undrained conditions from 1.2 m below the ground level. Figures 7 and Figure 8 present the results of a similar test for disturbed samples taken from 3.6 m and 6.6 m, respectively. Based on the results of the triaxial tests, the effective values of internal friction angle and cohesion, as fundamental parameters for the failure criterion of the soil were derived. The results are presented in the collective summary of the laboratory test results in Table 1.

Figure 6

Mohr circles obtained during a triaxial UU test of a disturbed sample in undrained conditions from 1.2 m below the ground level.

Figure 7

Mohr circles obtained during triaxial UU test of a disturbed sample in undrained conditions from 3.6 m below the ground level.

Figure 8

Mohr circles obtained during triaxial UU test of a disturbed sample in undrained conditions from 6.6 m below the ground level.

In the case of Magura sandstone, the Rock Quality Designation Index (RQD) [20], presented in Table 2, was found. One may notice that the quality of the rock mass is assessed as an average.

The strength of the rock was also measured using a Schmidt hammer. These values were obtained for six rock samples and are presented in Table 3. These rocks are characterized by homogeneous strength parameters in the range of 39.5-45.5 MPa. The uniaxial compression tests some of the soil samples were performed in an unconventional manner [21]. Figure 9 presents the experimental setup. The compression force and the vertical displacement were measured with the force measurement resolution equal to 0.1 N, while the resolution of the vertical deformations measurements was equal to 0.002 mm [22]. Moreover, a digital high resolution camera was used in the experiment; furthermore, a digital image correlation method was performed.

Figure 9

Experimental setup of uniaxial compression test consisting of the high resolution camera and the static testing machine.

The resolution refers to the original value, which is obtained by means of an optical measurement, and, in this case, it is the displacement of a given point (marker). Thanks to the use of sub-pixel measurement, the measurement resolution was, depending on the test, from 0.01 to 0.02 px, yielding 0.0004 and 0.0008 mm, respectively. The collected image data was processed using an optical measurement system (CivEng Vision) [21]. According to ISO 17892 [23], the cross-sectional area of the specimen may either be circular or square, and is at least 1000 mm². For cylindrical samples, the ratio between height and diameter is between 1.8 and 2.5. The largest particle in the sample should not exceed 16 of the sample diameter for both cylindrical and square sample. Unfortunately, due to the quality of the samples, the abovementioned sample requirements were fulfilled only once. Figure 10 presents the stress-strain curves as received for uniaxial compression tests. At the beginning of the test performed for samples # 1 and # 2, the machine was adapting to the proper work, and thus, the stress is not visible, even though displacements theoretically contributed to more than 1% of the strains. Table 4 presents the Young’s modulus E and the Poison coefficient ν of the analysed samples in the range between 2% and 4% of strains and the values of soil density ρ.

Figure 10

Stress-strain curves received for the uniaxial compression tests [21].

Figure 11 presents the digital maps of the strain tensor components, namely ε_xx, ε_yy, ε_xy, put on the image of soil sample 1. From the first tensor component ε_xx map, one may notice a least three oblique cracks. Two of them are not visible in ε_yy strain components, while the ε_xy tensor component map visibly exhibits only two cracks. Figure 12 presents the digital maps of the strain tensor components put on the image of soil sample # 2. Once again, from the tensor component ε_xx map, one may notice a least three oblique cracks, while the ε_yy strain components map shows only a horizontal crack. Observing the ε_xy tensor component map, one may notice two cracks nucleating in the upper corners. Figure 13 presents the digital maps of the strain tensor components put on the image of soil sample # 3. Similarly to the previous samples on the first tensor component ε_xx map, one may notice the oblique cracks very clearly, while the ε_yy strain component map does not show them at all. The ε_xy tensor component map is complementary to the ε_xx component map and confirms visible cracks.

Figure 11

Strain tensor components of the final digital maps, put on the image of soil sample # 1.

Figure 12

Strain tensor components of the final digital maps, put on the image of soil sample # 2.

Figure 13

Strain tensor components of the final digital maps, put on the image of soil sample # 3.

The analysis of the uniaxial compression tests of the soil samples indicated that the behaviour of the soil is more similar to cohesive clay than it is to non-cohesive gravel. The use of the non-contact optical method, followed by the study of a digital image correlation (DIC) of the sample obtained during tests effectively completes the description of the behaviour of this type of soil material. The parameters determined during the tests shown in Table 4 were applied to the numerical models built subsequently, as well as to further analyses.

3.4 Surveying work

As part of the surveying work, a local situation and altitude map of the landslide was made using a Riegl Vz–400 terrestrial laser scanner. The ordinates were determined using the method of direct geodetic measurements in the field in the local coordinate system.

Figure 14 shows a terrestrial laser scan within overall view of the analysed landslide. Figure 15 shows the image of a road and the landslide developed using the Riscanpro package on a stretch of 75 meters. The image obtained, in the form of a cloud of points, may be compared with subsequent scans in order to monitor landslide movements.

Figure 14

Terrestrial laser scan within overall view of the analysed landslide.

Figure 15

Road laser surface scan overview within the analysed landslide.