F. E. Sánchez et al.: Stress-strain behavior of blue marls from the Guadalquivir river basin in Spain STRESS-STRAIN BEHAVIOR OF BLUE MARLS FROM THE GUADALQUIVIR RIVER BASIN IN SPAIN NAPETOSTNO DEFOR-MACIJSKI ODZIV MODRIH LAPORJEV IZ POVODJA GUADALQUIVIRJA (ŠPANIJA) Félix Escolano Sánchez Manuel Bueno Aguado Svetlana Melentijevic Polytechnical University of Madrid, PROES Consultores. Complutense University of Madrid, Civil Engineering Department General Yagüe, 39, 28020 Madrid, Spain Faculty of Geology Alfonso XII, 3 y 5 - 28014 Madrid, Spain E-mail: mbueno@proes.engineering Madrid, Spain E-mail: felix.escolano@upm.es E-mail: svmelent@ucm.es j https://doi.org/10.18690/actageotechslov.16.L30-42.2019 DOI Keywords Guadalquivir blue marls, sedimentary clay, triaxial test, undisturbed and remolded samples, lime stabilization, stress-strain behavior Ključne besede Guadalquivirski modri lapor; sedimentirana glina; triosni preizkus; neporušeni in pregneteni vzorci; stabilizacija z apnom; odziv napetost-deformacija Abstract This paper describes the geotechnical behavior of the Blue Marls from the Guadalquivir River Basin in Southern Spain, defined as high-plasticity clays that behave as a stiff, over-consolidated clayey soil. These Blue Marls are, in general, classified as an inadequate material for reuse in various earthworks. In this study the comparison between the geotechnical properties of its natural and compacted remolded state, with and without additives, is studied based on different laboratory tests. A great variety of laboratory tests were performed, including triaxial tests under consolidated undrained conditions. Three types of samples were tested in triaxial tests to define different stress-strain behaviors: undisturbed and remolded samples of the in-situ clay material, as well as remolded clay samples mixed with lime. Lime stabilization is frequently used to improve the geotechnical properties of clayey soils for application in different earthworks. The back-analysis of the slope stability is performed for a comprehensive assessment of the condition of the slopes consisting of these remolded materials mixed with lime as well as without lime. The results of the retrospective analysis of man-made slopes permit an analysis of the operational geotechnical strength parameters necessary in the design of slopes as well as its comparison with the results of the triaxial consolidated undrained tests. The improvement of the strength parameters for the slope-stability analysis is confirmed under remolded conditions with and without the addition of the small amount of lime. Izvleček V članku je opisano geotehnično obnašanje modrih laporjev iz porečja reke Guadalquivir (južna Španija), ki so opredeljeni kot gline z visoko plastičnostjo in se obnašajo kot trdna prekon-solidirana glinasta zemljina. Obravnavani modri laporji so na splošno opredeljeni kot neustrezni material za njihovo ponovno uporabo v različnih zemeljskih delih. V tej študiji smo na podlagi različnih laboratorijskih preiskav preučili primerjavo medgeotehničnimi lastnostmi njenega naravnega in zgoščenega pregnetenega stanja z dodatki in brez njih. Izvedle so se številne laboratorijske preiskave, vključno s triosnimipreizkusi v konsolidiranih nedreniranih pogojih. Da bi določili različno obremenitveno-deformacijsko obnašanje, smo v triosnih aparatih preizkusili tri vrste preizkušancev: neporušene in pregnetene preizkušance »in situ« glinastega materiala, kot tudi pregnetene preizkušance gline, pomešane z apnom. Stabilizacija z apnom se pogosto uporablja za izboljšanje geotehničnih lastnosti glinastih tal za uporabo v različnih zemeljskih delih. Za celovito oceno varnosti pobočij, sestavljenih iz teh predelanih materialov, zmešanih z apnom in brez apna, so bile izvedene povratne stabilnostne analize. Rezultati retrospektivne analize takšnih umetnih pobočij omogočajo analizo dejanskih geotehničnih parametrov strižne trdnosti, ki so potrebni pri načrtovanju pobočij ter primerjavo z rezultati triosnih konsolidiranih nedreniranih preizkusov. Izboljšanje parametrov strižne trdnosti v stabilnostni analizi pobočij je bila potrjena za pregnetene pogoje, tako z kot brez dodatka majhne količine apna. 30. Acta Geotechnica Slovenica, 2019/1 F. E. Sánchez et al.: Stress-strain behavior of blue marls from the Guadalquivir river basin in Spain 1 INTRODUCTION The Guadalquivir Blue Marls were studied at different points along the Guadalquivir River Basin in Southern Spain by different authors for the construction of various geotechnical structures and after a lot of damage that occurred in these materials after having this material as a foundation for specific structures. It was studied in relation to underground works associated with the construction of the subway for the city of Seville and for many slope-stability problems that happened during the construction of infrastructures (e.g., the slope-stability problem at Almodovar del Rio at the high-speed railway Madrid-Seville) [1,2,3]. The study of these marls was a crucial part of the analysis of the 1998 failure of the Aznalcollar tailing dam, which was founded on this formation and caused one of the major environmental disasters in Spain [4,5,6,7]. These marls were also studied for the slope-stability analysis at the Cobre Las Cruces open pit mine [8,9]. This material is characterized by fragile shear strength and degradation after drying and wetting processes [4,8,10] that could provoke a sudden geotechnical failure in these materials. For that reason these Blue Marls are usually not suitable for reuse in the construction of different earthworks such as embankments, earth dams, pavement bases, etc. according to the Spanish standard PG-3 [11]. These marls represent a fine-grained soil with an average percentage of the clay fraction ranging from 50 to 60%, classified as high-plasticity clay (CH) with a high carbonate content. Its behavior is like a stiff soil, defined as over-consolidated clays [4,8,9,14]. This study is a comparison of the behavior of the undisturbed blue marls and compacted remolded samples with and without additives that are analyzed in order to quantify the improvement of the geotechnical properties. The improvement of the geotechnical properties by lime is studied. Lime is one of the most commonly used binders for an improvement of the geotechnical properties of clayey soils, due to its effect on the modification of soil particle packaging, soil plasticity, workability, dispersity, permeability, swelling and shrinkage properties, compressibility, etc. [12]. It is widely used for the soil stabilization of different types of structures, such as a foundation base, slope protection, embankments, highway and railway sub-grade, etc. due to its low cost and ease of application in construction [13]. For this paper, the observational method is conducted over man-made slopes constructed in these Blue Marls over a period of 10 years, to deduce the resistance parameters based on a slope-stability back analysis. These values are compared to laboratory strength parameters deduced from the triaxial consolidated undrained (CU) tests performed over undisturbed and remolded samples (with and without the addition of lime) to compare the stress-strain behavior of the undisturbed and remolded samples of Blue Marls. This paper emphasizes the importance of the compacted remolded state of the material to be introduced in the earthworks in order to improve an inadequate clay material, such as blue marls (highly plastic and expansive) with and without lime for its application in different earthworks. In general, this kind of expansive clay can be encountered worldwide, and this subject could be of a great importance for the construction of different earthworks. According to [15], slopes and cut-offs in blue marls have undergone instabilities despite its behavior as a soft rock/stiff soil. A wide range of factors affect its instability, making it very challenging to select the most suitable stabilization measures once the instability has occurred, as well as processes for its mitigation. Some of the factors that affect its instability is its great weathering, which is increased due to decompression during excavation, attributed to the existence of micro-discontinuities with low resistance and strain-softening behavior that is increased during the process of humidity draught (i.e., a rainfall period), in this way exhibiting a great variability in behavior with time. The effect of the drying and wetting cycles process on the changes of macro- and microstructure of undisturbed Guadalquivir blue marls were studied by [10]. 2 RESEARCH METHODOLOGY The investigation was divided into three stages, as described in Fig. 1. During the first stage (5-year duration), a field research campaign was carried out based on a study of 41 slopes with different heights and inclination angles, on the unit of Guadalquivir Blue Marls. The geotechnical campaign was then supplemented by the rotary drilling of mechanical boreholes with the continuous taking of undisturbed samples. The samples were taken in the winter period after a long period of drought. These samples were taken during a wet season in which the area had received significant precipitation. During the second stage, a laboratory-test campaign was carried out on 89 samples for the identification of properties (i.e., sieve analysis (wet-sieve analysis and sedimentation analysis), natural moisture content, Atterberg 31. Acta Geotechnica Slovenica, 2019/1 F. E. Sánchez et al.: Stress-strain behavior of blue marls from the Guadalquivir river basin in Spain STAGE 1 Field campaign Preliminary analysis of slopes Initial data collection ►_ Drilling of boreholes and sample taking 41 Unaltered Samples 8 Remolded Samples without lime 8 Remolded Samples with lime STAGE 3 Retrospective analysis of slopes Obtaining operational parameters for slope design ||41 Man-made slopes analysis Figure 1. Research methodology. limits, dry unit weight, unit weight of solid particles, etc.). Mostly, the samples were taken at shallow depths (down to 4 meters from the surface) due to its application for the possible reuse in the construction of man-made slopes. In order to study the stress-strain behavior, 57 triaxial tests were performed on: undisturbed samples (41), remolded samples (8) and remolded samples mixed with lime (8). The purpose of the tests carried out on remolded samples with and without lime was to study the improvement of their geotechnical behavior regarding strength properties (friction angle and cohesion) for its possible use as a material for the construction of different geotechnical structures (man-made slopes, pavement base, earth dams, embankments, etc.). The small amount of lime (2% of the weight) was added to the mixture of remolded samples in order to study its influence on the geotechnical parameters. Finally, stage 3 is the retrospective analysis of the slopes studied during the first stage that makes it possible to reach conclusions regarding operational geotechnical strength parameters used for slope design, comparing them with the results obtained by laboratory tests (i.e., triaxial consolidated undrained tests). 3 GEOGRAPHIC AND GEOLOGICAL CLASSIFICATION OF THE GUADALQUIVIR BASIN The Guadalquivir Basin occupies a large area in southern peninsular Spain. It takes the form of an elongated wedge running ENE-WSW, with the Guadalquivir River flowing through it. Geographically, to the north it is bordered by the terrain of the Sierra Morena, to the south by the terrain of the Baetic Range, and to the west by the Gulf of Cádiz. Although its topography is very gentle, there is a gradual decline in altitude from the eastern sector (800 m) to the west (sea level) [16]. In geological terms, it is a Tertiary exoreic basin which was formed during the Neogene. The northwestern edge of the basin is represented by the Iberian Massif, with very clear contact in a SW-NE direction as well, made up of materials from the Paleozoic Era (600-250 My) [17]. The southern edge of the depression, whose boundaries are not as well defined, is represented by materials from the Mesozoic Era (approximately 250-90 My), part of the Bética Range (Fig. 2) [18]. This basin began filling more than 10 My, the result of the erosion of the mountainous terrain bordering it. The northwestern edge always remained stable (without significant tectonic movements), providing a continuous supply of sediment to the basin from the stripping of the terrain. In contrast, the southeastern edge was constantly rising, with frequent movements and slides, supplying a smaller amount of sediment. The largest amount of sediment was supplied by the erosion of the northwestern edge, with the deposits being more significant due to their representativeness: clayey or carbonated marls, in some cases slightly silty, bluish-gray in color (when they are intact) with zero stratification, and massive in appearance with concho- 32. Acta Geotechnica Slovenica, 2019/1 F. E. Sánchez et al.: Stress-strain behavior of blue marls from the Guadalquivir river basin in Spain Figure 2. Geographic and geological locations of the Guadalquivir Basin (modified from [18]). idal fracture. In an altered state (weathered), they take on a brown and yellowish gray-brown color due to the presence of iron oxides. Material is predominantly phyl-losilicates (more than 65%). The rest of its components are carbonates and quartz. The clay minerals are formed mainly by illite and smectite. The presence of gypsum may also occasionally be observed. There is extremely abundant microfauna, which has made it possible to date them to the Upper Miocene, specifically, the Messinian (6.3-5.2 My). This subject has already been a study of [18]. Due to their color (bluish-gray, when fresh), they are described in the geological and geotechnical literature as Guadalquivir Blue Marls. These deposits outcrop along some 800 km, from the province of Huelva, at the mouth of the Guadalquivir River, to the province of Jaén. The unit is present in the subsoil of major metropolitan areas such as the city of Seville. 4 GEOTECHNICAL CHARACTERIZATION OF THE GUADALQUIVIR BLUE MARLS The geotechnical properties of Guadalquivir blue marls were extensively defined by different authors [2,4,8]. According to identification tests performed over 89 samples for this study, the material is primarily of a clayey nature, with over 80% of material passing through the No. 200 ASTM sieve mesh. The liquid limit values vary from 40 to 70%, while the plasticity index ranges from 15 to 40%. Therefore, these materials fall within the CH to CL category, according to the ASTM classification [19]. The Casagrande plasticity chart is given in Fig. 3, also indicating the suitability of the application of this material for the construction of highway subgrades according to the Spanish Code PG-3 [11]. Due to the nature of this material, classified as marginal to tolerable, the compaction is performed in order to study the behavior of remolded samples with and without the addition of lime. The specific dry weight ranges from 14 to 16 kN/m3, and the specific weight of the particles ranges from 25 to 27 kN/m3. Regarding the carbonate content, most tested samples contain 20-55% CaCO3, thus defining them as clayey marls and marls (Fig. 4). Regarding the state conditions, Fig. 5 shows the initial moisture content and unit dry specific weight of both undisturbed and remolded samples. It is observed how the dry specific density is hardly modified with the remolding process. The unit dry weight is generally within a relatively narrow range from 14 to 16 kN/m3, while the range of values for the initial natural moisture content is somewhat broader and mostly from 20 to 30%. It can be observed that these materials can be defi- 33. Acta Geotechnica Slovenica, 2019/1 F. E. Sánchez et al.: Stress-strain behavior of blue marls from the Guadalquivir river basin in Spain Unsuitable Marginal o j. Tolerable > ° o ^^ ? * ^^ lp = 0.73(LI-20} Suitable o & . - qSfi" o o O* O Marginal ^^ O ° t 0 % O Selected o-1-1-1-^--■- o o ° Marginal Liquid limit (%) Figure 3. Plasticity chart of the studied undisturbed samples of blue marls. Sample classification according to the CaC03 content 1 1 > ( 3 O C a P o 3 O o 0 o o o O O O o > o o O « O O O o o Clays Clayey Maris Maris Calcareous Marl L mesto ne S 10 15 20 2S 30 35 JO 4S 50 S5 SO 65 70 75 30 85 90 95 101 % CaC03 Figure 4. Relationship between % CaCO3 and the depth of the undisturbed samples. ♦ Unidsturbed samples -O Síernolde-d samples with lim« ❖ Remo Wed samples - -Sr =90% =100% ♦ \ / / / 0,0 200.0 400 0 600,0 800,0 1000,0 1200,0 1400,0 1600.0 P(KPa) Figure 7. Stress paths of the undisturbed samples in the p-q diagram. P I KPa] Figure 8. Failure criteria of the undisturbed samples. 6 SHEAR-STRENGTH CHARACTERISTICS OF THE REMOLDED MARL SAMPLES Similar analyses to those described in Section 5 for undisturbed samples were performed on two types of remolded samples: a) Eight remolded samples compacted to 100% of their normal Proctor density according to the ASTM standard [23] with moisture levels close to their initial moisture content. In this case, there is only one stress-strain behavior mode corresponding to the elasto-plastic one (ductile behavior). All the tested samples reach a given deviatoric stress at 1 to 5% of the axial strain, corresponding to a consolidation pressure of 50 and 300 kPa. After reaching these values of the axial strain level, a small increase can be observed until reaching the maximum strain of 20% when the test is stopped. Both curves for different consolidation pressures (50 and 300 kPa) present a similar elastic modulus (10 MPa) [21]. The pore-water pressure relationship represents an increase in the same range of axial strain (1 and 5% respectively), after which it shows the general tendency of remaining constant till the end of the test, as was already observed for the undisturbed samples, Fig. 9. 36. Acta Geotechnica Slovenica, 2019/1 F. E. Sánchez et al.: Stress-strain behavior of blue marls from the Guadalquivir river basin in Spain Figure 9. Stress-strain relationship for the remolded samples with and without lime. b) Eight samples mixed with slake lime (Ca(OH)2), 2% lime by weight of the dry soil, and then remolded and compacted to 100% of their normal Proctor density (15.2 kN/m3) with moisture levels close to optimum (27%) (see Fig. 5). Samples had a maturation period of about 24 hours before testing. The tendency of the relationship curves is the same as for the remolded samples of the natural material, representing a greater deviatoric stress under the same confinement stress level necessary for the failure, as expected. The lime-treated samples again display the behavior pattern observed on undisturbed samples. The elastic stress-strain behavior is observed reaching axial strains at approximately 1%, and after that showing the small increase in deviatoric stress until the failure is reached (the test was stopped at 20% of the strain level). The elastic modulus [21] obtained is 25 and 48 MPa for different consolidation pressures of 50 and 300 kPa, respectively. These values of the initial elastic modulus are greater than the ones obtained for the natural material tested as a remolded sample under the same consolidation pressure. The pore-water pressure represents the contractive behavior until reaching the maximum strain level, after that representing slightly dilatant behavior until reaching the constant value until the end of the test (see Fig. 9). The stress paths of both types of remolded samples in turn follow paths similar to those of the undisturbed samples (see Fig. 10). These stress paths are typical of over-consolidated clays [24]. Lastly, plotting the failure points, the geometric location of the failure is established in order to obtain the Mohr-Coulomb model parameters on the p-q diagram (see Fig. 11). These are: - Remolded samples without lime: - Cohesion 15 kN/m2 - Angle of internal friction 28° 37. Acta Geotechnica Slovenica, 2019/1 F. E. Sánchez et al.: Stress-strain behavior of blue marls from the Guadalquivir river basin in Spain 700.0 -, 600.0 500.0 400.0 a. ^ cr 300.0 200.0 100.0 0.0 ...... Brwh point brlt* L * r»ak point uctlit Br»ak point bftt« ♦ * P i" 0.0 200.0 400,0 600.0 800.0 1000,0 1200,0 1400.0 1500.0 P (Kpa) Figure 10. Stress paths of the remolded samples with and without lime. 700.0 600.0 - 500.0 1Î 400.0 CL £ ET 300.0 200.0 100.0 0.0 0.0 .........i i i i i - . VMieutlbirt Mohr Coulrnt Í»llurt