Available online at www.sciencedirect.com ScienceDirect Procedia Earth and Planetary Science 9 ( 2014 ) 214 221 The Third Italian Workshop on Landslides Prediction of suction evolution of silty pyroclastic covers in flume tests and field monitoring Alfredo Reder a *, Guido Rianna b, Luca Pagano a a DICEA Department of Civil, Construction and Evironmental, University of Naples Federico II, Via Claudio 21, 80125, Naples, Italy b CIRA Italian Aerospace Research Center, Via Maiorise s.n.c., 81043, Capua (CE), Italy Abstract The paper shows the results of calibration, validation and blind prediction of two different test cases: the former is a flume test for experiment involving a homogenous soil; the latter is a field case for the layered natural soil coming from Cervinara site.the results have been carried out through hydraulic and mechanical simplified approach. 2014 Elsevier B.V. This is an open access article under the CC BY-NC-ND license 2014 The Authors. Published by Elsevier B.V. (http://creativecommons.org/licenses/by-nc-nd/3.0/). Selection and peer-review under responsibility of Dipartimento di Ingegneria Civile, Design, Edilizia e Ambiente, Seconda Selection Università and di Napoli. peer-review under responsibility of Dipartimento di Ingegneria Civile, Design, Edilizia e Ambiente, Seconda Università di di Napoli. Keywords:Infiltration; Matric suction; Rainfall; Slope stability; Unsaturated pyroclastic soils. 1. Methodology and soil hydraulic and mechanical characterization Calibration, validation and blind prediction tasks assigned for both the flume tests and the field case 1, have been carried out by adopting a simplified approach, modeling seepage in an unsaturated and rigid medium under isothermal conditions. For the field case neglecting thermal effects and related evaporation phenomena can lead to overestimation of predicted pore water pressures during the dry periods while it should represent a reliable hypothesis for predictions carried out during the wet periods 2. For flume tests the isothermal assumptions realistic since evaporation phenomena are negligible during the simulated rainfall event 3. Adopting a rigid-soil skeleton hypothesis corresponds to neglect the effects of possible changes in soil porosity due to soil collapse upon wetting. * Corresponding author. Tel.: +39 081 7685916; E-mail address: alfredo.reder@unina.it 1878-5220 2014 Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/). Selection and peer-review under responsibility of Dipartimento di Ingegneria Civile, Design, Edilizia e Ambiente, Seconda Università di Napoli. doi:10.1016/j.proeps.2014.06.024
Alfredo Reder et al. / Procedia Earth and Planetary Science 9 ( 2014 ) 214 221 215 Total stresses are instead supposed to remain unchanged for the problems in hand. The prediction of pore water pressure development over time has been carried out by solving Richards equation numerically through SEEP/W FEM code 4. The prediction of stability conditions has been carried out by referring to an infinite slope geometry. Under unsaturated conditions the safety factor F 0 S may be expressed as 5 : tan' s tanb FoS (1) tan z sin cos where is the soil friction angle, is the slope inclination, = is the soil unit weight, z is the vertical height, s is the soil suction, b is the friction angle due to suction. Once determined soil parameters experimentally, equation (1) has been used to quantify the slope safety factor over time corresponding to suction provided by numerical analyses. 2. Soil hydraulic properties Figure 1 plots water content against suction measured in laboratory and flume tests. The two trends result different, starting from the points associated with saturated conditions, which diverge significantly in soil porosities (equal to 65-68%for laboratory specimens and 75% for flume material). The first value of porosity is obtained from available e o measurement performed at beginning of testes on undisturbed samples, for which a value of n about equal to 0.67 is retrieved; in simplified way s is assumed equal to n. In the second case the value of porosity is furnished by the Committee 1. These two experimental trends have been fitted through the van Genuchten relationship 6 : s r 1 s r n m (2) where r is the residual water content, s is the saturated water content, is the inverse of air entry suction, n and m are fitting parameters related to the pore-size distribution The best-fitting achieved (Figure 1) yields the parameters reported in Table 1. Fitting parameters n, m and r result similar for both experimental trends, consistently with the fact that they are traditionally considered as intrinsic soil properties 6 ; on the other hand, air entry values are different and decrease with increasing void ratio, according with both experimental and numerical experiences 7,8,9. Table 1. Calibration parameters of the hydraulic characteristic curves of the investigated soils. s r A n m K 0 Flume tests - Top Soil 0.750 0.235 0.750 2.75 0.50 3.50E-5-1.05 Laboratory specimens - Ashes 0.670 0.300 0.125 3.00 0.55 1.65E-6-1.25 Ashes with pumices 0.540 0.210 0.470 2.75 0.64 5.00E-6-1.85 Pumices (Damiano et al. 12 ) 0.550 0.250 0.337 7.00 0.30 1.00E-5 0.50 The two water retention curves have been turned into hydraulic conductivity functions by referring to the following relationship 10 : K K r 0 1 1 s r r s r 1 n 1 n n1 2 (3)
216 Alfredo Reder et al. / Procedia Earth and Planetary Science 9 ( 2014 ) 214 221 Equation (3) introduces two additional parameters: the fitted matching point at saturation K 0, and the empirical parameter. Schaap and Leij 11 show how the best-fitting of 235 soil samples indicates that K 0 values are about one order of magnitude lower than the hydraulic conductivity at saturation K s, while arises with negative values. They conclude that these two parameters should be merely considered as fitting parameters. Fig.1. Retention curves carried out by laboratory and flume tests. 3. Determination of permeability functions from back-analysis of flume tests and monitoring data Figure 2 draws the discretized geometry adopted in the analyses. It has been refined near the slope surface in order to accommodate the high gradients here assumed by hydraulic variables due to the presence of boundary flows. The top surface normal to the slope development and lowermost surface parallel to the slope development have been modeled as impervious; the down-slope surface (normal to slope development)has been modeled as a seepage surface, in order to simulate the capillary barrier effects induced by the geosynthetic material. Fig.2. Mesh of flume tests D3-D4.
Alfredo Reder et al. / Procedia Earth and Planetary Science 9 ( 2014 ) 214 221 217 Figure 3 plots the development over time of pore water pressures measured during the flume tests D3 and D4. In order to obtain parameters K 0 and of the permeability function, this observed behaviours have been reproduced following two distinct trial-and-error back-analysis procedures. Since the porosity was equal to 0.75 in both tests, back-analyses yielded parameters for the permeability function associated with s =0.75. Fig.3. Development over time of pore pressure for test D3 (a) and D4 (b). The first trial value of K 0 was obtained from the experimental relationship K s -e (Figure 4), provided by permeability constant head tests,as function of e=3. The first trial value of was assumed equal to 0.5 9. Fig.4. Saturated hydraulic conductivity against void ratio. The best-fitting resulted satisfactory (Figure 3). It yielded the parameters reported in Table 1 and the permeability function reported in Figure 5 (blue curve).
218 Alfredo Reder et al. / Procedia Earth and Planetary Science 9 ( 2014 ) 214 221 Fig.5. Hydraulic conductivity curves. The blind prediction referred however to flume test C4 conducted with material put in place at a porosity (0.65) well lowerthan that associated with the previously back-analyzed tests.porosity at C4 test resulted consistent, instead,with that of laboratory test, as indicated by the s value associated with the retention curve fitting laboratory experimental points (Figure 1, red curve).the blind prediction test was therefore carried out by referring to this retention curve.the two additional parameters, K 0 and, needed to characterize the permeability function, were quantified by interpreting numerically field monitoring data, since they contain a layer made of the same material and characterized by the same void ratio value. The back analysis of field data was carried out under a one dimensional flow hypothesis, by adopting the discretized geometries A and B reported in Figure 6.The vertical surfaces have been modeled as impervious, while at the lowermost surfaces an unit gradient has been assumed to governthe flow at the contact with the fractured limestone. Fig.6. Mesh A (a) and mesh B (b). The hydraulic parameters previously obtained from back-analyses of flume tests D3 and D4 have been assigned to the top-soil layers (schemes A and B of figure 6), since top-soil material is featured by high porosity. The hydraulic parameters provided by Damiano et al. 12 for the Cervinara site have been assigned to the pumiceous layer (scheme B of figure 6)(see Table 1). The parameters of the water retention curve of ashes with pumices (schemes A
Alfredo Reder et al. / Procedia Earth and Planetary Science 9 ( 2014 ) 214 221 219 and B of figure 6) have been obtained from the field experimental data of suction (Jet-fill measurements) and volumetric water content (TDR measurements) (Figure 7). The back-analysis at the two different section yielded K 0 and parameters for ashes with pumices and ashes. Fig.7. Ashes with pumices retention curve. The results of the two back-analyses indicate (Figure 8) that during the wet period the fitting is satisfactorily, with prediction matching quite satisfactorily the experimental points. During the dry period prediction departs instead from observations, overall because of the not accounting for evaporation phenomena. Parameters associated with the best fitting are reported in Table 1 and in Figure 5. 4. Blind predictions 4.1. Flume test Fig.8. Development over time of pore water pressure for field cases. Figure 9 plots the evolution of pore water pressures and safety factorpredicted for the flume blind test. Equation 1, set with the input data provided by the organizers ( = 40 ;' = 38 ; s tg b against s), yields a safety factor just less than unit (failure condition)when suction at the depth of 0.1 m reaches the value of 1 kpa. This suction threshold is reached 36-37 minutes after the beginning of the test.
220 Alfredo Reder et al. / Procedia Earth and Planetary Science 9 ( 2014 ) 214 221 4.2. Field case Fig.9. Blind prediction of development over timeof pore water pressure and of F 0S for test C4. For the field case the blind prediction of pore water pressures is reported in figure 10. Results, referred to the time periodfrom 1 January 2012 to 12 February 2012, were obtained by adopting the mesh A. References Fig.10. Blind prediction of development over time of pore water pressure from 1 January 2012 to 12 February 2012. 1. Bogaard TA, Greco R, Olivares L, Picarelli L. The Round Robin test on landslide hydrological modeling at IWL2013. Procedia Earth Planet Sci 2014;xxx:yyy-zzz. 2. Pagano L, Picarelli L, Rianna G, Urciuoli G. A simple numerical procedure for timely prediction of precipitation-induced landslides in unsaturated pyroclastic soils. Landslides 2010;7:273-289. 3. Pagano L, Zingariello MC, Vinale F. A large physical model to simulate flowslides in pyroclastic soils. In: Proceedings of the 1st European Conference on Unsaturated Soils, E-UNSAT, Unsaturated Soils: Advances in Geo-Engineering; Durham, United Kingdom; 2-4 July 2008; Code 83657. 2008. p. 111-115.
Alfredo Reder et al. / Procedia Earth and Planetary Science 9 ( 2014 ) 214 221 221 4. GEO-SLOPE International. SEEP/W User's Manual. Calgary, Alberta, Canada: Geo-Slope International LTD; 2004. 5. Fredlund DG, Rahardjo H. Soil Mechanics for unsaturated soils. Toronto: Wiley & sons; 1993. 6. van Genuchten MT. A closed form equation for predictiong the hydraulic conductivity. Soil Sci Soc Am J 1980; 44:892-898. 7. Wise WR, Clement TP, Molz FJ. Variably saturated modeling of transient drainage: sensitivity to soil properties. J Hydrol 1994;16(1):91-108. 8. Buenfil C, Romero E, Lloret A, Gens A. Experimental study on the hydro-mechanical behaviour of a silty clay. In C. Mancuso & A. Tarantino (eds.).unsaturated soils. Advances in Testing, Modelling and Engineering Applications.. London: Taylor & Francis Group; 2005. p. 15-28. 9. Nicotera MV, Papa R, Urciuoli G. An experimental technique for determining the hydraulic properties of unsaturated pyroclastic soils. Geotechnical Testing Journal 2010; 33(4):263-285. 10. Mualem Y. A new model for predicting the hydraulic conductivity of unsaturated porous media. Water Resour Res 1976; 12:513-522. 11. Schaap MG, Leij FJ. Improved prediction of unsaturated hydraulic conductivity with the Mualem-van Genuchten model. Soil Sei Soc Am J 2000;64:843-851. 12. Damiano E, Olivares L, Picarelli L. Steep-slope monitoring in unsaturated pyroclastic soils. Eng Geol 2012;137-138:1-12.