- Email: [email protected]
Assessment of friction properties at geotextile encapsulated-sand systems' interfaces used for coastal protection
Geotextiles and Geomembranes 44 (2016) 278e286
Contents lists available at ScienceDirect
Geotextiles and Geomembranes journal homepage: www.elsevier.com/locate/geotexmem
Assessment of friction properties at geotextile encapsulated-sand systems' interfaces used for coastal protection Andreia Moreira*, Castorina Silva Vieira, Luciana das Neves, Maria Lurdes Lopes Department of Civil Engineering, Faculty of Engineering, University of Porto, R. Dr. Roberto Frias, 4200-465 Porto, Portugal
a r t i c l e i n f o
a b s t r a c t
Article history: Received 21 June 2015 Received in revised form 8 November 2015 Accepted 24 December 2015 Available online xxx
Friction properties at geotextile interfaces play an important role in the stability of geotextile encapsulated-sand systems in coastal protection. The stabilising action of the frictional force is important to oppose sliding along the contact surface and to prevent deformations, which may lead to structural failures. This paper presents a set of experiments, based on large-scale direct shear tests performed under both cyclic loading and cyclic displacement conditions, examining friction at interfaces between geotextile specimens, sand-filled geosystem elements, and between a sand layer and a sand-filled geosystem element. The results presented here indicate that the friction parameters (i.e., shear strength and friction angle) derived from geotextile specimens are below those obtained for sand-filled elements, which suggests that using the former for the stability analysis is conservative. The tests carried out with a sand layer surface showed that a modification to the shear plane slope is likely to occur e for the sand-filled geosystem element buries into the sand layer. This deformation can result in toe instability, ultimately leading to progressive damage or even collapse of the entire structure. © 2016 Elsevier Ltd. All rights reserved.
Keywords: Geosynthetics Cyclic loading Shear strength Shear testing Sandbags Coastal protection
1. Introduction The use of geotextile encapsulated-sand systems in coastal engineering has grown significantly since its first application in the 1970s (Hornsey et al., 2011), yet, in this context, it is still considered to be innovative. The main obstacle in their wider application is the lack of understanding about key aspects of design (see, Pilarczyk, 2000; and das Neves, 2011) e for the design of geosystems is often based on rather vague experience than on generally accepted calculation methods, as very little systematic research is available on this topic. A few exceptions are Oumeraci and Recio (2009); van Steeg and Breteler (2008); van Steeg and Vastenburg (2010); das Neves (2011); Dassanayake (2013); and das Neves et al. (2015). The stability of sand-filled geosystems under wave loading depends, among others, on the shape of the element (and thus the deformation), for which the degree of filling, the properties of the filling material, and the properties of the element fabric are also important e as shown by Recio (2007), van Steeg and Vastenburg (2010), and Dassanayake and Oumeraci (2009), and on the structural connections in a multilayer element structure. Research by
* Corresponding author. Tel.: þ351 916 307 331. E-mail addresses: [email protected] (A. Moreira), [email protected] (C.S. Vieira), [email protected] (L. das Neves), [email protected] (M.L. Lopes). http://dx.doi.org/10.1016/j.geotexmem.2015.12.002 0266-1144/© 2016 Elsevier Ltd. All rights reserved.
Dassanayake and Oumeraci (2012) suggests that the contribution of friction between sand-filled elements for the hydraulic stability is still not fully clarified e for it has been little investigated. It mainly depends on variables such as the geotextile material itself (both the short term and the long term properties), contact area between elements, sand fill ratio and overlapping length (Dassanayake and Oumeraci, 2012). A number of stability formulae have been derived based on the available research results into the stability of geosystems under wave loading, of which the ones proposed by Recio (2007) and by van Steeg and Vastenburg (2010), consider the contribute of the friction coefficient between elements. However, the stability relations can be more complicated than the simplifications and limits of application assumed in those formulations are, furthermore it should be beared mind that some important parameters for stability change in time due to deformation. Until recently, most of the available data to assess the friction between sand-filled elements was obtained from direct shear testing on geotextile specimens (see, e.g., Moreira et al., 2013; and Moreira et al., 2014) rather than on sand-filled geosystems directly. Since most of the studies focused only on the friction behaviour of geotextileegeomembrane, geotextileegeonet and geomembraneegeonet interfaces (see De and Zimmie, 1998), there are relevant limitations. A number of new research have appeared in recent years focussing on the study of the friction properties at
A. Moreira et al. / Geotextiles and Geomembranes 44 (2016) 278e286
sandegeotextile interfaces (see Vieira et al., 2013) and sand/soil bags' interfaces by means of direct shear tests and vertical compression tests (see Krahn et al., 2007; Matshushima et al., 2008; Lohani et al., 2006; and Aqil et al., 2006). Krahn et al. (2007) concluded that the interface shear strength between sandbags tends to be larger than that deduced from geotextile specimens. A different relative shear movement (with respect to the direct shear test), was investigated in Dassanayake and Oumeraci (2012), with the performance of pullout tests to study the effect of the sand fill ratio and of the friction properties on the pullout forces of geotextile sand containers constructed of woven and non-woven geotextile materials. According to their results, both types of geotextile sandbags exhibit relative friction forces more than 30% higher than the theoretical relative friction forces of the geotextile materials based on direct shear tests. Moreover, there is a clear difference in pullout resistance, which is roughly proportional to the friction coefficients. Ansari et al. (2011) conducted a numerical analysis to investigate the performance of granular materials wrapped on polyethylene bags under vertical compression and cyclic shearing. Despite some noteworthy research, further investigation is needed to establish a relationship between direct shear testing of geotextileegeotextile interfaces and the friction between sand-filled elements. In this paper, an experimental study carried out at the Geosynthetics Laboratory (LGS) of the Faculty of Engineering of the University of Porto (FEUP) is presented. The study is part of a larger research program on the stability analysis of geotextile encapsulated-sand systems under wave loading, with focus on scour development. It is based on direct shear testing carried out under cyclic loading and cyclic displacement conditions. Interfaces between geotextile specimens and sandbags were considered. The tests performed on geotextile specimens aim at characterising the friction properties of the geotextile material itself, whilst the tests carried out on sandbags assess the friction on a more complex interface, because of the additional potential for surface deformation. Sand-geotextile interfaces, simulating the interface at the structures' foundation were also tested. The influence of frequency and amplitude of load variation and continuous cyclic shear on the friction behaviour are addressed herein.
2. Experimental study 2.1. Device and test materials The direct shear device used in the experiments was designed and built at FEUP. It consists of a shear box, a support structure, five hydraulic actuators and respective fluid power unit, an electrical cabinet, internal and external transducers and a dedicated computer (Vieira et al., 2013). The upper shear box is 0.3 m wide, 0.6 m long and 0.15 m high and is fixed in the horizontal direction. The lower shear box dimensions are 0.34 m in width, 0.8 m in length and 0.10 m in height. It is seated on a set of rollers that allow the shear displacement. A rigid base can be inserted in the lower box to make the apparatus suitable for direct contact-area shear testing. If a rigid ring is put in place, a reduced-contact-area (0.3 m 0.6 m) is achieved. Fig. 1 presents a general view of the equipment. The following materials were used in the testing: a needlepunched non-woven (NW) geotextile; two woven geotextiles (W and Wp); and silicate sand (D50 ¼ 273 mm; rs ¼ 2550 kg/m3). Relevant properties of the geotextile materials are listed in Table 1. The non-woven geotextile (NW) and one of the wovens (W) have been used in the physical model testing of encapsulated-sand systems under wave loading within the current research project and past research (e.g., das Neves, 2011 and das Neves et al., 2015).
279
The other woven geotextile (Wp) is commonly used in geosystems for coastal protection. Interfaces are as follows (refer to Fig. 2): NW-NW: interface between two non-woven geotextile specimens; W-W: interface between two woven geotextile specimens; Wp-Wp: interface between two woven geotextile specimens (prototype material); NWb-NWb: interface between sandbags (non-woven geotextile as container); NW-sand: interface between a non-woven geotextile specimen and a sand layer; and NWb-sand: interface between a sandbag (non-woven geotextile as container) and a sand layer. In the tests with geotextile specimens a rigid base is placed inside the lower shear box and the specimens are gripped to the upper and lower boxes, outside of the shear area, by screwed rigid bars (Fig. 2). The upper shear box is then filled with a 5 cm thick sand layer. Tailor-made sandbags are placed in the upper shear (one sandbag with 0.3 m 0.6 m dimensions) and the lower shear (two sandbags with 0.4 m 0.34 m dimensions) boxes (Fig. 3). Chosen configuration of sandbags simulates the overlapping effect in sloped structures. In the lower shear box, the sandbags are positioned so to carefully adjust the alignment of the interface with the shear plane of the test apparatus. Each test is done with unused materials. The filling degree of the sandbags was set at 80%, as commonly used (with some exceptions like in Australia) in prototype bags (Pilarczyk, 2000; Recio, 2007 and PIANC, 2011 e cited in Dassanayake and Oumeraci, 2013). For all tests with a sand layer, the lower and upper shear boxes are, respectively, filled with wet sand (~0.1 L/kg), regularly compacted to the shear plan level, and a geotextile specimen or sandbag. In the tests with sandbags a soft membrane is placed inside the upper box to promote a uniform distribution of normal load over the area of contact between the sandbag and the loading plate. 2.2. Testing program The testing under controlled conditions is essential to be able to extrapolate the results from the test series into an equivalent prototype situation. Main features of the prototype include the cyclic loading due to wave action, and wet and dry conditions and/or alternating dry. Friction between sand-filled elements in a coastal structure is then expected to develop under varying vertical loading conditions as the confining pressure varies with the incoming waves. The confining pressure at the interfaces of an emerged sandfilled geosystem corresponds to the weight of the overlying layers; whilst when submerged it is the submerged weight that should be considered. The orbital motion of the water particles due to the waves progress, also requests the sand-filled elements of a submerged structure repeatedly in seaward and landward directions, potentially abrading the geotextile material. This dynamic effect is reproduced in the cyclic direct shear testing by applying either cyclic variation of the normal stress or displacement-controlled shear movements in the form of sinusoidal wave forms. The amplitude of normal stress variation was defined based on the load variations induced by theoretical regular waves, of a given height and period, over a submerged geosystem. Test conditions A (Table 2) represent typical conditions along the northwest Portuguese coast (Hs ¼ 2 m and Tp ¼ 10 s) according to the statistical analysis in Coelho (2005). These wave loading conditions have been already used in the physical model testing of the stability of geotextile encapsulated-sand systems under wave loading conducted by das Neves (2011). The amplitude and frequency of load variation
280
A. Moreira et al. / Geotextiles and Geomembranes 44 (2016) 278e286
Fig. 1. General view of the direct shear test equipment.
Table 1 Geotextile properties. Property
Unit
Non-woven (NW)
Woven (W)
Woven (Wp)
Raw material (color) Mass per unit area Thickness Tensile strength MDa CDa Elongation at nominal strength MDa CDa Characteristic opening size Water permeability
e g/m2 mm
Polypropylene PP, (white) 300 1.6
Polyethylene PE, (black) 300 e
Polypropylene PP, (sand) 100 e
kN/m
50 30
40 20
198 189
%
50 40 70 70
20 20 230 65
15 11 416 20
a
mm l/m2s
MD: machine direction, CD: cross direction.
65 kPa are tested, whilst in test conditions C only the normal stress of 50 kPa is tested. Under cyclic displacement-controlled testing conditions, two amplitudes of displacement are defined. In test condition D, the amplitude of displacement is not expected to mobilise peak shear stress, whereas in test conditions E, it is expected to mobilise peak shear resistance. Both amplitudes were derived from preliminary results of monotonic direct shear testing. To cross-validate testing results, three tests are performed for each test condition. Mean values are reported and coefficients of variation are presented in the following paragraphs. 3. Discussion of results 3.1. Geotextileegeotextile interface Fig. 2. Geotextiles used in the experimental study.
can influence the direct shear results. By testing conditions B and C, defined to be within a range of possible wave climate conditions along the northwest Portuguese coast, that influence is investigated. In test conditions A and B the normal stresses of 35, 50 and
The shear stress evolution under vertical cyclic loading and cyclic displacement-controlled conditions on geotextileegeotextile interfaces is shown in Fig. 4. Note that only some selected test results are presented. The stress-displacement curves for the NW-NW interface show a well-defined peak shear stress, which occurs for shear displacements that slightly increase
A. Moreira et al. / Geotextiles and Geomembranes 44 (2016) 278e286
281
Fig. 3. Aspects of experiments: (a) sandbag in the upper shear box and levelling control; (b) interface shear testing of sandbags (NWb-NWb); (c) interface shear testing of geotextile specimens (Wp-Wp); and (d) interface shear testing of a sandbag on sand (NWb-sand).
Table 2 Testing conditions. Test conditions
Ab Bb Cb Dc Ec a b c
Vertical cyclic loading
Cyclic shear displacement
Normal stress variation (kPa)
Frequency (Hz)
Amplitude (mm)
±10 ±10 ±20 monotonic
0.1 0.05 0.1
monotonic
±2/±3 ±4/±6
No. of testsa
Interfaces
Frequency (Hz)
0.1 0.1
NW-NW, NW-NW, NW-NW, NW-NW, NW-NW, Total
W-W, NWb-NWb, NW-sand, NWb-Sand W-W, NWb-NWb W-W, NWb-NWb W-W W-W
45 27 9 6 6 93
Including tests under repeated conditions. Maximum shear displacement of 60 mm. Shear displacement rate of 1 mm/min. Constant normal stress (50 kPa)/30 shear cycles.
with normal stress. The results for the W-W interface show a different pattern of curves and higher shear stresses. This is caused by a higher surface roughness of the woven geotextile, which fibers exhibited imbrication and show signs of damage during shear. The results obtained for test conditions A, B and C on NW-NW and W-W interfaces indicate that an increase in the loading frequency results in slightly higher shear strength; and an increase in the amplitude of normal load variation from ±10 kPa to ±20 kPa result in a slight decrease in the peak shear strength. The prototype geotextile (interface Wp) tested under test conditions A, exhibited different shear behaviour, as the shear stress varies upward and downward after the peak. This can be attributed to the geotextile fibers arrangement which follow a repeated pattern every 2 cm approximately (refer to Fig. 2). The shear stresses obtained for this specimens are of the same order as those obtained with the NW-NW interface.
As far as the cyclic displacement-controlled direct shear test results are concerned, it was observed that for small amplitudes of displacement (±2 mm and ±3 mm for NW-NW and W-W interfaces, respectively) the shear stress increased with time, especially within the first cycles, after which it continuous to increase yet at a lower rate. For larger amplitudes of displacement (±4 mm and ± 6 mm for NW-NW and W-W interfaces, respectively) higher shear stresses are achieved within the first cycles. In the following cycles a near-steady shear stress is achieved, similarly to what happens for small amplitudes of displacement. The increase in peak shear stress from the first semi-cycle to the second was substantial demonstrating how an initial shear can influence the surface roughness. The interface peak friction angle, fpeak, can be obtained from direct shear testing, with the adhesion parameter neglected, as follows:
282
A. Moreira et al. / Geotextiles and Geomembranes 44 (2016) 278e286
Fig. 4. Stress e displacement curves for test conditions A, B and C: (a) NW-NW interface; (b) W-W interface; (c) Wp-Wp. displacement-controlled direct shear test results: (d) NWNW interface (test conditions D); (e) NW-NW interface (test conditions E); (f) W-W interface (test conditions D and E).
. fpeak ¼ tan1 tmax sref where tmax is the maximum shear stress recorded during the test and sref is the reference normal shear stress (35, 50 or 65 kPa).
In the cyclic displacement-controlled tests the mean peak shear strength for each cycle (positive and negative shear directions) is used to compute the peak dynamic friction angle at each cycle. Residual friction angles are also of interest to designers because it is likely that critical mechanisms develop in structures incorporating
A. Moreira et al. / Geotextiles and Geomembranes 44 (2016) 278e286
geotextiles. To characterise the residual response, the average of the shear strength registered between 50 and 60 mm of shear displacement is derived from the stress-displacement curves (see Table 3). The derived friction angles of the NW-NW interface are quite similar for test conditions A and B; whereas, an increase in the frequency of excitation from 0.05 Hz to 0.1 Hz leads to the decrease in the friction angles of the interface W-W, especially in residual conditions. A higher amplitude of normal stress variation seems to increase the peak friction angles, yet only the realisation of direct shear tests under values of normal load in excess to 50 kPa could provide a greater assuredness to that observation. Fig. 5 shows the evolution of the peak friction angle for tests performed under test conditions D and E. The friction angle tends to stabilise after a number of cycles, similarly to what occurs for the shear stress. The observed increase in the peak friction angle for the lower amplitudes of displacement is likely due to an increase in surface roughness caused by cyclic shear. When higher amplitudes of displacement are applied a reduction in the friction angle is observed, possibly linked to more widespread damage. In general, the friction angles obtained after 30 cycles of cyclic direct shear are higher than those obtained under test conditions A, B and C. This is related with the geotextiles' abrasion, as was visually observed at the end of the cyclic testing. Regarding the variability of results obtained within the repeated test (as explained earlier, three in each test condition), measures of statistical dispersion are calculated, namely the coefficient of variation. Table 4 shows a summary of the computed coefficients of variation of the peak shear stress for the tested interfaces. Test results of the latter interfaces are presented in the following sections. Most of the coefficients of variation are found to be small, corresponding to a satisfactory degree of repeatability.
283
Fig. 5. Variation of the dynamic peak friction angle with the number of cycles for test conditions D and E.
Table 4 Coefficients of variation of the peak shear stress. Interface
Test conditions A
B
C
D
E
NW-NW W-W Wp-Wp NWb-NWb NW-sand NWb-sand
2.5 5.4 5.4 5.5 6.6 4.6
4.8 4.2 e 4.0 e e
5.2 1.4 e 2.9 e e
7.6 7.6 e e e e
1.8 1.3 e e e e
3.2. Sandbagesandbag interface Sandbags made with the same non-woven geotextile fabric (NW in the testing with specimens) are tested under vertical cyclic loading conditions. Due to the rounded geometry of the sandbags, the contact are between the sandbag positioned inside the upper shear box and the loading plate is reduced with respect to the shear box dimensions. This reduction changes the vertical pressure on the sandbags interface by increasing it slightly. To take this into account, the contact area was measured at the end of each test, as a verification of repeatability. Fig. 6 shows the stress-displacement curves obtained for the interface NWb-NWb. As observed in earlier research studies (e.g., Dassanayake and Oumeraci, 2012; yet, it is important to note that this study is rather different from the one presented here), the shear stress between sandbags is significantly higher than that estimated for geotextile specimens (approximately 35% higher, under the same test conditions). In addition to this, other differences are noticeable: the shear strength loss after the peak is substantially lower; and the peak shear stress is achieved at higher shear displacements (9e11 mm), as compared to the peak shear stress of the material only (4e5 mm). This may be due to internal sand movement as the bags deform to adjust themselves to the rectangular shear box within the first millimeters of shear.
The influence of load frequency in peak shear stress is limited; yet results indicate that the residual shear stresses decreases with the increase in load frequency. With respect to the amplitude of normal stress, a variation from ±10 kPa to ±20 kPa results in a decrease in the peak shear stress. The interface friction angles of sandbags are higher as well. The obtained peak friction angles for test conditions A, B and C are, respectively, 26.8 , 27.6 and 25.3 , and 20.3 , 24 and 22.3 for large displacement conditions (50e60 mm), accordingly. In the shear behaviour of a sandbagesandbag interface, two aspects (not present in geotextileegeotextile interfaces) shall be considered: the sand can move inside the bags during the test, modifying the interface roughness and the sandbags shape; and the sandbags can move inside the shear boxes because they are not attached to them as geotextile specimens are in a conventional direct shear test. Both aspects are likely responsible for the higher shear stresses obtained for the NWb-NWb interface. Fig. 7 shows an aspect of bag deformation at the end of testing, happening inside the upper box at half bag length, over the boundary of the two sandbags in the lower box. When sandbags are sheared against each other the shear surfaces' flatness is influenced by the arrangement of sediments inside the container. The irregularities that developed in the shear interface increase the shear stress
Table 3 Friction angles of the geotextileegeotextile interfaces. Interface
NW-NW W-W Wp-Wp
Test conditions A
Test conditions B
Test conditions C
fpeak ( )
f50e60 mm ( )
fpeak ( )
f50e60 mm ( )
fpeak ( )
f50e60 mm ( )
15.5 31.2 20.2
6 23.1 14.8
16.6 33.1 e
6.7 30.2 e
18 37 e
7.3 20.5 e
284
A. Moreira et al. / Geotextiles and Geomembranes 44 (2016) 278e286
Fig. 6. Stress-displacement curves for the NWb-NWb interface: (a) test conditions A and C; (b) test conditions B.
stress increase after the peak of strength is sharpest as the normal stress increases. Due to the sand layer deformation (associated with the sandbag burial), tests with the NWb-sand interface were interrupted at a shear displacement of approximately 35 mm, in order to safeguard the equipment. This type of interface, which exists in prototype at the subsoil foundation of sand-filled geotextile structures, may experience the above described deformation mechanisms which are known to cause global structural instabilities. Scour mechanisms add to the weakening of stabilising forces. 4. Conclusions
Fig. 7. Deformation due to internal sand movement (after testing). Geotextile-sand and sandbag-sand interfaces.
beyond what is expected from the geotextile interfaces. In a direct shear test between two geotextile specimens the shear surface is horizontal, aligned by the shear boxes. The results of cyclic direct shear tests on geotextile specimens and sandbags on top of a sand layer surface (NW-sand and NWbsand interfaces) are shown in Fig. 8. The obtained results show an increase in the shear strength when sand is sheared against a geotextile specimen or a sandbag, as compared to that achieved in tests with equal surfaces (e.g. NWNW or NWb.NWb). In the NW-sand interface the peak shear stress is achieved at approximately 9 mm of displacement and the stress-softening after the peak is barely noticeable. The shear stress-displacement graph for the NWb-sand interface exhibits a specific feature that is a continued increase of the shear stress with shear displacement. This increase is at an initial stage very sharp after which it continues to increase, yet at a much lower rate. To compute the interface peak friction angle, the peak shear stress is taken equal to the shear stress at the end of the sharp phase (at approximately 8 mm shear displacement, see Fig. 8). The continuous growth in the shear stress of this interface should be carefully considered as it is likely related with modifications in the shear interface configuration during shear; at the start of testing it is planar, but it changes as the sandbag buries into the sand layer and the sediments are dragged (Fig. 9). Not unexpectedly, the shear
The frictional behaviour of 6 interfaces has been tested and results presented. The shear stresses and friction angles of the interfaces have been estimated based on cyclic direct shear tests. For purposes of comparison, the peak failure envelope graphs and friction angles of all interfaces are shown in Fig. 10. The plotted lines do not cross the origin, but y-intercepts, normally named adhesion parameter, are close to zero. Because these values fell out of the range of tested normal stresses (35e65 kPa) its extrapolation is bound to great uncertainty and they are not reported. The shear strength and friction angles of the interface between sandbags are larger than that deduced from geotextile specimens interface. It is estimated that is larger by ~35%, attributed to the internal movement of sand inside the bags along with their deformation during shear, as visually observed at the end of the tests. The use of results from direct shear tests on geotextile specimens when designing a sand-filled geosystem in coastal protection is likely conservative. From peak to residual conditions a reduction of ~25% in the shear stress and friction angle is observed for the sandbags interface. That reduction is clearly higher in the tests with geotextileegeotextile interfaces. Cyclic direct shear tests performed under different frequencies of excitation showed that the shear behaviour of interfaces between geotextile specimens and sandbags is not greatly influenced by it within the range of frequencies tested (0.1 and 0.05 Hz), which is the one with higher relevance in the present context of application, as justified in Section 2.2. Notwithstanding, it appears that an increase in the loading frequency results in the slightly decrease of the shear stress and friction angle for residual conditions. An increase in the amplitude of vertical load variation results in a decrease in the shear strength of these interfaces, for peak
A. Moreira et al. / Geotextiles and Geomembranes 44 (2016) 278e286
285
Fig. 8. Stress-displacement curves for test conditions A: (a) interface NW-sand; (b) interface NWb-sand.
Fig. 9. Sand layer surface after testing of the NWb-sand interface.
In the cyclic displacement-controlled shear tests performed under test conditions D (small amplitude of displacement), the shear stress and friction angles tend to increase during the first cycles, whilst for test conditions E (high amplitude of displacement) some degradation of the interface shear strength is observed. During 30 cycles of shearing, the dynamic friction angle increased by approximately 33% for test conditions D and 9.5% for test conditions E. It is likely that the repeated cycling abraded the geotextiles' surface, changing the interface frictional properties. The NWb-sand interface exhibits a higher shear resistance as compared to the interface between two sandbags. For this interface the shear stress increased monotonically, rapidly in the first ~8 mm, and at a lower rate from then on. Yet, this increase does not indicate a higher stability of the sandbag, as there are factors such as bag and sand deformation, under these the initial planar shear interface becomes non-horizontal and the estimated shear stresses are higher. Such a deformation also occurs in the NWb-NWb interface, as the sandbag inside the upper shear box may deform while it buries into the sandbags placed in the lower shear box. This is also a model effect, for which the different shear boxes' areas contribute. Deformation at the subsoil layer is known to induce general structural instability. The knowledge about the development of failure mechanisms in such coastal protection structures is extensively addressed in literature but it should be supplemented with the study on friction development as it influences the friction parameters to be adopted in design. The monitoring of lateral displacements for each sandbag layer during operation can help to define a maximum shear displacement after which excessive structural deformation compromises the overall stability. Acknowledgements The authors are grateful to the Portuguese Science and Technology Foundation (FCT) for their financial support through the PTDC/FEDER (European funds for Regional Development) programme (research Project ScourCoast PTDC/ECM/122760/2010).
Fig. 10. Peak failure envelopes of the different geotextile interfaces (test conditions A).
conditions. The friction angle seems to be influenced by loading amplitude, although better assuredness of this result can only be obtained with the performance of tests under different values of normal stress.
References Ansari, Y., Merifield, R., Yamamoto, H., Sheng, D., 2011. Numerical analysis of soilbags under compression and cyclic shear. Comput. Geotech 38, 659e668. Aqil, U., Matsushima, K., Mohri, Y., Yamazaki, S., Tatsuoka, F., 2006. Lateral shearing tests on geosynthetic soil bags. In: Proceedings of the 8th International Conference on Geosynthetics, Yokohama, Japan, vol. 4. Millpress, Rotterdam, pp. 1703e1706.
286
A. Moreira et al. / Geotextiles and Geomembranes 44 (2016) 278e286
~o de frentes urbanas para diferentes intervenço ~es Coelho, Carlos, 2005. Riscos de exposiça de defesa costeira. PhD Thesis. Universidade de Aveiro, Portugal (in Portuguese). das Neves, L., 2011. Experimental Stability Analysis of Geotextile Encapsulated-sand Systems under Wave-loading (PhD thesis). University of Porto, Faculty of Engineering, Portugal, p. 290. das Neves, L., Moreira, A., Taveira-Pinto, F., Lopes, M.L., Veloso-Gomes, F., 2015. Performance of submerged nearshore sand-filled geosystems for coastal protection. J. Coast. Eng. ISSN: 0378-3839 95, 147e159. http://dx.doi.org/10.1016/j. coastaleng.2014.10.005. Dassanayake, D., 2013. Experimental and Numerical Modelling of Hydraulic Sta€t bility of Geotextile Sand Containers for Coastal Protection (PhD thesis). Fakulta Architektur, Bauingeni eurwesen und Umweltwissenschaften der Technischen €t Carolo-Wilhelmina zu Braunschweig, Germany, p. 182. Universita Dassanayake, D.T., Oumeraci, H., 2009b. Planned Research on the Hydraulic Stability of Geotextile Sand Containers, 7. FZK-Kolloquium, Potenziale für die Maritime Wirtschaft, Hannover, Germany. Dassanayake, D., Oumeraci, H., 2012. Important engineering properties of geotextile sand containers and their effect on the hydraulic stability of GSC-structures. Terra Aqua J. 127, 3e11. Dassanayake, D., Oumeraci, H., 2013. Hydraulic stability formulae and monograms for coastal structures made of geotextile sand containers. In: Proceedings of the 7th International Conference on Asian and Pacific Coasts, Bali, Indonesia, pp. 180e186. De, A., Zimmie, T.F., 1998. Estimation of dynamic interfacial properties of geosynthetics. Geosynth. Int. 5 (1e2), 17e39. Hornsey, W.P., Carley, J.T., Coghlan, I.R., Cox, R.J., 2011. Geotextile sand container shoreline protection systems: design and application. Geotext. Geomembr. 29, 425e439. Krahn, T., Blatz, J., Alfaro, M., Bathurst, R.J., 2007. Large-scale interface shear testing of sandbag dyke materials. Geosynth. Int. 14 (2), 118e126. http://dx.doi.org/ 10.1680/gein.2007.14.2.119.
Lohani, T.N., Matshushima, K., Aqil, U., Mohri, Y., Tatsuoka, F., 2006. Evaluating the strength and deformations characteristics of soil bag pile from full-scale laboratory tests. Geosynth. Int. 13 (6), 246e264. Matshushima, K., Aqil, U., Mohri, Y., Tatshuoka, F., 2008. Shear strength and deformation characteristics of geosynthetic soil bags staked horizontally and inclined. Geosynth. Int. 15 (2), 119e135. Moreira, A., Vieira, C.S., das Neves, L., Lopes, M.L., 2013. Influence of the displacement rate on the direct shear behavior of geotextile interfaces. In: Proceedings of the 1st Geosintec Iberia, Seville, Spain, pp. 171e178. Moreira, A., Vieira, C.S., das Neves, L., Lopes, M.L., 2014. Friction behaviour of geotextile interfaces under cyclic conditions. In: 10th International Conference on Geosynthetics, Berlin, Germany. Oumeraci, H., Recio, J., 2009. In: Kim, Y.C. (Ed.), Geotextile Sand Containers for Shore Protection. Handbook of Coastal and Ocean Engineering. World Scientific Publishing Co. Pte. Ltd, pp. 553e600. PIANC, 2011. The Application of Geosynthetics in Waterfront Areas-PIANC Report No. 113 by MarCom Working Group 113. Pilarczyk, K.W., 2000. Geosynthetics and Geosystems in Hydraulic and Coastal Engineering. A.A. Balkema, Rotterdam, the Netherlands. Recio, J., 2007. Hydraulic Stability of Geotextile Sand Containers for Coastal Structures e Effect of Deformations and Stability Formulae (PhD thesis). Leichtweib Institute for Hydraulic Engineering and Water Resources, Germany. van Steeg, P., Breteler, K., 2008. Large Scale Physical Model Tests on the Stability of Geocontainers. DELTARES. van Steeg, P., Vastenburg, E., 2010. Large Scale Physical Model Tests on the Stability of Geotextile Tubes. DELTARES. Vieira, C.S., Lopes, M.L., Caldeira, L.M., 2013. Sandegeotextile interface characterization through monotonic and cyclic direct shear tests. Geosynth. Int. 20 (1), 26e38.
Contents lists available at ScienceDirect
Geotextiles and Geomembranes journal homepage: www.elsevier.com/locate/geotexmem
Assessment of friction properties at geotextile encapsulated-sand systems' interfaces used for coastal protection Andreia Moreira*, Castorina Silva Vieira, Luciana das Neves, Maria Lurdes Lopes Department of Civil Engineering, Faculty of Engineering, University of Porto, R. Dr. Roberto Frias, 4200-465 Porto, Portugal
a r t i c l e i n f o
a b s t r a c t
Article history: Received 21 June 2015 Received in revised form 8 November 2015 Accepted 24 December 2015 Available online xxx
Friction properties at geotextile interfaces play an important role in the stability of geotextile encapsulated-sand systems in coastal protection. The stabilising action of the frictional force is important to oppose sliding along the contact surface and to prevent deformations, which may lead to structural failures. This paper presents a set of experiments, based on large-scale direct shear tests performed under both cyclic loading and cyclic displacement conditions, examining friction at interfaces between geotextile specimens, sand-filled geosystem elements, and between a sand layer and a sand-filled geosystem element. The results presented here indicate that the friction parameters (i.e., shear strength and friction angle) derived from geotextile specimens are below those obtained for sand-filled elements, which suggests that using the former for the stability analysis is conservative. The tests carried out with a sand layer surface showed that a modification to the shear plane slope is likely to occur e for the sand-filled geosystem element buries into the sand layer. This deformation can result in toe instability, ultimately leading to progressive damage or even collapse of the entire structure. © 2016 Elsevier Ltd. All rights reserved.
Keywords: Geosynthetics Cyclic loading Shear strength Shear testing Sandbags Coastal protection
1. Introduction The use of geotextile encapsulated-sand systems in coastal engineering has grown significantly since its first application in the 1970s (Hornsey et al., 2011), yet, in this context, it is still considered to be innovative. The main obstacle in their wider application is the lack of understanding about key aspects of design (see, Pilarczyk, 2000; and das Neves, 2011) e for the design of geosystems is often based on rather vague experience than on generally accepted calculation methods, as very little systematic research is available on this topic. A few exceptions are Oumeraci and Recio (2009); van Steeg and Breteler (2008); van Steeg and Vastenburg (2010); das Neves (2011); Dassanayake (2013); and das Neves et al. (2015). The stability of sand-filled geosystems under wave loading depends, among others, on the shape of the element (and thus the deformation), for which the degree of filling, the properties of the filling material, and the properties of the element fabric are also important e as shown by Recio (2007), van Steeg and Vastenburg (2010), and Dassanayake and Oumeraci (2009), and on the structural connections in a multilayer element structure. Research by
* Corresponding author. Tel.: þ351 916 307 331. E-mail addresses: [email protected] (A. Moreira), [email protected] (C.S. Vieira), [email protected] (L. das Neves), [email protected] (M.L. Lopes). http://dx.doi.org/10.1016/j.geotexmem.2015.12.002 0266-1144/© 2016 Elsevier Ltd. All rights reserved.
Dassanayake and Oumeraci (2012) suggests that the contribution of friction between sand-filled elements for the hydraulic stability is still not fully clarified e for it has been little investigated. It mainly depends on variables such as the geotextile material itself (both the short term and the long term properties), contact area between elements, sand fill ratio and overlapping length (Dassanayake and Oumeraci, 2012). A number of stability formulae have been derived based on the available research results into the stability of geosystems under wave loading, of which the ones proposed by Recio (2007) and by van Steeg and Vastenburg (2010), consider the contribute of the friction coefficient between elements. However, the stability relations can be more complicated than the simplifications and limits of application assumed in those formulations are, furthermore it should be beared mind that some important parameters for stability change in time due to deformation. Until recently, most of the available data to assess the friction between sand-filled elements was obtained from direct shear testing on geotextile specimens (see, e.g., Moreira et al., 2013; and Moreira et al., 2014) rather than on sand-filled geosystems directly. Since most of the studies focused only on the friction behaviour of geotextileegeomembrane, geotextileegeonet and geomembraneegeonet interfaces (see De and Zimmie, 1998), there are relevant limitations. A number of new research have appeared in recent years focussing on the study of the friction properties at
A. Moreira et al. / Geotextiles and Geomembranes 44 (2016) 278e286
sandegeotextile interfaces (see Vieira et al., 2013) and sand/soil bags' interfaces by means of direct shear tests and vertical compression tests (see Krahn et al., 2007; Matshushima et al., 2008; Lohani et al., 2006; and Aqil et al., 2006). Krahn et al. (2007) concluded that the interface shear strength between sandbags tends to be larger than that deduced from geotextile specimens. A different relative shear movement (with respect to the direct shear test), was investigated in Dassanayake and Oumeraci (2012), with the performance of pullout tests to study the effect of the sand fill ratio and of the friction properties on the pullout forces of geotextile sand containers constructed of woven and non-woven geotextile materials. According to their results, both types of geotextile sandbags exhibit relative friction forces more than 30% higher than the theoretical relative friction forces of the geotextile materials based on direct shear tests. Moreover, there is a clear difference in pullout resistance, which is roughly proportional to the friction coefficients. Ansari et al. (2011) conducted a numerical analysis to investigate the performance of granular materials wrapped on polyethylene bags under vertical compression and cyclic shearing. Despite some noteworthy research, further investigation is needed to establish a relationship between direct shear testing of geotextileegeotextile interfaces and the friction between sand-filled elements. In this paper, an experimental study carried out at the Geosynthetics Laboratory (LGS) of the Faculty of Engineering of the University of Porto (FEUP) is presented. The study is part of a larger research program on the stability analysis of geotextile encapsulated-sand systems under wave loading, with focus on scour development. It is based on direct shear testing carried out under cyclic loading and cyclic displacement conditions. Interfaces between geotextile specimens and sandbags were considered. The tests performed on geotextile specimens aim at characterising the friction properties of the geotextile material itself, whilst the tests carried out on sandbags assess the friction on a more complex interface, because of the additional potential for surface deformation. Sand-geotextile interfaces, simulating the interface at the structures' foundation were also tested. The influence of frequency and amplitude of load variation and continuous cyclic shear on the friction behaviour are addressed herein.
2. Experimental study 2.1. Device and test materials The direct shear device used in the experiments was designed and built at FEUP. It consists of a shear box, a support structure, five hydraulic actuators and respective fluid power unit, an electrical cabinet, internal and external transducers and a dedicated computer (Vieira et al., 2013). The upper shear box is 0.3 m wide, 0.6 m long and 0.15 m high and is fixed in the horizontal direction. The lower shear box dimensions are 0.34 m in width, 0.8 m in length and 0.10 m in height. It is seated on a set of rollers that allow the shear displacement. A rigid base can be inserted in the lower box to make the apparatus suitable for direct contact-area shear testing. If a rigid ring is put in place, a reduced-contact-area (0.3 m 0.6 m) is achieved. Fig. 1 presents a general view of the equipment. The following materials were used in the testing: a needlepunched non-woven (NW) geotextile; two woven geotextiles (W and Wp); and silicate sand (D50 ¼ 273 mm; rs ¼ 2550 kg/m3). Relevant properties of the geotextile materials are listed in Table 1. The non-woven geotextile (NW) and one of the wovens (W) have been used in the physical model testing of encapsulated-sand systems under wave loading within the current research project and past research (e.g., das Neves, 2011 and das Neves et al., 2015).
279
The other woven geotextile (Wp) is commonly used in geosystems for coastal protection. Interfaces are as follows (refer to Fig. 2): NW-NW: interface between two non-woven geotextile specimens; W-W: interface between two woven geotextile specimens; Wp-Wp: interface between two woven geotextile specimens (prototype material); NWb-NWb: interface between sandbags (non-woven geotextile as container); NW-sand: interface between a non-woven geotextile specimen and a sand layer; and NWb-sand: interface between a sandbag (non-woven geotextile as container) and a sand layer. In the tests with geotextile specimens a rigid base is placed inside the lower shear box and the specimens are gripped to the upper and lower boxes, outside of the shear area, by screwed rigid bars (Fig. 2). The upper shear box is then filled with a 5 cm thick sand layer. Tailor-made sandbags are placed in the upper shear (one sandbag with 0.3 m 0.6 m dimensions) and the lower shear (two sandbags with 0.4 m 0.34 m dimensions) boxes (Fig. 3). Chosen configuration of sandbags simulates the overlapping effect in sloped structures. In the lower shear box, the sandbags are positioned so to carefully adjust the alignment of the interface with the shear plane of the test apparatus. Each test is done with unused materials. The filling degree of the sandbags was set at 80%, as commonly used (with some exceptions like in Australia) in prototype bags (Pilarczyk, 2000; Recio, 2007 and PIANC, 2011 e cited in Dassanayake and Oumeraci, 2013). For all tests with a sand layer, the lower and upper shear boxes are, respectively, filled with wet sand (~0.1 L/kg), regularly compacted to the shear plan level, and a geotextile specimen or sandbag. In the tests with sandbags a soft membrane is placed inside the upper box to promote a uniform distribution of normal load over the area of contact between the sandbag and the loading plate. 2.2. Testing program The testing under controlled conditions is essential to be able to extrapolate the results from the test series into an equivalent prototype situation. Main features of the prototype include the cyclic loading due to wave action, and wet and dry conditions and/or alternating dry. Friction between sand-filled elements in a coastal structure is then expected to develop under varying vertical loading conditions as the confining pressure varies with the incoming waves. The confining pressure at the interfaces of an emerged sandfilled geosystem corresponds to the weight of the overlying layers; whilst when submerged it is the submerged weight that should be considered. The orbital motion of the water particles due to the waves progress, also requests the sand-filled elements of a submerged structure repeatedly in seaward and landward directions, potentially abrading the geotextile material. This dynamic effect is reproduced in the cyclic direct shear testing by applying either cyclic variation of the normal stress or displacement-controlled shear movements in the form of sinusoidal wave forms. The amplitude of normal stress variation was defined based on the load variations induced by theoretical regular waves, of a given height and period, over a submerged geosystem. Test conditions A (Table 2) represent typical conditions along the northwest Portuguese coast (Hs ¼ 2 m and Tp ¼ 10 s) according to the statistical analysis in Coelho (2005). These wave loading conditions have been already used in the physical model testing of the stability of geotextile encapsulated-sand systems under wave loading conducted by das Neves (2011). The amplitude and frequency of load variation
280
A. Moreira et al. / Geotextiles and Geomembranes 44 (2016) 278e286
Fig. 1. General view of the direct shear test equipment.
Table 1 Geotextile properties. Property
Unit
Non-woven (NW)
Woven (W)
Woven (Wp)
Raw material (color) Mass per unit area Thickness Tensile strength MDa CDa Elongation at nominal strength MDa CDa Characteristic opening size Water permeability
e g/m2 mm
Polypropylene PP, (white) 300 1.6
Polyethylene PE, (black) 300 e
Polypropylene PP, (sand) 100 e
kN/m
50 30
40 20
198 189
%
50 40 70 70
20 20 230 65
15 11 416 20
a
mm l/m2s
MD: machine direction, CD: cross direction.
65 kPa are tested, whilst in test conditions C only the normal stress of 50 kPa is tested. Under cyclic displacement-controlled testing conditions, two amplitudes of displacement are defined. In test condition D, the amplitude of displacement is not expected to mobilise peak shear stress, whereas in test conditions E, it is expected to mobilise peak shear resistance. Both amplitudes were derived from preliminary results of monotonic direct shear testing. To cross-validate testing results, three tests are performed for each test condition. Mean values are reported and coefficients of variation are presented in the following paragraphs. 3. Discussion of results 3.1. Geotextileegeotextile interface Fig. 2. Geotextiles used in the experimental study.
can influence the direct shear results. By testing conditions B and C, defined to be within a range of possible wave climate conditions along the northwest Portuguese coast, that influence is investigated. In test conditions A and B the normal stresses of 35, 50 and
The shear stress evolution under vertical cyclic loading and cyclic displacement-controlled conditions on geotextileegeotextile interfaces is shown in Fig. 4. Note that only some selected test results are presented. The stress-displacement curves for the NW-NW interface show a well-defined peak shear stress, which occurs for shear displacements that slightly increase
A. Moreira et al. / Geotextiles and Geomembranes 44 (2016) 278e286
281
Fig. 3. Aspects of experiments: (a) sandbag in the upper shear box and levelling control; (b) interface shear testing of sandbags (NWb-NWb); (c) interface shear testing of geotextile specimens (Wp-Wp); and (d) interface shear testing of a sandbag on sand (NWb-sand).
Table 2 Testing conditions. Test conditions
Ab Bb Cb Dc Ec a b c
Vertical cyclic loading
Cyclic shear displacement
Normal stress variation (kPa)
Frequency (Hz)
Amplitude (mm)
±10 ±10 ±20 monotonic
0.1 0.05 0.1
monotonic
±2/±3 ±4/±6
No. of testsa
Interfaces
Frequency (Hz)
0.1 0.1
NW-NW, NW-NW, NW-NW, NW-NW, NW-NW, Total
W-W, NWb-NWb, NW-sand, NWb-Sand W-W, NWb-NWb W-W, NWb-NWb W-W W-W
45 27 9 6 6 93
Including tests under repeated conditions. Maximum shear displacement of 60 mm. Shear displacement rate of 1 mm/min. Constant normal stress (50 kPa)/30 shear cycles.
with normal stress. The results for the W-W interface show a different pattern of curves and higher shear stresses. This is caused by a higher surface roughness of the woven geotextile, which fibers exhibited imbrication and show signs of damage during shear. The results obtained for test conditions A, B and C on NW-NW and W-W interfaces indicate that an increase in the loading frequency results in slightly higher shear strength; and an increase in the amplitude of normal load variation from ±10 kPa to ±20 kPa result in a slight decrease in the peak shear strength. The prototype geotextile (interface Wp) tested under test conditions A, exhibited different shear behaviour, as the shear stress varies upward and downward after the peak. This can be attributed to the geotextile fibers arrangement which follow a repeated pattern every 2 cm approximately (refer to Fig. 2). The shear stresses obtained for this specimens are of the same order as those obtained with the NW-NW interface.
As far as the cyclic displacement-controlled direct shear test results are concerned, it was observed that for small amplitudes of displacement (±2 mm and ±3 mm for NW-NW and W-W interfaces, respectively) the shear stress increased with time, especially within the first cycles, after which it continuous to increase yet at a lower rate. For larger amplitudes of displacement (±4 mm and ± 6 mm for NW-NW and W-W interfaces, respectively) higher shear stresses are achieved within the first cycles. In the following cycles a near-steady shear stress is achieved, similarly to what happens for small amplitudes of displacement. The increase in peak shear stress from the first semi-cycle to the second was substantial demonstrating how an initial shear can influence the surface roughness. The interface peak friction angle, fpeak, can be obtained from direct shear testing, with the adhesion parameter neglected, as follows:
282
A. Moreira et al. / Geotextiles and Geomembranes 44 (2016) 278e286
Fig. 4. Stress e displacement curves for test conditions A, B and C: (a) NW-NW interface; (b) W-W interface; (c) Wp-Wp. displacement-controlled direct shear test results: (d) NWNW interface (test conditions D); (e) NW-NW interface (test conditions E); (f) W-W interface (test conditions D and E).
. fpeak ¼ tan1 tmax sref where tmax is the maximum shear stress recorded during the test and sref is the reference normal shear stress (35, 50 or 65 kPa).
In the cyclic displacement-controlled tests the mean peak shear strength for each cycle (positive and negative shear directions) is used to compute the peak dynamic friction angle at each cycle. Residual friction angles are also of interest to designers because it is likely that critical mechanisms develop in structures incorporating
A. Moreira et al. / Geotextiles and Geomembranes 44 (2016) 278e286
geotextiles. To characterise the residual response, the average of the shear strength registered between 50 and 60 mm of shear displacement is derived from the stress-displacement curves (see Table 3). The derived friction angles of the NW-NW interface are quite similar for test conditions A and B; whereas, an increase in the frequency of excitation from 0.05 Hz to 0.1 Hz leads to the decrease in the friction angles of the interface W-W, especially in residual conditions. A higher amplitude of normal stress variation seems to increase the peak friction angles, yet only the realisation of direct shear tests under values of normal load in excess to 50 kPa could provide a greater assuredness to that observation. Fig. 5 shows the evolution of the peak friction angle for tests performed under test conditions D and E. The friction angle tends to stabilise after a number of cycles, similarly to what occurs for the shear stress. The observed increase in the peak friction angle for the lower amplitudes of displacement is likely due to an increase in surface roughness caused by cyclic shear. When higher amplitudes of displacement are applied a reduction in the friction angle is observed, possibly linked to more widespread damage. In general, the friction angles obtained after 30 cycles of cyclic direct shear are higher than those obtained under test conditions A, B and C. This is related with the geotextiles' abrasion, as was visually observed at the end of the cyclic testing. Regarding the variability of results obtained within the repeated test (as explained earlier, three in each test condition), measures of statistical dispersion are calculated, namely the coefficient of variation. Table 4 shows a summary of the computed coefficients of variation of the peak shear stress for the tested interfaces. Test results of the latter interfaces are presented in the following sections. Most of the coefficients of variation are found to be small, corresponding to a satisfactory degree of repeatability.
283
Fig. 5. Variation of the dynamic peak friction angle with the number of cycles for test conditions D and E.
Table 4 Coefficients of variation of the peak shear stress. Interface
Test conditions A
B
C
D
E
NW-NW W-W Wp-Wp NWb-NWb NW-sand NWb-sand
2.5 5.4 5.4 5.5 6.6 4.6
4.8 4.2 e 4.0 e e
5.2 1.4 e 2.9 e e
7.6 7.6 e e e e
1.8 1.3 e e e e
3.2. Sandbagesandbag interface Sandbags made with the same non-woven geotextile fabric (NW in the testing with specimens) are tested under vertical cyclic loading conditions. Due to the rounded geometry of the sandbags, the contact are between the sandbag positioned inside the upper shear box and the loading plate is reduced with respect to the shear box dimensions. This reduction changes the vertical pressure on the sandbags interface by increasing it slightly. To take this into account, the contact area was measured at the end of each test, as a verification of repeatability. Fig. 6 shows the stress-displacement curves obtained for the interface NWb-NWb. As observed in earlier research studies (e.g., Dassanayake and Oumeraci, 2012; yet, it is important to note that this study is rather different from the one presented here), the shear stress between sandbags is significantly higher than that estimated for geotextile specimens (approximately 35% higher, under the same test conditions). In addition to this, other differences are noticeable: the shear strength loss after the peak is substantially lower; and the peak shear stress is achieved at higher shear displacements (9e11 mm), as compared to the peak shear stress of the material only (4e5 mm). This may be due to internal sand movement as the bags deform to adjust themselves to the rectangular shear box within the first millimeters of shear.
The influence of load frequency in peak shear stress is limited; yet results indicate that the residual shear stresses decreases with the increase in load frequency. With respect to the amplitude of normal stress, a variation from ±10 kPa to ±20 kPa results in a decrease in the peak shear stress. The interface friction angles of sandbags are higher as well. The obtained peak friction angles for test conditions A, B and C are, respectively, 26.8 , 27.6 and 25.3 , and 20.3 , 24 and 22.3 for large displacement conditions (50e60 mm), accordingly. In the shear behaviour of a sandbagesandbag interface, two aspects (not present in geotextileegeotextile interfaces) shall be considered: the sand can move inside the bags during the test, modifying the interface roughness and the sandbags shape; and the sandbags can move inside the shear boxes because they are not attached to them as geotextile specimens are in a conventional direct shear test. Both aspects are likely responsible for the higher shear stresses obtained for the NWb-NWb interface. Fig. 7 shows an aspect of bag deformation at the end of testing, happening inside the upper box at half bag length, over the boundary of the two sandbags in the lower box. When sandbags are sheared against each other the shear surfaces' flatness is influenced by the arrangement of sediments inside the container. The irregularities that developed in the shear interface increase the shear stress
Table 3 Friction angles of the geotextileegeotextile interfaces. Interface
NW-NW W-W Wp-Wp
Test conditions A
Test conditions B
Test conditions C
fpeak ( )
f50e60 mm ( )
fpeak ( )
f50e60 mm ( )
fpeak ( )
f50e60 mm ( )
15.5 31.2 20.2
6 23.1 14.8
16.6 33.1 e
6.7 30.2 e
18 37 e
7.3 20.5 e
284
A. Moreira et al. / Geotextiles and Geomembranes 44 (2016) 278e286
Fig. 6. Stress-displacement curves for the NWb-NWb interface: (a) test conditions A and C; (b) test conditions B.
stress increase after the peak of strength is sharpest as the normal stress increases. Due to the sand layer deformation (associated with the sandbag burial), tests with the NWb-sand interface were interrupted at a shear displacement of approximately 35 mm, in order to safeguard the equipment. This type of interface, which exists in prototype at the subsoil foundation of sand-filled geotextile structures, may experience the above described deformation mechanisms which are known to cause global structural instabilities. Scour mechanisms add to the weakening of stabilising forces. 4. Conclusions
Fig. 7. Deformation due to internal sand movement (after testing). Geotextile-sand and sandbag-sand interfaces.
beyond what is expected from the geotextile interfaces. In a direct shear test between two geotextile specimens the shear surface is horizontal, aligned by the shear boxes. The results of cyclic direct shear tests on geotextile specimens and sandbags on top of a sand layer surface (NW-sand and NWbsand interfaces) are shown in Fig. 8. The obtained results show an increase in the shear strength when sand is sheared against a geotextile specimen or a sandbag, as compared to that achieved in tests with equal surfaces (e.g. NWNW or NWb.NWb). In the NW-sand interface the peak shear stress is achieved at approximately 9 mm of displacement and the stress-softening after the peak is barely noticeable. The shear stress-displacement graph for the NWb-sand interface exhibits a specific feature that is a continued increase of the shear stress with shear displacement. This increase is at an initial stage very sharp after which it continues to increase, yet at a much lower rate. To compute the interface peak friction angle, the peak shear stress is taken equal to the shear stress at the end of the sharp phase (at approximately 8 mm shear displacement, see Fig. 8). The continuous growth in the shear stress of this interface should be carefully considered as it is likely related with modifications in the shear interface configuration during shear; at the start of testing it is planar, but it changes as the sandbag buries into the sand layer and the sediments are dragged (Fig. 9). Not unexpectedly, the shear
The frictional behaviour of 6 interfaces has been tested and results presented. The shear stresses and friction angles of the interfaces have been estimated based on cyclic direct shear tests. For purposes of comparison, the peak failure envelope graphs and friction angles of all interfaces are shown in Fig. 10. The plotted lines do not cross the origin, but y-intercepts, normally named adhesion parameter, are close to zero. Because these values fell out of the range of tested normal stresses (35e65 kPa) its extrapolation is bound to great uncertainty and they are not reported. The shear strength and friction angles of the interface between sandbags are larger than that deduced from geotextile specimens interface. It is estimated that is larger by ~35%, attributed to the internal movement of sand inside the bags along with their deformation during shear, as visually observed at the end of the tests. The use of results from direct shear tests on geotextile specimens when designing a sand-filled geosystem in coastal protection is likely conservative. From peak to residual conditions a reduction of ~25% in the shear stress and friction angle is observed for the sandbags interface. That reduction is clearly higher in the tests with geotextileegeotextile interfaces. Cyclic direct shear tests performed under different frequencies of excitation showed that the shear behaviour of interfaces between geotextile specimens and sandbags is not greatly influenced by it within the range of frequencies tested (0.1 and 0.05 Hz), which is the one with higher relevance in the present context of application, as justified in Section 2.2. Notwithstanding, it appears that an increase in the loading frequency results in the slightly decrease of the shear stress and friction angle for residual conditions. An increase in the amplitude of vertical load variation results in a decrease in the shear strength of these interfaces, for peak
A. Moreira et al. / Geotextiles and Geomembranes 44 (2016) 278e286
285
Fig. 8. Stress-displacement curves for test conditions A: (a) interface NW-sand; (b) interface NWb-sand.
Fig. 9. Sand layer surface after testing of the NWb-sand interface.
In the cyclic displacement-controlled shear tests performed under test conditions D (small amplitude of displacement), the shear stress and friction angles tend to increase during the first cycles, whilst for test conditions E (high amplitude of displacement) some degradation of the interface shear strength is observed. During 30 cycles of shearing, the dynamic friction angle increased by approximately 33% for test conditions D and 9.5% for test conditions E. It is likely that the repeated cycling abraded the geotextiles' surface, changing the interface frictional properties. The NWb-sand interface exhibits a higher shear resistance as compared to the interface between two sandbags. For this interface the shear stress increased monotonically, rapidly in the first ~8 mm, and at a lower rate from then on. Yet, this increase does not indicate a higher stability of the sandbag, as there are factors such as bag and sand deformation, under these the initial planar shear interface becomes non-horizontal and the estimated shear stresses are higher. Such a deformation also occurs in the NWb-NWb interface, as the sandbag inside the upper shear box may deform while it buries into the sandbags placed in the lower shear box. This is also a model effect, for which the different shear boxes' areas contribute. Deformation at the subsoil layer is known to induce general structural instability. The knowledge about the development of failure mechanisms in such coastal protection structures is extensively addressed in literature but it should be supplemented with the study on friction development as it influences the friction parameters to be adopted in design. The monitoring of lateral displacements for each sandbag layer during operation can help to define a maximum shear displacement after which excessive structural deformation compromises the overall stability. Acknowledgements The authors are grateful to the Portuguese Science and Technology Foundation (FCT) for their financial support through the PTDC/FEDER (European funds for Regional Development) programme (research Project ScourCoast PTDC/ECM/122760/2010).
Fig. 10. Peak failure envelopes of the different geotextile interfaces (test conditions A).
conditions. The friction angle seems to be influenced by loading amplitude, although better assuredness of this result can only be obtained with the performance of tests under different values of normal stress.
References Ansari, Y., Merifield, R., Yamamoto, H., Sheng, D., 2011. Numerical analysis of soilbags under compression and cyclic shear. Comput. Geotech 38, 659e668. Aqil, U., Matsushima, K., Mohri, Y., Yamazaki, S., Tatsuoka, F., 2006. Lateral shearing tests on geosynthetic soil bags. In: Proceedings of the 8th International Conference on Geosynthetics, Yokohama, Japan, vol. 4. Millpress, Rotterdam, pp. 1703e1706.
286
A. Moreira et al. / Geotextiles and Geomembranes 44 (2016) 278e286
~o de frentes urbanas para diferentes intervenço ~es Coelho, Carlos, 2005. Riscos de exposiça de defesa costeira. PhD Thesis. Universidade de Aveiro, Portugal (in Portuguese). das Neves, L., 2011. Experimental Stability Analysis of Geotextile Encapsulated-sand Systems under Wave-loading (PhD thesis). University of Porto, Faculty of Engineering, Portugal, p. 290. das Neves, L., Moreira, A., Taveira-Pinto, F., Lopes, M.L., Veloso-Gomes, F., 2015. Performance of submerged nearshore sand-filled geosystems for coastal protection. J. Coast. Eng. ISSN: 0378-3839 95, 147e159. http://dx.doi.org/10.1016/j. coastaleng.2014.10.005. Dassanayake, D., 2013. Experimental and Numerical Modelling of Hydraulic Sta€t bility of Geotextile Sand Containers for Coastal Protection (PhD thesis). Fakulta Architektur, Bauingeni eurwesen und Umweltwissenschaften der Technischen €t Carolo-Wilhelmina zu Braunschweig, Germany, p. 182. Universita Dassanayake, D.T., Oumeraci, H., 2009b. Planned Research on the Hydraulic Stability of Geotextile Sand Containers, 7. FZK-Kolloquium, Potenziale für die Maritime Wirtschaft, Hannover, Germany. Dassanayake, D., Oumeraci, H., 2012. Important engineering properties of geotextile sand containers and their effect on the hydraulic stability of GSC-structures. Terra Aqua J. 127, 3e11. Dassanayake, D., Oumeraci, H., 2013. Hydraulic stability formulae and monograms for coastal structures made of geotextile sand containers. In: Proceedings of the 7th International Conference on Asian and Pacific Coasts, Bali, Indonesia, pp. 180e186. De, A., Zimmie, T.F., 1998. Estimation of dynamic interfacial properties of geosynthetics. Geosynth. Int. 5 (1e2), 17e39. Hornsey, W.P., Carley, J.T., Coghlan, I.R., Cox, R.J., 2011. Geotextile sand container shoreline protection systems: design and application. Geotext. Geomembr. 29, 425e439. Krahn, T., Blatz, J., Alfaro, M., Bathurst, R.J., 2007. Large-scale interface shear testing of sandbag dyke materials. Geosynth. Int. 14 (2), 118e126. http://dx.doi.org/ 10.1680/gein.2007.14.2.119.
Lohani, T.N., Matshushima, K., Aqil, U., Mohri, Y., Tatsuoka, F., 2006. Evaluating the strength and deformations characteristics of soil bag pile from full-scale laboratory tests. Geosynth. Int. 13 (6), 246e264. Matshushima, K., Aqil, U., Mohri, Y., Tatshuoka, F., 2008. Shear strength and deformation characteristics of geosynthetic soil bags staked horizontally and inclined. Geosynth. Int. 15 (2), 119e135. Moreira, A., Vieira, C.S., das Neves, L., Lopes, M.L., 2013. Influence of the displacement rate on the direct shear behavior of geotextile interfaces. In: Proceedings of the 1st Geosintec Iberia, Seville, Spain, pp. 171e178. Moreira, A., Vieira, C.S., das Neves, L., Lopes, M.L., 2014. Friction behaviour of geotextile interfaces under cyclic conditions. In: 10th International Conference on Geosynthetics, Berlin, Germany. Oumeraci, H., Recio, J., 2009. In: Kim, Y.C. (Ed.), Geotextile Sand Containers for Shore Protection. Handbook of Coastal and Ocean Engineering. World Scientific Publishing Co. Pte. Ltd, pp. 553e600. PIANC, 2011. The Application of Geosynthetics in Waterfront Areas-PIANC Report No. 113 by MarCom Working Group 113. Pilarczyk, K.W., 2000. Geosynthetics and Geosystems in Hydraulic and Coastal Engineering. A.A. Balkema, Rotterdam, the Netherlands. Recio, J., 2007. Hydraulic Stability of Geotextile Sand Containers for Coastal Structures e Effect of Deformations and Stability Formulae (PhD thesis). Leichtweib Institute for Hydraulic Engineering and Water Resources, Germany. van Steeg, P., Breteler, K., 2008. Large Scale Physical Model Tests on the Stability of Geocontainers. DELTARES. van Steeg, P., Vastenburg, E., 2010. Large Scale Physical Model Tests on the Stability of Geotextile Tubes. DELTARES. Vieira, C.S., Lopes, M.L., Caldeira, L.M., 2013. Sandegeotextile interface characterization through monotonic and cyclic direct shear tests. Geosynth. Int. 20 (1), 26e38.