Khalid Farrag Geosynthetic Engineering Research Laboratory (GERL), Louisiana Transportation Research Center (LTRC), Gourier Road, Baton Rouge, Louisiana 70803, USA Yal c i n B. Acar* Civil Engineering Department, Louisiana State University, Baton Rouge, Louisiana 70803, USA & Il an Juran Department of Civil Engineering, Polytechnic University, 333 Jay Street, Brooklyn, New York, New York 11201, USA (Received 30 September 1991: accepted 25 November 1991) ABS TRACT Testing equipment, specimen preparation and testing procedures are presented f or load-controlled and displacement-rate-controlled pull-out tests f or geosynthetic reinforcements in granular soils. The influence of the test type, confining pressure, soil density, boundary conditions, and geotextile characteristics on pull-out load-displacement response of selected geogrids embedded in sand are evaluated. Implications f or testing procedure and analysis are discussed. I NT RODUCT I ON A wi de var i et y o f geot ext i l es a n d geogr i ds ar e n o w a va i l a bl e f or ci vi l e n g i n e e r i n g a p p l i c a t i o n s ( Bo n a p a r t e et al. , 1987; Koer ner , 1990). I n *To whom correspondence should be addressed. 133 GeotextUes and Geomembranes 0266--1144/93/$06.00 1993 Elsevier Science Publishers Ltd. England. Printed in Great Britain 134 Khalid Farrag. Yalcin B. Acar, l l an Juran selecting a specific geotextile or geogrid for reinforcement of embank- ments and slopes, the following performance aspects need to be assessed: (i) st ress-st rai n-st rai n rate behavior of the confined reinforcement- soil composite system, (ii) pull-out performance and the associated load transfer mechanism. A st andard testing procedure for confined properties ofgeosynthetics does not yet exist. The large number of factors that affect the interface properties of the confined reinforcement raise major difficulties in providing comparabl e test results. There is a wide scatter in the available pull-out test results. These differences in results are due to the use of different types of pull-out devices, the associated boundary effects, testing procedures and soil placement and compaction schemes (Juran e t a l . , 1988). The increasing use of geosynthetics in soil reinforcement prompts the need to develop and standardize methods in evaluating the in-soil mechani cal characteristics and interface properties of these materials. It becomes necessary to design adequate testing equipment, establish reliable testing procedures, and develop appropriate inter- pretation schemes. This paper presents the pull-out testing program implemented at the Geosynthetic Engineering Research Laboratory (GERL), a laboratory of Louisiana Transport at i on Research Center (LTRC) and Civil Engineer- ing Depart ment of Louisiana State University, to develop reliable testing procedures and interpretation schemes in evaluation of the short-term and long-term pull-out performance ofgeosynthetic reinforcements. The fundament al aspects of pull-out testing equi pment and procedure are reviewed. A pull-out box that incorporates schemes which overcome most of the current limitations is presented. The factors that affect the measured interaction properties are evaluated by performance evaluation tests. Implications for engineering analysis are discussed. FUNDAMENTAL ASPECTS OF PULL-OUT TESTING The shear stress-strain relationship developed along the soil-reinforce- ment interface is commonl y tested in a direct shear box and/ or a pull-out box. In the direct shear box, tests are usually conducted in accordance with the conventional procedure used in investigating interface properties (Acar e t a l . , 1982). Commonl y, a soil sample resting on a geotextile is sheared al ong the interface and the shear force-displacement behavi or is recorded. In the pull-out box, a geotextile confined in soil is pulled out and often the pull-out load and the front displacement are recorded. In Pull-out resistance of geogrid reinforcements 135 bot h tests, results are oft en expressed in t erms of a friction ratio, t an 6/ t an ~, (also called t he efficiency factor) where 6 is t he soi l -rei nforcement i nt erface friction angl e a nd O is t he soil friction angle. Effi ci ency factors r angi ng from 0.6 to 1.0 for geotextiles a nd val ues l arger t han one for geogrids are report ed ( Jur an et a l . , 1988; Koerner, 1990). Fi gure 1 present s a compar i s on of efficiency factors obt ai ned from tests conduct ed usi ng pul l -out a nd direct s hear testing equi pment . The frictional resi st ances report ed for di fferent types of i ncl usi ons in dense sands are f ound to be great er in pul l -out tests t han t hose obt ai ned in direct s hear tests (Jewell, 1979; Schl osser & Gui l l oux, 1979). I ngol d a nd Te mpl e ma n (1979) f ound compar abl e val ues in bot h tests for geogrids tested under low conf i ni ng pressures. However, under hi gher nor mal stresses, hi gher s hear strength val ues were obt ai ned i n pul l -out tests. Koer ner (1986) report ed hi gher s hear resi st ance for geogrids in pul l -out tests while, under hi gher conf i ni ng pressures, direct s hear tests gave hi gher shear resistance. Rowe e t a l . (1985) report ed t hat bot h tests give appr oxi mat el y equal val ues of s hear resi st ance for geotextiles tested in dense gr anul ar fill while for geogrids in a loose fill, pul l -out tests r ender ed si gni fi cant l y l ower values. Di rect s hear a nd pul l -out tests are associ at ed with di fferent testing procedures, l oadi ng paths, failure mechani sms, and boundar y conditions. Consequent l y, the i nt erface frictional par amet er s obt ai ned from bot h 3.5 ~ Direct Pull-out Reference Rein fon:emem Soil Shear Type Type 3.0 t'-'O------'" + lngot,J(l~83) TemmrSX-~ ~ S ~ 1 ---4~--- A Colins(1980) Nonwovm Gr ovel (cnu4t~l) I---O--. + xo~,,~ 09s6) T ~ SX-2 mm s ~ 2.s ~ , | . . . . . . . . x . . . . .9s6) ~z-100 D~S, ~ 2.0 ~ Row,: (1985) T e n s a r $R-2 Loose sand ~ 1.5 1.0 A . . . . . . . . L ~ . . . . . . . A C] o.sl ~ 0.0 ~ 0 50 100 150 200 250 300 NORMAL STRESS (kN/m2) Fi g. I . Co mp a r i s o n o f t he e f f i c i e nc y f act or s o b t a i n e d f or s o i l - g e o s y n t h e t i c s ys t ems i n di r ect s h e a r a n d p u l l - o u t boxes . 136 Khalid Farrag, Yalcin B. Acar, llan Juran tests can vary and often provide conflicting results. The reasons for the reported differences between the pull-out and direct-shear test results can be categorized as: (1) Rest rai ned dilatancy: Cohesionless soils will dilate at low to medi um confi ni ng pressures and at high densities. If this dilatancy is restrained in testing, the confi ni ng pressure al ong the interface will increase until a state is reached where failure can be achieved without any volume change (critical state). The amount of this restraint and the magnitude of the increase in confi nement will depend upon the type and geometry of the imposed boundary conditions. These distances are often selected arbitrarily in the boxes used. Furthermore, the pull-out boxes are often designed with fixed boundari es while displacements are allowed in the top platten in the direct shear box. Fixed boundari es may promote restrained di l at ancy and increased confinement. Therefore, even when the initial states are identical, different stress-deformation behavi or will be developed due to confi nement induced by boundary conditions. (2) Fai l ure and interaction mechani sm: The interaction mechani sm is different in direct shear and pull-out. In direct shear tests, when the box di mensi on is large, the mobilized shear strain is postulated to be uni forml y distributed al ong the soil-geosynthetic interface. While in the pull-out tests, mobilized strain is a combi nat i on of the interface shear strain and reinforcement extension. This coupled mechani sm results in a non-uniform shear strain-stress distribution along the reinforcement. A realistic experimental model of pull-out behavi or ofgeosynthetics in the field shoul d incorporate extensibility. Pull-out boxes are more appropriate when extensibility is to be incorporated. A standard design for pull-out testing devices does not yet exist. Therefore, box di mensi ons and testing procedures differ from one box to the other. The di mensi ons of the box are usually selected as large as technically possible with the ant i ci pat i on that boundary effects will be minimized. An evaluation of most of the available equi pment and testing parameters demonstrates that the differences in testing equi pment and procedures make comparisons of available test results extremely difficult (Juran et al., 1988; Farrag, 1990). The following is a list of desirable features in pull-out testing procedure and equipment: (1) The pull-out tests reported are often conducted at a controlled displacement rate. Few pull-out tests are reported under load- PuU-out resistance of geogrid reinforcements 13 7 controlled mode (e.g., Tzong & Cheng-Kuang, 1987). Pull-out testing equi pment should have the capability to provide load- controlled tests to facilitate investigation of time-dependent, in-soil creep behavior. ( 2 ) In displacement-rate controlled tests, a range of pull-out rates are reported varying from 0.1 ram/ rai n to 20 mm/ mi n. Myles (1982) studied the frictional resistance of the geotextile interface under different strain rates (10-75 mm/ mi n) in a direct shear box and showed t hat results had little sensitivity to this range of displace- ment rates. In wide strip tests, Rowe and Ho (1986) showed that the geogrid tension strength varies with the applied rate of strain. It is necessary to establish the effect of pull-out displacement-rate on test results in different types of soils. ( 3 ) The interaction between the soil and the side walls of the pull-out box can affect test results. Soil confi nement is usually applied by means of flexible air-bags to insure uniform distribution of normal stress al ong a pl ane (Christopher, 1976; Palmeira & Milligan, 1989; Ingold, 1983). The applied confining stress is partially carried out by the side wall friction causing a reduction in the normal pressure applied at the reinforcement level. Some investigators inserted lubricated membranes along the walls in an attempt to provide a low friction boundary (Jewell, 1980). Alternatively, the specimen width/box width ratio may be chosen so as to minimize the effect of friction by side-walls. (4) The interaction between the reinforcement-soil system and the rigid front wall can also influence test results. As the reinforcement is pulled out of the box, the lateral earth pressure developed on the front face can result in an increase in the pull-out resistance. Palmeira and Milligan (1989) used a front wall with different degrees of roughness to investigate the effect of friction on the front wall on pull-out test results. Chri st opher (1976) incorporated sleeves around the pull-out slot to transfer the poi nt of appl i cat i on of the pull-out load far behi nd the rigid front wall. Ot her investigators (Williams & Houl i han, 1987) used flexible front faces to mi ni mi ze the rigid front wall effect. The front wall interaction should be mi ni mi zed in a pull-out box. (5) The thicknesses of the soil above and below the reinforcement (referred as the soil thickness), differ according to the available clear height of the box. If soil thickness is small, the interaction between the soil-reinforcement system and the boundaries will significantly influence the shear resistance along the interface. Brand and Duffy (1987) studied the effect of soil thickness on pull- 138 Khal i d Farrag, Yalein B. Acar, Ilan Juran out resistance of a geogrid in clays. Thei r results demonstrate that as the soil thickness increases, pull-out resistance decreases until a mi ni mum pull-out load is obtained. It is essential to decide upon the soil thickness in conducting a standard pull-out test. (6) Different specimen preparation and compaction procedures are utilized to insure uniform soil density. Soil compaction is carried out by means of an electric j ack hammer (Johnston, 1985), standard proctor hammer (Saxena & Budiman, 1985), hand t ampi ng devices (Elias, 1979) and by mechanical tamping (Anderson & Neilsen, 1985). A hopper with flexible tube is also used to insure uniform soil placement (Palmeira & Milligan, 1989; Jewell, 1980). Different specimen preparation procedures will result in differences in fabric leading to significantly different stress-deformation behavior. It then becomes essential to standardize specimen preparation procedures in commercial testing of confined behavi or of geosynthetic reinforcements. (7) Several investigators (Christopher, 1976; Koerner, 1986; Brand & Duffy, 1987) clamped the reinforcement outside the box. This technique may result in an unconfi ned front portion of the reinforcement. Consequently, the effective interface area will vary and confi ned/ unconfi ned properties of the geosynthetic will couple during the pull-out test. (8) The existing pull-out testing equi pment are often instrumented to moni t or only the displacement and the pull-out resistance at the front face. For extensible reinforcements such as geogrids and geotextiles, it is essential to moni t or the displacements along the inclusion in order to interpret the load transfer mechanism and estimate the pull-out resistance in the field. The measured pull-out resistance is influenced by factors which include the details of testing equi pment and procedures discussed above, and also the following compositional and environmental variables influencing the behavior of the composite system: (i) the compositional characteristics of the geosynthetic reinforce- ment such as its type, geometry and configuration leading to different extensibility and load-strain-strain rate behavior, (ii) compositional characteristics of the reinforced soil such as its grain size distribution, and void ratio, (iii) environmental variables defining the initial state of the composite system such as the overburden pressure and the nature of the imposed deformation (load controlled or displacement rate controlled). Pull-out resistance of geogrid reinforcements 139 Two pull-out boxes which permit an evaluation of different factors influencing pull-out response are designed and constructed. A testing program is i mpl ement ed in performance evaluation of the boxes. TESTI NG PROGRAM The pri mary objective of the performance evaluation study was to assess the sensitivity of test results obt ai ned in the designed equi pment to the changes in the fundament al testing parameters (geogrid type, specimen width, sleeve length, soil thickness, displacement rate, soil density and confi ni ng pressure) and to establish a data base for development of reliable testing and interpretation procedures. Table 1 presents the i mpl ement ed testing program. Equipment A schematic diagram of the box and associated i nst rument at i on are presented in Fig. 2. Figure 3 presents a view of the pull-out box and the sand hopper system. The different component s and characteristics of the box are described below: (i) The i nner di mensi ons of the pull-out box are 1-52 m (60 in) long, 0-90 m (36 in) wide and 0.76 m (30 in) high. The box is constructed in modul ar units to allow changes in box di mensi ons and front wall openi ng size. A sleeve is used at the facing to transfer the interface pull-out load behi nd the rigid front wall. An air bag is used at the top to provide a uni forml y distributed vertical pressure. (ii) Pull-out load is applied by a hydraulic loading system through cl ampi ng plates that extend inside the sleeve. This arrangement ascertains that the geosynthetic specimen remains confined t hroughout the test. The l oadi ng system can appl y either a const ant pull-out rate or a constant pull-out load. Load controlled mode is used to evaluate the creep behavior. (iii) The front displacements, the pull-out rate and the pull-out load are acqui red/ moni t ored at the cl ampi ng plates by means of a LVDT, velocity transducer and a load cel, respectively. Displacements al ong the reinforcement are measured by tell-tale wires connected to LVDTs. Two earth pressure cells are placed on the front wall. 140 Khalid Farrag, Yalein R Aear, llan Juran Ta b l e I Testing Program and Testing Parameters Purpose ~ Geogrid Dimension Confining Unit Displacement type L/ W pressure weight rate, 6 (m) (kN/m 2) (kN/m 3) (mm/min) Repeatability Specimen width Tensar SR2 1.00/0.30 48-2 16.5 Conwed i.00/0-30 9027 Tensar SR2 1.00/0-30 48.2 16.7 1.00/1~45 i.00/0.60 1-00/0-75 !.00/0-30 48.2 16.4 Sleeve length Tensar SR2 (0, 20 and 30 cm) Soil thickness Tensar SR2 (20, 40, 60 and 70 cm) Displacement Tensar SR2 rate 6.0 4.0 20.0 1.00/0.30 48.2 16.4 6-0 1.00/ff30 48.2 16.4 Soil density Tensar SR2 1.00/0.30 48.2 15.7 16.4 16.7 17-0 Confining Tensar SR2 1.00/1~ 30 34-0 16.5 pressure 48.2 16.5 69-0 16.5 Conwed 1.00/0.30 48-2 16.7 9027 96.0 16.7 140-0 16.7 Effect of Conwed 1.00/0-30 48.2 16.7 transverse 9027 fibs 4-0 6.0 10-0 20-0 4-0 6-0 6-0 6.0 4.0 4.0 4.0 2.0 ~Unless otherwise noted, all tests are conducted at a displacement rate of 6 mm/min, at a soil thickness of 60 cm (30 cm top and 30 cm bottom) and a sleeve length of 30 cm. ( i v) An e l e v a t e d s a n d h o p p e r wi t h v a c u u m e x t r a c t i o n is u s e d t o f a c i l i t a t e s a n d p l a c e me n t a n d r e mo v a l f r o m t h e box. ( v) A d a t a a c q u i s i t i o n s y s t e m mo n i t o r s t h e i n p u t p a r a me t e r s a n d r e c o r d s t h e r e s p o n s e p a r a me t e r s (e. g. d i s p l a c e me n t r at e, a p p l i e d Pull-out resistance of geogrid reinforcements 141 I 30.5 l m l 30.5 l c m ~ cm [ AI ~ i V l T M 5cm 153 cm 5 cm Fig. 2. A s c h e ma t i c d i a g r a m o f t h e p u l l - o u t box. Fi g. 3. A v i e w o f t h e b o x a n d t h e s a n d l o a d i n g s ys t em. ~.~2 Khahd Farrag, Yalcin B. Acar, llan Juran 100 80 ~ - 20 0 .01 q t o n . . . . . . I ~ 1 . . . . . . I . . . . . . . .1 1 10 GRAIN SIZE (nun) Fig. 4. Grai n size distribution of the sand used in the study. vertical pressure, pul l -out l oad, soil pressure at t he walls, and di spl acement s at t he front and at l ocat i ons al ong t he rei nforce- ment). Mat eri al s A l ocal l y avai l abl e commer ci al bl ast i ng sand is used. The grai n size di st ri but i on of' this s and is pr esent ed i n Fig. 4. Thi s sand is poor l y gr aded 9,it h a n eftective di amet er , Duo of 0-26 mm a nd ma xi mum and mi n i mu m densi t i es of l 7.4 kN/ m 3 ( 110-9 pcf) and 15.6 kN/ m 3 (99.0 pc0, respectively. l vt~ di fferent types of geogrids, Tensar SR2 and Conwed-9027, are s cl eoed in per l or mance assessment eval uat i on tests. Geogr i d speci mens ol 0. t 5 rt~, 0~30 m, 0.45 m and 0.75 m i n wi dt h and 0.90 m i n l engt h are ~c'~(t'd i n ~he box. "~ e s t pr oc e dur e i he sand i~ pour ed i nt o t he pul l -out box from t he el evat ed hopper ~hrough a flexible outlet. The el evat ed hopper moves al ong t he box to : t*~am a relatively uni f or m s and fill. The s and is pl aced i n four layers of ~ cm ~0 m) each, l evel ed and t hen compact ed to t he desi red relative ~s i v, , Compact i on effort is i mpar t ed manual l y usi ng a vi brat i ng ' c~, , , hammer . After compact i on, t he densi t y is meas ur ed with a . ~.~.. clc~s4ty gauge. When t he sleeve el evat i on is reached, t he ~:~, , w~ ~ pl aced on the compact ed bot t om l ayer of sand. Geogr i d Pull-out resistance of geogrid reinforcements 143 specimens are bolted to the cl ampi ng plates which are then inserted through sleeves of 30 cm (1 ft) length on the front wall. The inextensible tell-tale wires used for displacement measurement along the reinforce- ment are connected to the five LVDTs placed on a rear end table. The wires are encased in a polyethylene tube placed along the reinforcement. Subsequently the top layer of sand is placed in layers of 15 cm and compacted. A st andard test is defined as a test conducted on a Tensar SR2 geogrid of 30 cm (1 ft) width and 92 cm (3 ft) length at a confi ni ng pressure of 48.2 kN/ m 2 (7 psi), sand density of 16.5 kN/ m 3 (105 pcf), soil thickness of 61 cm [30 cm placed on the top hal f and 30 cm placed on the bottom halt], a displacement-rate of 6 mm/ mi n, and a sleeve length of 30 cm (1 ft). ANALYSIS OF TEST RESULTS Repeatability of tests and load transfer mechanism The repeatability of the pull-out response in standard tests is demon- strated in Fig. 5a while the repeatability of displacements obt ai ned at the front and rear ends of the reinforcement are displayed in Fig. 5b. Geosynthetics are extensible reinforcements. Consequently, methods developed in evaluating the shear resistance of rigid reinforcements cannot be applied in interpreting pull-out test results for geosynthetics. The displacement distribution al ong the geogrid is indispensable in interpretation of test results and in evaluation of pull-out resistance. The interface shear distribution al ong the confined geogrid is established by nodal -di spl acement measurements. The locations of the nodes for displacement measurement are displayed in Fig. 6. The displacement distribution al ong the geogrid is measured at different pull-out load levels. Figure 7 presents the development of displacements along the reinforcement with pull-out load. The progressive movement of the geogrid nodes duri ng testing is demonstrated. The displacements recorded al ong the geogrids demonstrate the non- linearity in the displacement distribution along the geogrid due to coupl i ng of the shear strain at the interface with material elongation resulting in a hi gher extension at the front part. It is necessary to i mpl ement a load-transfer procedure to determine the confined extension properties and interface shear stress-strain behavior. Such a procedure is presented by Juran et al. (1990). 144 Khalid Farrag, Yalcin B. Acar, / / an J uran 8O E 6 0 Z a I 0 4 0 . J I-- 0 ! . J "J 2 0 =) n GEOGRI D: TENSAR SR2 O'n = 4 8 k N/ m 2 { 7 p s i ) Vn = 1 6 . 5 k N/ m 3 { 1 0 5 p c f ) = 6 r am/ r ai n 0 I I I I I I 20 4 0 6 0 80 I 0 0 120 140 F RONT DI S P L A CE ME NT ( r a m} ( a ) 140 E 120 I - Z "~ I 0 0 I.iJ U . ~ 8 0 . J n or) 6 0 . J . ~ 4 0 E3 0 Z 20 0 GEOGRID: TENSAR SR2 o- n = 4 8 k N/ m2 { 7 p s i ) )'n = 16.5 k N/ m 5 ( 105pc f ) 2 0 0 4 0 0 6 0 0 8 0 0 I 0 0 0 1200 T I ME ( s e c ) ( b ) Fi g. 5. Re p e a t a b i l i t y o f tests. ( a ) P u l l - o u t l o a d v e r s u s f r o n t d i s p l a c e me n t , ( b) d i s p l a c e - m e n t s d i s t r i b u t i o n a l o n g t h e r e i n f o r c e me n t . Geogrid speci men width The de ve l opme nt o f soi l - wal l fri cti on al ong the box si de wal l s can i nf l ue nc e t he test results. The appl i e d c o nf i ni ng pressure c an be parti al l y c ar de d by the fri cti on al ong the si de walls. As a result, the c o nf i ni ng pressure appl i e d o n the s pe c i me n wi l l be reduced. Thi s reduct i on wi l l be Pull-out resistance of geogrid reinforcements Disolocement Meosurement Locotions 145 Fig. 6. Displacement measurement locations along a geogrid. o ~- o tt.I , ~ U 0 = = ~ , ~ GEOGRID: TENSAR SR2 o" n = 4 8 k N / m 2 ( 7ps i ) ~C~\ ~ , , ~ - - P e o k L o o d I 2 3 4 GE OGRI D N OD E S F i g . 7. D i s t r i b u t i o n o f d i s p l a c e m e n t s a l o n g t h e g e o g r i d . mor e pr e domi na nt near t he si de walls. Speci men wi dt h/ box wi dt h rat i o s houl d be so sel ect ed as to keep t he r ei nf or cement a nd t he si de wal l at a di st ance whi ch will mi ni mi z e t hi s effect. I n or der to eval uat e t he effect of t he si de fri ct i on i n t he desi gned box, (a) l oad cells are pl aced at t he cent er a n d cl ose to t he wal l of t he box at geogr i d level a nd t he vert i cal r esponse to vert i cal pr essur e is r ecor ded, a nd (b) pul l - out tests are conduct ed usi ng di fferent s peci men wi dt hs. Fi gur e 8a demons t r at es t hat l ower cell pr essur es are r ecor ded cl ose to t he wal l t han at t he cent er. Fi gur e 8(b) shows t he results of tests on Tens ar geogr i d s peci mens of di fferent wi dt hs. These tests were conduct ed unde r a conf i ni ng pr essur e of 48-2 k N/ m 2 (7 psi), soil densi t y of 16.7 k N/ m 3 (106 pcf) a n d pul l - out rat e of 4 mm/ mi n. A si gni f i cant r educt i on i n t he pul l - out r esi st ance is not ed whe n s peci men wi dt h is i ncr eas ed to 76 cm (2.5 ft). Tests c onduc t e d wi t h geogr i d s peci mens of 30 cm to 61 cm (1 ft to 146 Khal i d Farrag, Yalcin B. Acar, l l an Juran @ i i 1.10 1.00 0 . 9 0 0.80 ~:- 3 4 kPa 4 8 k P a [ ] 7 0 k P a g B O X WI D T H = 2 L = 0 . 9 0 m i i i t t i t i i = i J i 0 0 . 1 0 . 2 0 . 3 0 . 4 0 . 5 0 . 6 0 . 7 NORMALIZED DISTANCE FROM SIDE WALL, x/L ( a ) 8O "E 6 0 S p e c i me n s wi t h 0 . 5 0 , 0 . 4 5 8~ 6 0 m Wi d t h ~ / / I z ~. ~- - - ~- - ' - ~- - - - ~ o ..J 40 t - 0 . 7 5 m W i d t h ? _J J 2 0 o. GEOGRI D: TENSAR SR2 o- n = 4 8 k N / m 2 ( T p s i ) ~n = 16.7 k N/ m3 ( 1 0 6 p c f ) = 4 mm/ mi n I I I 0 20 40 6O 80 F RONT DI S P L A CE ME NT ( r a m) ( b ) Fig. 8. (a) The change in vertical pressure ratio at the geogrid level along the width of the box. (hi The effect of geogrid specimen width on the pull-out response. 2 fl) rendered results within the variability of tests. It is concl uded that a dJsta nee of at least 15.0 cm is left between the speci men edge and the box wall (i.e. a maxi mum speci men width 2 ft for the designed box) to mi ni mi ze the side wall effect on test results. Pull-out resistance of geogrid reinforcement,~ 147 Sleeve length The i nt er act i on bet ween t he soi l -i ncl usi on system and t he rigid front wall of t he pul l -out box can affect t he meas ur ed pul l -out resistance. As t he r ei nf or cement is pul l ed out of t he box, lateral eart h pressure devel ops agai nst t he ri gi d front face and results i n an appar ent i ncrease i n t he geogri d pul l -out resistance. The rigid boundar y effect can be r educed by means of a sleeve i ncor por at ed ar ound t he slot on t he front wall. Sleeves t ransfer t he poi nt of appl i cat i on of t he pul l -out l oad i nsi de t he soil mass far beyond t he ri gi d front wall. The effect of sleeve l engt h is i nvest i gat ed by tests conduct ed with no sleeve at t he front face, and sleeve l engt hs of 20 cm (8 in) and 30.5 cm (12 in). Fi gure 9a shows t he effect of sleeve l engt h on t he peak pul l -out l oad. I n t hese tests, t he lateral pressures devel oped on t he faci ng were also meas ur ed usi ng t he two eart h pressure cells fixed on t he rigid front wall of t he box. The r el at i onshi p bet ween t he sleeve l engt h and t he lateral pressure devel oped on t he front wall is pr esent ed in Fig. 9b. Cell 01 is l ocat ed hi gher t han Cell 02 (see Fig. 2). Bot h cells di spl ay a decr easi ng lateral eart h pressure as t he sleeve l engt h is i ncreased. The results from bot h cells demonst r at e t hat t he i ncrease in t he sleeve l engt h l eads to a r educt i on i n t he eart h pressure devel oped at t he front rigid wail: consequent l y, a r educt i on i n t he pul l -out resistance. The effecl of t he rigid front wall is mi n i mu m when a sleeve l engt h of at least 30 cm is used. It is deci ded to use a sleeve l engt h of 30 cm in st andar d pul l -out tesling. Soil thickness The ri gi d boundar i es above and bel ow t he r ei nf or cement can affect t he i nt er act i on me c ha ni s m bet ween t he soil and t he geogrid. These boundar i es can l ead to i ncreases i n t he nor mal stresses in t he vicinity of t he geogri d surface, speci fi cal l y when t he soil t hi ckness is smal l and t he soil di l at ancy is rest rai ned. Moreover, friction can devel op bet ween t he soil and t he bot t om ri gi d boundar y. Thi s friction will also updat e t he mobi l i zed soi l -geogri d shear stress at t he interface. Tests are per f or med with di fferent t hi ckness of top and bot t om soil. Top and bot t om soil t hi cknesses of 10 cm, 30 cm and 40 cm are used. In one test, a top soil t hi ckness of 30 cm and bot t om soil t hi ckness of 40 cm (i.e. a total soil t hi ckness of 70 cm) is also used i n an at t empt to i nvest i gat e t he i nf l uence of di fferent soil t hi cknesses. Tests are conduct ed with t he Tens ar geogri d under a conf i ni ng stress of 48.2 kN/ m 2 ('7 psi) and an average soil densi t y of 16.4 kN/ m 3 (104 pcf). The effect of soil thickness, on t he pul l - out r esponse of t he geogri d is shown in Fig. 10. The f ol l owi , : observat i ons are made: 148 Khalid Farrag, Yalcin B. Acar, flan Juran I 0 0 " ~ 8O Z o 6 0 <~ o J I - : ~ 4 0 o i ,,-I J n 2 0 0 GEOGRI D: T ENSAR SR2 cr n = 4 8 k N/ m 2 ( 7 p s i ) X n 16.4 k N/ m 3 ( 1 0 4 p c f ) = 2 0 mm/ r a i n - ~ " ~ " " - ' ' - - - ~ N o -s le e v e f -- ~ m { 8 i n l S le e v e 3 0 c m { 1 2 i n ) 1 I I 50 I00 150 F RONT DI S P L A CE ME NT ( mm) (a) 2 0 0 120 , , , i " X G E O G R I D : T E N S A R $ R 2 100| ~ Confining P ressure: 4 8 . 2 kN Im 2 - - - . Ear t h pressure a t te s t c o m p le ~ n l u 'CELL01 / ~m 4 0 " l - " J C E L L 0 2 f ' 2 0 t , I , I , g 0 10 20 30 S LEEVE L E NGT H (cm) (b) Fig. 9. Th e ef f ect o f s l eeve l e ngt h o n t he pul l - out r es pons e. (a) Pul l - out l oad; (19) l at er al ear t h pr e s s ur e o n t he f r ont wal l . (a) The decrease i n soil t hi ckness results i n an appar ent i ncrease i n t he pul l -out st rengt h of t he geogrid. (b) The i ncrease i n t he soil t hi ckness above t he geogri d f r om 30 cm to 40 cm does not have any si gni fi cant effect on t he pul l -out resi st ance of t he geogrid. Pull-out resistance of geogrid reinforcements 149 z J 0 . J k- 0 .5 ._1 a. 80 / r . 2O c m 601- / " . . . - - ~- - ~ / T = 40era / I i ~'----../.... / / ' / / f " ~ ' .
I # ( / / , . 7 . . . . . . 1 1 t 7 20 [ GEOGRI D: TE_NSAR SR2 I ~ . 4.8 kNlm ~ 17p$i) . I Y.n " 16,4 kN/mi (104pcf) 0 8 = 6ram/rain 50 IO0 150 200 FRONT DI SPLACEMENT ( r am) Fig. 10. The effect of soil t hi ckness on pul l -out response. These results suggest that a soil thickness of at least 30 cm (1 fi) above and below the reinforcement (a total of 2 ft soil thickness) is necessary in uniform sands in order to eliminate the effect of the boundary and confi nement effects on the pull-out response. Displacement rate ASTM D4595-86 recommends a st andard rate of strain of 10 + 3% per mi n for wide strip method of test. However, this rate is established by experience obt ai ned in unconfi ned tests. In confined tests, reinforcement experiences its peak pull-out strength at much lower levels of strain. Therefore, there is a need to establish the displacement-rate for confined tests. Tests are conducted to investigate the effect of displacement-rate on pull-out response using the Tensar geogrid. Figure 11 shows the pull-out response of the geogrid under four different pull-out displacement-rates (2 mm/ mi n, 6 mm/ mi n, 10 mm/ mi n, and 20 mm/ mi n). An order of magni t ude increase in the displacement-rate from 2 mm/ mi n to 20 mm/ mi n results in a reduction of 25% in the peak pull-out resistance. It is noted that hi gher displacement rates mobilize lower strains along the reinforcement. Pull-out l oad will mobilize both the interface shear along the geosynthetic-soil interface and the passive resistance along the transverse ribs. Lower strains al ong the reinforcement i mpl y a hi gher cont ri but i on from the interface friction. Passive resistance al ong the transverse ribs will lead to hi gher pull-out resistance and mobilize hi gher strains al ong the reinforcement. The test results with this specific type of soil and the Tensar geogrid 150 Khal i d Farrag, Yalcin B. Acar. 11an Juran 60 i 40 < 3 0 0 TENSAR SR2 Conf i ni ng Pr essur e: 48.2 kN/ m2 Soi l D~Mty: 16.4 kN/ m3 m l * I * I , I i 5 10 15 20 25 DISPLACEMENT RATE (ram/rain) (a) 1. 0l 0. 8 0. 4 0.2 0. 0 0 . . . . . ~ ' ~ ma ~ s ~ ' " Displacements correspond I 20 mm(m!n ~ 10 ram/rain - - 6 ram/rain i , I i i 1 2 3 4 5 6 (b) Fig. II. The effect of displacement rate on the pull-out response. (a) Pull-out load: (b) displacements along the reinforcement. indicate that displacement rates of less than 6 mm/ mi n influence test results to a lesser degree. Displacement rates of less than 6 mm/ mi n are recommended in standard pull-out testing of geotextiles in uniform sands. Pul l -out resistance o f geogrid reinforcements 151 Soi l density The frictional resistance al ong the inclusion is hi ghl y influenced by soil density. Dense soils, under moderate confi ni ng pressures, tend to dilate when shear stresses are mobilized al ong the reinforcement interface. The soil surroundi ng the soil-reinforcement interface develops a condition of restrained di l at ancy as the reinforcement is pulled out. The magnitude of this restraint depends upon the type of test (displacement controlled or stress controlled), soil density, soil thickness and confi ni ng stress. When soil dilation is restrained, the confi ni ng stress at the interface increases and results in an apparent increase in the pull-out resistance. The effect of the relative soil density on the pull-out response are evaluated by tests conducted with soil compacted at different densities (15-7 kN/ m 3, 16-4 kN/ m 3 and 16.7 kN/ m 3) and under identical confi ni ng pressures of 48.2 kN/ m 2. Figure 12 shows the effect of soil density on Tensar geogrid. An increase in the soil density results in an increase in the peak pull-out resistance and significantly higher interface stiffness moduli are obtained. Soil compaction mobilizes the lateral earth pressure resistance on the geogrid and increases the frictional resistance mobilized at the interface. The restraint on geogrid movement and slippage increases the interface modulus and peak pull-out resistance. The increase in density concentrates the strains to the point of load application on the front face and consequently, the mode of failure can be due to breakage of the reinforcement rather t han slippage along the reinforcement. Confining pressure The effect of confi ni ng pressure on the frictional resistance of the reinforcement has been demonstrated by several investigators (McGown e t a l . , 1982; Juran & Chen, 1989). For geogrid reinforcements, confined extension duri ng pull-out testing is restrained by the passive resistance of the soil and particles interlocking within transversal geogrid elements. The confined extension results in an apparent increase in the tensile strength and stiffness modulus. Figure 13 shows the effect of the confi ni ng pressure on the pull-out response for Tensar geogrid measured under confi ni ng pressures of 34 kN/ m 2 (5 psi) and 48-2 kN/ m 2 (7 psi). Figure 14 presents test results on Conwed geogrid tested under confi ni ng pressures of 48.2 kN/ m 2 (7 psi), 96.4 kN/ m 2 (14 psi) and 140 kN/ m 2 (20 psi). An increase in the confi ni ng pressure results in significant increases in 152 Khalid Farrag, Yalcin B Acar, 11an Juran G E O G R I D : T E N S A R S R 2 Or. = 4 8 k N / m 3 ( 7 p s i ) = 4 r a m / r a i n Yn = 1 6 . 6 k N / m 3 ,,~Yn = 1 6 . 4 k N / m 3 . . . . . . . . . . . Yn = 1 5 . 4 k N / m 3 6oS I / , ~ . ur. u KNI m o _ j 4 0 o 2 0 O r I I I I 2 0 4 0 6 0 8 0 F R O N T D I S P L A C E M E N T ( r a m ) ( a ) I 0 0 C O N F I N I N G ~ = 48 kPa 70 IMspimmmmt Rate = 3.8 - 6 ram/rain I " 4O 1 0 15 15.5 16 16.5 17 1"/.5 DRY DENSI TY (kNl m3) (b) the pull-out resistance of these geogrids. Although the increase in the confi ni ng pressure reduces soil t endency to dilate, it leads to an increase in the passive soil resistance on the transverse ribs: consequently, the geogrid pull-out resistance increases. The effect of the confining pressure on the mobilized load transfer al ong the reinforcement is also illustrated. Pull-out resistance of geogrid reinforcements 153 0 a c ~ Z , , 6 U 0 t Z . J ~ O. (/) Q 0. 8 0 . 6 0 . 4 0 . 2 GEOGRID T E NS A R SR2 0" n = 4 8 k N / m 21, 7psi ) fYn = 16.4 k N/ m 3 / / ~ Y n ~ 16. -6 k N/ m3 I I I T I 2 3 4 5 G E O G R I D N O D E S ( c ) Fig. 12. The effect of density on the pull-out response. (a) Development of pull-out load: (b) the change of peak pull-out resistance with soil density: (c) displacements along the reinforcement when the peak pull-out load is reached. The soil-geogrid interface shear stress is more uni forml y mobilized al ong the geogrid under low confi ni ng pressures. The increase in the confi ni ng pressure restrains the geogfid displacement and results in a hi gher mobi l i zat i on of the soil-geogrid interface shear stresses near the pull-out appl i cat i on point, a lower mobi l i zat i on of t he shear stresses at the rear end, and consequently a need for shorter effective adherence length of reinforcement. These results suggest that the design criteria for geogrid reinforced soil structures shoul d take into account the in-soil confi ned extension properties derived from pull-out tests rather t han the material properties obt ai ned from unconfi ned extension tests. The unconfi ned behavi or of the geogrid is substantially more ductile and its strain level at peak load is hi gher t han t hat obt ai ned under confi ned conditions. The use of unconfi ned properties may result in over-conservative design values for admissible geogrid extension and corresponding structure displacements. Interacti on mechani s m The magni t ude of the mobilized shear resistance al ong the soil- reinforcement interface depends upon the type of reinforcement. In case of geogrids, the interface shear strength is pri mari l y mobilized by the skin friction and the passive resistance against transversal fibs. For 154 Khalid Faring Yaicin R Acar, I lan J uran 80 E ~- 60 z 0 1 40 F- o I .J .J 20 o. O 0.8 0. 6 N 0. 2 GEOGRID TENSAR SR? 3 ?'n = 16.4 kN/ m 3 (104 p c f ) : 6 r a m / r a i n / " ~ " . ~ . . . . O - n : 48k N/ m 2 / I I I 50 I 00 150 FRONT DI S P L A CE ME NT ( r a m) ( a ) 20~ p o " n = 5 4 k N / m 2 G ~ p R e oSkRp2u 1,- o u t l o a d - ~ I I I I 0 I 2 3 4 5 GEOGRI D NODES ( b ) Fig. 13. The effect of confining pressure on pull-out response ofTensar SR2. (a) Pull-out load: (b) displacements along the reinforcement. coars e gr ai ne d soi l s, t he o pe ni ng s i n t he geogri ds ma y a l l o w s oi l part i cl es to i nt erl ock be t we e n t he ribs; t hus , i nc r e as i ng its s he ar strength. Th e i nt erf ace f ri ct i on be t we e n t he soi l a nd t he geogri d de pe nds o n t he t ype o f soi l and t he s urf ace r oughne s s o f t he g e o g d d whi l e t he c ont r i but i on o f pas s i ve soi l res i s t ance a l o ng t he t ransverse ri bs to t he overal l pul l - out st rengt h d e p e n d s o n ma n y factors s uc h as c o n f i n i n g pressure, geogri d Pull-out resistance of geogrid reinforcements I 00 155 E z E:] 0 _J ,0- n = 1 4 0 k N/ m 2 8 0 1 ~ . . % 'n = 9 6 k N / m 2 L / " ~ . / ~ ' n = 4 8 k N / m 2 0 I 0 2 0 3 0 F R O N T D I S P L A C E M E N T ( r am) ( a ) 4 0 / o n = 4 8 k N / m 2 oa:~ 0.8 0.4 GEOGRID: CONWED 0.2 Ya = 16" 7kN/ m3 ( 106pcf ) "- 2 ram/rain 0 I I I I I 2 3 4 5 G E O G R I D N O D E S ( b ) Fig. 14. T h e effect o f conf i ni ng pressure on pul l - out response o f Conwed 9027. ( a) Pul l - out l o a d : ( b ) d i s p l a c e m e n t s a l o n g t h e r e i n f o r c e m e n t . geomet ry a nd t he di amet er rati o ( t he rati o o f t he me a n grai n si ze o f the soi l to t he si ze o f t he geogri d openi ng) . Fi gure 15 presents the resul ts o f pul l - out tests on Conwe d geogri ds perf ormed wi t h a nd wi t hout t hei r transversal ribs. The resul ts provi de a s ens e o f t he cont ri but i on o f t he passi ve resi st ance on t he transverse ribs to t he overal l pul l - out l oad. The f ri ct i onal resi st ance al ong the 156 Khalid Farrag Yalcin B. Acar, nan Juran E z Q o ._1 ..~ 0 .5 _ J o . GEOGRI D: CONWED Yn = 16. 7 k N/ m 5 ( 1061: ) cf ) = 9 6 k N/ m 2 o - - - oe o ~ O ~ o ~ o oo . ~ a ; e g o " CT n = 4 8 k N/ m 2 o No T r a n s v e r s a l Ri bs o Wi t h T r a n s v e r s a l Ri bs i I I I I I I 0 I0 20 30 4O FRONT DI SPLACEMENT ( r nm) Fig. 15. The effect of transverse ribs on the pull-out response of Conwed geogrid. l ongi t udi nal ribs constitutes an average of 75% of the total pull-out load for this specific geogrid. It is also noted that the passive resistance al ong the transverse ribs is mobilized at higher levels of strain resulting in a higher post-peak pull-out resistance. The strain level required to mobilize the passive resistance al ong the transverse elements would vary with the geogrid and the soil type. SUMMARY AND CONCLUSIONS There exists significant differences in pull-out or direct shear tests used in experimental modelling of soil-geosynthetic interaction mechani sm and/ or performance evaluation of in-soil properties of geosynthetics. Standard testing equi pment and testing procedures are needed to obtain comparabl e results and to develop appropriate methodologies in modelling the load-transfer mechanism. Pull-out boxes are designed and constructed at GERL/LTRC-LSU. The component s of these equi pment and the procedure used in testing are presented. Tests conducted in performance assessment of this equi pment employed a uniform sand, and TENSAR SR2, CONWED 9027 geogrids. The following conclusions are obt ai ned for the uni form sand and the geogrids used in this study: Pull-out resistance of geogrid reinforcements 157 (1) Side frictions on the side walls of the box decrease the peak pull- out load. A mi ni mum clearance of 15 cm (6 in) is necessary in order to avoid any side-wall effect in case the friction along the wall is not mi ni mi zed by some other means. (2) The increase in sleeve length decreases the effect of front wall on the pull-out resistance. The tests conducted in this study demonstrate t hat sleeve lengths of at least 30-5 cm (12 in) of length are necessary. (3) Increased thickness of soil cushi oni ng the geosynthetic decreases the effects of the top and bottom boundaries. A soil thickness of at least 30 cm above and 30 cm below the geogrid is needed to el i mi nat e the influence of these boundaries. (4) Hi gh di spl acement rates affect both the pull-out load and the displacement distribution al ong the reinforcement. The displace- ment rate effects are mi ni mi zed if rates of the order of less t han 6 mm/ mi n are used. (5) Increased densities lead to mobilization of the interlocking mechani sm and consequently, result in a hi gher contribution of the passive resistance exerted on the transverse ribs. At hi gher densities, pull-out resistance increases and displacements move closer to the point of appl i cat i on of the pull-out load. (6) Similar to the effects of increased densities, confi nement increases the peak pull-out resistance and moves the displacements al ong the reinforcement toward the point of application of the load. (7) Pull-out resistance is developed by a combi nat i on of the passive resistance developed al ong the transverse ribs and interface shear mobilized al ong the geogrid. (8) Met hods developed and employed in evaluation of rigid reinforcements shoul d not be used in det ermi ni ng the confined extension properties and shear stress-strain behavi or ofgeogrids. The results presented in this study demonstrate the significant variability obt ai ned in pull-out load-displacement response of geogrids due to differences in equi pment and testing procedures. It is necessary to realize the influence of these parameters on pull-out response in st andard commercial testing ofgeosynthetics and in interpretation of the load-transfer mechani sm using pull-out test results. ACKNOWLEDGEMENTS Design, construction, and performance assessment of the pull-out testing facility at GERL are supported by funds awarded and provided to 158 Khalid Farrag Yaicin B. Acar, llan Juran LSU and by t he Loui si ana Tr anspor t at i on Resear ch Cent er of Loui si ana Depar t ment of Tr anspor t at i on and Devel opment , Federal Hi ghway Admi ni st r at i on, t he Boar d of Regent s of t he State of Loui si ana, Conwed Plastics Inc. and Civil Engi neer i ng Depar t ment of Loui si ana State University. The funds awar ded and pr ovi ded by t hese agenci es and i nst i t ut i ons are gratefully acknowl edged. Dr I l an Jur an of Pol yt echni c Uni versi t y has had si gni fi cant cont r i but i ons in est abl i shment of GERL in cooper at i on and col l abor at i on with t he second aut hor. Dr Juran' s efforts and cont r i but i ons are grat eful l y acknowl edged by GERL. LTRC proj ect officer, Mr Paul Gri ffi n and Conwed Plastics Product Devel opment Manager , Mr Lee Ri c hmond are acknowl edged for t hei r col l abor at i on and cooperat i on. We acknowl edge and appreci at e t he efforts of Ex-di rect or of LTRC, Mr Ara Ar man, t he current director, Dr Pet er St opher, ex- chai r man of t he Civil Engi neer i ng Depar t ment , Dr Roger IC Seals and t he current c ha i r ma n Dr Ri char d Avent in est abl i shment and e nha nc e me nt of GERL. Any opi ni ons, fi ndi ngs, and concl usi ons or r ecommendat i ons expressed in this mat eri al are t hose of t he aut hor s and do not necessari l y reflect t he views of t he sponsors. REFERENCES Acar, Y. B., Durgunoglu, T. & Tumay, M. T. (1982). Interface properties of sand. ASCE, JGE, 108(GT3), 648-54. Anderson, L. R. & Nielsen, M. R. (1984). Pull-out resistance of wire mats embedded in soil. Report for the Hilfiker Company, Eureka, CA. Bonaparte, R., Holtz, R. D. & Giroud, J. P. (1987). Soil reinforcement design using geotextiles and geogrids. Geotextile Testing a n d the Design Engineer, ASTM STP 952, pp. 69-116. Brand, S. R. & Duffy, D. M. (1987). Strength and pullout testing of geogrids. Geosynthetics Conference, New Orleans, Vol. 1, pp. 226-36. Christopher, B. R. (1976). Tensar SS2 geogrid evaluation. Evaluation report to Tensar Corporation. Elias, V. (1979). Friction in reinforced earth utilizing fine grained backfills. International Conference on Soi l Reinforcement, Paris, France, pp. 435-8. Farrag, K. (1990). Interaction properties of geogrids in reinforced soil walls - - testing and analysis. Dissertation submitted to Graduate School of Louisiana State University, Baton Rouge, LA. lngold, T. S. & Templeman, J. (1979). The comparative performance of polymer net reinforcement. International Conference on Soi l Reinforcement, Paris, Vol. 1. Ingold, T. S. (1983). Laboratory pull-out testing of grid reinforcement in sand. Geotech. Testing J., 6(3 ), I 01 - 11. Jewell, R. A. (1979). Discussion presented to Section 8, Proceedings o f VII European Conference on Soi l Mechanics and Foundat i on Engineering, Brighton, UK. Pull-out resistance of geogrid reinforcements 159 Jewell, R. A. (1980). Some effects of rei nforcement on the mechani cal behavi or of soils. Dissertation submi t t ed to Cambri dge University, Cambridge, UK. Juran, I., Knochenmus, G., Acar, Y. B. & Arman, A. (1988). Pull-out response of geotextiles and geogrids (Synthesis of Available Experimental Data). Proceedings o f Sympos i um on Geotextiles f o r Soi l Improvement, ASCE, Geotech. Special Publ i cat i on 18, pp. 92-111. Juran, I., Guermazi, A., Chen, C. L. & Ider, M. H. (1988). Model i ng and si mul at i on of load transfer in reinforced soil: Part 1. Int. J. Numer. Anal. Meth. Geomech, 12, pp. 141-55. Juran, I., lder, H. & Farrag, Kh. (1990). Strain compatibility analysis for geosynthetic reinforced soil walls. ASCE, JGE, 116(2), 312-29. Koerner, R. M. (1986). Direct shear/ pul l -out tests on geogrids, Report No. 1, Dept. of Civil Eng., Drexel Univ., Phi l adel phi a, PA. Koerner, R. M. (1990). Designing with Geosynthetics. 2nd edn, Prentice Hall, Engl ewood Cliffs, NJ. McGown, A., Andrawes, K. Z. & Kabir, M. H. (1982). Load-extension testing of geotextiles confi ned in-soil, 2nd Int. Conf. on Geotextiles. Las Vegas, Vol. 3, pp. 793-8. Myles, B. (1982). Assessment of soil fabric friction by means of shear, 2nd Int. Conf. on Geotextiles, Las Vegas, Vol. 3, pp. 787-91. Palmeira, E. & Milligan, G. (1989). Large scale direct shear tests on reinforced soil. Soil and f oundat i ons , Japan. Soc. Soi l Mech. Found Eng., 29(1), 18-30. Rowe, R. K. & Ho, S. K. (1986). Det ermi nat i on of geotextile stress-strain characteristics using a wide strip test, 3rd Int. Conf. on Geotextiles, Vienna, pp. 885-90. Rowe, R. K., Ho, S. K. & Fisher, D. (1985). Det ermi nat i on of soil-geotextile interface strength properties, 2nd Canadi an Sympos i um on Geotextiles, Montreal, Canada, pp. 25-34. Saxena, S. K. & Budi man, J. S. (1985). Interface response ofgeotextiles, Proc. 1 l t h Int. Conference on Soi l Mechanics and Foundat i on Engineering, San Francisco, Vol. 3, pp. 1801--4. Schlosser, F. & Guilloux, A. (1979). Friction between soil and strips in reinforced earth structures, Int. Conf. on Soi l Reinforcement, Paris, France, Vol. 1. Tzong, W. H. & Cheng-Kuang, S. (1987). Soil-geotextile interaction mechani sm in pull-out test, Geosynthetics Conference. New Orleans, Vol. 1, pp. 250-9. Williams, N. D. & Houl i han, M. F. (1987). Evaluation of interface friction properties between geosynthetics and soils, Geosynthetics Conference. New Orleans, Vol. 2, pp. 616-27.