Synthesis of Geosynthetic Reinforced Soil (GRS) Design Topics

CHAPTER 7. REDUCTION FACTORS

Current design methods for GRS walls and abutments (e.g., AASHTO guideline, FHWA NHI manual, and NCMA manual) typically stipulate that design strength of geosynthetic reinforcement should be determined by applying reduction factors for installation damage, creep, and durability to the ultimate strength of geosynthetic reinforcement.^(1,2,3) The equation in figure 56 expresses the allowable reinforcement design strength, T_a:

Figure 56. Equation. Allowable reinforcement design strength, T subscript a.

T_ult is the ultimate wide-width strip tensile strength of the geosynthetic (in accordance with ASTM D4595) based on the minimum average roll value (MARV) for the product. (MARV is commonly defined as the strength that is two standard deviations below the mean tensile strength). FS is an overall safety factor to account for uncertainty. RF is a combined factor to account for geosynthetic strength loss during the wall design life and is equal to RF_ID times RF_CR times RF_D, where RF_ID is a reduction factor for installation damage, RF_CR is a reduction factor for creep, and RF_D is a reduction factor for chemical and biological degradation.

In the FHWA NHI manual, RF_ID = 1.1 to 3.0, RF_CR = 1.6 to 5.0, and RF_D = 1.1 to 2.0 are recommended, resulting in K = 3 to 50 percent.⁽²⁾ Different state transportation departments have adopted some variations for the reduction factors and safety factor, hence the K-value range. CDOT, for example, has adopted the following values: the allowable design strength is on the order of 10 to 24 percent of MARV for preapproved products and 5 percent of MARV for not preapproved products. This type of practice has practically excluded the use of all geotextiles in reinforced soil wall applications, even though many GRS walls and abutments constructed with geotextiles as reinforcement have performed successfully.

This synopsis addresses the issue of geosynthetic reinforcement reduction factors. The available research on the use of cumulative reduction factors is reviewed, and current practices are discussed. In addition, brief comments about the K-stiffness method, a working stress design method intended to address long-term behavior of geosynthetic reinforcement of reinforced soil walls, are provided.

RESEARCH AND CASE HISTORIES

Degradation

Aging of buried geosynthetics by processes such as hydrolysis, oxidation, and abrasion may result in long-term strength loss of geosynthetic reinforcement. Elias evaluated 24 geosynthetics from 12 retrieval sites, and confirmed that little, if any, chemical degradation had occurred in the geosynthetic reinforcement in those full-scale structures, some of which were up to 25 years old.^(95,96) This has been further confirmed by Allen and Bathurst.⁽⁹⁷⁾ They indicated that the greatest contributors to strength loss and reduced wall performance (i.e., increased deformation) for the geosynthetic reinforcement products in use today are installation damage and possibly creep.

Installation Damage

Numerous field studies regarding the installation survivability of geosynthetics have been performed. They have shown that the level of installation damage depends on polymer type, manufacturing method, and geosynthetic coating.^(96,98) The level of damage has also been shown to be affected by the weight, type, and number of passes of the construction and compaction equipment, the graduation, angularity, and condition of the fill material and the lift thickness. (See references 99 through 102.)

Allen and Bathurst show that installation damage has limited impact on the initial working stress performance of geosynthetic walls.⁽¹⁰³⁾ For most geosynthetics used as reinforcement (i.e., woven geotextiles and geogrids), their load-strain-time behavior is not significantly affected by installation damage at typical or even relatively high working strains for the levels of installation damage observed in full-scale walls. The data provided by Allen and Bathurst indicate that installation damage does not severely affect modulus, if at all, until damage levels become quite high for woven geotextiles and geogrids.⁽¹⁰³⁾

Bathurst et al. describe how to compute bias statistics from project-specific installation damage trials for use in reliability-based design for the reinforcement rupture limit state, or by using data from multiple sources for LRFD calibration.⁽¹⁰⁴⁾ A database of results from field installation damage trials on 103 different geosynthetic products was collected from 20 different sources. The computed reduction factors for installation damage (RF_ID) of their study confirm earlier recommendations by Elias that woven and nonwoven geotextiles with mass per unit area less than 7.96 oz/yd² (270 g/m²) should not be used in combination with Type 1 soil (with D₅₀ > 19 mm (.74 inches).⁽⁹⁵⁾ For a certain soil type, there were detectable differences in the calculated RF_ID values depending on the geosynthetic type.

Hufenus et al. conducted a series of field installation tests and concluded that, in applications where only the tensile strength at relatively low elongations is relevant, the effects of moderate installation damage is very limited, and the factor RF_ID can be designated as very close to unity (1.0 to 1.1) based on product-specific test data.⁽¹⁰⁵⁾

Allen and Bathurst investigated the combined effect of polymeric creep and installation damage using a database of constant sustained load (creep) data for both undamaged and installation damaged geosynthetic specimens.⁽¹⁰⁶⁾ They concluded that multiplication of creep reduction (RF_CR)and installation damage factors (RF_ID) may be conservative and hence results in errors on the safe side for current ASD practice. Greenwood came to the same conclusion based on stepped isothermal creep-rupture tests performed on a polyester geosynthetic material in undamaged and damaged states.⁽¹⁰⁷⁾

Creep

Major types of geosynthetics used in GMSE and GRS structures are polypropylene, polyethylene, and polyester. These geosynthetics, manufactured with various types of polymers, are creep-sensitive. Stress level, polymer type, manufacturing method, and temperature have been known to affect the creep potential of a geosynthetic material. In general, polypropylene and polyethylene exhibit larger creep deformation than polyester and polyvinyl alcohol under otherwise identical conditions.

As noted above, the allowable reinforcement tensile load employed in current design guidelines is determined by applying a safety factor and a combination of reduction factors to a limiting strength determined from short- and/or long-term laboratory tests. There tests are conducted by applying uniaxial tensile forces directly to the geosynthetic reinforcement (in a confined or unconfined condition) without regard to the soil-geosynthetic interaction behavior. It is important to point out that conducting uniaxial creep tests in the confinement of soil does not mean soil-geosynthetic interaction is accounted for. Wu indicates that long-term creep behavior of geosynthetic reinforcement in a reinforced soil structure must be determined by allowing soil-geosynthetic interaction to take place in a manner mimicking the field conditions.⁽¹⁰⁸⁾

McGowan et al. developed a fairly sophisticated uniaxial tension test device to measure creep behavior of geotextiles under soil confinement.⁽¹⁰⁹⁾ Wu and Ling et al. developed a simplified confined creep test method, in which a constant sustained tensile force is applied to a membrane-confined geosynthetic specimen (without inducing artificial soil-geosynthetic interface friction).^(110,111) Various geotextiles under different confining pressures have been tested. The results indicated that pressure confinement gave various degrees of improvement in creep behavior for different geotextiles. The greatest improvement was for needle-punched nonwoven geotextiles, while the improvement in woven geotextiles and geogrids was negligible.

Boyle manufactured a plane strain test device similar to that developed at Massachusetts Institute of Technology by Abramento and Whittle.^(112,113) In the test, a geosynthetic specimen was embedded in soil and subjected to a plane strain loading condition. The loads at both ends of the geosynthetic material were measured. Creep tests using sand as confining soil indicated that the geosynthetic material experienced stress relaxation. After creep deformation had diminished, the force in the geosynthetic material would reduce with time. The practical implication is that such GRS structures will have increasing safety margins as time progresses.

Crouse and Wu synthesized measured field behavior of seven reinforced soil walls that had been monitored for extended periods of time for assessment of their long-term performance characteristics.⁽¹¹⁴⁾ The GRS walls represented a variety of wall types using granular backfill. The walls were: (1) the Glenwood Canyon wall; (2) the Tanque Verde-Wrightown-Pantano Roads project wall; (3) the Norwegian Geotechnical Institute (NGI) project, NGI wall; (4) the Japan Railway Test Embankment project, JR wall; (5) the Highbury Avenue, London Ontario project, Highbury wall; (6) the FHWA Algonquin wall; and (7) the Seattle Preload Fill project, Seattle wall. (See References 115 through 124, 7, 125, 54, 126, and 127.) The maximum creep strains in the reinforcements measured by strain gauges for all the walls were less than 1.5 percent, and the creep strain rate in all cases decreased with time in that there was a linear relationship between log-[creep rate] and log-[time]. Crouse and Wu also proposed a creep equation for predicting long-term creep deformation of GRS walls.⁽¹¹⁴⁾

Allen and Bathurst analyzed the long-term creep data from 10 full-scale geosynthetic wall case histories.⁽⁹⁷⁾ Post-construction, long-term wall face deformation data show that geosynthetic wall face deformations, if the wall is properly designed, will generally be less than 0.98 to 1.17 inches (25 to 30 mm) during the first year of service and less than 1.37 inches (35 mm) during the design lifetime for walls lower than 42.64 ft (13 m). Allen and Bathurst et al. studied reinforcement strains measured in geosynthetic-reinforced walls and slopes.^(128,129) The maximum strains were on the order of 1 to 2 percent or less.

Helwany and Wu performed finite element analysis on two 9.84 (3-m) high geosynthetic-reinforced retaining walls.⁽¹³⁰⁾ The walls were identical in every respect except that one was with a clayey backfill and the other with a granular backfill. In the clay-backfill wall, the maximum strain in the geosynthetic reinforcement increased by 3.5 percent from the end of construction to 15 years later. In the granular-backfill wall, the increase in maximum strain over the same time period was negligible. Note that the difference occurs in spite of comparable levels of load for the two walls. The analysis strongly suggests that the backfill played a very important role in creep deformation of a soil-geosynthetic composite. The finding was supported by studies conducted by Li and Rowe, Skinner and Rowe, Rowe and Taechakumthorn, Bergado and Teerawattanasuk, Liu and Won, Liu et al., and Li et al. (See references 131 through 139.)

Recognizing that the interaction between soil and geosynthetic reinforcement must be used as the basis for a rational design, Wu and Helwany developed a soil-geosynthetic long-term performance test, also referred to as the Soil-Geosynthetic Interactive Performance (SGIP) test.⁽¹⁴⁰⁾ The test has two important features. First, the stresses applied to the soil are transferred to the geosynthetic material in a manner similar to the typical load transfer mechanism in GRS structures (i.e., loads in reinforcement are transferred from soil through soil-geosynthetic interface bonding to geosynthetic material, rather than being applied directly to geosynthetic material). Second, both the soil and geosynthetic material are allowed to deform in an interactive manner under plane strain conditions. Two carefully conducted long-term performance tests, one using a clayey backfill and the other a granular backfill, have been reported. An element test on the geosynthetic material alone underestimated the maximum strain by 250 percent in the clay-backfill test, and overestimated the maximum strain by 400 percent in the sand-backfill test. It is noteworthy that creep deformation essentially ceased within 100 min after the sand-backfill test began; whereas the clay-backfill test experienced creep deformation over the entire test period (18 days), at which time shear failure occurred in the soil.

The SGIP test subsequently evolved to accommodate taller specimens, as shown in figure 57.⁽¹⁴¹⁾ A number of geosynthetic/granular road base composites have been tested, with a few under elevated temperatures to accelerate creep of the geosynthetic materials. For a nonwoven geotextile embedded in a road base material (with 20 percent of fines, prepared at 95-percent R.C. and 2-percent wet-of-optimum moisture) subject to an average surcharge of 15 psi (103 kPa) and in a 125 °F (52 °C) environment (note that the creep rate of the geotextile was about 150 times faster at 125 °F (52 °C) than at the ambient temperature, as measured in a series of long-term creep tests), creep deformation of the soil-geotextile composite was very small and decreased rapidly with time. Creep deformation ceased completely in 12 days and gave an accumulative average strain of 0.58 percent.

©K. Ketchart and J.T.H. Wu
1 mm = 0.039 inches
Figure 57. Illustration. Schematic diagram of a modified soil-geosynthetic interactive performance (SGIP) test.⁽¹⁴¹⁾

DISCUSSION

This discussion addresses five points regarding long-term design consideration of GRS walls and abutments: (1) long-term degradation, (2) construction damage, (3) creep, (4) current practice, and (5) the K-stiffness method.

Long-Term Degradation

Studies conducted by Elias and Allen and Bathurst have indicated that long-term degradation of geosynthetics in reinforcement applications during exposure to the in-soil environment appear to be very small.^(95,97)

A full-scale experiment conducted by the PWRI of Japan is enlightening in terms of long-term degradation. The experiment was to examine the failure mechanism of a 19.7-ft- (6.0-m-) high GRS wall. The wall face was segmental concrete blocks, and the backfill was a sandy soil reinforced with six layers of 11.5-ft- (3.5-m-) long polymer grid. The reinforcement inside the wall was severed at selected sections after the wall was constructed (for details, see chapter 3). The maximum horizontal movement due to cutting of the reinforcements was nearly zero until the cut was approximately 0.2H from the wall face, at which time the movement increased from 30 mm to 40 mm. The result of the PWRI experiment reveals the fundamental concept of GRS-the reinforcement serves not as tiebacks but as improvements to stiffness and strength properties. It also suggests that long-term degradation of the reinforcement is not a design issue. The cutting of reinforcements can be viewed as an extremely severe state of degradation in that the reinforcement has been degraded into small pieces and is not continuous.

Installation Damage

Most study on installation damage indicates that such damage does not severely affect load-strain behavior for woven geotextiles and geogrids, if at all, until damage levels become quite high. In applications where only the tensile strength at relatively low elongations is relevant, the effects of moderate installation damage is very limited. It has been suggested that RF_ID should be designated as being close to unity, on the order of 1.0 to 1.1.⁽¹⁰⁵⁾

Allen and Bathurst provided strong evidence that installation damage would have little, if any, effect on creep strains and rates for typical levels of installation damage in full-scale structures.⁽¹⁰⁶⁾ They stated that "in many cases, installation damage will have a negligible effect on the long-term strength at working stress levels (i.e., the geosynthetic behaves as if it is not damaged)."

Creep

Most geosynthetic materials are susceptible to creep under sustained loads. Stress level, polymer type, manufacturing method, and temperature have been known to affect the creep potential of a geosynthetic material. However, it can be drastically misleading to evaluate the creep potential of geosynthetic reinforcement based on tests performed by applying a sustained tensile force to geosynthetic specimens, as has been done by current design guides. If the confining soil has a tendency to deform faster than the geosynthetic reinforcement along its direction of elongation, the geosynthetic material will impose a restraining effect on the time-dependent deformation of the soil through the interface bonding forces. Conversely, if the geosynthetic reinforcement in isolation tends to deform faster than the confining soil, then the confining soil will restrain creep deformation of the reinforcement. This restraining effect is a direct result of soil-reinforcement interaction wherein redistribution of stresses in the confining soil and changes in tensile forces in the reinforcement occur over time in an interactive manner. Field-measured data reveal that geosynthetic creep deformation is not a design issue when well-compacted granular fill is used as backfill in the reinforced soil zone.

Current Practice

In an attempt to accommodate the effects of installation damage, creep, and durability, current practice applies a combined reduction factor to short-term reinforcement strength in design. The combined factor is obtained by multiplying individual reduction factors for each effect, which is unwarranted for closely spaced GRS. First, soil-geosynthetic interaction is critical to the susceptibility of geosynthetic creep; the current design approach is based on tests in which forces are applied directly to geosynthetic materials without regard to the soil. Second, creep is a deformation problem; the current design approach uses a somewhat arbitrary reduction factor for creep and treats it as a strength problem. Third, installation damage has been shown to have little effect on long-term creep, i.e., there is no compounding effect of installation damage and creep; also RF_D for geosynthetic materials has been found to be near unity.^{(106, 105)}

When well-compacted granular fill is employed, a single safety factor can be used to account for long-term effects, uncertainty, and ductility. It is an indisputable fact that creep of geosynthetic reinforcement in a GRS is strongly affected by the time-dependent behavior of the fill. When time-dependent properties of a given fill material are in question, a laboratory test similar to the SGIP test (see figure 57) may be conducted to evaluate potential creep "deformation" of a GRS wall or abutment.