The results of the PT research performed on GRS composites leads to several important conclusions. General conclusions include the following:
The performance test can be used to model the load-deformation behavior of a particular GRS composite and is repeatable. Based on equivalency comparisons, the ratio of the measured vertical capacity in a PT to that of the same GRS in plane strain is near unity, and the stiffness of what a plane strain GRS composite might experience (such as an abutment) is up to 3 times higher (on the conservative side) compared to what is measured in a GRS PT (section 5).
A two-post frame with single hydraulic jacks produces more even load distribution than a two bolted channel beams in combination with four hydraulic jacks.
The load-deformation relationship of GRS composites depends on several parameters including preloading, aggregate angularity, compaction level, presence of bearing bed reinforcement, and facing confinement.
Figure 32 indicates that preloading of the GRS composite (TF-6 with Sv = 7⅝ inches, Tf = 4,800 lb/ft, 21A material, CMU facing) resulted in strain hardening to produce a stiffer response during two unload-reload cycles. The reload response is similar during both cycles, with a ratio of stress to strain equal to approximately 3,000 ksf, compared to the ratio for the initial, virgin loading equal to approximately 750 ksf, an increase by a factor of four. Similar rsults were found for an open-graded material as well (TF-1 with Sv = 7⅝ inches, Tf = 2,400 lb/ft, AASHTO No. 8 material, CMU facing), as shown in figure 30. The effect of preloading was not explored for other reinforcement spacing in this study.
Angularity impacts the strength (e.g., friction angle) of the backfill used and therefore the GRS composite. The GRS composite using a rounded pea gravel aggregate (DC-2) had both a lower ultimate strength (qult,emp) and service limit capacity (q@ε=0.5% ) than that using the more angular aggregate (DC-1) meeting the same gradation specifications for an AASHTO No. 8 material (table 10).
The results indicate a similar ultimate vertical capacity (qult,emp)between an uncompacted (DC-5) and compacted (DC-1) GRS composite, but a much softer response with no compactive effort. For the particular GRS composite tested (table 11 and figure 58), at an in-service dead load of 4,000 psf, an uncompacted abutment will experience an initial strain about three times more than a compacted abutment; however, near failure, at about 20,000 psf, an uncompacted abutment will experience about 1.7 times the strain as a compacted abutment.
The results indicate that the bearing bed provides some added vertical capacity; however, vertical deformation is not improved at low strain levels (figure 60). The modulus for primary compression is similar whether the bearing bed is present or not (table 12). At both low and high applied normal stresses, representing bridge loads (figure 61) and near failure loads (figure 62), respectively, the bearing bed reinforcement serves to limit lateral deformation in the zone of its placement.
The frictionally connected CMU facing has an impact on the performance of GRS; it provides confinement, leading to a stiffer response and an increased capacity compared to a GRS composite with no facing element. From table 20, the facing more than doubles the initial stress-strain ratio as compared to the PTs without any facing. In terms of capacity, the facing plays the biggest and smallest role for the largest spaced (Sv = 15¼ inches) and the closest spaced (Sv = 3 13/16 inches) system tested, respectively. For the 5 pairs of tests conducted at TF, including the CMU facing produced an improved ultimate capacity between 1.2 and 2.2 times greater than the GRS composite without any facing (table 21). A similar trend to the strain at the current 4,000-psf service limit is also found. The design assumption to not include the effect of confinement from the face in determining the capacity and required reinforcement strength is therefore conservative.(1)
In addition, at the current service stress limit (applied stress, q, of 4,000 psf), the ratio of service vertical strain with no facing to service vertical strain with a CMU facing ranges from 1.2 to 2.2 (table 21). The largest impact was for the 15⅝-inch spaced GRS composite (TF-9 and TF-10). At ultimate failure, the ratio is considerably closer, ranging from 0.8 to 1.2.
For the same Tf/Sv ratio of 3,800 lb/ft2, there is a linear relationship between the reinforcement spacing (Sv) and the ratio of capacity with CMU facing (qult,emp CMU) to the capacity without facing (qult,emp no CMU) (figure 75). Similarly, there is a linear relationship with reinforcement strength (figure 76).
Both open-graded and well-graded aggregates can be used as the reinforced backfill in GRS composites; each has their advantages and disadvantages. The PT results further distinguished the effect aggregate selection can have on the behavior of GRS.
The well-graded material is considerably stiffer than the open-graded material. At an applied pressure of 4,000 psf, the vertical strain is about 1.1 percent for TF-1 (open-graded) and 0.4 percent for TF-2 (well-graded). The modulus of the composite tested in TF-1 is 320 ksf
(table 13) compared to 710 ksf for the composite tested in TF-2; TF-1 (open-graded) is 55 percent less stiff than TF-2 (well-graded). In terms of bearing capacity, TF-1 was 20 percent less strong than TF-2. The results indicate that the gradation, and perhaps cohesion (at the same friction angle), impacts the modulus more so than the strength, giving indication that well-graded fills have an advantage with respect to serviceability.
As discussed in section 6.1, isolating the effect of cohesion and other soil parameters on the performance of GRS is difficult using PTs, although based on the soil-geosynthetic capacity equation (figure 8), it will serve to improve capacity, although its contribution should not be considered in design.
The response indicates that the higher reinforcement strength (4,800 lb/ft) produces a stiffer and stronger response than the lower reinforcement strength (2,400 lb/ft) for open-graded backfill (table 15); the same is true for the composites with no facing element. By doubling the reinforcement strength, the results indicate an increase of measured capacity by a factor of 1.14 for capacity and 1.34 for the initial stress-strain ratio.
For the well-graded aggregate, increasing the reinforcement strength increases the capacity by a factor of 1.5 and 1.7 for no facing and CMU facing (table 16), respectively, but does not significantly impact the stiffness.
The relationship between reinforcement strength and spacing was investigated through a series of PTs with the same Tf/Sv ratio.
The response for experiments conducted with a Tf/Sv ratio of 3,800 lb/ft2 showed that as reinforcement spacing increased, the vertical capacity decreased for the same Tf/Sv ratio, whether a CMU facing was absent (figure 66) or present (figure 67). This suggests that the relationship between reinforcement strength and spacing is not proportional to capacity as outlined in current MSE design (AASHTO 2012); a GRS abutment with a given Tf and Sv will not have the same strength as a GRS abutment with twice the strength (2Tf) and reinforcement spacing (2Sv).
At the same Tf/Sv ratio of 3,800 lb/ft2, increasing the reinforcement spacing by a factor of two from 7⅝ to 15¼ inches and the reinforcement strength by a factor of two from 2,400 to 4,800 lb/ft, resulted in a reduction of the capacity by a factor of 0.9 and 0.6 for CMU facing and no facing, respectively. The relationship is therefore not directly proportional, as indicated for MSE design (AASHTO 2012).
The primary purpose of PTs is to provide a designer with the unique stress-strain properties of a particular GRS composite for use in design. There is an empirical method and an analytical method currently available for GRS by FHWA.(1)
Note that only 3 GRS composites tested in this study meet both the design limits and the material specifications (e.g., Tf≥ 4,800 lb/ft, dmax ≥ ½ inches) for GRS abutments; however, based on the results of this study, perhaps the criteria can be amended to reduce the reinforcement strength limitation, provided the reinforcement strength meets internal stability design requirements.(1)
Based on the results of the PTs presented in this report, the average bias for the capacity equation is 0.88 with a COV of about 35 percent (table 23). Including the results from other sources in the literature, the average bias is 0.95 with a COV of about 32 percent. These numbers indicate relatively good agreement between figure 8 and the measured ultimate capacity (table 24).
The 5-percent vertical strain limit can significantly reduce the allowable stress placed on a GRS composite; figure 82 shows that at 5-percent strain, the applied pressure is between 50 and 85 percent of the measured ultimate capacity (table 8).
The results indicate that by using 10 percent of the predicted design capacity (figure 8), the 0.5-percent vertical strain as required by FHWA for the service limit of GRS abutments can largely be satisfied. The mean bias of the ratio between the measured results at 0.5-percent vertical strain from this series of PTs and the predicted allowable stress at 10 percent of the design capacity is 1.15 with a COV of 0.51 (table 25). This is on the conservative side yet offers another tool to estimate deformation in lieu of conducting a performance test on a particular GRS composite.
Finally, a reliability analysis for the soil-geosynthetic capacity equation (figure 8) was performed on the 16 PTs taken to failure in this study, along with previous results from additional GRS testing found from the literature.
A target reliability index of 2.5 is reasonable for the strength limit models for GRS composites designed according to Adams et al. (2011). A lower target reliability index is warranted for closely-spaced GRS composites because of the redundancy in the reinforcement and because no catastrophic collapse was observed in any of the performance tests at failure. In addition, Bathurst et al. (2008) suggest that a reliability index of 2.3 is appropriate for the internal stability of reinforced soil walls.(1,41)
Performing a reliability analysis using the FOSM approach produces a resistance factor around 0.45, similar to that found through calibration by fitting to ASD methods.
Page Owner: Office of Research, Development, and Technology, Office of Infrastructure, RDT