Strength Characterization of Open-Graded Aggregates for Structural Backfills

CHAPTER 5. CONCLUSIONS

Sixteen AASHTO M43 gradations, ranging from an AASHTO No. 10 to an AASHTO No. 5 aggregate, were characterized in this study, including their gradation, minimum and maximum density, repose angle, angularity, and texture. In addition, LSDS and LDTX tests were conducted on the samples, compacted at a relative density of approximately 30 percent, to evaluate the strength. The results provide insight into the strength behavior of manufactured OGAs.

5.1 FRICTION ANGLE

Transportation agencies use OGAs as structural backfill for road and bridge construction, yet their strength parameters are not fully understood. To address this need, a comprehensive study was initiated, with the primary objectives to quantify the strength parameters, evaluate different conventional testing devices, and determine the influence of various factors on strength. The results indicate that the typical default friction angle of 34 degrees is a conservative measure (table 23). It is important to note that the reported friction angles are from samples tested at 30 percent relative density. In the field, compaction would result in added strength; therefore, the friction angles reported herein are considered conservative.

Table 23 . Summary of LSDS and LDTX testing.

CV = Constant volume.
LDTX = Large-diameter triaxial.
LSDS = Large-scale direct shear.
MC = Mohr-Coulomb.
ZDA = Zero dilation angle.
- = Not measured.

Two methods for determining the friction angle were investigated for comparison: (1) the slope of the best-fit linear MC failure envelope as in current practice, and (2) the ZDA approach. When combining all the results from every aggregate, the tangent friction angle was 48.2 and 38.4 degrees for LSDS and LDTX testing, respectively. Similarly, the CV friction angle from the ZDA method was 46.1 and 38.9 degrees for LSDS and LDTX testing, respectively (figure 67). Based on a typical COV of 5 to 10 percent for lab testing, and using the ZDA method for LSDS testing, the recommended default friction angle for AASHTO OGAs is 41 degrees.⁽⁸⁵⁾

Figure 67. ZDA Approach for LDTX and LSDS testing.

Regardless of the approach, the results are similar; however, the ZDA method offers more confidence as a conservative estimate for strength. The linear MC method resulted in more data scatter and produces a measured cohesion value for a cohesionless material. The ZDA method also provides a consistent basis for obtaining the friction angle at the critical state but requires an extra step in the analysis to determine the dilation angle. Other methods such as measuring the secant friction angle at a given confining stress are available but rely on a design assumption that confining stress will remain constant throughout the life of the structure.

5.2 TEST METHOD

The selection of the test device to measure strength plays an important role. The LDTX device resulted in strengths that were considerably lower than those measured in the LSDS device. The difference in results generally increases with increasing aggregate size. The dilation angles determined from the LDTX device are also lower than those measured with the LSDS device (about 60 percent) because there is a forced failure plane during the LSDS test, causing more dilation. The difference in dilatancy from each test device helps explain the lower friction angles measured with LDTX because dilation is a component in the strength of aggregates. Because the contribution of dilation is different depending on whether the LDTX or LSDS test is employed (which suggests that there is a trend with dilation but not with tangent or CV friction angles between LDTX and LSDS results), it is clear that more factors are contributing to the strength between the two devices.

The SDS device was also investigated to evaluate a common practice of scalping the sample. To use a 2.5-inch-diameter SDS device, aggregates larger than 0.25 inches must be removed from each test sample. The SDS results of the scalped samples indicated friction angles that were still considerably greater than the default of 34 degrees. There was no clear relationship between the scalped and unscalped results determined from SDS and LSDS testing, respectively, with friction angles largely between 10 and 20 percent different. While the overall strength values were similar, the scalping process changes the dilation properties of the aggregates, with relative differences ranging from 15 to 65 percent; no trend was found. In addition, the initial void ratio also changed, which has an impact on strength. Overall, scalped samples are unrepresentative of field behavior and lead to uncertainty in the results.

The effect of saturation was evaluated in the LSDS device. For two-thirds of the aggregates (10 out of 15), the difference in the tangent friction angles between saturated and dry conditions was within 10 percent. The results were similar for a larger population-almost 90 percent of the samples (13 out of 15)-when using the ZDA method. It is thought that saturation reduces the dilative behavior, which is why the ZDA approach shows more negligible results. Overall, the difference between friction angles under dry and saturated conditions was within 20 percent for all samples regardless of test method. While the test should ideally mimic in-service conditions, the effect of saturation for OGAs appears minimal.

To take advantage of the strength of these engineered aggregates for more cost-effective designs, large-scale testing is recommended. The selection of LSDS versus LDTX testing depends on the use of the results and other factors. For example, if the friction angle is the only parameter needed, then the LSDS device is quicker and simpler to perform. The LDTX test can offer more information about the deformation behavior but is more complex, time consuming, and costly to perform.

Both the LSDS and LDTX devices produce friction angles that are theoretically less than PS conditions. The result from either device is appropriate in design; however, in most cases, the LSDS results would produce more cost-effective structures. In typical applications utilizing OGAs, such as in reinforced soil retaining walls and bridge foundations, the aggregates will be confined in a PS condition, which the LSDS device would more adequately mimic as compared with the LDTX device

5.3 STRENGTH CORRELATIONS

Correlations between various soil properties and parameters were investigated, but no direct relationship was found that designers can use in the absence of strength testing. LDTX results were independent of the various parameters; however, there were general trends found for LSDS testing. The strongest relationships were observed when comparing the aggregate size, angularity, void ratio, and sphericity to strength. There was no relationship between the measured friction angle and the coefficient of uniformity, maximum unit weight, repose angle, and aggregate texture; however, these poor correlations could potentially be the result of the relatively small differences between the physical parameters for the aggregates tested in this study.

As the mean aggregate size increased, the friction angle also increased. The tangent and CV friction angles measured with the LSDS device show an inverse relationship with the average angularity index. While this result is contrary to expectations, it is postulated that as the aggregates gets more angular, there is less contact area between the particles and a higher void ratio (i.e., less dense), and higher void ratios typically result in lower friction angles. There is a positive trend with sphericity, however, whereby as the average sphericity index increases, the friction angle also increases. With the forced failure plane in LSDS testing, sphericity affects the amount of dilation that occurs.

5.4 FUTURE RESEARCH

This study was the first to systematically test and characterize AASHTO OGAs. While many insightful results were produced, additional research is needed before concrete design recommendations can be proposed. A database of multiple sources of each OGA is needed to quantify the material variability to determine appropriate confidence intervals. The impact of relative density on the strength will also be investigated. In addition, the results presented utilized material that was retested for each confining stress; the influence and practicality of using virgin material for each test will be studied. Until this range of research is conducted, readers are encouraged to understand and take advantage of the strength of these OGAs in design. Testing can then be used to facilitate the use of higher than 34-degree default values.