Test SM-S v 12-C4 (Adoption of Sand Cushions)

Figure 6.25 shows the pictures of the test SM-Sv12-C4. The wall did not move until the first tension crack developed at the topmost layer at 88 mins after the rainfall started.

The crack quickly propagated downward to the lower layers and 20 minutes after, more tension cracks were found at the topmost layer. After six and a half hours, water started to fill into the tension cracks, causing further slight wall displacement.

(a)  Start of the test (t = 0 min)

(b)  Water reached the bottom and first tension crack developed (t = 88 mins)

εxy

(c)  Second tension crack developed (t = 107 mins)

(d)  End of the test (t = 1080 mins) Figure 6.25 Pictures of the model test SM-Sv12-C4

tension cracks

0.0

-0.2

-0.4

-0.6

0.0

-0.2

-0.4

-0.6 εxy

Figure 6.26 presents the change of the volumetric water content with time elapse.

The first tension crack developed when the water reached the bottom of the wall. VWC3 was cut off because the development of the tension crack resulted in the exposure of the volumetric water content gauge, causing the measured value beyond saturation. VWC3, VWC2, and VWC1 sensed the water in succession as the infiltrated water flowed along the direction of the gravity. VWC2 and VWC1 eventually reached the same value because the permeability of the marginal backfill was relatively low; the water was indeed confined at the top layers as the water was ponded in the tension cracks as observed.

Figure 6.26 VWC with time elapse

Figure 6.27 displays the wall displacement profile. Several sudden displacements were observed within t = 90 ~ 600 mins. The maximum wall displacement at the end of the test was found in the topmost layer with a value of 15% wall height. The test without sand cushions (SM-Sv12) had a maximum wall displacement of 33% wall height at t = 127 minutes and the wall fully collapsed. In the test SM-Sv12-C4 however, even the reinforcement had large Sv, the wall did not collapse owing to the adoption of the sand cushions. The large interface friction between the sand cushions and the geogrid reinforcement constrained the excessive displacement.

Figure 6.27 Wall displacement profile

Figure 6.28 presents the maximum wall displacement and settlement versus time.

The major wall displacement commenced at t = 60~120 mins after the rainfall started.

The development of the major wall displacement in the first 2 hours resulted from the loss of matric suction and soil shear strength upon the infiltration of rainfall. The displacement became gradual afterward. Furthermore, the rainfall infiltration had led to the accumulation of water in the topmost layer, causing the saturated soil bulged out.

However, the soil not collapse because of the resistance force from the geogrid reinforcement and the strong interface friction between the sand cushions and the geogrid reinforcement.

Figure 6.28 Maximum wall displacement and settlement versus time

Figure 6.29 presents the mobilized reinforcement tensile strain in each layer. The maximum strain was mobilized at the topmost layer where the largest wall displacement ensued. The maximum mobilized strain was 5%, which is very close to the ultimate tensile strain of the geogrid reinforcement (εu = 6.5%), and this can be correlated to the relatively small factor of safety for breakage (FSbreakage = 1.6). By connecting the locus of the maximum tensile strain, a critical failure surface can be plotted. Notably, there were two regions where larger tensile strains were mobilized in layers 3, 4, and 5. The regions coincided with the location of the tension cracks. Moreover, the strain contour from the PIV analysis indicated large strains at both the critical failure surface and the location of tension cracks. The maximum strain determined from the PIV analysis was 65%, which was much larger than the soil failure strain. Failure was indeed observed along the maximum strain contour.

Figure 6.30 displays the maximum reinforcement tensile force profile. The maximum reinforcement tensile strain was a cantilever type, causing the strain at the top layers be underestimated while the strain at the lower layers was overestimated by the earth pressure method.

Figure 6.30 Maximum reinforcement tensile force profile

In summary, the test results from the test SM-Sv12-C4 demonstrates that the installation of sand cushions can effectively prevent the collapse of the wall, even though excessive deformation was observed at the end of the test. The low permeability of the backfill caused the accumulation of rainfall, further induced the loss of matric suction and soil shear strength. Tension cracks, large deformation, and the critical failure surface therefore developed. Nevertheless, the high interface friction between the soil and geogrid

6.2.4 Comparison

Table 6.5 summarizes the factor of safety against breakage and pullout for the tests SM-Sv12 and SM-Sv12-C4, and Figure 6.31 presents the pictures of the two tests. Notably, Figure 6.31 (a) was taken at the moment before the wall fully collapsed (t = 127 mins).

The thickness of the sand cushions was 4 cm, accounting for 26% of the backfill. The factor of safety against pullout was significantly larger in the test SM-Sv9-C12 owing to the high interface shearing characteristic between sand-geogrid interface. However, the factor of safety against breakage was relatively small.

Table 6.5 Factor of safety

Model test FSbreakage FSpullout (dry) FSpullout (wet)

SM-Sv12 1.6 2.7 1.9

SM-Sv12-C4 1.6 3.3 3.2

(a) (b)

Figure 6.31 Pictures of tests: (a) SM-Sv12; (b) SM-Sv12-C4

Figure 6.32 shows the wall displacement profile of the test SM-Sv12 and SM-Sv 12-C4. The maximum wall displacement achieved a value of 33% wall height in the test SM-Sv12 before the reinforced wall totally collapsed. Regarding the test SM-Sv12-C4, though the wall displacement is quite large (18% wall height), the wall did not collapse. The interface interaction with the geogrid was enhanced and the sand cushions effectively improved the deformation characteristic. With sand cushions adopted, the wall displacement was reduced by 50% with only 26% sand replaced as sand cushions.

Figure 6.32 Wall displacement profile

Figure 6.33 shows the development of the maximum wall displacement in the test SM-Sv12 and SM-Sv12-C4. The figure indicated that for the test without sand cushions (SM-Sv12), the maximum wall displacement dramatically increased to 11 cm after the tension cracks developed (t = 94 mins). The wall eventually collapsed in an interlayer sliding mode. Contrarily, after the adoption of sand cushions (SM-Sv12-C4), the wall displacement had turned into a more progressive development. The major displacement induced by the loss of matric suction and soil shear strength was 8 cm, followed by a 3 cm successive displacement owing to the development of soil shear strain. The effect of sand cushions can be clearly verified. With such large Sv, the wall experienced large deformation but did not collapse.

Figure 6.33 Maximum wall displacement versus time

Figure 6.34 presents the mobilized reinforcement tensile strain in the tests SM-Sv12 and SM-Sv12-C4. The maximum reinforcement tensile strain mobilized in the two tests were around 5% since the maximum wall displacement observed in the two tests were similar. Notably, the mobilized reinforcement strain in the test SM-Sv12 was calculated according to the wall displacement at the moment before the wall collapsed (t = 127 mins).

The rather large reinforcement tensile strain mobilized can correspond to the relatively low factor of safety against breakage (FSbreakage = 1.6).

Figure 6.34 Reinforcement tensile strain

Figure 6.35 displays the mobilized maximum reinforcement tensile force profile.

The mobilized reinforcement tensile force was similar in both tests since the wall displacement in the both tests were similar. The maximum tensile force occurred at the topmost layer, resembling the cantilever pattern as the wall displacement. The earth pressure method significantly underestimated the tensile force at the top layers and overestimated that at the bottom layers for reinforced walls with marginal backfill.

Figure 6.35 Maximum reinforcement tensile force profile

In summary, the adoption of sand cushions can effectively prevent the reinforced wall with large spacing from failure by improving the soil-geogrid interface friction and enhancing the deformation characteristics. Table 6.6 summarizes the performance of the tests SM-Sv12 and SM-Sv12-C4. The stability of the wall was dramatically improved by only replacing 26% of the backfill as sand cushions. Sand cushions can provide drainage and prevent the accumulation of water; the displacement induced by water was therefore restrained. In addition, the deformation characteristic of sand cushions was much better than that of marginal backfills. Specifically, sand cushions can withstand the large overburden pressure and contribute to the wall stability. With that, the test SM-Sv12-C4 did not collapse even with such large Sv.

Table 6.6 Summary of test results

Test SM-Sv12 SM-Sv12-C4

Final stage Fully collapsed Excessive

deformation

Failure mode Sudden Progressive