# In Situ Test of Traffic-Load-Induced Settlement of Alluvial Silt Subsoil Treated by Unslaked Lime

## Abstract

The soft wet alluvial silt is widely distributed material in the world. In order to improve the bearing capacity and decrease the traffic-load-induced settlement of silt subsoil, the shallow subsoil always treats with the unslaked lime. However, the mitigating effect of this ground treatment method on traffic-load-induced settlement of alluvial silt subsoil is inconclusive. Therefore, with the developed falling weight simulation equipment of traffic load, in situ tests are carried out on the natural and unslaked lime treated alluvial silt ground in the Yellow River delta of China to study traffic-load-induced settlement. Furthermore, Chai-Miur cumulative deformation model of soil is employed to numerically simulate the long-term traffic-load-induced cumulative settlement. In situ test results indicate that because unslaked lime treatment enhances wave impedance of the reinforced soil layer, the wheel-load-induced dynamic stress and excess pore water pressure in the substratum decreases. The decrease of excess pore water pressure reduces the cumulative settlement of unslaked lime treated subsoil. For short-term cumulative settlement, there are differences between tested and calculated results, but not much. The calculated results imply that after opening to traffic for 10 years, compared with the natural ground, the cumulative settlement of the unslaked lime treated subsoil reduces by about 21%, and the change of transverse slope of pavement induced by cumulative settlement decreases by 1/3. In situ test and numerical calculation results demonstrate that shallow layer treatment with unslaked lime can effectively mitigate the cumulative settlement of alluvial silt subsoil.

## 1 Introduction

Alluvium is loose soil, which has been eroded and reshaped by water in some form, and typically made up of a variety of materials, including fine particles of silt and clay and larger particles of sand and gravel. Alluvium is widely distributed in the world. In China, the Yellow River delta is the youngest large river delta due to frequent river diversion in history, and in the Yellow River delta, the alluvium is mainly low liquid limit silt. The alluvial silt is newly formed under consolidated deposit and has poor engineering performances with low liquid limit and plasticity index, small cohesion, low strength, intensive capillarity, poor graduation and water stability and high liquefaction potential. Furthermore, because the alluvial silt is liquefiable and the groundwater level is very high in the region, the traffic-load-induced cumulative settlement of the road with low embankment is significant after opening to traffic, and this can cause serious pavement diseases. In order to improve the bearing capacity of the soft wet subsoil in the region and reduce its cumulative settlement, shallow layer treatment with unslaked lime is usually employed. The improving effect of unslaked lime treatment on the bearing capacity of soft wet subsoil is undeniable (Kamon 1992; Narasimha Rao and Rajasekaran 1996). However, at present the mitigating effect of hard crust formed by unslaked lime treated subsoil on the traffic-load-induced cumulative settlement is inconclusive. Note that in this study, the pavement means the hard layered structure that forms a road carriageway, the embankment means the compacted soil layer below the pavement of a road, and the subsoil means the natural soil below a road embankment; the term subgrade includes all layers above the natural ground surface.

Previous studies on the cumulative settlement of subsoil focus on dynamic deformation test of soil in the laboratory and numerical simulation. In the early 1950s, Seed and his co-workers (Seed et al. 1955; Seed and McNeill 1956) studied the settlement behaviour of road under repeated loads by normal compression tests. Monismith et al. (1975) analyzed the characteristics of permanent deformation of subsoil due to repeated loadings based on repeated-load triaxial compression tests. Yildirim and Ersan (2007) studied consolidation settlements of soft clay by undrained cyclic simple shear tests in the laboratory. Shahu and Yudhbir (2008) studied cumulative plastic strain of a quasi-saturated compacted silty clay under cyclic load by monotonic and cyclic undrained triaxial tests. Cyclic triaxial tests were conducted by Liu and Xiao (2010) to study the behaviour of silt subsoil under various physical states and stress conditions. Chai and Miura (2002) modified the Li-Selig model (Li and Selig 1996) and calculated the permanent settlement of road with a low embankment on soft subsoil. Akira et al. (2003) analyzed the traffic-load-induced settlement of low embankment road on silty-clay. Abdelkrim et al. (2003) adopted a general structure analysis approach to predict the traffic-load-induced residual settlement. Fujiwara and coworkers (Fujiwara et al. 1985; Fujiwara and Ue 1990) researched the effect of preloading on the cumulative settlement of clay under cyclic loadings, and they concluded that the amount of settlement that occurs after construction depends strongly on the soil over consolidation ratio, degree of consolidation at the time of unloading, static loading magnitude, and repeated loading magnitude. Chai and Miura (2002) calculated the permanent settlement of the improved subsoil with soil-cement columns. Indraratna et al. (2010) carried out a large-scale triaxial testing on clay-drain complex subjected to cyclic loading representing a typical track environment to predict the behaviour of the soft estuarine subsoil with short vertical drains. However, currently there has been no studies on the cumulative settlement of unslaked lime treated subsoil.

Though the laboratory tests can evaluate the traffic-load-induced cumulative deformation of subsoil, they are not accurate because the stress and boundary conditions are more complex in real practices. The numerical method, although it has the advantage of analyzing the subsoil settlement under complicated stress conditions, has its shortcomings, such as long calculating time and uncontrollable cumulative calculation errors, which limit its practical use in engineering. In addition, the road settlement observed in the field after opening to traffic is the sum of the settlement induced by the weight of subgrade and traffic-load-induced settlement. However, there is no effective method to separate them and analyze individually traffic-load-induced cumulative settlement. Therefore, it is significant to conduct in situ tests for directly simulating the traffic-load-induced settlement of subsoil. In this paper, with the developed Falling-Weight Traffic Load Simulation Equipment (FWTLE), in suit tests are performed to study the short-term traffic-load-induced cumulative settlement of unslaked lime treated alluvial silt subsoil in the Yellow River delta. And by the numerical calculations, the long-term cumulative settlement is studied. Comparing the cumulative settlement of the natural ground with that of unslaked lime treated subsoil demonstrates the validity of unslaked lime treatment method.

## 2 In Suit Tests

### 2.1 Description of the Site

Geomechanical parameters of natural subgrade

Soil layer | Liquid limit (%) | Plasticity index (%) | Water content (%) | Void ratio | Saturation | Cohesion (kPa) | Internal friction angle (°) | Compression modulus (MPa) |
---|---|---|---|---|---|---|---|---|

0–0.6 m | 25 | 7.4 | 17.6 | 0.602 | 0.78 | 31.8 | 22.9 | 7.45 |

0.6–10 m | 27 | 9.8 | 18. | 0.603 | 1.00 | 26.6 | 20.5 | 15.86 |

Physical and mechanical parameters of unslaked lime treated soil

Lime content | Optimum moisture content (%) | Maximum dry density (g/cm | Elastic modulus (MPa) | Cohesion (kPa) | Internal friction Angle (°) |
---|---|---|---|---|---|

6% | 15 | 1.81 | 77 | 115.4 | 34.7 |

### 2.2 Test Devices and Parameter Settings

For in situ simulated test of cumulative settlement, the parameter settings of the FWTLE are critical. These parameters include the weight and height of falling weight, the stiffness of air spring and the size of loading plate. By setting different parameters, the different stress response of subsoil can be simulated.

The height of the road embankment has a great effect on traffic-load-induced vertical stress on the ground surface. According to the earlier investigation (Cui 2012), for the vehicle with standard axle load of 100 kN, the common highway pavement structure (the thickness is approximately 70 cm) and zero-fill embankment (the height of the embankment is 0 m), the vertical stress on the ground surface is about 22 kPa. But when the height of the embankment is 0.8 and 1.5 m, the vertical stresses reduce to 7.8 and 5.6 kPa, respectively. Simultaneously, for zero-fill embankment, when the horizontal distance from the center of wheel gap is more than 0.7 m, the vertical stress changes slowly with the horizontal distance, and this is basically irrelevant to the action time of wheel load (i.e. vehicle speed). Therefore, the size of square loading plate used in situ test is determined as 1.2 m × 1.2 m, and its thickness is 3 mm. And the radius of its equivalent area circle is 0.68 m, which can meet the test precision requirement in the case of the zero-fill embankment. The falling weight in the test is the hammer used in the standard penetration test (SPT) for the engineering geological investigation and it is 62.5 kg in weight. Before in situ test, by adjusting the height of falling weight and the stiffness of air spring, different vertical stress under the loading plate can be obtained to meet the requirements of different engineering cases. The amplitude and duration of vertical dynamic stress response obtained in tests should be approximately consistent with the computed results in the cases of different heights of the embankment and different vehicle speeds.

Because the cumulative settlement of the zero-fill road embankment is significantly larger, in this study, the falling distance and the stiffness of air spring were adjusted to simulate the case of the zero-fill embankment.

### 2.3 Layout of Sensors

Before in situ tests, remove weeds and cover soil on the ground. In order to study the mitigating effect of unslaked lime treatment method on the cumulative settlement, natural and unslaked lime treated subsoil were tested, respectively. In the process of cyclic loading, the parameters such as stress, displacement and pore pressure were tested. Because the vehicle-load-induced excess pore pressure need a long time to dissipate, the settlement and pore pressure after cyclic loading were also tested. The total cumulative settlement was the sum of the settlement during and after loading.

### 2.4 Test Results and Analyses

#### 2.4.1 Dynamic Stress

Compared with the natural silt subsoil, it can be seen from Fig. 4b that the stress amplitude on the unslaked lime treated subsoil surface increases. This is because after mixing the soft wet soil with the reasonable amount of unslaked lime, a series of reactions take place: water absorption, exothermic action and expansive action; ion exchange action; carbonation (chemical cementation reaction); pozzolanic action (chemical gelation reaction) and crystallization action. These reactions can make the moisture in soil reduce, slit particles coagulate to form larger aggregates. With the gradual hardening of subsoil, soil particles bonded together and the physical and mechanical properties of the subsoil are improved. With the subsoil treated by unslaked lime, the wave impedance of shallow subsoil increase, and the hard curst effect causes the increase of the dynamic stress amplitude on the subsoil surface.

#### 2.4.2 Excess Pore Water Pressure

*N*during the loading process. It can be seen that, compared with the natural ground, the excess pore water pressure in the soft substratum of the unslaked lime treated ground is obviously lower. This is because with the ground treated by unslaked lime, the dynamic stress level in the substratum decreases (as shown in Fig. 5).

At the initial stage of loading, the pore water pressure in the soft substratum of the unslaked lime treated ground develops faster than that in the natural ground. Nevertheless, after certain load numbers, the pore water pressure in unslaked lime treated ground does not increase but gradually declines to a stable state at last. By contrast, the pore water pressure in natural ground continuously increases with the load numbers. The difference above is mainly because of the different drainage conditions of the unslaked lime soil treated and natural grounds during the loading process. Because the permeability coefficient of unslaked lime treated soil is smaller than the natural soil, the soil-lime hard crust can block the dissipation of pore water pressure in the substratum. Therefore, at the beginning of cyclic loading, the pore water pressure in unslaked lime treated ground increases faster than the one in the natural ground. But as the pore water pressure increases, the strength of the substratum decreases, and this induces the cracking of the soil-lime hard crust under the traffic load (observed in tests). The cracks provide good drainage channels. As mud comes out through the channels, the pore water pressure in the substratum of hard crust does not increase, but decreases. However, for the natural ground, the flexibility of top soil is larger, so its main failure mode is not local cracking but the plastic deformation. Therefore, the seepage paths change little in the process of cyclic loading, and this makes the pore water pressure continuously increase.

After terminating loading, compared with the natural ground, the dissipation of excess pore water pressure in unslaked lime treated ground needs more time. The dissipation of excess pore water pressure in natural ground needs seven hours, but the one in unslaked lime treated ground takes more than twenty hours. This is because after terminating cyclic loading, the mud hardens in the cracks of soil-lime hard crust induced by cyclic loading, and this makes the hard crust becomes impervious. However, the practical traffic load is cyclic and continual, the cracks in hard crust are not blocked. So the delay effect of dissipating the pore water pressure does not occur.

#### 2.4.3 Cumulative Settlement

The traffic-load-induced cumulative deformation of soil is mainly composed by undrained shear deformation and consolidation deformation. Under practical traffic loads, these two deformations occur simultaneously. However, in model tests, only in the process of cyclic loading, are these two deformations concurrent. After terminating loading, the settlement of subsoil is mainly induced by consolidation deformation of soil. Although the deformation paths of subsoil under the practical and simulated vehicle load are different, the sum of the undrained shear and consolidation deformations obtained in the tests can be regarded as the total cumulative deformation induced by traffic load (Cui 2012).

## 3 Numerical Simulations of Cumulative Settlement

Traffic-load-induced settlement is a long-term cumulative process. In situ tests as above can be employed to study the cumulative settlement. However, its conduction is costly and time-consuming, so it is only suited for studies of the initial fast-developing cumulative settlement. Therefore, numerical method is widely employed to predict the traffic-load-induced long-term cumulative settlement. At the early age, Monismith et al. (1975) proposed a power model of cumulative deformation of soil, and then it was modified by Li and Selig (1996). Chai and Miura (2002) made further improvement to Li-Selig model. The Chai-Miura model considered not only the dynamic deviatoric stress and static strength of soil but also the effect of initial static deviatoric stress on cumulative deformation. The Chai-Miura model is widely used to calculate the cumulative deformation of soil.

*J*

_{2s}is the second deviatoric stress invariant of initial static stress; \( q_{\text{d}} = \sqrt {3J_{{ 2 {\text{d}}}} } \) is the dynamic deviatoric stress,

*J*

_{2d}is the second deviatoric stress invariant of dynamic stress peaks in all directions;

*q*

_{f}is the static strength of soil;

*a*,

*b*,

*m*and

*n*are parameters of soil;

*N*is the load numbers.

*q*

_{f}can be determined by strength index

*c*

_{cu}and

*φ*

_{cu}of the consolidation undrained total stress:

*K*

_{0}is the static soil pressure coefficient at rest;

*σ*

_{cz}is overlying soil pressure.

*a*, *b*, *m* and *n* in Eq. (1) reflect the combined effect of stress state, physical state and types of soil. Series of triaxial tests, the constants in Eq. (1) were obtained: *a *= 0.64, *b *= 0.10, *m *= 1.70 and *n *= 1.00.

### 3.1 Comparisons and Analyses of the In Situ Test and Numerical Results

Before calculating long-term cumulative settlement of subsoil, in order to prove the validity and feasibility of calculation method, the cumulative settlement of natural ground was numerically calculated by simulating in situ test.

Firstly, the static module of the finite difference program Flac3D was used to calculate initial static deviatoric stress *q*_{s} in the subsoil. Secondly, the dynamic module of Flac3D was employed to simulate the dynamic response of subsoil. In the dynamic calculation, the vertical dynamic stresses obtained by soil pressure sensors (Fig. 3a) under the loading plate in the tests were loaded on the subsoil surface. The dynamic deviatoric stress *q*_{d} can be obtained from the peaks of the dynamic stresses in three orthogonal directions. Finally, *q*_{s} and *q*_{d} were taken into Eq. (1), and cumulative deformation of subsoil can be calculated. The static strength of soil *q*_{f} is calculated by Eq. (2). The cumulative settlement on the subsoil surface can be obtained by integrating the vertical cumulative deformation of soil along depth.

### 3.2 Numerical Simulations and Analyses of Long-Term Cumulative Settlement

Road structure and material parameters

Structural layer | Materials | Thickness (cm) | Elastic modulus (MPa) | Poisson ratio |
---|---|---|---|---|

Upper surface | SMA | 4 | 1400 | 0.3 |

Middle surface | Mesograin modified asphalt concrete | 6 | 1200 | 0.3 |

Lower surface | Coarse graded asphalt concrete | 8 | 1000 | 0.3 |

Upper base | Large grain size asphalt gravel | 12 | 1400 | 0.35 |

Lower base | Cement stabilized gravel | 36 | 1500 | 0.35 |

Subbase | Lime-ash soil (30% additive gravel) | 20 | 800 | 0.35 |

Roadbed | Low liquid limit silt (96% compaction degree) | 80 | 30 | 0.35 |

Embankment | Low liquid limit silt (94% compaction degree) | 60 | 17 | 0.35 |

*t*is time;

*F*

_{max}is the wheel load peak, 65 kN for the front wheel and 35 kN for the rear wheel;

*T*

_{s}is the duration time of single vehicle load, and has an inverse relationship with the vehicle speed. Herein

*T*

_{s}is taken as 0.031 s, representing the equivalent vehicle speed of 120 km/h (Huang 1993).

*N*

_{1}= 13463 is the annual average daily traffic at the early stage after opening to traffic;

*r*= 5.775% is the average annual growth rate;

*t*is time in year.

## 4 Conclusions

- (1)
Compared with the natural ground, the wave impendence of unslaked lime treated ground increases. This makes traffic-load-induced dynamic stress in substratum significantly reduce. Moreover, the excess pore water pressure in the substratum of unslaked lime treated subsoil is less than that in the natural ground.

- (2)
In comparison with the natural ground, the cumulative settlement of the unslaked lime treated ground obviously decreases. This proves the mitigating effect of unslaked lime treatment on the cumulative settlement of alluvial silt subsoil.

- (3)
The cumulative settlement of subsoil aggravates the development of pavement diseases. Unslaked lime treatment of subsoil can effectively decrease the harms of traffic loads on the road with low embankment.

## Notes

### Acknowledgements

This work was supported by the Chinese Natural Science Foundations (Nos. 51279094, 51078222 and 50708056), the Natural Science Foundations of Shandong Province, China (No. ZR2011EEM012) and the Independent Innovation Foundation of Shandong University (IIFSDU) (No. 2012HW003).

## References

- Abdelkrim, M., Bonnet, G., Buhan, P.D.: A computational procedure for predicting the long term residual settlement of a platform induced by repeated traffic loading. Comput. Geotech.
**30**(6), 463–476 (2003)CrossRefGoogle Scholar - Akira, S., Lawalenna, S., Norihiko, M.: Partially-drained cyclic behavior and its application to the settlement of a low embankment road on silty-clay. Jpn. Geotech. Soc.
**43**(1), 33–46 (2003)Google Scholar - Chai, J.C., Miura, N.: Traffic-load-induced permanent deformation of road on soft subsoil. J. Geotech. Geoenviron. Eng.
**128**(11), 907–916 (2002)CrossRefGoogle Scholar - Cui, X.Z.: Traffic-induced settlement of subgrade of low liquid limit silt in Yellow River delta. China Civil Eng. J.
**45**(1), 154–162 (2012). (in Chinese)Google Scholar - Fujiwara, H., Ue, S., Yasuhara, K.: Consolidation of alluvial clay under repeated loading. Soils Found.
**25**(3), 19–30 (1985)CrossRefGoogle Scholar - Fujiwara, H., Ue, S.: Effect of preloading on post-construction consolidation settlement of soft clay subjected to repeated loading. Soils Found.
**30**(1), 76–86 (1990)CrossRefGoogle Scholar - Huang, Y.H.: Pavement Analysis and Design. Pearson Education, Delhi (1993)Google Scholar
- Indraratna, B., Rujikiatkamjorn, C., Ewers, B.: Class A prediction of the behavior of soft estuarine soil foundation stabilized by short vertical drains beneath a rail track. J. Geotech. Geoenviron. Eng.
**136**(5), 686–696 (2010)CrossRefGoogle Scholar - Kamon, M.: Recent developments of soil improvement. In: Proceedings of International Symposium on Soil Improvement and Pile Foundation, Nanjing, China, vol. I, pp. 1–16 (1992)Google Scholar
- Li, D., Selig, E.T.: Cumulative plastic deformation for fine-grained subgrade soils. J. Geotech. Eng.
**122**(12), 1006–1013 (1996)CrossRefGoogle Scholar - Liu, J.K., Xiao, J.H.: Experimental study on the stability of railroad silt subgrade with increasing train speed. J. Geotech. Geoenviron. Eng.
**136**(6), 833–841 (2010)CrossRefGoogle Scholar - Monismith, C.L., Ogawa, N., Freeme, C.R.: Cumulative deformation characteristics of subsoil due to repeated loading. Transp. Res. Rec.
**537**, 1–17 (1975)Google Scholar - Narasimha, R.S., Rajasekaran, G.: Reaction products formed in lime-stabilized marine clays. J. Geotech. Eng.
**122**, 329–336 (1996)CrossRefGoogle Scholar - Seed, H.B., Chan, C.K., Monismith, C.L.: Effects of repeated loading on the strength and deformation of compacted clay. Highw. Res. Board Proc.
**34**, 541–558 (1955)Google Scholar - Seed, H.B., McNeill, R.L.: Soil deformation in normal compression and repeated loading test. Highw. Res. Board Bull.
**141**, 44–53 (1956)Google Scholar - Shahu, J.T., Yudhbir, Hayashi S.: Cumulative plastic strain and threshold stress of a Quasi-saturated compacted silty clay. Lowland Technol. Int.
**10**(2), 10–20 (2008)Google Scholar - Shen, Z.J.: Earth pressure of clay based on effective consolidation stress theory. Chinese J. Geotech. Eng.
**22**(3), 353–356 (2000). (in Chinese)Google Scholar - Wang, X.: Test on dynamic stress of roadbed and pavement under heavy loads. J. Vib. Shock
**26**(2), 169–173 (2007). (in Chinese)Google Scholar - Yildirim, H., Erşan, H.: Settlements under consecutive series of cyclic loading. Soil Dyn. Earthq. Eng.
**27**, 577–585 (2007)CrossRefGoogle Scholar