1 Introduction

1.1 History of Cement Stabilized Base in Thailand

With the limited resource of the crushed rock used for granular base course and an increase in the transportation cost, the highway construction budget in the northeastern part of Thailand becomes increasing substantially. In 1965, the Siam Cement Company and the Department of Highways (DOH), Thailand constructed the first 5 km soil–cement road by stabilizing lateritic soil with cement as a base course in the northeastern part of Thailand. After the first soil–cement road had been constructed, the DOH continued to construct the soil–cement roads. Until 1972, over 1,400 km of soil–cement roads have been constructed (Rananand et al. 1983).

Unfortunately, the transverse cracks were later detected along the pavement surface of several soil–cement roads. Those cracks were believed to reflect from the cement stabilized base course. Since then, the DOH stopped constructing the soil–cement road. From 1973 to 1985, the DOH officials led by Mr. Nibon Rananand and Dr. Therachatri Ruenkrairergsa performed a number of comprehensive studies to evaluate the long-term performance of the existing soil–cement roads in order to alleviate those reflective cracks. After every issue had been resolved, the cement stabilized bases were adopted again by the DOH (Rananand 2001; Ruenkrairergsa 1989) and many soil–cement roads were built successfully.

In addition to the cement stabilized lateritic soil or “soil–cement base”, the DOH also applied the cement stabilization method in the conventional crushed rock base, which is called “modified crushed rock”. Moreover, the cement stabilization method was applied for pavement rehabilitation on the old highways in such a way that the crushed rock base and the reclaimed asphalt pavement were mixed with the Portland cement. Such rehabilitation technique was commonly known as pavement recycling and has been adopted in Thailand since 1994.

1.2 Mix Design for Cement Stabilized Bases

The first soil–cement road was initially designed based on the guidelines according to the Siam Cement Company, which required a minimum CBR value of 120% for the cement stabilized base. Later, the DOH adopted the unconfined compressive strength (UCS) instead of the CBR. The UCS after 7 days of curing time for cement stabilized lateritic soil and cement stabilized crushed rock shall not be less than 1.7 MPa (17.5 ksc) and 2.4 MPa (24.5 ksc), respectively. It should be noted that the unconfined compression test in the DOH was performed in accordance with the DH-T 105/1972, which is equivalent to ASTM D-1502.

To perform the mix design, the selected marginal highway materials were prepared by mixing with various cement contents at optimum moisture content based on a modified Proctor compaction. In order to avoid the reflective cracks on the asphalt concrete pavement, those selected marginal highway materials shall meet the standard specifications in Table 1. The amount of cement added in each selected highway material is the cement content that shall meet the minimum UCS of the cement stabilized base. To compensate for soil–cement mixing plant efficiency, approximately five to twenty percent of cement content at a minimum UCS value for 7 days of curing time shall be added. Figure 1 illustrates an example of mix design for the cement stabilized bases from two DOH construction projects in Thailand.

Table 1 Standard specifications for highway materials mixed with cement used as base and subbase course
Full size table
Fig. 1
figure 1

Example of soil–cement mix design from two highway construction projects in Thailand (after Apimeteetamrong et al. 2005)

Full size image

1.3 Strength Development of Cement Stabilized Bases

Strength development in concrete material primarily depends on the ratio of the amount of free water to cement, in which it can be referred to the pioneer work by Prof. Abrams in 1918 (a.k.a. Abrams’s law). Minimum amount of free water to cement of 0.25 was required to achieve the completed hydration reaction (Murdock et al. 1991). Extra water was added for the workability purpose, however, the permeability, weathering, and shrinkage increased with the increase in the water to cement ratio.

In the case of a compacted cement stabilized highway material, several investigators also indicated the strength gained with the amount of the cement content for a given amount of water and curing time (Ruenkrairegsa and Sanguandeekul 1977; Ruenkrairergsa 1989; Apimeteetamrong et al. 2005; Horpibulsuk et al. 2007; Sunitsakul and Sawatparnich 2008). The UCS for the coarse-grained soils stabilized with cement was dependent on the water to cement ratio if the samples were prepared on the wet side of compaction (Horpibulsuk et al. 2007). In addition, a unique relationship between UCS and the water to cement ratio was developed in Eq. 1 if samples were prepared at the optimum moisture content (Ruenkrairegsa and Sanguandeekul 1977; Horpibulsuk et al. 2007; Sunitsakul and Sawatparnich 2008).

$$ {\text{UCS}}_{\text{i}} = \frac{\text{A}}{{\left( {{\raise0.7ex\hbox{${\text{W}}$} \!\mathord{\left/ {\vphantom {{\text{W}} {\text{C}}}}\right.\kern-\nulldelimiterspace} \!\lower0.7ex\hbox{${\text{C}}$}}} \right)^{\text{B}} }} $$
(1)

where

A and B:

fitting parameters determined from the statistical analysis

W/C:

water to cement ratio

2 Mix Design Database for Cement Stabilized Bases

A series of mix design data for the cement stabilized bases were collected from various DOH construction projects. Data used in this study were consisted of 520 samples from different sites. All of the marginal highway materials were low-plastic lateritic soils and non-plastic crushed rock. The UCS was plotted against the cement content, water to cement ratio (W/C), dry density, and soaked CBR as shown in Figs. 2, 3, 4, and 5, respectively. Those plots indicated that there existed some trends in UCS-cement content, UCS-W/C, and UCS-soaked CBR relationships.

Fig. 2
figure 2

UCS-cement content relationship

Full size image
Fig. 3
figure 3

UCS-W/C relationship

Full size image
Fig. 4
figure 4

UCS-dry density relationship

Full size image
Fig. 5
figure 5

USC-soaked CBR relationship

Full size image

A soaked CBR was selected to represent the basic properties (i.e., liquid limit, plasticity index, water content, fine content, gradation, activity etc.) of marginal highway materials. The soaked CBR of those marginal materials used is 27–60% for the lateritic soils, 16–96% for the reclaimed highway materials, and 84–99% for the crushed rock base. On the other hand, the cement content and dry density represented the properties of cement stabilized materials. The cement content and dry density of the cement stabilized samples are 1–6% and 2.03–2.37 g/ml (i.e., greater than 95 percent of the maximum dry density based on the modified Proctor compaction), respectively. The corresponding UCS of the cement stabilized bases is ranged from 0.46 to 5.60 MPa. Figure 6 shows a plot of UCS with W/C for a given soaked CBR. Good correlation was observed between the UCS and W/C for a given soaked CBR value. The UCS-W/C relationship gradually moved upward as the soaked CBR increased.

Fig. 6
figure 6

UCS-W/C relationship for a given soaked CBR value

Full size image

3 Proposed Model

3.1 Statistical Analysis

The influencing factors such as W/C ratio, dry density (γDRY), and soaked CBR on UCS were evaluated in the study as shown in Eq. 2. A nonlinear multi-variable regression was performed in order to evaluate the relative effect of such parameters on the UCS of the cement stabilized bases. The ultimate goal of this study was to develop a refined, practice-oriented model that contains minimum numbers of soil parameters.

$$ {\text{UCS}}_{\text{i}} = {\text{ f}}\,({\text{CBR}}_{\text{i}} ,\,{\text{W}}/{\text{C}}_{\text{i}} ,\,\gamma_{\text{DRYi}} ) $$
(2)

where

UCS:

unconfined compressive strength at 7 days of curing time

CBR:

soaked CBR at 95 percent of maximum dry density based on modified Proctor compaction and is measured at 0.1 inch penetration

W/C:

water to cement ratio

γDRY :

dry density

Based on the statistical analysis, three alternative models were obtained as follows:

$$ {\text{UCS}}_{\text{i}} = {\text{A}}\frac{{\left( {\text{CBR}} \right)^{\text{B}} }}{{\left( {\text{W/C}} \right)^{\text{C}} }}\left( {{\upgamma}} \right)^{\text{D}} $$
(3)
$$ {\text{UCS}}_{\text{i}} = {\text{A}}\frac{{\left( {\text{CBR}} \right)^{\text{B}} }}{{\left( {\text{W/C}} \right)^{\text{C}} }} $$
(4)
$$ {\text{UCS}}_{\text{i}} = {\text{A}}\left( {\frac{\text{CBR}}{\text{W/C}}} \right)^{\text{B}} $$
(5)

where

A, B, C, and D:

fitting parameters determined from statistical analyses

3.2 Results of Statistical Analysis

The fitting parameters in Eq. (3) to Eq. (5) are summarized in Table 2. The statistical analysis indicated that the coefficient of determination (R2) obtained from the models in Eq. (3) to Eq. (5) were closed. The model described in Eq. 5 was more desirable because it had less variables and fitting parameters. However, the data contained some outliers as shown in Fig. 7. To exclude those outliers, the attempt was made by plotting the difference between the optimum moisture content (OMC) and the water content of cement stabilized samples against the UCS as shown in Fig. 8. It was found that those outliers exhibited large difference in OMC and water content of cement stabilized samples. The fitting parameters without those outliers are also presented in Table 2 and the new regression was developed in Fig. 9. A developed model clearly indicated the significance of the soaked CBR and W/C ratio on the UCS of cement stabilized bases.

Table 2 Results of statistical analysis
Full size table
Fig. 7
figure 7

UCS model for cement stabilized bases

Full size image
Fig. 8
figure 8

Difference between optimum moisture contents and water content of cement stabilized samples for a given UCS

Full size image
Fig. 9
figure 9

UCS model for cement stabilized bases (excluding outliers)

Full size image

To verify a developed model, the UCS from another series of 91 mix design data collecting from highway construction projects after 2005 were plotted against the UCS estimated from the developed model as shown in Fig. 10.

Fig. 10
figure 10

Comparison between predicted and tested UCS

Full size image

4 Conclusions

The unconfined compressive strength (UCS) model of cement stabilized bases was developed based on a series of mix design data collected from various highway construction sites in Thailand. The dry density of cement stabilized base was found to be insignificant if the mix design requirement can be achieved. The key influencing factors on the UCS of cement stabilized bases are the soaked CBR and W/C.