Recalibration and Recommendation for the Prediction of the California Bearing Ratio (CBR) from Dynamic Cone Penetrometer (DCP)

1. Introduction

The California Bearing Ratio (CBR) is one of the most important characteristics representing the strength of subgrade or base material in pavement structure. In order to conduct the CBR test, samples must be transported, prepared, compacted, soaked, then penetrated with a CBR equipment in the laboratory. Consequently, realistic CBR is difficult to obtain because it takes a long time and is not readily determined in the field. In addition, civil engineers often faced with the urgent need of the CBR of soil in a short amount of time, a survey of large amount of material resources for road construction is a good example. Therefore, in the literature, a number of correlations between CBR and other strength properties of soil were established [1,2,3,4,5,6,7,8,9,10,11,12]. One of the tests that can provides a high reliable correlation with the CBR is the Dynamic Cone Penetration (DCP) test [3,4,13,14,15,16,17,18,19,20]. Notably, the DCP equipment is considered as the compact and lightweight equipment as demonstrated in Figure 1.

The correlations between CBR and DCP for local material proposed by many researchers are usually in the form of log-log or exponential equation as shown in Table 1. Although all equations in Table 1 have the same trend, The big errors were found among these equations. Feleke and Araya (2016) [14] proposed equation for material having low DCP ranging from 1 to 15 mm/blow. Surprisingly, the equation proposed by Fernando et al. (2018) [16] for material having high DCP ranging from 20 to 60 mm/blow was well consistent with Feleke and Araya (2016) [14]. However, these two equations provide much lower value than the test results conducted by Al-Refeai and Al-Suhaibani (1996). This implies that the correlation from local material can be problematic for other material in general engineering practice. Therefore, existing equations are need to be recalibrated and the proper correlations for CBR and DCP for general material are required.

In this paper, the study composed of two steps: (1) evaluation of the existing correlations between CBR and DCP in the literature by four test results from distinct researchers (three existing data from Al-Refeai and Al-Suhaibani (1996), Felek and Araya (2016) and Fernando et al. (2018) and one additional data set from the authors); and (2) derivation of the proper correlations for CBR and DCP that can be used in general material rather than the local material.

Figure 2. Graphs of the existing correlation between California Bearing Ratio (CBR) and Dynamic Cone Penetration Test.

Figure 2. Graphs of the existing correlation between California Bearing Ratio (CBR) and Dynamic Cone Penetration Test.

2. Evaluation of the Existing Equations

The existing equations in Table 1 were investigated by four set of different data set. The results were expressed by the coefficient of determination ($R^2$).

From Table 2, It is obvious that the existing correlations are not able to predict the CBR from DCP test for all material in all locations. Therefore, the new correlations are required.

3. Proposed correlation between CBR and DCP

After we analyzed the data between existing relationships from various studies including experimental results data. We then derived new two correlations for a general proposed as shown in Figure 3 and Equations (1) and (2).

In Figure 3, two new correlations are proposed to gain higher R². The obtained equation for soil with no or less sand and equation for sandy soil are expressed as shown in Equations (1) and (2), respectively.

CBR = 190/(DCP) + 0.53704 (Soil win no or less Sand)

(1)

CBR = 235/(DCP-5.9) + 1.15 (Sandy Soil)

(2)

Equation (1) was validated by the test results of soil from 5 provinces in Thailand (Maha-Sarakham, Khon Kaen, Chaiyaphum, Roi-Et and Kalasin)- the orange triangle symbols. It was found that Equation (1) provides a high value of R² to the test results of soil in Thailand up to 0.904.

4. Conclusions and Discussion

The California Bearing Ratio (CBR) representing the strength of soil subgrade in highway engineering. A high performance to gain the reliable value of CBR still required in the design and the construction processes. However, in some process such as the material surveying of the large amount of subgrade borrow pits is different. The test of the reliable value of CBR can be replaced by the estimation with less time-consuming in this case. The objective becomes the simple test with high predictive capability in correlation. The Dynamic Cone Penetration Test (DCP) is one of the tests that can provide the high R² to CBR when compared to other tests for soil strength. Therefore, in this study we focus on determination of correlations that can predict the CBR from the DCP test. In this study, the evaluation of a number of existing CBR-DCP correlations was performed in first stage. The findings show that no existing correlation provide a high value of R² to all of the published test results. A recalibrated for existing equations was required. After investigation of all test data, two new correlations were proposed in this paper i.e., equation for soil which has no sand of less sand and equation for sandy soil and the R² for these two cases are 0.86 and 0.82, respectively. In addition, equation for soil with less sand was validated by soil from 5 provinces in Thailand and the obtained R² was up to 0.904. This confirmed that calibrated equation can be improve the prediction. It was recommended that a simple visual inspection of engineer is enough in the distinguishing between soil has less sand and sandy soil.