Cbr Valu E.pdf

100

International Journal of Earth Sciences and Engineering ISSN 0974-5904, Volume 04, No 06 SPL, October 2011, pp 100-103

Moisture and Compaction Based Statistical Model for Estimating CBR of Fine Grained Subgrade Soils

Dharamveer Singh

Graduate Research Assistant, School of Civil Engineering and Environmental Science, University of Oklahoma, 202 W. Boyd Street, Room 334, Norman, Oklahoma, USA, 73019, Email: [email protected]

K. S. Reddy Professor, Department of Civil Engineering, Indian Institute of Technology Kharagpur, India, Email: [email protected]

Laxmikant Yadu Assistant Professor, Department of Civil Engineering, National Institute of Technology, Raipur, India, 492010, Email: [email protected]

ABSTRACT: The present study was undertaken to develop regression-based models for estimating soaked and unsoaked California Bearing Ratio (CBR) values for fine-grained subgrade soils. Five locally available soils were collected from different zones of West Bengal. The samples were compacted at four different levels of compaction (i.e., 50, 56, 65, and 75 blows) and at five different levels of moisture contents on dry and wet sides of an optimum moisture content (OMC) of a soil (i.e., ± 2% OMC, ± 1% OMC, and OMC). A total of 100 samples were prepared in the laboratory. Soaked and unsoaked CBR tests were conducted on each sample. Regression models were developed considering different independent parameters namely, index properties of soils, degree of compaction, and moisture content. The models were validated using a soil that was not used in the development phase of the models. Analyses of the results show that the developed models give a reasonable estimate of CBR values. Furthermore, it was observed that variation in the moisture content and compaction efforts has significant effect on the soaked and unsoaked CBR of a soil. KEY WORDS: CBR, Subgrade, OMC, Compaction, Fine-grained soils. INTRODUCTION A proper compaction of subgrade soil is necessary for building long last lasting pavements (Huang, 1993). California Bearing Ratio (CBR) is considered as an indication of strength of subgrade soil. Furthermore, CBR value is used to estimate the resilient modulus of a soil. Recognizing the importance of this test, several prediction models were developed to estimate CBR value of a soil (Black, 1962; Agarwal and Ghanekar, 1970; Kin et al., 2006; NCHRP, 2001). First, Black (1962) correlated CBR with grain size distribution of soil and plasticity index (PI). A graph was developed to estimate CBR value using liquid limit (LL) and PI of a soil. In another study, Agarwal and Ghanekar (1970) developed a CBR model considering optimum moisture content (OMC) and LL of soils. Similarly, the Mechanistic Empirical Pavement Design Guide (NCHRP, 2001) proposed a CBR model based on soil passing on a 75 micron sieve and PI. Recently, Kin et al. (2006) developed a model for Malaysian soils. A total of 65 different soils samples of coarse and fine grained soils were collected. The CBR was corrected with OMC and maximum dry density (MDD) of soils. All above mentioned models are based on the index properties of soil, MDD, and OMC of a soil. So far, limited studies have been conducted to consider the effect of degree of compaction (i.e., no. of blows) and moisture content on CBR.

MATERIAL Five fine grained subgrade soils were collected from different zone of West Bengal (i.e., Kharagpur, Nachipur, Narayangarh, Amarda). Preliminary laboratory tests, such as grain size distribution, LL, and plastic limit (PL) were conducted in accordance with Indian Standard Codes (IS: 2720). Table 1 summarizes index properties and classification of the soils. Five soils: Kharagpur Reddish, Nachipur Reddish, Narayangrah Reddish, Narayangrah Blackish, and Amarda Blackish were classified as CL, CL, CI, CH, and CH, respectively (Table 1). The classification of the soils was done as per Indian standard soil classification system (IS: 1498-1970). The PI of soils varied from 14% to 44%. The moisture-density relationship for each soil was determined in accordance with IS 2720-Part 8: 1983 (Table 1). OMC and MDD values of the soils vary from 7.8% to 15.5%; and from 17.22 to 20.90 kN/m3, respectively. Table 1 Properties of Different Types of Soils Soil Sample CL-1 CL-2 CI

OBJECTIVE The main objective of the present study was to develop regression-based models for estimating CBR value of fine grained subgrade soils, considering degree of compaction, moisture content, and various index properties of a soil.

CH-1 CH-2

Soil Source Kharagpur Reddish Nachipur Reddish Narayangarh Reddish Narayangarh Blackish Amarda Blackish

LL (%)

PI (%)

Soil Type

OMC (%)

MDD (kN/m3)

26

14

CL

7.8

20.90

33

19

CL

11.2

18.98

46

34

CI

15.5

17.62

65

44

CH

16.0

17.22

55

36

CH

16.5

17.49

#020410125 Copyright © 2011 CAFET-INNOVA TECHNICAL SOCIETY. All rights reserved

101

Moisture and Compaction Based Statistical Model for Estimating CBR of Fine Grained Subgrade Soils

The regression models developed for unsoaked CBR and soaked CBR are shown in Equation (1) and Equation (2), respectively. The model developed for unsoaked CBR has correlation coefficient (R2) = 0.70, indicating a reasonable fit to the data. A fair correlation was obtained for a model developed for soaked CBR (R2 = 0.48). Unsoaked CBR Model  MC   Density  × 100  + 0.239 ×  × 100  − 2.004 × PL UCBR = 104.71 − 0.671 ×   OMC   MDD 

(1)

(R2=0.70)

where, UCBR = Unsoaked CBR (%), MC = Moisture Content (%), OMC = Optimum Moisture Content (%), Density = Measured or calculated density (gm/cc), MDD = Maximum Dry Density (gm/cc), and PL = Plastic Limit of soil (%).

The developed models were further checked for all the soils. Figure 3 shows the relationship between the predicted and the measured unsoaked CBR values (R2 = 0.750). Similarly, Figure 4 depicts the graph of the predicted and the measured soaked CBR values for all the soils (R2 = 0.60). 50

y = 1.036 x R2 = 0.91 40

30

20

10

0 0

10

20

30

40

50

Measured UCBR (%)

Fig. 1 The Measured and the Predicted Unsoaked CBR value for CH-1 Soil 6

y = 0.40 x R2 = 0.70

5

4

3

2

1

0 0

2

4

6

8

10

12

Measured SCBR (%)

Soaked CBR Model  Density   MC  SCBR = −2.213 − 0.055 ×  × 100  − 1.147 × PL × 100  + 0.328 ×   MDD   OMC 

VALIDATION OF MODELS The developed models (Equations (1) and (2)) were validated using CH-1 soil (Table 1). This soil was not used in the development phase of the models. In addition, models were checked using a combined database of all the soils. Figure 1 shows the graph between the predicted and the measured unsoaked CBR values (R2= 0.910). Similarly, Figure 2 shows the plot of the measured and the predicted soaked CBR values (R2=0.70). The model predictions show a good agreement with the measured CBR values.

Predicted UCBR (%)

DEVELOPMENT OF MODELS Soaked and unsoaked CBR models were developed considering various independent variables: OMC, MC, Density, MDD, PL, LL, and PI. A commercial software, called Statistical Package for the Social Science (SPSS) was used to develop the models. A total of four soils (i.e. CL-1, CL-2, CI, and CH-2) (Table 1) were used to develop the models. The outliers from the data were eliminated using a box plot method. The presence of collinearity among the independent variables was measured using Pearson’s correlation (Rahim, 2005). It was found that percentage passing on a 75 micron sieve is highly correlated with LL (Pearson correlation = 0.515, p<0.001). Similarly, LL and PL are highly correlated (Pearson correlation = 0.846, p<0.001). Therefore, only PL of the soils was considered for developing the regression models. The MC and density values were normalized by dividing them with OMC and MDD, respectively.

where, SCBR = Soaked CBR (%).

Predicted SCBR (%)

SPECIMEN PREPARATION AND TESTING The modified proctor procedure was used to prepare CBR samples. The compaction efforts used in present study were: 50, 56, 65, and 75 blows, while the moisture levels were ± 2% OMC, ± 1% OMC, and OMC. Each CBR sample was compacted in five layers of soil using predetermined number of blows and water content. A total of 100 soil samples were prepared for CBR testing (20 samples each soil x 5 soils). The dry density of the compacted sample was estimated using bulk density and moisture content of the sample. The degree of compaction was calculated as ratio of measured dry density and MDD. Furthermore, actual moisture content (MC) of each compacted sample was determined to check any variability from the pre-determined water content. Unsoaked and soaked CBR values of the compacted sample was determined in accordance with IS 2720-Part 16:1987.

(2)

Fig. 2 The Measured and the Predicted Soaked CBR values for CH-1 Soil

(R2=0.48)

International Journal of Earth Sciences and Engineering ISSN 0974-5904, Volume 04, No 06 SPL, October 2011, pp 100-103

102

Dharamveer Singh, K. S. Reddy, Laxmikant Yadu

80

Effect of Degree of Compaction

y = 0.99 x R2 = 0.75

At constant moisture content, as compaction effort increases the unsoaked CBR values also increases (Figure 5). For example, at 90% of OMC (dry side of OMC), increase in compaction effort from 80% to 90% would result in 6% increase in unsoaked CBR value. Similarly, at 110% of OMC (wet side of OMC), increase in degree of compaction from 80% to 100%, resulted in approximately 18% increase in unsoaked CBR (Figure 5).

Predicted UCBR (%)

60

40

20

0 0

20

40

60

70

80

Measured UCBR (%)

80 % MDD 90 % MDD 100 % MDD

60

Fig. 3 The Measured and the Predicted Unsoaked CBR Values for all Soils UCBR (%)

50

20 y = 1.01 x R2 = 0.60

40 30 20

Predicted SCBR (%)

15

10 0

10

70

80

90

100

110

120

(MC/OMC) (%)

Fig. 5 Variation of Unsoaked CBR value with Compaction and Moisture Content for CH-1 Soil

5

16

0 0

5

10

15

Measured SCBR (%)

VARIATION OF CBR WITH MOISTURE AND COMPACTION: The developed models (Equations (1) and (2)) were used to evaluate the effect of moisture content and the degree of compaction on CBR. For brevity, only soil Narayangarh Blackish (i.e., CH-1) (Table 1) soil was selected. Unsoaked CBR Effect of Moisture Content It can be seen from Figure 5 that as moisture content increases the unsoaked CBR value of soils decreases. For example, at a constant degree of compaction say 90%, sample prepared at 90% of OMC (dry side of OMC) resulted approximately 19% higher unsoaked CBR compared to sample compacted at OMC. Similarly, if soil is compacted at 120% of OMC (wet side of OMC), the unsoaked CBR value decreases approximately 38% compared to CBR value at OMC.

12 SCBR (%)

Fig. 4 The Measured and the Predicted Soaked CBR Values for all Soils

80 % MDD 90 % MDD 100 % MDD

14

20

10 8 6 4 2 0 70

80

90

100

110

120

(MC/OMC) (%)

Fig. 6 Variation of Soaked CBR value with Compaction Moisture Content for CH-1 Soil Soaked CBR Effect of Moisture Content Figure 6 shows the effect of moisture and the degree of compaction on soaked CBR. It is evident that as moisture content increases, the soaked CBR values of soil decreases for each level of compaction. For example, at a constant degree of compaction say 90%, a sample prepared at 90% of OMC (dry side of OMC) resulted approximately 6.8% higher soaked CBR compared to soil compacted at OMC. Similarly, if a sample is compacted at 120% of OMC (wet side of OMC), the soaked CBR value

International Journal of Earth Sciences and Engineering ISSN 0974-5904, Volume 04, No 06 SPL, October 2011, pp 100-103

Moisture and Compaction Based Statistical Model for Estimating CBR of Fine Grained Subgrade Soils decreases approximately 13.6% compared to soaked CBR value at OMC (Figure 6). Effect of Degree of Compaction At constant moisture content, as compaction effort increases, the soaked CBR values also increases (Figure 6). For example, at 90% of OMC (dry side of OMC), increase in compaction effort from 80% to 90% would result in 62% increase in soaked CBR value. Similarly, at 110% of OMC (wet side of OMC), an increase in degree of compaction from 80% to 100%, would result approximately 156% increase in soaked CBR (Figure 6). Effect of compaction on soaked CBR is more dominant compared to unsoaked CBR. CONCLUDING REMARKS The present study was undertaken to develop regressionbased models to estimate soaked and unsoaked CBR values of fine grained subgrade soils. A total of 100 samples were tested for soaked and unsoaked CBR values for five different soils. Regression-based models were developed and validated. A good agreement was observed between the measured and the predicted CBR values. Furthermore, it was observed that the CBR value, both soaked and unsoaked significantly affected by change in moisture content and compaction effort. The effect of both moisture and compaction effort is more significant on the soaked CBR value. It is recommended that the developed regression models be validated on large range of soils.

103

REFERENCES [1] Agarwal, K.B. and Ghanekar, K.D. (1970). Prediction of CBR from Plasticity Characteristics of Soil. Proceeding of 2nd South-east Asian Conference on Soil Engineering, Singapore, June 11-15, 1970, Bangkok: Asian Institute of Technology, 571-576. [2] Black, W.P.M. (1962). A Method of Estimating the CBR of Cohesive Soils from Plasticity Data. Geotechnique, Vol. 12: 271-272. [3] Huang, Y.H., Pavement Analysis and Design, 1993 (Prentice-Hill, Inc. New Jersey). [4] Kin, M.W., California Bearing Ratio Correlation With Soil Index Properties. Master of Engineering Thesis, University of Technology, Malaysia, 2006. [5] National Cooperative Highway Research Program (2001). Guide for Mechanistic and Empirical- Design for New and Rehabilitated Pavement Structures. [6] Rahim, A.M. Subgrade soil index properties to estimate resilient modulus for pavement design. The International Journal of Pavement Engineering, Vol. 6, No. 3, September 2005, 163-169.

International Journal of Earth Sciences and Engineering ISSN 0974-5904, Volume 04, No 06 SPL, October 2011, pp 100-103