![]() |
![]() |
A Laboratory Study to Establish
Department. of Civil Engineering, University of Peradeniya, and Senior Lecturer, Department of Civil Engineering, University of Peradeniya, |
Abstract
This paper explains a laboratory investigation carried out to develop some useful relationships for the use of Dynamic Cone Penetrometer (DCP) for road subgrade evaluation.
Series of laboratory tests were performed by varying the moisture content and the dry density. Clayey or silty gravel (GF) was used for the study. Experimental procedure, statistical analysis of the results and the equations developed are presented in this paper.
KEYWORDS: Dynamic Cone Penetrometer, Soaked and Unsoaked CBR, Moisture content, Dry density
Introduction
Dynamic Cone Penetrometer (DCP) is an instrument, which can be used to evaluate California Bearing Ratio (CBR) value of soil. The DCP has many advantages over the traditional CBR test. It is an insitu test, simple to use and inexpensive. Hence it is possible to introduce this instrument to local road authorities that deal with rural road construction and maintenance work. It is very rarely that a local authority in a developing country like Sri Lanka, evaluates the subgrade properties in a road construction project. They do not have facilities to evaluate the existing pavements in maintenance work or designing of overlay thickness. However, the DCP, which can also be produced by local authorities for themselves, with very low cost, can be introduced for those rural road projects as an effective road evaluation tool.
This study was conducted in two phases. In the first phase, three correlations between the DCP and CBR were developed for some selected soil types of Sri Lankan residual soils. The second phase consists of a series of laboratory tests. The variation of DCP with the moisture content and the dry density were studied. Relationships between the DCP and soaked CBR and other important soil parameters were developed.
The dynamic cone penetrometer
A.J. Scala originally developed the DCP in 1956 in Australia. After that various researchers developed both the testing instrument and the testing procedure. The relationships were developed between DCP and CBR, shear strength, soil index properties etc..
There are various types of DCPs available in the world. They are operated on the same principle. A DCP made locally to the specifications of the Transportation and Road Research Laboratory (TRRL) DCP was used for this study. It consists of an 8 kg weight dropping through a height of 575 mm and a 600cone having a base diameter of 20 mm as shown in Figure 1. The penetration of the cone is measured using a calibrated scale. It is possible to measure up to 800 mm depth without an extension rod and up to 1200 mm depth when fitted with an extension rod. It needs three operators, one to hold the instrument, one to raise and drop the weight and the other to record the penetration.
Figure 1. The Dynamic Cone Penetrometer
THE PREVIOUS WORK
The previous research work under this study was focused to develop a correlation between;
· The DCP and disturbed unsoaked CBR (DUCBR)
· The DCP and disturbed soaked CBR (DSCBR)
· The DCP and undisturbed unsoaked CBR (UUCBR)
for clayey or silty sand.
Tests were carried out for two rural road projects under the Central Provincial Council and the Peradeniya Engineering Faculty premises. The DCP tests were carried out at the selected sections. Soil samples were also collected from the same locations for laboratory tests. Undisturbed Unsoaked CBR, Disturbed Unsoaked CBR, Disturbed Soaked CBR, Moisture Content test, Particle Size Distribution test and Compaction test were carried out as laboratory tests.
The disturbed CBR values were obtained by testing remoulded samples at the optimum moisture content and the maximum dry density, which are the values used for road design purposes. Undisturbed unsoaked CBR values were obtained by testing a sample, which was extracted directly to a standard CBR mould from the field.
Regression analysis of the results shows that there is an inverse relationship between the DCP and the CBR for the tested soils. The data was analyzed with the Linear, logarithmic, Exponential and Power (Log/Log) models. Out of the four models, power model is the best model to describe this relationship. Three correlations were established between the DCP and the disturbed unsoaked CBR, the DCP and the Disturbed soaked CBR and the DCP and the Undisturbed Unsoaked CBR for clayey or silty sand. The equations derived are
1. Log CBR = 2.182 – 0.872 Log DCP For DCP/DUCBR,
2. Log CBR = 1.145 – 0.336 Log DCP For DCP/UUCBR
3. Log CBR = 1.671 – 0.577 Log DCP For DCP/DSCBR.
Where DCPis in mm/blow. The data limits of the equation derived are 7 mm/blow <DCP < 75 mm/blow.
OBJECTIVES
1. To find out the relationships between the unsoaked CBR and the DCP and soaked CBR and the DCP for evaluating existing pavement subgrade.
2. To develop relationships between the DCP and the other important soil parameters for road subgrade evaluation.
Hence to widen the usage of the DCP in the field of subgrade evaluation.
SAMPLE PREPARATION
The DCP test was carried out in a mould with a diameter of 280 mm. Samples were prepared by combining gravel, sand and fine to pre-decided proportions. First the particles were separated in to three main groups.
· Gravel (20 – 2 mm),
· Sand (2 – 0.06 mm) and
· Fine (< 0.06 mm).
The limitation of gavel for 20 mm was due to the CBR test. Those particles were remixed with five different proportions to make five different soil samples, which fall in to very clayey or silty gavel (GF) according to the British Soil Classification System (BSCS). Details of the samples are shown in the Table 1.
Table 1. Details of Samples Used in Laboratory Tests
![]() | ||||
Sample | Gravel / [%] | Sand / [%] | Fine / [%] | Soil Type |
![]() | ||||
ST1 | 50 | 15 | 35 | GF |
ST2 | 55 | 20 | 25 | GF |
ST3 | 60 | 20 | 20 | GF |
ST4 | 65 | 15 | 20 | GF |
ST5 | 55 | 15 | 30 | GF |
![]() |
EXPERIMENTAL PROCEDURE
Standard proctor compaction test was carried out for each soil sample according to the BS 1377; Part 4; 1990 specifications, to find out the relationship between the dry density and the moisture content of samples. Relationships between the moisture content and the dry density were obtained with relevant equation. The graph of the moisture content vs. the dry density for soil sample ST1 is shown in Figure 2.
Figure 2. Compaction Test Results of Sample ST1
DCP and CBR tests were carried out by varying the moisture content and the dry density of the sample. About five numbers of tests were carried out for each soil type by varying the moisture content and the dry density.
Five points from the graph of the dry density vs. moisture content was selected, including the maximum dry density and the optimum moisture content. The points were selected starting from about 6.0 % of the moisture contents, which was being increased at 3.0 % intervals. The relevant dry density was calculated using the equation represents the moisture content dry density relationship. Bulk density was calculated using the equation;
gb = gd (1 + m)
Where gb, gd and m are the bulk density, dry density and the moisture content respectively.
The selected moisture contents, dry densities and bulk densities for the sample ST1 are shown in Table 2.
Table 2. Basic properties of ST1
![]() | |||
No | Moisture Content [%] | Dry Density [kg/m3] | Bulk Density [kg/m3] |
![]() | |||
1 | 8.0 | 1674.9 | 1808.8 |
2 | 11.0 | 1740.2 | 1931.6 |
3 | 14.0 | 1779.0 | 2028.1 |
4 | 16.9 | 1791.5 | 2094.3 |
5 | 19.0 | 1785.1 | 2124.3 |
![]() |
The total weight for a single DCP test was calculated considering the bulk density and the volume of the DCP mould (Weight = Bulk Density ´ Volume of the Mould). Before preparing the sample, the moisture content of the bulk sample was obtained by the difference of the moisture content between the bulk sample and the oven-dried sample. The amount of water that should be added was decided considering the target moisture content. The required amount of water was added to the pre weighed sample and mixed thoroughly. Extreme care was taken during the mixing process to ensure a uniform mixture. The prepared sample was well covered and left for 24 hrs.
Then the sample was compacted in five layers. A surcharge load that gives the same intensity as in CBR test (147.2 kg/m3) was applied on the top of the mould. A schematic diagram of the sample with the surcharge load is shown in Figure 3 illustrate a compacted sample with the surcharge load. The DCP was placed over the hole of the surcharge load, which lies at the center of the mould. The initial value of the DCP was taken and the test was carried out until it reached 20 – 30 mm above the mould base. The DCP was withdrawn and the samples were taken from the top, middle and bottom of the mould for moisture content test. For each stage, two similar tests were carried out and the average of the two tests was taken. A test being carried out in a mould is shown in Figure 4.
Figure 3. A Schematic Diagram of Compacted Sample with Surcharge Load
Figure 4. A DCP Test Being Carried Out in a Mould
The CBR test was carried out in the standard CBR mould. Samples were prepared in a similar way as described above. The sample size was decided according to the CBR mould size. It was compacted in three layers by manually to achieve the relevant bulk density. Two similar samples were prepared: one for the unsoaked CBR test and the other for the soaked CBR test. The soaked CBR test was carried out after a four-day soak. Any way the soaked CBR could not be carried out for the first step of the every soil sample as they were failed when taken out from the water bath due to low density.
TEST RESULTS
About 23 sets of tests were carried out for the prepared soil samples. The soil type is clayey or silty gravel. The moisture content, the dry density relationship of each DCP test and the CBR test was compared with the compacted test results. The percentage difference of the dry density for the DCP test and compaction test for ST1 sample lies in between 0.65 and 0.028 where as the same for the CBR test and the compaction test is for 0.19 and 0.43. The percentage difference of the dry density for the DCP test and the compaction test for all the test series lies in between 1.705 and 0.001 and for the CBR test and the compaction test 0.676 and 0.0129. The comparison of tested conditions of the DCP and the CBR tests with the compaction test for sample ST1 is shown in Table 3 and Figure 5.
Table 3. Comparison of Compaction for Sample ST1
![]() | |||||||
No | MC [%] | Dry Density / [kg/m3] | Percentage Difference | CBR | |||
Compaction | DCP | MC [%] | DD [kg/m3] | Percentage Difference | |||
![]() | |||||||
1 | 7.9 | 1672.3 | 1683.3 | -0.6592 | 8.1 | 1675.6 | -0.1987 |
2 | 11.5 | 1748.5 | 1758.5 | -0.5730 | 11.1 | 1740.8 | 0.4393 |
3 | 13.7 | 1776.3 | 1784.0 | -0.4314 | 14.1 | 1775.4 | 0.0527 |
4 | 16.1 | 1790.5 | 1797.7 | -0.4004 | 16.7 | 1790.3 | 0.0129 |
5 | 19.1 | 1784.5 | 1784.0 | 0.0288 | 19.1 | 1786.8 | -0.1281 |
![]() |
MC – Moisture Content DD – Dry Density
Figure 5. Comparison of Tested Conditions for Sample ST1
DATA ANALYSIS
The simple linear regression and the multiple regression methods were used to analyze the data obtained. The DCP value was used as the independent variable and CBR value, moisture content, dry density, compaction level and MC/OMC were used as dependent variables. The Mean Square Error (MSE) and the Coefficient of Determination (R2) were used to find out the goodness of fit. Mean Square Error (MSE), which is a measure of how well the model fits the data, should be minimum. Coefficient of Determination (R2) is a standardized measure of the goodness of fit and should be highest. The coefficient of determination (R2) ranges from 0 to 1. If R2 is greater than 0.5, the determination is considered as acceptable.
VARIATION OF UNSOAKED CBR WITH DCP
The laboratory investigation, found a strong relationship between the DCP and the unsoaked CBR. Four models were analyzed including the linear, logarithmic, exponential and power (Log/Log) model. Out of the four models the best-fit model was the power model. It has the smallest MSE and the highest coefficient of determination (R2). The relationship obtained is
Log CBRU = 1.966 – 0.667 Log DCP R2 = 0.95, MSE = 0.112, N = 23 | (1) |
Where CBRU is the unsoaked CBR value, and Nis the number of tests. DCP is in mm/blow. The data limit of the equation derived are 11 mm/blow < DCP < 386 mm/blow.
This equation can be used to evaluate the unsoaked CBR value of an existing subgrade. The graph of the variation of unsoaked CBR with the DCP is shown in Figure 6.
Figure 6. Variation of Unsoaked CBR with DCP
VARIATION OF SOAKED CBR WITH DCP
The analysis of the test results show that a clear relationship between the soaked CBR and the DCP value could not be defined as for the unsoaked CBR and the DCP. The graph of the variation of the unsoaked CBR and the Soaked CBR with the initial moisture content of the samples showed that the difference between the two values decreased with the test number. The graph of the variation of the unsoaked CBR and the Soaked CBR with the initial moisture content for ST1 is shown in Figure 7.
Figure 7. Variation of Unsoaked and Soaked CBR
with the Initial Moisture Content of the Sample
As the DCPincreased with the moisture content, it was used to develop a relationship between the DCPand the soaked CBR value. It was found that there is a significant correlation between the DCP and the difference between the unsoaked and the soaked CBR value (CBRU - CBRS). The best-fit model was the logarithmic model. The relationship is
CBRU - CBRS = 25.611 – 11.5 Log DCP R2 = 0.89, MSE = 2.58, N = 11 | (2) |
Where CBRS is the soaked CBR value. The data limit that the equation derived is 11 mm/blow <DCP < 104 mm/blow.
Using the Equation 1, it is possible to obtain the relevant unsoaked CBR value (CBRU). Then the soaked CBR value can be calculated based on the predicted unsoaked CBR value. Figure 8 shows the variation of (CBRU - CBRS) with the DCP.
Figure 8. Variation of (CBRu - CBRs) with the DCP
A better correlation was obtained considering the effects of the moisture content.
CBRU - CBRS = 67.12 – 1.48MC – 30.64DCP1/MC R2 = 0.91, N = 11 | (3) |
Where MC is the moisture content. The data limit that the equation derived is 11 mm/blow <DCP < 104 mm/blow.
The 3D surface graph of the variation of (CBRU - CBRS) with the DCP and the MC is shown in Figure 9.
Figure 9. Variation of (CBRu – CBRs) with the DCP and Moisture Content
The Equation 2 is advantageous, as it needs only the DCP test results to obtain the soaked CBR value. However the Equation 3 needs both the DCP and the moisture content test results in order to obtain the soaked CBR value.
VARIATION OF MOISTURE CONTENT WITH DCP
The DCP value increases with the moisture content. A significant correlation exists between the DCP and the moisture content. Out of the linear, logarithmic, exponential and power models, the logarithmic model is the best-fit model.
MC= 0.5 + 6.9 LogDCP R2 = 0.71, MSE= 4.510, N = 23 | (4) |
The data limit of the equation derived are 11 mm/blow <DCP < 386 mm/blow.
If the DCP value of a point is known, the moisture content of that point can be predicted using the Equation 4. Anyway it was observed that the accuracy of the DCP value is very low when the moisture content is high. Figure 10 shows the variation of moisture content with the DCP.
Figure 10. Variation of Moisture Content with DCP
VARIATION OF DRY DENSITY WITH DCP
Analysis of the test results shows the relationship between the DCP and the dry density is unacceptable. However, including the effects of moisture content it was possible to develop a significant correlation between the dry density and the DCP. The equation obtained is
DD = 1940.75 - 1783.34[1/(1 + MC)] – 0.058DC R2 = 0.77, N = 23 | (5) |
Where DD is the Dry Density in kg/m3. The data limit of the equation derived are 11 mm/blow <DCP < 386 mm/blow. 5 <MC < 20.
This can be used to evaluate an existing subgrade. And also it can be used as a quality control test at construction level. The 3D surface graph of the variation of the dry density with the DCP and the moisture content is shown in Figure 11.
Figure 11. Variation of Dry Density with DCP and Moisture Content
VARIATION OF DCP, MOISTURE CONTENT AND COMPACTION LEVEL (DD/MDD)
The analysis of the laboratory experiment results indicated that a significant correlation between the DCP and the compaction level could be obtained by taking the effect of the moisture content in to account.
DD/MDD = 1.126 + 0.064MC + 0.126DCP1/MC R2 = 0.87, N = 23 | (6) |
Where MDD is the Maximum Dry Density in kg/m3. The data limit of the equation derived are 11 mm/blow <DCP < 386 mm/blow. 5 <MC < 20.
The compaction level of a point can be obtained, if the moisture content and the DCP value are known. This also can be used as a quality control test at a construction level. Figure 12 shows the 3D surface graph of the variation of DD/MDD with the DCP and the moisture content.
Figure 12. Variation of DD/MDD with the DCP and Moisture Content
VARIATION OF DCP, MOISTURE CONTENT AND OPTIMUM MOISTURE CONTENT
The multiple regression analysis shows that there is a good correlation between the DCP, moisture content and the optimum moisture content.
MC/OMC = -0.186 + 0.064MC + 0.126DCP1/MC R2 = 0.97, N = 23 | (7) |
Where OMC is the Optimum Moisture Content. The data limit of the equation derived are 11 mm/blow <DCP < 386 mm/blow.5< MC < 20.
The 3D surface graph of the variation of MC/OMC with DCP and moisture content is shown in Figure 13.
Figure 13. Variation of MC/OMC with DCP and Moisture Content
COMPUTER PROGRAM TO CALCULATE THE PARAMETERS
A Visual Basic (VB) program was prepared to calculate the UCBR, SCBR, moisture content, dry density, DD/MDD and MC/OMC. The main form includes the instructions. The program allows the user to select the parameter. Then it displays the relevant equation. Some equations need both the DCP and the moisture content to calculate the parameter. In such a case both the DCP and the MC text boxes are appearing. When it needs the DCP only for the calculations it shows only the DCP text box. It also facilitates the user to take the print out of the form if necessary. The main form and the calculation forms are shown in the Figures 14(a) and 14(b).
Figure 14(a). The Main Form of the VB Program to Calculate CBR and Other Soil Parameters
Figure 14(b). The Calculation Form of the VB Program to Calculate CBR and Other Soil Parameters
CONCLUDING REMARKS
It was found that there is an inverse relationship between the DCP and the CBR. The best-fit model is the power (Log/Log) model. It was also shown that there is an inverse relationship between the DCP and (CBRU – CBRS).
In addition to the correlations between the DCP and the CBR, significant relationships between DCP/Moisture content, DCP/Dry density/Moisture content, DCP/Moisture content/Dry density/Maximum dry density and DCP/Moisture content/Optimum moisture content were also developed. The soil type used is very clayey or silty gravel. The results of the laboratory investigation indicated that the DCP could be used as an effective tool in road construction and maintenance process.
As the correlations were derived using one soil type, further research should be carried out using wide range of soil types with various strength levels.
ACKNOWLEDGEMENTS
The two funds provided by the Asian Development Bank (ADB) and the National Science Foundation (NSF – Grant No: RG/2001/E/05) and the assistance provided by the Central Provincial Highway Department to make this research a success are gratefully acknowledged.
REFERENCES
![]() | |
© 2002 ejge |