Soil classification based on cone penetration test (CPT) data in Western Central Java

Arwan Apriyono, Yanto, Purwanto Bekti Santoso, and Sumiyanto

Citation: AIP Conference Proceedings 1941, 020004 (2018); doi: 10.1063/1.5028062

View online: https://doi.org/10.1063/1.5028062
View Table of Contents: http://aip.scitation.org/toc/apc/1941/1
Published by the American Institute of Physics

Articles you may be interested in

Possibility of wax control techniques in Indonesian oil fields
AIP Conference Proceedings 1941, 020001 (2018); 10.1063/1.5028059

Assessment of shrimp farming impact on groundwater quality using analytical hierarchy process
AIP Conference Proceedings 1941, 020003 (2018); 10.1063/1.5028061

Analysis of maizena drying system using temperature control based fuzzy logic method
AIP Conference Proceedings 1941, 020005 (2018); 10.1063/1.5028063

Production, purification and utilization of biogas as fuel for internal combustion engine
AIP Conference Proceedings 1941, 020009 (2018); 10.1063/1.5028067

Competency test result of vocational school teacher’s majoring light vehicles subject in East Jakarta
AIP Conference Proceedings 1941, 020006 (2018); 10.1063/1.5028064

The influence of activities and nutrition status to university students’ achievements

AIP Conference Proceedings 1941, 020008 (2018); 10.1063/1.5028066
 Soil Classification Based on Cone Penetration Test (CPT)
Data in Western Central Java
Arwan Apriyonoa), Yanto, Purwanto Bekti Santoso and Sumiyanto

Civil Engineering Department, Jenderal Soedirman University

a)
corresponding author: arwan.apriyono@unsoed.ac.id; arwan.unsoed@gmail.com

Abstract. This study presents a modified friction ratio range for soil classification i.e. gravel, sand, silt & clay and peat,
using CPT data in Western Central Java. The CPT data was obtained solely from Soil Mechanic Laboratory of Jenderal
Soedirman University that covers more than 300 sites within the study area. About 197 data were produced from data
filtering process. IDW method was employed to interpolated friction ratio values in a regular grid point for soil
classification map generation. Soil classification map was generated and presented using QGIS software. In addition, soil
classification map with respect to modified friction ratio range was validated using 10% of total measurements. The
result shows that silt and clay dominate soil type in the study area, which is in agreement with two popular methods
namely Begemann and Vos. However, the modified friction ratio range produces 85% similarity with laboratory
measurements whereby Begemann and Vos method yields 70% similarity. In addition, modified friction ratio range can
effectively distinguish fine and coarse grains, thus useful for soil classification and subsequently for landslide analysis.
Therefore, modified friction ratio range proposed in this study can be used to identify soil type for mountainous tropical
region.

INTRODUCTION
Many places in Indonesia are landslide prone area resulting in numerous severe landslide hazards. In recent
decade, landslide events have caused 141 fatalities and 112 injured. Western Central Java is one of the most
vulnerable area to landslide hazard in Indonesia1. Due to large number of population in this region, early
identification of landslide prone zone is important.
Landslide mechanism is associated with basic mechanical properties of soil by which each soil type has its
unique characteristic that directly connected to soil strength, key variable in landslide analysis2. One of important
soil mechanical properties is friction ratio (Rf), defined as ratio between sleeve friction and cone resistance. Both
sleeve friction and cone resistance can be gained from Cone Penetration Test (CPT). However, friction ratio data are
publicly limited in term of number and access. Hence, finding soil feature that easily obtained and has correlation
with friction ratio is potentially useful for landslide analysis.
Soils classification is widely known across field studies. Many studies have been aimed to determine empirical
correlation between soil classification and CPT data3,4,5,6. Previous researcher3,4,5 have recently developed empirical
equation to analyze soils classification using CPT data in Sudan, India and Sweden respectively. In addition,
Begemann6 found correlation between sleeve friction, cone resistance and soil type. However, most of the studies
concerning in this subject were conducted in subtropical regions. The number of study on this matter in tropical
region is limited or even none, particularly in Indonesia.
This study presents correlation analysis of CPT data and soil classification in Western Central Java. In particular,
using existing correlation of Frictio Ratio (Rf) value and soil type, a map of soil type is created from Rf value that
interpolated using Inverse Distance Weighting (IDW) method. A new Rf range for each soil type is proposed to
yield nearly accurate soil type distribution. Once the soil type validated, it will be very useful for landslide analysis.

Engineering International Conference (EIC2017)

AIP Conf. Proc. 1941, 020004-1–020004-10; https://doi.org/10.1063/1.5028062
Published by AIP Publishing. 978-0-7354-1633-8/$30.00

020004-1
 SOIL CLASSIFICATIONS BASED ON CPT
Study on correlation of soils classification and CPT data can be traced from a study by Begemann6 in 1965.
Begemann reported that the soil type is not a sole function of either cone resistance or sleeve friction. Instead, it is a
function of combination of these values. The correlation between sleeve friction, cone resistance and soils type
suggested by Begemann is shown in FIGURE 1.
.

FIGURE 1. Correlation between sleeve friction, cone resistance and soil type6

As can be seen in FIGURE 1, soil type is a function of ratio between sleeve friction and cone resistance, which
is called friction ratio (Rf). In addition, Begemann4 suggested that each soil type has a range of Rf as presented in
TABLE 1.
TABLE 1. Correlation between Rf and soil type6.
No Soil Type Friction Ratio
1 Coarse sand with gravel through fine sand 1.2 % - 1.6 %
2 Silty sand 1.6 % - 2.2 %
3 Silty sandy clayey soils 2.2 % - 3.2 %
4 Clay and Loam, and loam soils 3.2 % - 4.1 %
5 Clay 4.1 % - 7.0 %
6 Peat >7%

In 1980’s, electrical CPT was proposed for soil profiling. Application of electrical CPT on soil profiling resulted
in more accuracy on soil parameter due to the existence of automatic sensor embedded in the equipment, the
accuracy of soil parameter increased. Using electrical CPT, Douglas and Olsen7 suggested unified soil classification
system to the soil type zones as shown in Fig. 2.

020004-2
 FIGURE 2. Correlation between friction ratio, cone resistance and soils classification Douglas and Olsen7.

In addition to Douglas and Olsen7, Vos8 suggested Rf range for soil type based on Dutch soils as presented in
TABLE 2. The percentage values are similar but not identical to those recommended by Begemann6.
TABLE 2. Correlation between FR and soil type in Dutch Vos8.
No Soil Type Friction Ratio
1 Coarse sand and gravel < 0.5 %
2 Fine sand 1.0 % - 1.5 %
3 Silt 1.5 % - 3.0 %
4 Clay 3.0 % - 7.0 %
5 Peat >7%

In early 1990s, CPT was modified by adding piezometer named piezocone penetrometer test (CPTu). In addition
to CPTu, pore water pressure value was also considered to develop soil profiling chart. Robertson9 is the first to
present a chart based on the piezocone with the cone resistance corrected for pore pressure. Moreover, Fellenius and
Eslami10 proposed correlation between sleeve friction, effective cone resistance and soil type as shown in Fig. 3.

020004-3
 FIGURE 3. correlation between sleeve friction, effective cone resistance and soil type10.

DIGITAL GEOTECHNICAL MAP

By nature, soil mechanical properties and soil types have spatial variability2. Hence, understanding spatial
distribution of correlation between soil type and soil mechanical properties is helpful for landslide analysis. To do
this, geotechnical map is useful tool to visually describe spatial variation of correlation between soil mechanical
properties and soil type. Rapid increase of computer technology has stimulated the development of digital
geotechnical map. The most popular computer software to present spatial data is ArcGIS software developed by
Environmental System Research Institute (ESRI). Given spatial data (usually consists of latitude, longitude and
corresponding values), ArcGIS can be easily used to produce thematic map such as flood risk zone map, landslide
prone area map, etc. This tool has been employed to visualize spatial distribution of natural resources in the specific
area 11,12,13.

METHODOLOGY

Data Collecting and Filtering

Two types of data are required in this study; i) CPT data and ii) soil type data. The data was collected from Civil
Engineering Laboratory UNSOED. Roughly over 300 data were collected in the last ten years. Most of the data
describe CPT and soil type data of Western Central Java which covers 10 districts in this region as shown in Fig. 4.
However, some data are missing important variables or do not reasonably depict common feature of regional soil
properties. To overcome this situation, we filtered the data such as only reasonable and complete data are used.

020004-4
 FIGURE 4. Area of research

Data Analysis
The mechanical properties used in this study is friction ratio (Rf) which is computed as:

ܴ݂ ൌ  ‫ ܿݍ‬Τ݂‫ݏ‬ (1)

where Rf is friction ratio (%), qc is measured cone resistance (kg/cm2) and fs is sleeve friction (kg/cm2).

Once Rf of measured CPT data was calculated, interpolation of this value on a regular grid point is required to
create a digital geotechnical map of soil classification. Inverse Distance Weighting (IDW) method was employed for
Rf interpolation. In IDW method, interpolated values were estimated as a weighted linear combination of the most
proximal observations, with the weights proportional to the inverse of the distance between neighboring
observations and interpolation location14, as presented below.

ͳ
σ୬୧ୀଵ “
”୧஑ ୧
“ො ൌ (2)
ͳ
σ୬୧ୀଵ ஑
”୧

where q ̂ is the interpolated (gridded) value, q_iis the observed value in station i, r_iis the euclidean distance
between interpolated station and station i; α is the power of distance and n is the total number of stations interpolated
per grid. This method has been successfully applied to interpolate daily precipitation over Java island 15.
Soil classification was derived from Rf. In this study, soil was classified into 5 classes i.e. gravel, sand, silt, clay
and peat. For each soil type, a respective range of Rf suggested by Begemann and Vos was employed. The produced
soil type was ten validated using measurement data. Based on this validation, a new Rf range for each soil type was
proposed in order to improve the accuracy of soil classification. This new Rf range can be defined as
characterization of soil classification based on CPT data for mountainous tropical region.

Soil Type Validation

Validation is required to assess the accuracy of soil type spatial distribution. In this study, validation was
performed by remove out 10% of the data and compare interpolated soil type with measurement data. In addition,
the soil type spatial distribution is compared side by side with two other soil classification methods namely
Begemann and Vos.

020004-5
 Developing Soils Classification Map by QGIS
To perform validation using aforementioned method, a soil classification map needs to be generated. In this
study, QGIS v.2.18.2, an open source mapping software developed by qgis community, was used in developing the
soil classification map. Taking the advantage of this software, IDW interpolation for soil mechanical properties was
also conducted in QGIS v.2.18.2. Generated soil type map was then overlaid with administrative boundary map
downloaded from info-geospatial.com.

RESULT AND DISCUSSION

CPT Data Recapitulation

Out of approximately 300 CPT data, about 197 data were considered to be feasible and reasonable during
filtering process, hence they are used in this research. This data are distributed unevenly within the administrative
boundaries. The number of data for each district can be seen in TABLE 3.
TABLE 3. CPT data for each district
Year Total
District
2005 2006 2007 2010 2011 2012 2013 2014 2015 2016 Data
Banjarnegara 1 1 5 1 5 13
Banyumas 5 2 2 4 2 9 1 1 9 61
Batang 2 2
Brebes 4 4 9 5 3 1 26
Cilacap 3 1 3 2 3 6 1 19
Kebumen 1 1 1 2 8 13
Pemalang 7 4 1 12
purbalingga 2 2 1 6 3 1 8 5 28
Tegal 1 4 5 5 2 1 18
Temanggung 1 1
Wonosobo 2 1 1 4
Grand Total 8 1 5 14 9 38 38 16 40 18 197

Table 3 shows that Banyumas has the most number of CPT data (79), meanwhile Temanggung has the smallest
one (1). Spatial distribution of this CPT data in the study area is presented in Fig. 5.

FIGURE 5. Spread of CPT data in research area.

020004-6
 Figure 7 shows that all data are distributed throughout the study area. The middle area (Banyumas and
Purbalingga) have the majority of the data. In addition, northern side has relatively equal number of data to the
southern side. Moreover, western side has also quite similar number and distribution compared to eastern side.

Correlation analysis between soils type and Rf

In order to simplify the analysis, soils grouped into 5 type of classification i.e. gravel, sand, silt, clay and peat
with respect to soils classification test result. Boundary of Rf for each soil type is determined from Rf range.
Distribution of count and percentile of Rf for each of soil type is shown in TABLE 4.
TABLE 4. Spread of Rf data for each soil type
Percentile
Class Quantity
0.25 0.5 0.75
Gravel 5 0.302 0.502 1.478
Sand 9 1.422 2.296 2.759
Silt 55 2.948 5.036 7.558
Clay 68 2.456 4.372 6.692
Peat 40 3.585 7.138 10.930

Table 4 show that soil in the study area is dominated by clay and silt. Moreover, both the soil types have similar
range of Rf. On the other hand, coarse grain soil i.e. gravel and sand, have limited number of data. Based on
distribution of Rf shown in Table 4, range of Rf for each soil in the study area is proposed and presented in TABLE
5.
TABLE 5 Correlation between soils type and Rf.
No Soil Type Friction Ratio
1 Gravel < 0.7 %
2 Sand 0.7 % - 2.75 %
3 Silt & Clay 2.75 % - 7.0 %
5 Peat >7%

Soils Classification Map

The map of soil classification based on TABLE 5 can be seen in Fig. 6. It shows that almost all across the study
area silt and clay dominate the soil type. Peat soil type exists in the coastal zone as a result of deposition process of
eroded materials.

FIGURE 6. Soil classification map.

020004-7
 FIGURE 7 shows soil classification map of the study area using soil classification proposed by Begemann and
Vos whereby respective range of each soil type is given in TABLE 1 and TABLE 2 for Begemann and Vos method
respectively.
.

FIGURE 7. Begemann and Vos soil classification maps.

It can be seen from Fig. 7 that silt and clay dominate soil type in the study area, similar to the friction range
proposed in this study. It is also noticed that some area in the northern Java coastal zone have peat soil type.
Meanwhile, very limited area in the middle and southern have coarse grain fraction of soils. Begemann method has
the most soil distribution variability compared to another two methods.
In order to compare among the three classification methods, we conducted validation test by checking the
similarity between soil type derived from the interpolation and from laboratory test. Validation data were randomly
chosen laboratory test data. Distribution of validation location is presented in Fig. 8.

FIGURE 8. Spead of validation data location.

On each validation location, soil type was extracted and the result is shown in

020004-8
 TABLE 6. Soil classification map validation result
Map Classification Validation
No x y Lab Class
Result Begemann Vos Result Begemann Vos
-
1 108.86 6.95 MH or OH Peat Peat Peat yes Yes Yes
-
2 108.99 6.91 CH or OH Peat Peat Peat yes Yes Yes
-
3 109.03 7.21 MH or OH Silt and Clay Clay Clay no No No
-
4 109.05 7.08 SC Silt and Clay Clay Clay yes Yes Yes
Clay
and Loam,
- and loam
5 109.07 7.44 ML or OL Silt and Clay soils Clay yes Yes No
-
6 109.13 7.54 CH or OH Peat Peat Peat yes Yes Yes
-
7 109.14 6.98 MH or OH Silt and Clay Clay Clay no No No
-
8 109.14 7.62 SC Silt and Clay Clay Clay yes Yes Yes
-
9 109.22 7.42 ML or OL Silt and Clay Clay Clay yes No No
-
10 109.28 7.59 CL Silt and Clay Clay Clay yes Yes Yes
-
11 109.29 7.45 ML or OL Silt and Clay Clay Clay yes No No
-
12 109.35 7.40 SC Silt and Clay Clay Clay yes Yes Yes
-
13 109.37 7.07 MH or OH Silt and Clay Clay Clay no No No
-
14 109.37 6.92 ML or OL Silt and Clay Clay Clay yes No No
-
15 109.42 6.89 SC Silt and Clay Clay Clay yes Yes Yes
Clay
and Loam,
- and loam
16 109.51 7.66 CL or OL Silt and Clay soils Clay yes Yes Yes
-
17 109.53 7.32 CL Silt and Clay Clay Clay yes Yes Yes
-
18 109.63 7.43 CL Silt and Clay Clay Clay yes Yes Yes
-
19 109.75 7.38 CL or OL Silt and Clay Clay Clay yes Yes Yes
-
20 109.90 7.21 SC Sand Silty sand Silt yes Yes Yes
Total Valid Data 17 14 13
Percentage of Accuracy 85% 70% 65%

Table 6 shows that proposed Rf range proposed in this study yields in 85% similarity between interpolation and
laboratory measurement. This value is higher than two another method that have similarity of 70% and 65% for
Bagemann and Vos method respectively. This indicates that Rf range for each soil type proposed in this study better
represents the classification of fine grain (silt and clay) and coarse grain (sand and gravel) compared to two other
discussed method. Hence, Rf range proposed in this study can be used to near accurately classify soil type and
subsequently useful for landslide analysis.

CONCLUSION
Soil classification map was generated using IDW method and presented using QGIS software. In addition, soil
classification map with respect to modified friction ratio range was validated using 10% of total measurements. On
the validation, the modified friction ratio range produces 85% similarity with laboratory measurements. Within the
study area, silt and clay dominate soil type, which is in agreement with two popular methods namely Begemann and

020004-9
Vos. The result shows that modified friction ratio range can effectively distinguish fine and coarse grains, thus
useful for soil classification.

ACKNOWLEDGMENTS
The author thanks to Jenderal Soedirman University for supporting this study through Institutional Research
Grant and to Soil Mechanic Laboratory of Jenderal Soedirman University for providing high quality CPT data.

REFERENCE
1. H. Z. Abidin, H. Andreas, M. Gamal, Surono and M. Hendrasto, J. Eng. Tech. Sci. 36(2), 109–123(2004).
2. A. Verruijt, Soil mechanics (Delft: Delft University of Technology, 2001).
3. A. O. Mohamed and O. A. El Fatih, J. Build. Road Res. 5(August), 37–46(2003).
4. S. S. De Dalal and R. B. Sahu, Indian Geotech. J. 41(4), 206–214(2011).
5. A. A. Shahri, A. Malehmir and C. Juhlin, Eng. Geology 189, 32–47(2015).
6. H. K. S. Begemann, ‘The friction jacket cone as an aid in determining the soil profile.’, in 6th International
Conference on Soil Mechanics and Foundation Engineering. (Montreal, 1965), pp. 17–20.
7. B. J. Douglas and R. S. Olsen ‘Soil classification using electric cone penetrometer’, in Symposium on Cone
Penetration Testing and Experience, Geotechnical Engineering Division, ASCE. St. Louis, Missouri, (Missouri,
1981), pp. 209–227.
8. J. D. Vos, ‘The Pratical Use of CPT in soil Profiling’, in The Second European Symposium on Penetration
Testing, EOSPT-2. (Amsterdam, 1982), pp. 933–939.
9. P. K. Robertson, Canadian Geotech. J. 27(1), 151–158(1990).
10. B. H. Fellenius and A. Eslami, ‘Soil Profile Interpreted from CPTu Data’, in Geotechnical Engineering
Conference, , November 27 - 30, 2000. (Bangkok, 2000), pp. 1–19.
11. R. Schimmer, ‘A remote sensing and GIS method for detecting land surface areas covered by copper mill
tailings’, in Pecora 17 - The Future of Land Imaging Going Operational. (Denver, Colorado, 2008), pp. 1–13.
12. S. Maryati, H. Shimada, T. Sasaoka, A. Hamanaka, K. Matsui and H. Nagawa, J. Geographic Inf. Sys. 4(1), 62–
70(2012).
13. S. M. Ahmadi, A. Jafarirad and N. Nezafati, J. Geographic Inf. Sys., 7(December), 677–691(2015).
14. D. Shepard, ‘A two-dimensional interpolation function for irregularly space data’, in Proceedings of the 23rd
National Conference ACM, (1968), p. 1968.
15. Yanto, B. Livneh and B. Rajagopalan, Sci. Data, 4, 170072(2017).

020004-10