Some issues related to applications of the CPT

N. Ramsey
Sinclair Knight Merz, Melbourne, Australia

ABSTRACT: This paper reviews some issues related to the use of Cone Penetration
Testing for geotechnical applications. Some of the areas that are considered include:
a) The advantages and disadvantages of Cone Penetration Testing (CPT)
b) The advantages of integrating CPT with laboratory testing.
c) Identification of similar geological units using statistics.
d) A review of published classification/behaviour charts, using a diverse and
highly dependable database.
e) The importance of using correct cone calibration and cone zero values in
normally consolidated fine-grained soils.

1 INTRODUCTION

The Burland Triangle (Burland, 1987), shown in Figure 1, provides a useful

framework for the majority of geotechnical problems. The Cone Penetration Test
(CPT) can provide valuable input to this framework, by providing cost-effective and
useful information for the “Ground profile” and “Soil behaviour” aspects of the
triangle.

Figure 1: The Burland Triangle (Burland, 1987)

The main purposes of this paper are to review the contribution of the CPT, in terms
of ground profiling and the assessment of soil behaviour. The paper concentrates
primarily on sites containing normally consolidated (NC) fine grained soils, as these
soils tend to be relatively difficult to analyse, and because published correlations can
be less reliable in these soils. Practical examples are presented from several widely
distributed sites, brief details of which are presented in Table 1.
 Site Location Idealised soil Water Liquid Limit Plasticity Fines Content
description Content (%) (%) Index (%) (%)

X Offshore loose to medium - - - <35

Europe dense silty sand
Y Offshore NC clay 80 - 90 90 - 100 50-60 100
North Africa
Z Offshore NC clay 70 - 80 70 - 80 40-50 100
West Africa
V Onshore NC clay over 75 – 85 70 – 90 50 – 60 90 – 100
Australia OC clay 25 – 45 40 - 80 20 – 50 80 - 100
S Onshore FILL overlying - - - -
Australia NC clay 55-60 55-60 30-40 25-35
Note: NC denotes “normally consolidated” and OC denotes “overconsolidated”

Table 1: Approximate details of example sites

2 CPT FOR ASSESSING SOIL LAYERING AND VARIABILITY

The capability of the CPT (and CPTu) for assessing soil layering and variability is
well documented (e.g. Lunne et al., 1997). As a practical example, Figure 2 presents
CPT data from Site V, indicating sand, overlying normally consolidated clay,
overlying overconsolidated clay. As well as highlighting the strength of the CPT for
differentiating soil layers and transition zones, Figure 2 also illustrates the CPT’s
ability to identify variations in soil macro-fabric.

Figure 2: Site V - Differentiation of soil layers and identification of macro-fabric

 The value of the CPT for identifying variability within similar soil units is
illustrated in Figure 3, using data from Site Z. Although the eight CPTs at this site
were separated laterally by approximately 25km, the cone resistance profiles at seven
of the eight locations are similar – indicating a relatively uniform depositional
environment over the area of interest. The anomalous cone resistance profile at the
eighth site was later attributed to the presence of a nearby salt intrusion (diapir),
which had increased the lateral effective stresses in the vicinity of the test.

Figure 3: Site Z - Identification of anomalies

Figure 4 illustrates the ability of the CPT to differentiate even relatively subtle
differences is geological units. Site Y is located offshore North Africa in an area
containing a large variety of geohazards, including mud volcanoes and areas that
have previously experienced major mass movements. As a consequence, the
depositional environment is extremely variable. To provide a reference framework, a
statistical approach was used to assess individual soil units and grouping of similar
soils. Although the four locations shown in Figure 4 are spread over an area of
approximately 4km by 8km. the results clearly suggest that the depositional
environment, at all four locations, is similar.

Figure 4: Site Y - Variation of net cone resistance

At the site, the quantity of site investigation was sufficient to enable statistical
characterization of the CPT results to identify statistical signatures for each soil unit.

The statistical signature was based on:

• The average normalised cone resistance, Qt.
• The coefficient of variation of Qt.
• The calculated “best fit” slope of the net cone resistance, qnet, values with
respect to depth.
• The regression coefficient, “r”, representing the closeness of the depth and
qnet data to a straight line.
This approach enabled:
• A quantitative objective method of recognising individual units and unit
boundaries to be established.
• A quantitative objective method for assessing soils that may have been
subjected to mass-movement to be established.
• A calibrated estimate of soil parameter values from cone penetration tests, at
locations where no geophysical or laboratory data were available - including
the undrained shear strength, remoulded shear strength, overconsolidation
ratio, carbonate content, constrained modulus and small strain shear modulus
(Gmax).

The ability of CPTs to assess variations in soil fabric has already been mentioned.
Published information normally suggests that the order of precedence for maximising
resolution and reliability is typically; pore-pressure ratio, Bq, followed by friction
ratio, Rf, followed by cone resistance, qc. To illustrate, the capability of these
sensors, Figures 5 to 7 present a sample of composite results between 2m and 6m
below mudline at Site Z. It may be seen, by inspection, that the Bq profile can be
used to detect thin layers of relatively permeable soil that cannot be discerned using
the other the other sensors. These results confirm that the Bq parameter is the
premier parameter for assessing variations in soil macro-fabric, at least for offshore
sites, where sensor saturation is more reliable.

Figure 5: Site Z, qc profiles Figure 6: Site Z, Rf profiles

Figure 7: Site Z, Bq profiles

One of the limitations of performing CPTs in fine-grained soils containing granular

inclusions, such as coarse-gravel and cobbles, is that these inclusions can distort the
soil interpretation by causing sharp reductions in pore-water pressure (pwp) that
temporarily impair the performance of the cone sensor, when the cone sensor is
located on the cone shoulder. These rapid reductions in pwp are caused by the
inclusion being pushed aside by the cone, thus creating local suctions adjacent to the
pwp sensor. The time taken for the sensor to return to normal performance depends
on the stiffness and permeability of the soil being tested, and on the magnitude of the
reduction in pwp. Typically, the pwp sensor recovers within 0.5m to 1m penetration,
but in severe cases, the pwp sensor may take metres of penetration to recover, or may
not recover for the whole test – this behaviour is often referred to as “cavitation”.
When this behaviour is caused by coarse gravel or cobbles, it is common to observe
complementary “spikes” in the friction sleeve readings, which are caused by local
increases in lateral stress, as the inclusion is pushed to one side. If the inclusion is
dilatant sand, spikes in the friction sleeve readings are seldom observed. Figure 8
presents cavitation data obtained from a variety of sites, with water depths ranging
from less than 20m to almost 250m. It may been that, in general, the pwp sensor
appears to cavitate when the pwp reduces to the ambient pressure at the cone
shoulder, regardless of whether glycerine or silican oil has been used to saturate the
cone sensor.

Another disadvantage of performing CPTs is the limited thrust that can be applied
to the cone, either due to the limitation on the magnitude of reactive force, or due to
the risk of damage to the cone. As a consequence, CPTs tend not to be the tool of
choice for accurately assessing the depth to rock-head (i.e. bedrock), or deep soil
profiling in coarse-grained soils, or heavily overconsolidated fine grained soils. In
volcanic soils, such as the basaltic clays found around Melbourne, Australia, the
problem is exacerbated by the presence of isolated boulders (floaters) of relatively
unweathered material in a soil matrix. It would be wrong to suggest, however, that
boreholes always produce better results. This is illustrated in Figure 9, where a
series of boreholes performed over a site failed to identify a zone of relatively
unweathered basalt beneath a critical part of a proposed structure. In this case,
geophysics would have been the most reliable approach for assessing this boundary.

Figure 8: Measured whole-test cavitation pressure versus depth of water for pwp
sensor on cone shoulder

Figure 9: Photo showing unweathered Basalt rock near the surface at a site in
Victoria, Australia
3 PREDICTING SOIL PARAMETER VALUES

Published correlations provide a useful means of predicting soil property values from
CPTs, and there are many excellent papers on this subject, most recently Robertson
(2009). Robertson (2009) correctly points out that for all but low-risk projects CPT
should be supplemented with other types of in-situ testing and/or more advanced
project-specific laboratory testing. Two examples are provided in Figure 10 and
Figure 11 to illustrate why this approach is recommended. Figure 10 presents
remoulded shear strength profiles from Site Y – including a profile generated from
the design undrained shear strength profile using sensitivity calculated using the
approach recommended by Robertson (2009), where sensitivity is calculated using
Equation 1.

St = 7.1/Rf (1)

It may be seen that there is good agreement between the measured remoulded
strengths and the generated profile. It is also interesting to note that the author’s
typical approach, of assuming that the remoulded strength is two-thirds of measured
sleeve friction, also produces good agreement.

Figure 10: Site Y – Remoulded shear strength compared to CPT sleeve friction
 In contrast, at the same site, the estimation of Gmax/qnet directly from CPTs, using
the approach suggested by Robertson (2009) and shown in Equation 2, is not in good
agreement with values inferred from laboratory tests (Figure 11). Note that the
laboratory tests were performed on test specimens inferred to be of good quality, so
sample disturbance is not considered to be the cause of the discrepancy between the
measured and inferred values. It is also worthy of note that Robertson (2009) points
out that the formula presented in Equation 2 is less reliable in NC fine-grained soils.

Gmax = [10(0.55 Ic + 1.68)] (qnet – σv) (2)

where:

Ic = [3.47 – log Qt)2 + (log Rf + 1.22)2]0.5 (Robertson, 2009) (3)

Figure 11: Site Y - Comparison between estimated Gmax/qnet based on Robertson

(2009) and measured values on laboratory samples
4 INTEGRATING IN-SITU AND LABORATORY TESTING

The advantages and disadvantages of CPTs and laboratory testing are summarised in
Table 2. It may be seen that the pros and cons of the two techniques are, in most
cases, complementary. As a consequence, the reliability of each is significantly
increased when performed in combination with the other.

CPT Lab Testing

Pros • Provide precise boundary elevations • Enable the effects of increasing

oftern missed by drilling techniques. strain levels to be assessed on soil
behavior – particularly changes in
• Provide accurate profiling effective stress, strength and
information, including the proportion stiffness.
of soil inclusions and the proportion
• Soil samples can be reconsolidated to
of intermittent layers of higher
take account of the increases in
permeability.
effective stresses caused by the
• Provide information on natural soil foundation loading.
variability within geological units.
• Tests are performed in a controlled
• Provide information on the environment of strain-rate, drainage
characteristics of the soil matrix. and temperature.
• Soil behavior under cyclic loading
conditions can be assessed.

Cons • Do not provide information on the • A much smaller amount of soil is

effect of the foundation on soil being tested so macro effects may
behavior. not be representative of macro-
behaviour
• Do not provide information on the
• Sensitive to the quality of the test
effect of different rates of loading on
specimen and the method of
soil behavior..
specimen preparation (particularly
sand soils).
• Do not provide information on soil
behavior under cyclic loading
conditions.
•

Table 2: Advantages and disadvantages of CPT and laboratory testing

The improved reliability associated with combining CPTs and laboratory testing is
illustrated in Figure 12 Both conventional CPTs and seismic CPTs (SCPT) were
used to estimate and measure Gmax in a loose to medium dense silty sand layer,
beneath a proposed large Concrete Gravity Structure (CGS) at Site X in the North
Sea. The seismic cone results were considered to provide a reliable measure of the
in-situ Gmax prior to the CGS installation. However, due to the critical nature of Gmax
for foundation design, it was considered appropriate to use laboratory testing to
assess Gmax after the CGS had been installed. As a consequence, bender element and
resonant column laboratory tests were performed on soil specimens reconstituted to
relative densities inferred from the CPT data and then consolidated to take account of
the additional effective stresses induced by the structure loading. The results from
these tests indicated an increase in Gmax, which was used in design. In addition, the
laboratory test results were in good agreement with a modified Gmax profile,
calculated using the procedure recommended by Lunne et al (1997) to account for
the effect of increased foundation stresses. The good agreement between the two
approaches was considered to increase the reliability of the chosen Gmax design
profile.

Figure 12: Site X – Gmax, profiles

5 SOIL CLASSIFICATION METHODS

A crucial component for successful foundation design is accurate interpretation of

soil type to enable the practising design engineer make reliable assessments of soil
behaviour. To assist in this assessment a comparison of several published
approaches is presented in Figures 15 to 22.The assessment has been made using the
author’s database which comprises 439 data points, with tests depths ranging from
less than 0.3m to greater than 100m below ground level, at 101 testing and sampling
locations, spread over a wide geographical area. The database is taken only from
offshore piezocone data, to ensure that all tests were performed in saturated soils.

The Ramsey (2002) database was divided into nine soil categories, separated
according to the criteria presented in Table 3:
Category General Description St OC su / Fines Clay Relative to
(% by po’ Content Content A’ Line
(-) weight) (-) (%) (%)

1 Extra sensitive clay soils, i.e. soils >8 - - >35 >12 Above
with a sensitivity greater than 8.

2 Organic soils and peat, i.e. soils - >5 - - -

with an organic content greater
than 5% by weight.

3 Clay soils with a ratio of su / po’ ≤ - - ≤1 >35 >12 Above

1 (approximately corresponding
to normally consolidated to
slightly overconsolidated clay)

4 Clay soils with a ratio of su / po’ > - - >1 >35 >12 Above
1 (approximately corresponding
to significantly overconsolidated
clay)

5 clayey SAND - - - ≤35% ≥5 - 12% Below

6 sandy very clayey SILT - - - >35% - ≥5 - 12% Below

65%

7 sandy SILT - - - >35 – 65% <5% -

8 silty SAND - - - >12 – 35% <5% -

9 “clean” to slightly silty - - - ≤12% <5% -

SAND/GRAVEL
Notes:
(1) Zones 1 and 2 are based directly on the Robertson (1990) classification model.
(2) St = Sensitivity
(3) OC = Organic Content (by weight).
(4) ‘A’ line indicates the plotting position of Atterberg Limits results relative to the ‘A’ line

Table 3: Soil Zone Categorisation (Ramsey, 2002)

The Ramsey (2002) Model (Figure 13 and 14) evolved during the 1990s, when the
author was working at Fugro Limited in the UK. The database, presented in this
paper, was not used to develop the model; it was merely used to corroborate its
effectiveness in a quantitative way. This was achieved by asking a person with no
previous experience with CPT to look through Fugro’s archives to find situations
where CPT data were available adjacent to laboratory test results, then compare the
predicted soil category with the measured soil category. As a consequence, the
database is considered to be highly reliable. Full details of the methodology used to
verify the database, and the results of the assessment, are presented in Ramsey
(2002).

The database is presented on the following figures:

• Figure 15: Schmertmann (1969)
• Figure 16: Douglas and Olsen (1981)
• Figure 17: Jefferies and Davies (1991)
 • Figure 18: Schneider (2008)
• Figure 19: Robertson Rf plot (1990)
• Figure 20: Robertson Bq plot (1990)
• Figure 21: Ramsey Rf plot (2002)
• Figure 22: Ramsey Bq plot (2002)

Figure 13: Ramsey (2002) Rf classification plot

Figure 14: Ramsey (2002) Bq classification plot

 It is not the purpose of this paper to make detailed evaluations of the relative merits
of each model, and it is acknowledged that each of the models has strengths,
weaknesses and limits of application. Instead, it is left to the readers to assess the
information, and make their own decisions. However, some general comments are:

a) Methods that use appropriately corrected and normalized cone parameters,

tend to have a greater range of application than methods that apply no
correction or normalization..

b) Most of the models are relatively good at differentiating between clean sand
and normally consolidated to slightly overconsolidated clay (su/po’ < 1).

c) Intermediate soils tend to be much more difficult to differentiate.

d) None of the published friction ratio soil classification models, presented in

this paper apply for tests performed using cones with non-standard projected
areas of 2cm2 or 1cm2, as friction ratios measured using these cones tend to
differ considerably from values obtained using standard 10cm2 cones.

e) None of the models presented in this paper give reliable soil classification in
calcareous soils, due to the variations in grain compressibility, carbonate
content and degree of cementation, that often occur in these soils over very
short distances.

f) None of the models reliably predict peat or highly organic materials.

One area in which the Ramsey (2002) model differs from the other models is that
the soil classification zones were purposely chosen so that, if there is a discrepancy
between soil classifications according to the two models, the decision-making process
to assess the matrix soil classification is based on a single criterion. This criterion is
simply that whenever the models predict different zones then the zone with the lower
numerical value should be chosen. This is because it is generally more conservative
to interpret a soil with a higher fines content. So, for example, if the Qt/Rf model
indicates Zone 4 and the Qt/Bq model indicates Zone 5 (or any zone higher than 5), then
the matrix soil classification should be classified as Zone 4. It should be noted,
however, that in keeping with previous discussions, the Bq parameter, and hence the
Qt/Bq model, is much better at detecting thin secondary layers or inclusions within a
primary soil matrix, provided the pwp sensor remains saturated. Therefore, if it is
deemed important to define the extent and thickness of thin layers of coarser
material, within a fine grained soil layer, then the Qt/Bq model data should be
reassessed and, if appropriate, used to over-ride the Qt/Rf model.
 Key to symbols

Figure 15: Schmertmann (1969) Soil Classification Plot

Key to symbols

Figure 16: Douglas and Olsen (1981) Soil Classification Plot

 Key to symbols

Figure 17: Jefferies and Davies (1991) Soil Classification Plot

Key to symbols

Figure 18: Schneider (2008) Soil Classification Plot

 Key to symbols

Figure 19: Robertson (1990) Soil Classification Rf plot

Key to symbols

Figure 20: Robertson (1990) Soil Classification Bq plot

Figure 21: Ramsey (2002) Soil Classification Rf plot

Figure 22: Ramsey (2002) Soil Classification Bq plot

6 CPT IN SOFT SOILS

Figure 23 presents a CPT performed at Site S, in South Australia, in predominantly

normally consolidated clay soils. It may be seen that, at this site, calculating the net
cone resistance, using vertical total stress based on measured soil unit weights,
results in negative cone resistance between 12m and 17m depth. It is the author’s
experience that there are other similar cases in very soft normally consolidated soils.
In these instances, the author’s standard practice has been to recalculate the net cone
resistance using half the total stress, rather than the total stress. It is the author’s
experience that this approach generally results in much better correlations with other
in-situ and laboratory test results. This approach is in keeping with Wroth (1988)
who pointed out that the correct equation for net cone resistance should be as shown
in Equation 4. It is the author’s suggestion that it would be better to adopt Ko = 0.5,
as standard practice, as the only soils likely to be significantly affected by this
change would be those soils with Qt values less than 5, i.e. those soils where the Ko
value would tend to be closer to 0.5, rather than the standard value of 1.0 that is
currently used.

qnet = qt – Ko.σv (4)

The discussion above is explicitly based on the assumption that the measured cone
sensor readings are accurate. In the author’s experience, if the measured cone
resistance is zero, or lower, it is almost always due to one of the following:
a) Incorrect calibration.
b) Insufficient sensor resolution.
c) Incorrect zero load reading at the start of the test.
The only other obvious alternatives are that the soil layer is still consolidating after
the recent addition of overburden (i.e. under consolidated), or there are artesian
conditions.

In stiffer soils (Qt > 20), small errors in the assessment of zero load readings often
have a negligible influence on the inferred soil parameters, but in the case of soft
soils a small error can cause major errors. As an example of the implications of
incorrect calibration or incorrect zero load reading, Figure 24 presents two CPTs
performed within a distance of less than 2m, by two different CPT operators at
Site S. The cone resistance values measured by Operator E resulted in an
anomalously low net cone profile. In contrast, undrained shear strength profiles,
interpreted from a CPT and a T-bar test performed by Operator B, indicated
remarkably similar results. On the basis of the discrepancies between the operators’
data, the calibration and zero readings were rechecked, and Operator E confirmed
that it had inadvertently used an incorrect procedure to record the zero load reading.
Figure 23: Site S – Net cone resistance, soil type and shear strength (based on Ko =
1.0 and 0.5) and associated undrained shear strength profiles

Figure 24: Site S - Comparison of results from two different CPT operators
7 CONCLUSIONS

This paper presents and discusses a few selected issues related to the application of
CPTs. Emphasis has been placed on CPT profiles in predominantly normally
consolidated fine-grained soils. On the basis of the data and discussions presented,
the following observations and conclusions are considered appropriate.

1) The CPT is an excellent tool for assessing soil layers and soil variability.

2) The pore-pressure sensor, in particular, provides valuable information on thin

layers of more, or less, permeable material, within a soil matrix.

3) The CPT has limitations, in terms of guaranteeing target penetration in very

dense granular soils or in fine-grained soils containing coarse gravel-sized, or
larger, inclusions.

4) Granular inclusions, within a generally fine-grained soil matrix, can cause

temporarily impaired pore-pressure sensor readings.

5) Laboratory testing significantly enhances the benefits of CPT.

6) In some cases, statistical assessments can be used to identify similar

geotechnical units over wide geographical areas and, in these cases, soil
parameter values interpreted from CPTs, can be used with a good degree of
reliability.

7) The choice of soil classification model has a significant bearing on the

interpreted soils, if no complementary soil sampling and testing have been
performed.

8) In soft soils, in particular, it is extremely important to ensure that the cones have
sufficient resolution and have been correctly calibrated and zeroed prior to
testing.

9) There are some grounds for defining the net cone resistance as the total cone
resistance minus one half of the total overburden pressure, rather than the current
standard practice of subtracting the total overburden pressure.

8 ACKNOWLEDGEMENTS

The author would like to thank:

• All the energy companies that he has had the pleasure of working with over
the last three decades – particularly BP and Statoil.
• All the government departments that he has worked with in Australia,
particularly, DTEI and PoMC.
• Black In-situ Testing, of Melbourne, Victoria, for providing some of the
results presented in this paper.
 • Fugro, for providing the author with opportunity to work on numerous
interesting and challenging projects throughout the world.
• Davron Lu and Haran Janarthanan for assisting with producing the figures
• Liz Mooney for providing Figure 9.

8 KEY TO SYMBOLS AND TERMS

„ Bq is the excess pore-water pressure ratio, defined as the ratio of the measured
excess pore-water pressure to the net cone resistance.
„ Clay content is defined as the % of particles with a diameter less than 2 microns.
„ Coefficient of Variation (COV) of Qt (expressed as %) is obtained by dividing
the standard deviation of the Qt values by the average Qt value. It therefore
indicates the uniformity of the Qt values about the average.
„ Fines content is defined as the % of soil grains passing a 63 micron sieve.
„ fs is the measured sleeve friction
„ po’ is the effective overburden pressure, also referred to as σv’.
„ po is the total overburden pressure, also referred to as σv.
„ qc is the measured cone resistance
„ qt is the total cone resistance, i.e. the measured cone resistance, corrected for the
effects of cone shape and pore-water pressure distribution around the cone tip.
„ qnet is normally defined as the total cone resistance minus the total overburden
pressure relative to ground level.
„ Qt is the normalized cone resistance, defined as the ratio of net cone resistance to
effective overburden pressure.
„ “r” coefficient - The “r” value is the regression coefficient representing the
closeness of the depth and qnet data to a straight line. In general, an “r” value
greater than +/-0.8 represents an excellent fit, greater than +/-0.6 represents a
good fit, and a value greater than +/-0.4 represents a moderate fit.
„ Rf is the friction ratio, defined as the ratio of the measured sleeve friction to the
net cone resistance.
„ St is the sensitivity, defined as the ratio of the undisturbed undrained shear
strength to the remoulded undrained shear strength.
„ su is the undrained shear strength, as measured by unconsolidated undrained
triaxial compression tests.
„ σv’ is the effective overburden pressure, also referred to as po’.
„ σv is the total overburden pressure, also referred to as po.
9 REFERENCES

ASTM, 2004. Standard Method of Deep Quasi-Static Cone and Friction-Cone Penetration Tests of
Soil; ASTM Standard D 3441, ASTM International, West Conshohocken, PA, 7 pp.
ASTM D-5778 Standard Test Method for Performing Electronic Friction Cone and Piezocone
Penetration Testing of Soils.
Barentson, P. 1936. Short Description of a Field-testing Method with Cone-Shaped Apparatus, Proc.
1st International Conference on Soil Mechanics, Cambridge Mass. 1, 7-10
Begemann, H. K. S, 1965. The Friction Jacket Cone as an Aid in Determining the Soil Profile;
Proceedings, 6th ICSMFE, Montreal, Quebec, Canada, Vol 1, pp.17-20.
Burland, J.B. 1987. Nash Lecture: The Teaching of Soil Mechanics – a Personal View. Proceedings,
9th ECSMFE, Dublin, Vol. 3, pp 1427-1447.
De Reuter, J. 1971. Electric Penetrometer for Site Investigations; Journal of SMFE Division, ASCE,
Vol. 97, SM-2, pp. 457-472.
Douglas, B.J., and Olsen, R.S. 1981. Soil classification using electric cone penetrometers. Cone
penetration testing and experience. Proc., Cone Penetration Testing and Experience, ASCE, New
York, 209-227.
ISSMFE 1989. International Reference Test Procedure for Cone Penetration Test CPT, reproduced in
Lunne et al. 1997.
Janbu, N., and Senneset, K. 1974. Effective stress interpretation of in situ static penetration test. Proc.,
European Symp. On Penetration Testing, ESOPT, 181-193.
Jefferies, M. G. and Davies, M.O. 1991. Soil Classification by the cone penetration test: Discussion.
Canadian Geotechnical Journal, 281, 177 – 176.
Lunne, T., Robertson, P.K., and Powell, J.J.M. 1997. Cone penetration testing in geotechnical
practice, Blackie Academic and Professional, London .ISBN 0 751 40393 8.
Olsen, R.S., and Mitchell, J.K. 1995. CPT stress normalization and prediction of soil classification.
Proc. Int. Symp. on Cone Penetration Testing, CPT 95, Vol. 2, 257-262.
Ramsey, N 2002. A Calibrated Model for the Interpretation of Cone Penetration Tests CPTs in North
Sea Quaternary Soils. Proc SUT Conf, London
Robertson, P.K., and Campanella, R.G. 1988. Guidelines for use interpretation and application of the
CPT and CPTU, UBC Soil Mechanics Series No. 105. Civil Eng. Dept., Vancouver, B.C., Canada.
Robertson, P.K. 1990. Soil classification using the CPT. Can. Geotech. J., 271, 151-158.
Robertson, P.K. 2009. Interpretation of cone penetration tests – a unified approach. Can. Geotech. J.,
46, 1337-1355.
Schmertmann J. H. 1969. Dutch Friction Cone Penetration Exploration of Research Area at Field 5,
Eglin Air Force Base, Florida, US Army Waterways Experimental Station, Vicksburg, Mississippi,
Contract Report S-69-4, 1969
Schneider, J.A., Randolph, M.F., Mayne, P.W., and Ramsey, N. 2008. Analysis of factors
influencing soil classification using normalized piezocone tip resistance and pore pressure
parameters, ASCE Journal of Geotechnical and Geoenvironmental Engineering, 13411, 1569-
1586.
Torstensson, B 1975. Pore Pressure Sounding Instrument. American Society of Civil Engineers
Conference on In-situ Measurement of Soil Properties, 48-54, Raleigh
Wissa, A.E.Z., Martin, R.t. and Gaklanger, J.E. 1975. The Piezometer Probe. American Society of
Civil Engineers Conference on In-situ Measurement of Soil Properties, pp536-545, Raleigh
Wroth P. 1984. The interpretation of In-situ Tests - Rankine Lecture, Geotechnique, 344, pp 449 –
489. -
Wroth P. 1988. Penetration testing – A More Rigorous Approach to Interpretation, Proc. of the
International Symposium on Penetration Testing, ISOPT-1, Orlando, 303-311, Balkema Pub.,
Rotterdam.