3rd International Symposium on Cone Penetration Testing, Las Vegas, Nevada, USA - 2014

Introduction to CPT accuracy

J. Peuchen & J. Terwindt
Fugro, the Netherlands

ABSTRACT: CPT practice has an admirable history on benchmarking accuracy of parameter values. A
proposed Analytic CPT can serve as a benchmark for true values, independent of deployment-specific
CPT features. This paper provides an introduction to CPT accuracy with reference to the proposed Ana-
lytic CPT and with reference to CPT standard ISO 22476-1 and its application (accuracy) classes.
ISO 22476-1 defines accuracy, where (1) cone resistance, sleeve friction and pore pressure are uncoupled
from spatial position and (2) true values depend on permissible deployment-specific features. Uncertainty
estimates are presented for favourable and adverse conditions. The adverse case illustrates difficulty in
demonstrating compliance with the application classes of ISO 22476-1. Significant epistemic uncertainty
is the principal reason for this situation, notably prediction models for transient temperature and meas-
urement error. The uncertainty model provides what may be the first estimate that complies with
ISO 22476-1 procedures. Actual accuracy is probably significantly better than calculated.

1. IN SITU PENETRATION TESTING

A supreme in situ penetration tool:

provides zero measurement uncertainty,
gives unambiguous soil behaviour identification,
has a closed-form theoretical interpretation model directly linked to fundamental soil mechanics,
penetrates at high speed to desirable depths into any type of ground at temperatures between perma-
frost and volcanic outflow,
allows cheap operation from a small autonomous vehicle operating on any terrain, above and below
water,
is uniquely standardised.

No such tool exists. The (piezo)cone penetration test (CPT) is closer to this wish list than any of its
in situ rivals. A CPT measures soil resistance to penetration of a cone and friction sleeve and, optionally,
pore pressure, all at a fairly standardised geometry and a push-in penetration rate of about 20 mm/s.
Cone penetration testing in a real world demands concessions in investing in technology and limiting
operational cost. Successful CPT technologies are available in the market place, providing cost-effective

1
support to industry and society at large. Successful may be recognized by accurate measurements even
as upholding robustness under extreme physical demands. Is it possible to estimate CPT accuracy?
The following sections provide an introduction to CPT accuracy, exploring and illustrating what is
known. Significant epistemic uncertainty remains. Discussions are limited to friction cone and piezo-
cone penetration tests, using cone penetrometers with load cells as force measuring transducers, which
probably represents >95% of cone penetration tests in practice. Supplementary tests and systems such as
pore pressure dissipation, seismic downhole (seismic cone penetration test) and full displacement pres-
suremeter (cone pressuremeter test) are excluded from the discussions.

2. USE OF CPT RESULTS IN PRACTICE

Society and industry prescribe design and construction of facilities and the associated engineering risks
that are considered to be acceptable. These prescriptions are typically captured by standards and guide-
lines. Cone penetration tests can be part of design and construction processes, in that they provide ge-
otechnical measurements and are thus part of structure reliability (ISO 1998).
Widely used CPT standards are those published by ISO (2012a) and ASTM (2012), hereafter abbre-
viated to ISO 22476-1 and ASTM D5778. The following sections include critique on sections of these
CPT standards. This is hoped to be constructive. In this regard, the authors wish to stress their apprecia-
tion and recognition of the efforts made in standardisation. Work is typically done by only a small core
group of dedicated volunteers with limited backing. It is work in progress, never finished. Furthermore,
consensus is to be achieved within the core group and, usually, with an active peer group of similar size.

3. BENCHMARKING OF CPT ACCURACY

CPT practice has an admirable history on benchmarking accuracy of parameter values (Table 1). Such
ambitions are not yet seriously contemplated for parameter values inferred from soil sampling, sample
handling and geotechnical laboratory testing. For example, Ruiter (1975) noted, without special precau-
tions, under routine field conditions, the possible error in cone resistance is about 5 kg/cm2, with care-
ful calibration before and during the test this can be improved to 1 kg/cm2 cone resistance. This
statement applies to a friction cone penetrometer with strain-gauge load cells. Note that 1 kg/cm2 is ap-
proximately 100 kPa. Also note routine field conditions. This rightly recognises the complexity of ac-
curately measuring resistance values differing by typically four orders of magnitude, with no control on
measuring environment. This paper continues with this recognition of complexity and introduces the
terms favourable and adverse. Favourable may be compared to a setting that approaches a calibra-
tion laboratory. Adverse represents common hostile site conditions and/or a setting with real-world lim-
its on tool control.

Table 1. Historical notes on CPT standardisation with focus on accuracy.

Year Notes
1968 Recommended method for static sounding (static penetration test) (European Group of the Sub-committee for
Static and Dynamic Penetration Test Methods, 1968).
Description of method covers just over one page; 1,000 mm2 for cross sectional area of cone; no geometry require-
ments for friction sleeve; penetration rate should be constant and <20.8 mm/s
1977 Recommended standard for the cone penetration test (CPT) (ISSMFE 1977).
Advancement of 1968 work; detailed requirements for CPTs with friction cone penetrometer; reference cone pene-
trometer defined with nominal cross sectional area of 1,000 mm2 and friction sleeve area of 15,000 mm2; surface
roughness for friction sleeve of 0.5 m with tolerance of 50% in longitudinal direction; measurement requirements
for two precision classes: the larger of 5% of measured value and 1% of maximum value of range, respectively
10% and 2%; the term range is undefined: measuring range is probably intended; check is required on sensor ze-

2
Year Notes
ro-load drift; cylindrical height of cone defined at 2 mm to 5 mm; penetration rate defined at 20 mm/s 5 mm/s;
use of (undefined) smaller and larger diameter cone penetrometers permitted
1979 Deep, quasi-static, cone and friction-cone penetration tests of soil (ASTM 1979).
First (?) CPT standard published by a national standardisation institute; requirements for friction cone penetrometer
with cross sectional area of 1,000 mm2 and friction sleeve area of 10,000 mm2 or 15,000 mm2; bottom of friction
sleeve <10 mm above base of cone; use of (undefined) larger cone penetrometers permitted; penetration rate de-
fined at 20 mm/s 5 mm/s; accuracy of thrust measuring instrumentation to 5% of correct values; statement in-
cluded on precision estimates for qc and fs
1989 International reference test procedure for cone penetration test (CPT) (ISSMFE 1989).
Advancement of 1977 work; piezocone test included; cylindrical height of cone defined at 7 mm to 10 mm; maxi-
mum height of gap between cone and friction sleeve defined at 5 mm; piezocone filter in conical part of cone (u1
position) or in cylindrical extension (u2) with precautions to maintain full saturation; comment on influence of wa-
ter pressure and area ratios for cone and friction sleeve, requirement on cone penetrometer bending influence; sur-
face roughness of friction sleeve between 0.25 m < r < 0.75 m in longitudinal direction; precision of measure-
ments and zero drift to be better than the greater of 5% of the measured value or 1% of the maximum value of the
measured resistance in the layer under consideration (undefined); depth measurement to an accuracy of 0.1 m; in-
clinometer may be built into penetrometer; geometry scaling according to cone penetrometer diameter
1996 Dutch standard NEN 5140 Determination of the cone resistance and the sleeve friction of soil electric cone pene-
tration test (NNI 1996).
Advancement of 1989 work; restricted to friction cone penetrometers; requirements for bevelled ends of friction
sleeve; practical guidance on penetration interruptions including electronic heave compensator; time-based data re-
cording; four accuracy classes for cone resistance, sleeve friction, inclination and penetration depth; accuracy ex-
pressed as the larger of a threshold value (e.g. 50 kPa, 250 kPa or 500 kPa for cone resistance qc) and percentage of
measured value (e.g. 3% and 5% for qc); accuracy defined with reference to ISO metrological standards; resolution
should be better than one third of accuracy; guidance on uncertainty analysis; geodetic requirements; presentation
of digital tabular results
1999 International reference test procedure for the cone penetration test (CPT) and the cone penetration test with pore
pressure (CPTU) (ISSMGE 1999).
Advancement of 1989 and 1996 work; piezocone filter in u1, u2 or u3 (immediately above friction sleeve) position;
u2 filter may be in cylindrical part of cone or in gap between cone and friction sleeve; inclusion of pore pressure
dissipation test; four accuracy classes for cone resistance, sleeve friction, pore pressure, inclination and penetration
depth; accuracy classes with tighter values for qc (e.g. 50 kPa, 200 kPa, 400 kPa or 500 kPa for cone resistance) and
fs; threshold values for pore pressure of 5 kPa, 25 kPa or 50 kPa; presentation of corrected cone resistance qt and
pore pressure ratio Bq
2012 ISO standard: electrical cone and piezocone penetration tests (ISO 2012a).
Based on 1999 reference test procedure, four application/ accuracy classes (Fig. 18) with stricter values for qc
(35 kPa, 100 kPa, 200 kPa or 500 kPa for cone resistance) and fs; adjusted values for pore pressure: 10 kPa, 15 kPa,
25 kPa or 50 kPa; normative requirements on maintenance, checks and calibration
2012 ASTM standard: electronic friction cone and piezocone penetration testing of soils (ASTM 2012).
Periodic ASTM update, some matches with ISSMGE international reference test procedure; requirements for fric-
tion cone and piezocone penetrometers with cross sectional areas of 1,000 mm2 (reference); cylindrical height of
cone plus any filter to be more than 2 mm and less than 20 mm; bottom of friction sleeve between 5 mm and
15 mm above base of cone; friction sleeve with equal end areas; cross sectional areas of 500 mm2 and 1,500 mm2
are permitted; penetration rate defined at 20 mm/s 5 mm/s; precision of pore pressure sensor of better than
14 kPa; permissible zero drift of 2% of full scale output FSO of qc, fs and u measurements; depth accuracy of
0.1 m; normative annual calibration requirements for laboratory environment under ideal conditions with re-
quirements typically expressed as percentage of FSO: atmospheric axial calibration (no bending) for qc and fs, pres-
sure calibration for u, and ambient temperature stability; statement included on precision estimates for qc, fs and u
expressed in terms of standard deviation as percentage of FSO, refer to Table 2
2014? ISO standard: marine soil investigations (draft, ISO 2012b)
Possible publication year 2014; expected similarities with application classes of ISO (2012a) with adjustments to
suit offshore practice

A literature search conducted on behalf of the authors led to many hits on CPT accuracy claims. Early
and recent claims are by and large unsubstantiated. No publication provided an elementary estimation of
accuracy of cone penetration test results according to ISO (2003) and EA (1999) or equivalent.

3
 What is clear is that many in industry have overrated beliefs in accuracy of geotechnical measure-
ments. CPT measurements are no exception. These overrated beliefs are probably due to researchers and
practitioners correlating repeatability with accuracy. A lower accuracy is, in itself, usually not an issue
for practice. The belief in a better accuracy than actually achieved can lead to unsafe conclusions:
There is generally no relationship between precision of a sensor and accuracy of a measurement.
Geotechnical parameter values usually require combining results from a number of measurements.
Derived measurements may be dominated by a weakest-link principle.
Operating conditions at site often differ significantly from assumptions and calibration laboratory
conditions. The success of geotechnical measuring systems designed for a specific set of operating
conditions and site conditions triggers wider use, without appropriate recognition of potential limita-
tions.
Many publications tell of success stories for carefully selected sites showing exceptional simplicity in
soil conditions. Errors remain undetected and unexplained observations remain untold.
Many standards and detailed guidelines have implicit fuzziness about accuracy.

Usually, discussions on CPT accuracy take place without a satisfactory definition for accuracy. Some
use accuracy in metrological terms of resolution or repeatability. Others presumably refer to accuracy
under calibration laboratory conditions. Many are ambiguous about: (1) coupling of qc, fs and u with
spatial position xyz and (2) true value related to permissible equipment-specific and procedure-specific
features.
Defining accuracy requires a digression into metrology. Metrologists tend to be precise and accurate.
Their vocabulary is reflected in JCGM (2012) international vocabulary of metrology. Accuracy relates
to a true quantity value of a measurand; refer to Figure 1 and Glossary of Metrological Terms, Ta-
ble 5. The following sections show that this true value is generally described in an approximate manner.
Value

Accurate:
Precise
Unbiased
Penetration Depth

Inaccurate:
Precise
Biased

Inaccurate:
Imprecise
Unbiased

Inaccurate:
Imprecise
Biased
True
Value Bias
Figure 1. Accuracy, precision and bias.

The issue of true value related to permissible equipment-specific and procedure-specific features may
be demonstrated by a proposed Analytic CPT that serves as benchmark for true values (Fig. 2). The
proposed Analytic CPT complies with ISO 22476-1 and ASTM D5778 and features: (1) exact spatial

4
position xyz where penetration depth z equals penetration length l, (2) penetration rate v of 20 mm/s (3)
application of push rod thrust with no influence of thrust machine on ground conditions, (4) an imagi-
nary cone penetrometer with a solid body of uniform geometry with no gaps, no pore pressure filter in-
terference and surface roughness Ra = 5 m and (5) zero measurement uncertainty.

Figure 2. Example cone penetrometer from practice and proposed cone penetrometer for Analytic CPT.

4. CPT ACCURACY IN STANDARDS

The following sections focus on ISO 22476-1 and ASTM D5778. It is interesting to note that ISO
22476-1 uses the term accuracy and ASTM D5778 uses precision.
ISO 22476-1 is a performance-based standard, particularly defining CPT accuracy in terms of accura-
cy classes. ISO 22476-1 uses the term application class for accuracy class. The performance-based ap-
proach aims at a minimum accuracy of test parameter values within a framework of only essential re-
quirements. ISO (ISO/IEC 2011, 2013) encourages a performance-based approach. ISO 22476-1 is
remarkable, in that it refers normatively to the VIM (1993), by means of a normative reference to ISO
10012 (2003) measurement management principles. In comparison, the authors are not aware of other
standards for geotechnical tests defining important results in terms of accuracy as defined by VIM
(1993). Some that do are in preparation. Note that JGCM (2012) replaced VIM (1993).
ISO (2003) states, The measurement management system shall ensure that specified metrological
requirements are satisfied. ISO 22476-1 expresses metrological requirements in terms of accuracy and
error: If all possible sources of errors are added, the accuracy of the recorded measurements shall be
better than the largest of the values given in Table 2. The inaccuracy analyses shall include internal fric-
tion, errors in the data acquisition, eccentric loading, temperature (ambient and transient) effects and di-
mensional errors. The term accuracy assumes knowledge of a true value, i.e. what should be regarded
as true value. Note that ISO 22476-1 assigns accuracy requirements to defined parameters, e.g. meas-
ured cone resistance qc, measured sleeve friction fs, and measured pore pressure u. These measur-
ands depend on actual CPT system characteristics. ISO 22476-1 provides no explicit guidance on inevi-
table and allowable differences between measuring instruments and measuring practice. In other words,
true values obtained by different systems will not be directly comparable. The term error implies a refer-

5
ence quantity value. This probably refers to a value determined in a calibration laboratory or similar, as
it is generally impossible to obtain an in situ reference value. For example, calculation of cone resistance
requires axial force and cone base area. A reference value for dimensional error for cone base area can
be readily established. However, what would be the reference quantity for the influence of a permissible
change in cone apex? Also, ISO 22476-1 provides no explicit guidance on coupled measurands (parame-
ters). Presumably, ISO 22476-1 considers a true value of cone resistance to be uncoupled from a true
value of pore pressure u and uncoupled from a true value of spatial position xyz below ground surface.
Normally consolidated or slightly overconsolidated clay would represent a favourable setting for spatial
coupling. Layered soil would represent adverse conditions. In practice, all possible sources of error
will usually be interpreted as errors listed in Annex E of ISO 22476-1 uncertainties in cone penetration
testing. Note that Annex E offers no guidance on intentions for the words include but are not limited
to.
ASTM D5778 is prescriptive, method-based. Most standards for geotechnical tests are method-based.
This implies that the aim is to produce equivalent and competing systems through detailed descriptions
of specific apparatus and step-by step methodology. The accuracy of the acquired test parameter values
is generally unknown. ASTM D5778 states: Precision - There are little direct data on the precision of
this test method, in particular because of the natural variability of the ground. Committee D-18 is active-
ly seeking comparative studies. Judging from observed repeatability in approximate uniform deposits,
persons familiar with this test estimate its precision as follows (Table 2). ASTM D5778 also states:
Bias - This test method has no bias because the values determined can be defined only in terms of this
test method.

Table 2. CPT precision according to ASTM (2012).

15.1.1 Cone Resistance Provided that compensation is made for unequal area effects as described in 13.2.1, a
standard deviation of approximately 2% FSO (that is, comparable to the basic electrome-
chanical combined accuracy, nonlinearity, and hysteresis).
15.1.2 Sleeve Friction - Standard deviation of 15% FSO.
Subtraction Cones
15.1.3 Sleeve Friction - In- Standard deviation of 5% FSO.
dependent Cones
15.1.4 Dynamic Pore-water Strongly dependent upon operational procedures and adequacy of saturation as described in
Pressure 11.2. When carefully carried out a standard deviation of 2% FSO can be obtained.
Note: a subtraction-type penetrometer has two load cells in series. The lower load cell measures forces on the cone (C) and
the upper load cell measures the sum of forces on the cone and the friction sleeve (C+F). The force on the friction sleeve is
obtained by subtraction: (C+F)-C. An independent-type cone penetrometer has two load cells that measure forces on the
cone and friction sleeve independently.

ASTM (2011) and ASTM (2013) define bias and precision: respectively a systematic error that con-
tributes to the difference between the mean of a large number of test results and an accepted reference
value and the closeness of agreement between test results obtained under prescribed conditions. The
term uncertainty is mentioned under measurement, but not explained. ASTM (2011) mentions accuracy
but provides no definition. ASTM (2013) refers to bias when looking up accuracy. It may be noted here
that NIST (2013) distinguishes between bias and accuracy: Accuracy is a qualitative term referring to
whether there is agreement between a measurement made on an object and its true (target or reference)
value. Bias is a quantitative term describing the difference between the average of measurements made
on the same object and its true value.

6
5. CONE RESISTANCE
5.1 Geometry

ISO 22476-1 allows the use of cone penetrometers with cross sectional areas between 500 mm2 and
2,000 mm2. Cone penetrometer geometry should be adjusted proportionally to diameter, where dimen-
sions for a 1,000 mm2 penetrometer serve as reference. Similarly, ASTM D5778 considers cone pene-
trometers with cross sectional areas of 1,000 mm2 and 1,500 mm2. Test results may be used without the
application of correction factors, i.e. need not be accounted for in accuracy statements. However, scale
effects will inevitably influence CPT results (De Beer 1963, Diepstraten 2003, Hird & Springman 2006,
Meave Silva 1999, Peuchen et al. 2005, Peuchen 2012, Powell & Lunne 2005a, 2005b, Randolph 2004,
Titi et al. 2000, Tufenkjian et al. 2010, Vreugdenhil et al. 1994). General comments are as follows.
Cone resistance qc in low-permeability (clay) strata may be expected to be within 10% of the refer-
ence cone size.
It is much more difficult to make comparisons between cone penetrometers in sands. Build-up of qc
of a smaller cone penetrometer is faster than that for a larger penetrometer.
There is some evidence that a smaller cone penetrometer provides higher qc in strongly dilatant sands
and lower cone resistance in sands with contractive behaviour. Differences can partially be explained
by elastic theory.
CPT signature may be affected when effective particle size D50 exceeds about 10% of the diameter of
the cone penetrometer. Individual particles rather than the soil mass may contribute to the measure-
ments.
The loading response of a smaller penetrometer to soil layering is more rapid than that of a larger
cone penetrometer. This relates to soil failure mechanisms in layered soils. Depending on ground
conditions, the smaller probe may show higher peak qc values and lower base values.
Soil structure may cause failure mechanisms along zones of weakness. The larger cone penetrometer
affects a larger mass of soil. There is a greater potential for soil structure effects. Thus, the larger
cone penetrometer may exhibit lower qc in structured soils.
A fixed rate of penetration of 20 mm/s will affect transitions between drained, partially drained and
undrained soil response to penetrometer penetration. This may be approximately quantified by a non-
dimensional parameter V where V = vd/cv, where v is penetration rate, d is cone diameter and cv is
coefficient of consolidation of soil.

The following sections consider cone penetrometers with a nominal cross-sectional area of
1,000 mm2, unless indicated otherwise.
Measured cone resistance qc is derived from the cross sectional area Ac of the cone and the axial force
Qc acting on the cone.
Measurement of cross sectional area is typically done by vernier calliper achieving an uncertainty of
about 0.5% for 1,000 mm2 cross sectional area under favourable conditions. The 0.5% value considers
averaging of three vernier calliper measurements at 120o, each with a calliper measuring uncertainty of
0.1 mm. This results in about 0.3% measurement error for cross sectional area. Cross sectional area
will also be affected by a difference in temperature at the time of cross section measurement and tem-
perature of a cone penetrometer embedded in soil. This is because of steel expansion/contraction at
about 10-3%/oC. This influence is negligible.
Wear of a cone penetrometer is a non-random process. Approximate corrections for wear are feasible
but uncommon in practice.
ISO 22476-1 and ASTM D5778 prescribe tolerances for cone geometry. In practice, the diameter of a
new cone will be at a permissible (steel) geometry that provides maximum wear before cone replace-
ment, i.e. 36.0 mm for a nominal 1,000 mm2 cone penetrometer (ISO 22476-1 and ASTM D5778). This

7
provides a cross sectional area that is 1.8% larger than nominal. Note that ISO 22476-1 allows a diame-
ter of 36.1 mm for an u2 filter, giving 2.4% larger than nominal. ASTM D5778 excludes this permission.
A lower limit of 35.3 mm for a worn cone gives a 2.1% smaller area than nominal. These percentage
values for area tolerances are upper limits for proportional influence on cone resistance. The influence of
cross section variations will generally be less than proportional to cone resistance. This is because of net
area ratio, discussed below. CPT processing software commonly considers a fixed cross sectional area at
midpoint of wear. This approach results in a measurement error of up to about 2%. Note that this fixed
area differs from the nominal value of 1,000 mm2 and takes no account of net area ratio effects.
Tolerances not only affect cross sectional area but also, for example, cone apex and height of cylin-
drical extension (Fig. 3). ISO 22476-1 appears ambiguous on whether such changes in geometry may be
ignored for estimates of accuracy of cone resistance. Effects on cone resistance are primarily significant
for clay soils. For example, a 10 mm cylindrical height implies about 3.4% contribution to cone re-
sistance for a typical clay soil and soil-steel friction on the cylindrical height equal to 3% of the cone re-
sistance. ISO 22476-1 allows wear down to 7 mm cylindrical height. This would then imply a reduction
in cone resistance of 1% compared to the 10 mm case. ASTM D5778 allows a wider range of cylindrical
heights: 3 mm to 15 mm depending on penetrometer design. Contributions to cone resistance will vary
accordingly. Changes in cone apex will also affect cone resistance, but probably to a lesser degree than
cylindrical height. The authors are not aware of specific study results.

Figure 3. Wear of cones made of different materials after approximately same metres of penetration (Schaap &
Zuidberg 1982).

Effects of variations in surface roughness are probably relatively small for cone resistance. It is tenta-
tively estimated that variations in cone resistance are possibly less than 0.2% for favourable conditions
and less than 2% for adverse conditions. The authors are not aware of specific study results.
ISO 22476-1 specifies an average surface roughness Ra of <5m for steel to be determined by a surface
profile comparator. The requirement for cone surface roughness applies to the time of manufacture, with
the intention that the roughness at manufacture approaches the roughness of a cone acquired upon use in
common ground conditions. ASTM D5778 prescribes The cone is made of high strength steel of a type
and hardness suitable to resist wear due to abrasion by soil. In practice, the ASTM D5778 approach
will be equivalent to ISO 22476-1. The presence of a pore pressure filter will additionally affect surface
roughness as a function of material type, manufactured geometry tolerances and geometric variations
upon soil stresses acting on the filter.
The Analytic CPT has a solid body. In practice, all cone penetrometers incorporate a gap between the
cone and the friction sleeve. The gap varies in volume upon cone penetration into soil and affects cone
resistance because of deformation/ displacement of a flexible mechanical soil seal and an O-ring water
seal and because of stresses induced by soil particles, water and gas in the gap.

8
 A well-designed soil seal affects cone resistance by less than laboratory calibration uncertainty, i.e.
no difference is typically observed between calibrations with and without soil seals.
Soil particles in the gap can lead to force transfer from the cone to the friction sleeve, i.e. cone re-
sistance shows an apparent reduction. Sleeve friction will show an apparent increase by a factor 1/15,
representing the area ratio of the cone to the friction sleeve. Force transfer can probably be ignored un-
der favourable conditions, i.e. a well-designed soil seal and a nearly constant volume of the gap at a
nearly constant force on the cone and the friction sleeve. Highly variable gap volume represents poten-
tial for adverse conditions. Such conditions may be expected under strongly variable soil resistance and
with low-stiffness penetrometers. In extreme cases, errors in cone resistance may exceed 1 MPa, refer to
Section 14. Sleeve friction signature will then typically approach that of cone resistance and accurate
separation of cone resistance and sleeve friction will no longer be possible. These extreme cases can of-
ten be attributed to particles locking into the gap, possibly exacerbated by high steel-soil temperatures
generated during cone penetration.
Soil in the gap will be subject to radial forces and soil-soil shear forces acting in an axial direction.
Radial forces may cause particles to show Poissons effects, the magnitude of which is probably small.
The shear forces will largely transfer to the friction sleeve. This will be further discussed below. Schaap
& Zuidberg (1982) observed calibration errors for cone resistance in the order of 2% to 3% for cone
penetrometers returned to a calibration laboratory. Simple cleaning and maintenance reduced these er-
rors to about 0.4%, indicating significant influence of soil ingress into the gap. The observations by
Schaap & Zuidberg presumably applied to air-dried cone penetrometers. Better calibration errors could
possibly apply under in situ conditions, with access to water. Jekel (1988) compared 11 CPTs in sand
with removal of soil from gaps before start of a test with 7 CPTs without removal. No trend with qc, fs or
Rf was found. The authors are not aware of other specific studies on cone/friction sleeve interference by
soil particles at/in the gap.
Water and gas in the gap affect measured cone resistance. This may be explained by means of
qt = qc+(1-a)u2g. ISO 22476-1 uses the term corrected cone resistance for corrected cone resistance for qt.
ASTM D5778 use the terms corrected total cone resistance, estimated total tip resistance and total tip
stress. The term u2g represents water and/or gas pressure in the gap. It replaces u2, where u2 is pore water
pressure traditionally derived from an initially saturated pore pressure measuring system positioned near
but outside the gap. The term a represents net area ratio, i.e. the ratio of the cross-sectional steel area
at the gap between cone and friction sleeve to the cone base area. Net area ratio is typically between 0.5
and 0.85. It depends on cone penetrometer design and wear of the cone. Note that wear of a cone from
36.0 mm diameter to 35.3 mm diameter would reduce (1-a) from an initial design value of, say, 0.25 to
0.22, i.e. by 14%. CPT processing software may allow entry of actual cross sectional area and corre-
sponding adjustment in the theoretical value for (1-a).
It may be noted here that CPT standards prescribe measures promoting initial saturation of an u2
measuring system. No such requirements apply to the gap. The effect of u2g on cone resistance is im-
portant for normally and slightly overconsolidated low-permeability soils. The influence of u2g on qc is
very small for dense sands. This small influence is because of high qc values compared to u2g. Values for
u2g depend on:
initial degree of water/gas saturation in the gap, i.e. at start of test
soil-induced changes to saturation
soil seal behaviour
soil permeability
static groundwater pressure u0
transient pore water and/or gas pressures induced by cone penetration into soil
volume change of the gap by strongly variable soil resistance and with low-stiffness penetrometers.

9
 Accurate measurement of u2g will be extremely difficult. The authors are not aware of cone pene-
trometers equipped with sensors for u2g measurement, although recommended by ISO 22476-1 and al-
luded to by ASTM D5778. Under favourable conditions, u2g will approximate u2. Favourable conditions
may be expected when u2 values are relatively uniform and exceed about 2 MPa, i.e. at which point any
free gas in pore water will be forced into solution (Fig. 4). This point should be reached at the equivalent
of about 100 m to 200 m hydrostatic (ground) water head in high-permeability soil, i.e. rarely onshore
and frequently offshore. It may possibly be reached at shallower penetration in low-permeability soils
(clays and silts) with u2g response exceeding u0. However, low-permeability soils will allow only limited
inflow of water into the gap, causing a delay in saturation. Also, soil particles in the gap may act as a
barrier to water inflow. A deeper point applies to soil with u2g < u0, e.g. soil showing undrained and
highly dilatant response upon cone penetration.
Normalised Net Area Rat io [-]
0.9 0.92 0.94 0.96 0.98 1
0

6
Pressure [MPa]

10

12

14

16

18

20

Figure 4. Laboratory water pressure versus normalised net area ratio.

Figure 4 suggests that favourable conditions may be approximated by an equivalent uncertainty for
(1-a) and assuming u2 = u2g. Figure 4 presents applied pressure u versus normalised net area ratio, de-
fined as (1-ameasured)/(1-atheoretical). Values for atheoretical were derived from manufacturing specifications
and values for ameasured were derived from qc/u. An uncertainty in the order of, say, 2.5% for normal-
ised net area ratio would give uncertainties for qc of about 0.5% for (1-a) = 0.2 for normally consoli-
dated or slightly overconsolidated clay. The value of 0.5% would increase to 1.3% for (1-a) = 0.4.
Adverse conditions should be expected for most onshore CPTs.
An adverse setting can lead to a bias compared to a theoretical value of (1-a). An equivalent underes-
timate for (1-a) could then be in the order of, say, 12%. This may be mitigated to an uncertainty of say
8% by judicious selection of (1-a) values for data processing. The 8% value would give uncertainties
for qc of about 1.5% for (1-a) = 0.19. The 1.5% value considers data processing at 95% of the nominal
value for (1-a), i.e. 0.19 instead of 0.2. Again this value would apply to normally consolidated or slightly
overconsolidated clay. The corresponding uncertainty for qc would be about 3.8% for (1-a) = 0.38.
An extreme case would be normally consolidated or slightly overconsolidated clay with u2g remain-
ing at atmospheric pressure during penetration. This could possibly apply when low permeability of soil
prevents inflow of water into the gap. The corresponding bias would increase cone resistance by about
20% for (1-a) = 0.2 and by about 50% for (1-a) = 0.4, compared to favourable conditions.
Note that a penetration interruption may present an adverse condition. In an extreme case, it may pos-
sibly lead to an apparent increase of cone geometry. Examples of inevitable penetration interruptions are
adding a push rod and performing a pore pressure dissipation test. The apparent increase may develop

10
because of consolidation of low-permeability soil around a cone and associated to soil/penetrometer ad-
hesion that is sufficient to give an increase in cone diameter (Fig. 5). Note that a stationary cone can
apply local stresses that approach failure conditions, i.e. to close to net cone resistance or about two
times the in situ mean effective stress. Resumption of penetration will usually lead to loss of adhered
soil, perhaps within an equivalent distance of a few times the cone diameter. The authors are not aware
of specific studies on this topic.

Figure 5. Clay adhering to cone penetrometer after retraction from soil.

5.2 Measuring range

Cone penetrometers used in practice typically match push capacity of common thrust machines and al-
low penetration into a wide range of soil conditions. Typical nominal measuring ranges for qc are 0 to
50 MPa for cone penetrometers with cross-sectional areas of 1000 mm2 and 1500 mm2. Cone resistance
measurements to about 150 MPa may be considered for special cases, for example for penetration of
very dense overconsolidated sands by a cone penetrometer with a cross-sectional area of 500 mm2. Cone
penetrometer design typically incorporates an allowance for fully elastic behaviour to twice the nominal
range.
Cone penetrometer calibration may be to a lower measuring range than allowable for a load cell. This
can be useful for fine-tuning of linear sensitivity coefficients. A fitted curve may also be used. Note that
a lower measuring range was historically used for improving resolution limits. This is no longer neces-
sary. For example, a 24bit analogue to digital A/D conversion system would typically provide resolution
better than 1 N or 0.01 kPa.
So-called pressure compensation (for example, CEC 1986, Boggess & Robertson 2010) is occasional-
ly adopted for limiting the measuring range for offshore cone penetration tests. The interior of a pene-
trometer is filled with a fluid and a pressure compensator connects the fluid with tubing to the water at
seafloor. The practice of pressure compensation relates to hydrostatic water pressure generating signifi-
cant values of cone resistance. Results for uncompensated cone resistance require correction to zero at
seafloor. For example, the correction value for cone resistance at a water depth of 1000 m is about
7.5 MPa if the net area ratio of the cone penetrometer is 0.75. For comparison, normally consolidated
clay may show a cone resistance qc-gradient of 30 kPa/m depth, i.e. a cone resistance of 1.2 MPa at 40 m
depth. This would imply measured cone resistances that are a small fraction of the correction values.
In practice, a small-range penetrometer may compromise offshore project economics. Particularly, a
larger measuring range for qc of up to about 50 MPa can avoid penetrometer change-outs when encoun-
tering any deepwater hard grounds. For this case, the additional complexity of pressure compensation
would not be necessary.

11
5.3 Force measurement

A well-designed cone penetrometer may show laboratory uncertainty for force (expressed as qc) in the
order of [5 kPa + 0.005 qc] and an additional 0.005 qc,max for hysteresis effects. The parameter qc,max
is the maximum encountered hysteresis value of qc prior to reaching penetration depth.
A load cell is made of steel of a shape that suits a cone penetrometer. Its shortening upon loading is
measured by strain gauge sensors glued to the steel surface. The strain gauges are connected in a single
or double full Wheatstone bridge configuration. The bridge is excited with an electric voltage and the re-
sulting ratio of output voltage to excitation voltage is measured, converted to a digital signal and a force.
An accurate load cell requires a process of machining of the steel body, artificial aging by loading and
heating processes, meticulous application of the strain gauges and again aging of the adhesives used for
gluing the strain gauges. The space for the load cell in the cone penetrometer is typically filled with air,
initially at atmospheric pressure. Water seals prevent build-up of internal pressure by external stresses
induced by cone penetration. Variation in internal air pressure will occur because of secondary effects,
such as cone penetrometer deformation and temperature variation. These secondary effects are assessed
to be negligible for a well-designed penetrometer.
Accuracy and repeatability of a load cell are typically described by calibration error, repeatability,
non-linearity, hysteresis and zero-load error (ASTM D5778). The sensitivity coefficient of a load cell is
the relation of an applied load to the force output of the load cell. The sensitivity coefficient can be op-
timised by curve fitting or by applying a specific sensitivity coefficient for a specific portion of the nom-
inal measuring range. Internal friction is a main contributor to hysteresis of a load cell. Particularly,
transfer of soil resistance to an axial force on a load cell implies inevitable relative movements between
the various components of the cone penetrometer. Disturbance of force transfer will occur due to inter-
nal friction caused by the water pressure seal(s) on the load cell body. Zero drift of load cell output typi-
cally results from shifts in the total calibration curve including its zero load output.
Laboratory calibration for a cone penetrometer is typically for loading/unloading against a reference
load cell. Note that a high quality cone penetrometer has characteristics equivalent to a high-quality ref-
erence load cell, i.e. calibration results may be limited by the reference load cell. Low-range data may be
obtained by incremental application of mass. Low-range data and any loading/ unloading loops provide
information for soft soils and for offshore use, where cone resistance is taken as zero at seafloor, with an
offset axial force representing water pressure (Peuchen 2000).
Routine laboratory calibration takes place under atmospheric conditions. A high pressure environ-
ment affects soil seals and water seals of a cone penetrometer. These seals consist of rubber or similar
material. Their reduced elasticity upon in situ pressure increase can thus affect the inevitable relative
movements between cone penetrometer components. VanLoon and Schaareman (1991) conducted ex-
periments at an ambient pressure of 30 MPa, equivalent to a water depth of 3000 m. The experiments
indicated a measurement error equivalent to about qc = 4 kPa (2.5%) for an axial force of 250 N (equiva-
lent to qc = 166 kPa) on a subtraction-type cone penetrometer with a cone base area of 1500 mm2. In-
creased error may be expected for low-stiffness penetrometers.
Laboratory-type error for axial loading increases with applied force and with hysteresis. Uncertainty
analysis of actual CPTs may incorporate a prediction model for such errors. A simple model would con-
sider the highest force applied at any point before reaching the actual penetration depth. For example,
the model would assign low measurement uncertainty for a soil profile showing soft clay overlying
dense sand. Conversely, soft clay underlying dense sand would show high measurement uncertainty.
Axial loading is applied during routine laboratory calibration. This represents favourable conditions.
Non-axial loading of a cone penetrometer often occurs in practice. Figure 6 presents an example of
penetrometer bending moment M inferred from inclination measurements according to Ooi & Ramsey
(2003).

12
 o
qc [MP a] Inclinat ion i x, iy [ ] Deflection x, y [m] M [N.m]

0 10 20 30 40 -6 -4 -2 0 -1 -0.5 0 0 10 20 30 40
0

5
Penetration Depth [m]

10

15

20

25
Figure 6. Bending moment inferred from cone penetrometer inclination.

Adverse conditions may be represented by a bending moment in the order of 100 Nm, which approx-
imates a bending radius of 30 m. Figure 7 shows results of laboratory experiments for 1,000 mm2 sub-
traction-type cone penetrometers equipped with double-bridge and single-bridge strain gauges. The ex-
periments consisted of fixing a horizontal cone penetrometer just above the friction sleeve and hanging a
mass from the cone. No axial force is applied. The directional dependency relates to radial positions of
the strain gauges. As expected, higher errors may be observed for cone penetrometers equipped with
single-bridge strain gauges. In-house Fugro studies indicate typically 2 to 4 times worse behaviour for
single-bridge technology. Higher errors should also be expected for a penetrometer with low bending
stiffness, for example for a so-called sensitive or low-range cone penetrometer. A lower bending re-
sistance increases deformation. This can then increase internal friction between a load cell and body of
the cone penetrometer, particularly in combination with varying axial force. Note that the diameter of a
cone penetrometer is important for bending stiffness. Peuchen et al. (2005) speculated on bending stiff-
ness substantially affecting measurement uncertainty. They reported a bending stiffness ratio (EI) of
about 12:1 for subtraction-type cone penetrometers with 1000 mm2 and 100 mm2 cross sectional areas.
This ratio is for a force (F) normalised to cone base area: EI1000/F1000 = 12 EI100/F100.
10 1
Error for Cone Resistance [kPa/N.m]

Error for Sleeve Friction [kPa/N.m]

5 0.5

0 0

-5 -0.5

A - Single Bridge A - Single Bridge

-10 B - Double Bridge -1 B - Double Bridge
C - Double Bridge C - Double Bridge
-15 D - Single Bridge -1.5 D - Single Bridge
E - Single Bridge E - Single Bridge
-20 -2
0 45 90 135 180 225 270 315 360 0 45 90 135 180 225 270 315 360
o o
Orient at ion [ ] Orient at ion [ ]

Figure 7. Measurement error due to bending moment - cone resistance (left) and sleeve friction (right).

13
 Steel and strain gauges are subject to creep effects. Such effects are probably more than one order of
magnitude less than temperature effects discussed below. Also, they cannot be readily isolated from
temperature effects. Stress changes in a water pressure seal may possibly also contribute to apparent
creep. Effects are assessed to be negligible for a well-designed cone penetrometer.
Force measurement will include measurement of barometric and gravitational variations taking place
during a CPT. These influences can be ignored for a penetration phase of a CPT, which typically takes
less than 1 hour. For example, a barometric variation of 5 mbar (high value) and a tide change (gravita-
tional) of 1 m would theoretically affect cone resistance by less than 1 kPa and less than 10 kPa respec-
tively. A long penetration interruption represents a possible adverse condition.

5.4 Temperature

Temperature change causes straining of the steel of a load cell, its strain gauges and their adhesives. Dif-
ferential thermal strain shows as an apparent cone resistance. Temperature change of a well-designed
load cell is a major reason for drift and a minor one for change in sensitivity. Temperature change may
be ambient, where it is uniform throughout a penetrometer, with respect to time. Ambient temperature
change may result from external air/ water/ soil temperature around a cone penetrometer. Under favour-
able conditions, ambient temperature change during a penetration phase will be limited to a few degrees.
Note that near-surface onshore soils will typically show non-uniform temperature versus depth. A tem-
perature gradient in the order of 0.04oC/m may be expected for deep penetration. An adverse setting may
show ambient temperature change in the order of 20oC. Transient temperature gradients will be induced
by self-heating of a cone penetrometer and may result from frictional heat during cone penetration and
from thermal flux through push rods. Under favourable conditions, transient temperature change during
a penetration phase will be close to zero. Uniform soft clays provide a favourable setting. An adverse
setting may show transient temperature change in the order of 25oC. Transient temperature variations
exceeding 100oC may possibly occur for cone resistance values of >120 MPa during penetration of
dense sands (Post & Nebbeling 1995, Peuchen 2012).
Lunne et al. (1986) reported laboratory comparisons of ambient thermal zero drift for 12 cone pene-
trometers. Most, if not all, of the penetrometers were instrumented for temperature compensation. The
measurements showed qc zero drift ranging between -200 kPa to +760 kPa for T = 25oC. The better
penetrometers (6 of 12) showed qc zero drift between +46 kPa and +100 kPa. Fugro in-house experi-
ments conducted in 2011 and 2012 on penetrometers from three suppliers, including Fugro, showed val-
ues and spread comparable to the 1986 study results. Zero drift is approximately linear with temperature
(Boylan et al. 2008). Temperature-induced drift under in situ stress pressures and axial loading is proba-
bly of similar magnitude to zero drift at atmospheric conditions.
Ambient temperature variation also changes the sensitivity of a load cell. Schaap & Zuidberg (1982)
give explanations. They reported <1% change for T = 20oC. Fugro in-house experiments conducted in
2004 for three Fugro penetrometer types confirmed this value for a temperature drop from +20oC to -
20oC: -1.4%/40oC, -1.5%/40oC and -1.9%/40oC. The change is proportional to temperature change. Cor-
rections can be applied. To the knowledge of the authors, this is not done in practice.
Boylan et al. (2008) inferred transient temperature influence from laboratory data. Figure 8 shows a
laboratory example of variation in qc for an unloaded penetrometer at an initial temperature of 20oC be-
ing immersed in water of 5oC. Note that a period of 250 s corresponds to 5 m continuous penetration.
Boylan et al. also discuss in situ results for a cone penetrometer at an initial temperature of about 11oC
pushed into soil with a temperature of 7oC at 1 m depth. The results are compared to in situ results from
a similar cone penetrometer at an initial temperature of about 7oC. The comparison was expected to lead
to a qc offset of up to about 50 kPa based on laboratory simulations. Boylan et al. quoted 170 kPa offset
for the in situ measurements and speculated about reasons for this higher value.

14
 50
T = 15C, aft er Boylan et al. (2008)
Cone Resistance Output [kPa]

-50

-100

-150

-200
0 50 100 150 200 250
T ime [s]

Figure 8. Laboratory variation of qc upon transient temperature differential (Boylan et al. 2008).

Figure 9 presents ambient temperature influence on cone resistance. The results are for laboratory
conditions, no axial loading, with and without temperature compensation. The case of no temperature
compensation simulates adverse conditions that may apply in situ upon intense variation in transient
temperatures in a penetrometer.

200
Cone Resistance Output [kPa]

-200

-400
Compensat ed 1, McCallum (2012)
Compensat ed 2, McCallum (2012)
-600
Uncompensated, Schaap & Zuidberg (1982)
Compensat ed, Schaap & Zuidberg (1982)
-800

-20 -10 0 10 20 30 40 50
o
T emperat ure [ C]

Figure 9. Zero drift of cone load cell caused by ambient temperature change.

Strain gauges and electronics inside a cone penetrometer dissipate power that will lead to heating of
load cells. In addition, compression of steel will lead to very small energy dissipation. Resulting transi-
ent temperature effects are assessed to be insignificant for a semi-continuous penetration phase of a
CPT. Effects may possibly apply after a long interruption, for example after a pore pressure dissipation
test. The authors are not aware of dedicated experiments.
Post & Nebbeling (1995) focused on frictional heat. They presented results of field measurements
and laboratory simulation for transient temperature effects on cone resistance. The measurements were
performed with a Fugro subtraction-type cone penetrometer, routinely instrumented for ambient temper-
ature compensation and specifically equipped with three internal temperature sensors. Post & Nebbeling
inferred penetrometer heating in the order of 1oC per 1 MPa increase in cone resistance. This is for an
assumed linear relationship between temperature and qc. The relationship depends on penetrometer type.
Furthermore, Post & Nebbeling reported a temporary shift in cone resistance output of about 130 kPa for

15
penetration of a sand layer with a cone resistance of about 25 MPa. This adverse setting represents a
0.5% error for qc = 25 MPa and a 26% error for qc = 0.5 MPa, for example for any underlying soft clay.
Heat flux from push rods to a cone penetrometer may occur under adverse conditions. This scenario
is generally secondary to dominating effects of frictional heat on a cone penetrometer. It is ignored in
practice.
Transient temperature influence may be mitigated by operational adjustments that cause a more or
less ambient temperature setting in the cone penetrometer. An example would be a penetration interrup-
tion of around 5 minutes before cone penetration from dense sand into soft clay (Post & Nebbeling
1995). It would seem reasonable to mitigate temperature influence by measuring temperatures in load
cells and correcting measured data based on laboratory simulation. However, this is complex because of
temperature gradients T with respect to time t (T/t >>0) acting concurrently with temperature gradi-
ents with respect to space x, y, z (T /x, T /y, T /z >0). The authors are not aware of any successful
applications of this approach for strain gauge technology for cone penetrometers. In practice, this means
that an error prediction model according to ISO 22476-1 should account for substantial epistemic uncer-
tainty for heat accumulation/dissipation and for load cell response to temperature versus time. For ex-
ample, the prediction model would consider (1) conservative heat accumulation/ dissipation and (2)
maximum temperature bias (Fig. 8) as input for uncertainty.

6. SLEEVE FRICTION
6.1 Geometry

ISO 22476-1 requires a nominal perimeter surface area of 15,000 mm2 for the friction sleeve. This is for
a cone penetrometer with a cross sectional areas of 1,000 mm2 and implies a length to diameter ratio of
3.75 for the friction sleeve. ISO 22476-1 permits ratios of 3 to 5. ASTM D5778 considers cone pene-
trometers with cross sectional areas of 1,000 mm2 and 1,500 mm2, with nominal surface areas of
15,000 mm2 and 22,500 mm2 for the friction sleeve, respectively. Nominal surface areas of 20,000 mm2
to 30,000 mm2 are permitted for a cone penetrometer with a cross sectional area of 1,500 mm2. Scale ef-
fects will inevitably influence CPT results, as discussed for cone resistance. The following sections con-
sider cone penetrometers with a nominal perimeter surface area of 15,000 mm2.
Measured sleeve friction fs is derived from the perimeter surface area of the friction sleeve and the
axial force acting on the friction sleeve. The perimeter surface area is the cylindrical area. This excludes
any machined bevelled ends of the friction sleeve. Measurement of surface area is typically done by
measuring diameter and length by vernier calliper. Measurement uncertainty will be similar to that for
measurement of cross sectional area of the cone.
ISO 22476-1 and ASTM D5778 prescribe tolerances for geometry of the friction sleeve. The authors
are not aware of specific studies that quantify the effects of geometry tolerances. Tentative estimates are
given below.
In practice, the diameter of a new friction sleeve will be at a permissible (steel) geometry that pro-
vides maximum wear before replacement, i.e. 36.1 mm diameter and 135.0 mm length for a nominal
1,000 mm2 cone penetrometer (ISO 22476-1), giving a surface area that is 2.1% larger than nominal
(15,000 mm2). Lower limits of 35.3 mm and 132.5 mm are permissible in theory. Wear to these limits
will not be concurrently achieved in practice. Limits of 35.3 mm and 134 mm would give a surface area
that is about 1% smaller than nominal. ASTM D5778 prescribes 15,000 mm2 2%. CPT processing
software may allow entry of measured surface area.
Tolerances not only affect surface area but also, for example, cross-sectional areas of the ends of the
friction sleeve and cylindrical shape of the friction sleeve. As for cone resistance, ISO 22476-1 appears
ambiguous on whether such geometry variations may be ignored for estimates of accuracy.

16
 The effects of the cross-sectional areas may consider soil flow around the gaps below and above the
friction sleeve and a friction sleeve diameter up to 0.35 mm larger than the cone, a tolerance set by ISO
22476-1 and ASTM D5778. Tentative estimates may be inferred from considering cone resistance to ap-
ply to this 0.35 mm annulus area at the bottom end of the friction sleeve, i.e. assuming 1x cone re-
sistance would include reverse end bearing at the top of the friction sleeve. An initial friction ratio Rf of
3% for typical clay soil would then result in about 4.4% increase of fs compared to sleeve friction with
no cross-sectional area effects. Re-calculated Rf would be 3.1%. For clean sand, the increase in fs would
be about 19% for a friction ratio of 0.7%. Re-calculated Rf would be 0.83%. The 19% value contributes
substantially to a 55% value that triggered geometry studies by Jekel (1988) discussed below. Cemented
soils may show very high fs peaks. Damage to a cone penetrometer may result.
Wear of a friction sleeve will occur during penetration and during retraction. Schaap & Zuidberg
(1982) and Jekel (1988) reported studies of the mechanical wear of cone penetrometers and implications
for robust cone penetrometer design (Figs 10, 11).

Figure 10. Observed geometry after wear (in sand) of three different cone penetrometers (Jekel 1988).

Figure 11. Surface roughness of friction sleeves observed after use (Schaap & Zuidberg 1982).

17
 Jekel (1988) observed non-uniform wear of friction sleeves of cone penetrometers pushed into sand
strata. Figure 10 illustrates a diameter reduction from the bottom to the middle of the friction sleeve. The
reduced diameter is approximately constant for the upper part of the friction sleeve. Such wear will lead
to lower sleeve friction because of reduced radial stresses on a significant length of the friction sleeve,
compared to a nominal cylindrical shape. The influence of non-uniform wear is probably secondary to
effects of cross-sectional areas of the ends of the friction sleeve. The authors tentatively suggest that ad-
verse non-uniform wear may possibly lead to 10% reduction in sleeve friction. Note that ASTM D5778
limits the diameter of the top of the friction sleeve: it must not be smaller than the bottom diameter.
ASTM D5778 also prescribes friction sleeves designed for equal end areas. These combined require-
ments leave practice with little opportunity to mitigate Jekels observations with respect to a note by
ASTM D5778: Normally, the top of the sleeve will wear faster than the bottom. Adverse ground con-
ditions may imply replacement or machining of a friction sleeve after each test or a few tests in harsh
ground conditions.
Effects of variations in surface roughness of the friction sleeve may be expected to be higher than for
cone resistance (Jekel 1988, Uesugi & Kishida 1986, DeJong & Frost 2002). The authors are not aware
of study results specifically aimed at the requirements of ISO 22476-1 and ASTM D5778, i.e. an initial
average surface roughness Ra of 0.4 m +/- 0.25 m upon manufacture. In principle, sleeve friction will
depend on effective soil particle size/shape at the soil-steel interface relative to surface roughness. Parti-
cle size/shape in the interface zone may be influenced by particle breakdown and smearing upon cone
penetration. In an extreme case, clay may adhere to a friction sleeve. A penetration interruption may
promote clay adherence to a friction sleeve, as discussed for cone resistance. Clay adherence causes soil-
soil shear away from the friction sleeve instead of soil-steel shear at the surface of the friction sleeve. A
change in friction coefficient is likely. This change may cause unusually high sleeve friction. The influ-
ence of an increase in surface area is likely to be minor. If the increase in surface area is considered on-
ly, a uniform increase of 1 mm diameter would result in approximately 3% increase in measured sleeve
friction. Additional end effects would account for an additional sleeve friction in the order of 13% for
the unlikely case that clay would only adhere to the friction sleeve and not to the cone and not to the
shaft.
Sleeve friction will be affected by two gaps: (1) gap between the cone and the bottom of the friction
sleeve and (2) gap between the top of the friction sleeve and the shaft. The earlier discussion on cone re-
sistance presents principles on gap influence.
Note that the bottom gap volume depends on relative displacements between the cone and the friction
sleeve, each subject to variable axial forces. The volume of the top gap depends on friction sleeve dis-
placement relative to the shaft of the cone penetrometer. The shaft may be regarded as fixed. Gap vol-
ume also changes by pressure, notably because of compression and displacement of the water seal.
Soil-soil shear forces acting in an axial direction on the bottom gap will largely transfer to the friction
sleeve. A tentative estimate for increase in fs would be about 3.7%. This is for a gap height of 5 mm and
a soil-soil shear resistance assumed equal to fs. Soil-soil shear forces acting on the top gap may be ig-
nored. They will largely transfer to the shaft, apart from some Poisson's effects.
Water and gas in the gaps affect measured sleeve friction. Similar to cone resistance, this may be ex-
plained by means of a modified equation for ft, corrected sleeve friction:

f t = fs
(u 2g A sb u 3g A st ) (1)
As

where Asb is the cross sectional area in the gap between the friction sleeve and the cone, Ast is the cross
sectional area in the gap above the friction sleeve and As is the surface area of the friction sleeve.

18
 Note that the above equation assumes a cylindrical friction sleeve. For a non-cylindrical friction
sleeve, water and gas pressures will also act on sections in-between the top and bottom of the friction
sleeve. This effect is estimated to be secondary and may be ignored in practice. Equal end areas provide
a reasonable compromise on gap influence on sleeve friction. ISO 22476-1 is not prescriptive on equal
end areas and provides guidance on how to deal with their influence on sleeve friction. ASTM D5778
requires friction sleeve design with equal end areas (Asb = Ast), which fits a prescriptive approach.
ASTM D5778 states: This will remove the tendency for unbalanced end forces to act on the sleeve.
This probably holds for CPTs in highly permeable soils and for fortuitous situations with u2g = u3g.
Sleeve friction fs will be higher in soils with u2g > u3g and lower for u2g < u3g, compared to fs = ft for the
Analytic CPT.
Water and gas pressures will give an apparent increase in fs for equal end areas of the friction sleeve
and normally consolidated to slightly overconsolidated clay. For example, fs would be about 1.125 ft for
u2g/fs = 50, u2g = 1.2 u3g, Asb/As = Ast/As = 0.015 and favourable conditions, i.e. pore water saturation in
the gaps. For this case, it should be feasible to correct fs to ft to an uncertainty in the order of 5% of ft.
Estimation of ft may consider u3/u2 data processing relationships based on soil behaviour type, similar to
u1/u2 relationships proposed by Peuchen et al. (2010). Note that As, Asb and Ast will change upon wear of
the friction sleeve. This provides additional uncertainty if permissible geometry tolerances are consid-
ered with no test-specific measurement of geometry.
Lunne (1986) studied cone penetrometers with (Asb - Ast) / As ranging from -0.019 to +0.019. A posi-
tive value for (Asb - Ast) will show an apparent increase in fs compared to a friction sleeve with equal end
areas.

6.2 Measuring range

In practice, measuring range for sleeve friction will be designed to match that of cone resistance. This is
typically no issue for subtraction-type cone penetrometers. For independent-type cone penetrometers,
this would imply a measuring range for fs of, say, 0 to 1 MPa matching a typical nominal measuring
range for qc of 0 to 50 MPa and a nominal friction ratio of 2%. The corresponding axial forces would be
15 kN for the friction sleeve and 50 kN for the cone.
For offshore use, consideration should be given to force required to overcome (sea)water pressure
acting on any unequal end areas of a friction sleeve. For an independent-type cone penetrometer, tension
forces in a load cell will be induced if Asb < Ast. Tension forces are unlikely for a subtraction-type cone
penetrometer.

6.3 Force measurement

Principles on force measurement for sleeve friction are as for cone resistance. Independent-type cone
penetrometers derive sleeve friction from either compressive or tensile force measurement. Subtraction-
type cone penetrometers warrant additional comments: addition and subtraction (compensation) of un-
certainties apply to compressive forces measured by two load cells in series. Laboratory uncertainty for
fs may then be expressed as:

b QC 2 + QC+F 2 (2)

where b is Ac/As and QC and QC+F are force uncertainties for the C and C+F load cells. A well-
designed subtraction-type cone penetrometer shows very small or zero subtraction output under equal
loading of the C and C+F load cells. Equal loading applies to routine practice in a calibration laboratory.

19
During in situ testing, the force on the C+F load cell will exceed the C force. The two load cells will
then be at different points of a calibration curve. Hysteresis for the two load cells will be concurrent,
leading to subtraction of the two hysteresis errors. For a well-designed penetrometer, it is estimated that
the resulting error for fs will be less than about 1.5% under favourable conditions and about 3% un-
der adverse conditions.
As for cone resistance, laboratory-type error for axial loading increases with applied force. Uncertain-
ty analysis of actual CPTs may incorporate a prediction model for such error.
Figure 7 illustrates influence of non-axial loading on measured sleeve friction. The examples high-
light benefits from double-bridge strain gauge technology. Note that a subtraction-type cone penetrome-
ter may show improved force measurement for sleeve friction under non-axial loading compared to
force measurement for cone resistance. This improvement is possible when applying circumferential
alignment of the C and C+F strain gauges, causing partial compensation of errors.

6.4 Temperature

The earlier discussion on cone resistance provides the principles for temperature influence on sleeve
friction, particularly for independent-type cone penetrometers. Note that temperature influence results in
an apparent force in a load cell. The nominal ratio of sleeve friction area to cross-sectional area of the
cone is 15. If independent qc and fs load cells show equal influence and if Rf is 2%, then the relative er-
ror for fs would be 3.3 times that for cone resistance.
Sensitivity of fs to temperature effects is typically better than 15 kPa for T = 50oC (from -10C to
+40C) for a well-designed penetrometer.
As discussed for cone resistance, an error prediction model for transient temperature influence should
allow for substantial epistemic uncertainty for heat accumulation/ dissipation and for load cell response
to temperature versus time. An error prediction model for a subtraction-type cone penetrometer may
possibly allow for uncertainty reduction because of subtraction of concurrent uncertainties. For example,
predicted bias for cone resistance would apply almost concurrently with that for sleeve friction. Such
uncertainty reduction does not apply to independent-type cone penetrometers.

7. PORE PRESSURE
7.1 Geometry

ISO 22476-1 and ASTM D5778 aim at measuring in situ fluid pressure during penetration. ISO 22476-1
uses the term pore pressure. ASTM D5778 defines pore-water pressure. The requirement for measuring
in situ fluid pressure implies isolating fluid pressure from soil inter-particle stresses and influence of any
gas in soil. In practice, most piezocone penetrometers have measuring systems comprising an internal
membrane-type pressure sensor, internal saturation fluid for transmission of pressure and a cylindrical
porous filter positioned in a recess of the cone or shaft. One of the functions of the filter element is miti-
gation of mechanical damage to the pressure sensor. Further discussion will be limited to membrane-
type pressure sensors and cylindrical porous filters, i.e. will exclude slot filters (Elmgren 1995) and radi-
al button filters (Peuchen 1998).
Figure 2 presents positions for measured pore pressure. ISO 22476-1 and ASTM D5778 consider u1,
u2 and u3, with u2 commonly quoted as standard, preferred, recommended or reference position. For pore
pressure u2, ISO 22476-1 defines: u2 - pore pressure measured on the cylindrical section of the cone
(preferably in the gap between the cone and the sleeve) and recommends: To correct for pore pressure
effects on cone resistance the filter element should be located in the gap between the cone and the fric-
tion sleeve. Since this is not possible in practice, the filter should be located in the cylindrical part of the

20
cone as close as possible to the gap. The u2 position points toward two objectives: (1) characterisation
of soil behaviour and (2) corrections for cone resistance and sleeve friction related to the gap between
the cone and the friction sleeve. The second objective can only be partially met in practice, except under
favourable conditions as discussed above for cone resistance. Inherent difficulties with the second objec-
tive are often ignored in practice, potentially leading to unnecessary loss of information. For example,
onshore characterisation of highly stratified or overconsolidated sands and clays would typically im-
prove when using u1 data compared to u2 data. In these conditions, a pore pressure measuring system at
the u2 position can lose saturation where the u1 position would not. Loss of saturation leads to sluggish
pore pressure response and reliable fluid pressure data may not be consistently obtained (Fig. 12).

qc [MP a] u2 [MP a]

0 10 20 30 40 0 0.2 0.4 0.6 0.8

12

13 T est 1
T est 2
14
uo
Penetration Depth [m]

15

16

17

18

19

20

21

22

Figure 12. Pore pressure response for two closely spaced CPTs.

Robertson et al. (1986) illustrated typical pore water pressure distribution at the perimeter of cone pene-
trometers pushed in saturated soils (Fig. 13). Small variations in the precise vertical u1 and u3 positions
will usually lead to minor differences in recorded pore pressures. Small variations in the u2 position can
lead to major differences in recorded pore pressure during penetration, particularly in sensitive clays and
heavily overconsolidated clays. This implies high dependence of u2 on penetrometer design (Peuchen et
al. 2010).

21
Figure 13. Pore pressure distribution in saturated soil (adapted from Robertson et al. 1986).

ISO 22476-1 and ASTM D5778 provide no specific requirements for external surface area of the fil-
ter element. Cone geometry and resistance against penetrometer wear provide indirect limitations. Sur-
face areas for most piezocone penetrometers are determined by filter heights between about 3 mm and
6 mm.
The external surface of the filter element will not be perfectly flush with the steel of the cone or shaft
and its surface roughness will differ from that of steel. ISO 22476-1 specifies external geometric toler-
ances for filter element diameter. ASTM D5778 does not. The ISO 22476-1 tolerances are relative to di-
ameters of the cone and the friction sleeve. They apply to the start of a test, as geometry of a filter ele-
ment may change during a test as a result of soil/fluid/gas pressures, shear and friction-induced
temperature. Inevitably, a filter element represents a discontinuity in a flow field of soil around the cone
penetrometer. Discontinuities represent stress concentrations. Micro fractures and localised shear zones
cannot be excluded. This should generally be of little interest, but may possibly be important for any re-
distribution of dissolved or free gas in soil, as discussed below. Dissolved gas is defined here as gas dis-
solved in pore water with no occurrence of free gas. Free gas is defined as dissolved gas in pore water
plus gas contained in bubbles or vacuoles. The occurrence of dissolved gas and free gas depends on fac-
tors such as gas concentration and solubility in pore water and ambient pressure. Methane and carbon
dioxide are common types of gas in shallow soils below the ground water table.
Pore pressure may be affected by surface roughness of a u1 or u2 filter element, compared to steel of
the cone. Filter diameter and roughness will also affect cone resistance and sleeve friction. The section

22
on cone resistance provides comments. The authors are not aware of specific studies on these geometry
topics. It is tentatively estimated that these influences are secondary to the precise u1, u2 or u3 positions
discussed above and material type discussed below.
The response time of a pore pressure measuring system will be affected by shape and volume of the
internal space between the pressure sensor and the outside perimeter area of the filter element. Shape
and volume affect initial fluid saturation of internal voids, retention of fluid saturation during penetra-
tion, fluid displacement required for pressure measurement and fluid displacement for accommodating
inevitable system compliance effects such as deformation of the steel of the cone and the filter element
material.
ISO 22476-1 and ASTM D5778 provide guidance on filter material types, filter pore sizes and fluid
types for initial saturation of the measuring system. Choice of filter material type represents a compro-
mise of competing requirements on geometric stability, robustness and pore size. The requirements for
pore size typically compete on separation of in situ soil from saturation fluid, flow characteristics of sat-
uration fluid and mitigation of formation of free gas within the pore pressure measuring system. Con-
ventional Fugro practice is to use (ductile) high-density polyethylene HDPE cylindrical filter elements
with an initial average pore size of 10 m (Fig. 14). This type of filter material is robust and maintains
its geometry reasonably well under favourable conditions, notably soft soil. Filter material compression
and filter element deformation will inevitably occur upon penetration of competent soils. This will
change pore size of the filter element and fluid flow paths at the interface of the filter element and its
steel recess. Fugro trials with ceramic filter elements under adverse conditions showed frequent cracking
or complete absence of the filter element upon recovery of the penetrometer. Trials with sintered steel
showed a decrease in magnitude response due to pressure changes, compared to HDPE or ceramic fil-
ters.

HDPE Stainless Steel Ceramic

Figure 14. Filter elements (u2) after use.

7.2 Measuring range

A piezocone penetrometer typically incorporates an off-the-shelf pressure sensor. The sensor housing
may allow selection between sensors with different measuring ranges. Adverse conditions will typically
require a measuring range of -0.1 MPa to 5 MPa. The -0.1 MPa limit relates to cavitation taking place at
negative one atmosphere. Exceptionally harsh ground conditions may require a pressure sensor of
>10 MPa.
A larger measuring range may be required for offshore use. This is because of additional (sea)water
pressure. For example, a water depth of 100 m corresponds to a pressure of about 1 MPa. Some offshore
CPT systems allow use of a differential pressure sensor that negates (sea)water pressure.

23
7.3 Pressure measurement

A membrane-type pressure sensor is similar to a load cell for force measurement in that it relies on strain
gauge technology and typically incorporates compensation for ambient temperature variation. A well-
designed pressure measurement system differs favourably from axial force measurement with respect to
absence of influences from bending stresses and inner friction caused by water pressure seals. A high-
quality pressure sensor may show laboratory uncertainty in the order of [1 kPa + 0.002 u +
0.0025 umax] where umax is the maximum encountered hysteresis value of u prior to reaching penetration
depth.
As for cone resistance, laboratory-type error for pressure measurement increases with applied pres-
sure. Uncertainty analysis of actual CPTs may incorporate a prediction model for such error.
Routine laboratory calibration takes place in a pressure vessel filled with tap water, with no filter fit-
ted. This approach inevitably provides only partial simulation of in situ conditions. However, some in-
formation on any cross-talk for pressure measurement can be obtained by monitoring zero-load output of
the pressure sensor during axial force calibration of a cone penetrometer.
Measuring in situ fluid pressure is a major challenge in practice. What is measured may comply with
ISO 22476-1 or ASTM D5778, yet can fail to meet intentions. Particularly, accurate measurement of flu-
id pressure may not be feasible under adverse conditions that are not avoidable in practice. Examples of
adverse conditions are: (1) soil with free gas, where induced fluid pressures at the measuring location
(u1, u2 or u3) are less than required for bringing free gas into solution and (2) cone penetration into soil
with dissolved gas, where induced fluid pressures at the measuring location are lower than equilibrium
in situ pore pressures. The first situation may occur locally, for example near ground surface or within a
permeable soil layer with free gas in-between impermeable layers with no free gas. The second situation
refers to exsolution where dissolved gas forms free gas. Gas diffusion may be ignored as a possible ad-
verse situation. This is because of the relatively short time required for a CPT compared to a diffusion
process.
Response time may generally be regarded as immediate for a measuring system with saturated fluid
and no free gas (Rad 2003). However, rapid variations in pore pressure will cause inevitable expulsion
of saturation fluid and entry of in situ pore fluid or gas, as discussed above. Low permeability of soil
near the filter/soil interface may delay entry of fluid and cause a lag in pore pressure response.
Ideally, a pore pressure measuring system retains its initial fluid saturation during rapid variations in
pore pressure or when locally encountering free gas. A u1 position has the better prospects. Entry of free
gas can affect measured pressure and response time. Laboratory studies by Nageswaran (1983) suggest-
ed that the error of the measured pressure should be less than 5 kPa for a well-designed pore pressure
measuring system, even if the system becomes sufficiently gas charged for discontinuous fluid passage.
Observations of in situ test data suggest possible error exceeding 5 kPa. Effects of free gas on response
time depend on the characteristics of the pore pressure measuring system and the soil conditions. A re-
sponse delay of less than a few seconds will often remain unnoticed. Sluggish response may become vis-
ible in uniform soils, for example upon regain of saturation during further penetration or during a pene-
tration interruption.
If critical, loss of saturation of the pore pressure measuring system may be mitigated by an additional
CPT, preceded by pre-drilling or pre-pushing to below the soil zone causing loss of saturation.

7.4 Temperature

A high-quality pressure sensor has built-in temperature compensation. Its thermal stability compares fa-
vourably to that of a load cell. A thermal stability in the order of 0.5 kPa/C or better is feasible under
ambient temperature variation. Response time to transient temperature variation is assessed to be in the

24
order of seconds. This is because of small size and mass of a pressure sensor, which limit thermal differ-
entials.

8. INCLINATION

ASTM D5778 recommends measurement of inclination i, but provides no requirements. ISO 22476-1
limits inclination of the penetrometer to 15o from vertical and provides accuracy requirements of 2o for
Application Classes 1 and 2 and 5o for Application Class 3. Off-the-shelf inclinometers should meet the-
se requirements under adverse conditions including transient temperature variation. Fixing the verticality
of an inclinometer in a cone penetrometer is probably the principal source of uncertainty. A typical
measurement uncertainty in the order of 1o can be achieved in practice. Omni-directional inclinometers
and bi-directional inclinometers are common. Bi-directional inclinometers provide some opportunity for
mapping horizontal position of a cone penetrometer. Gyro measurement may be necessary for accurate
horizontal position, for cases with limited control of rotation of push rods. Figure 6 includes a spatial tra-
jectory for a CPT profile in terms of lateral deflection.

Directional drift cannot be avoided. Comments are as follows.

Engineered ground and young sedimentary deposits often exhibit soil layering that is close to hori-
zontal. Inclined sub-bedding may still apply, for example as a result of spigoting practice, wind or
wave action. Close-to-vertical penetration can often be observed.
Older sedimentary deposits and residual soils may show inclined bedding and associated soil anisot-
ropy. A cone penetrometer will incline toward a path of least resistance, which may not be vertical.
Non-axial loading of a cone penetrometer can be expected in soils with significant heterogeneity at,
say, 10 mm to 100 mm scale. Examples are locally cemented soils and soils containing gravel-size
particles and larger.
The conical shape of a cone penetrometer contributes to directional stability, albeit probably less than
a bullet shape.

9. SPATIAL POSITION
9.1 Horizontal position

CPT parameters such as cone resistance require a geodetic position xyz in space and time. Ideally, accu-
racy of spatial position is coupled to accuracy of a CPT parameter. In practice, this dependency is left to
geotechnical judgement.
ISO 22476-1 and ASTM D5778 provide no accuracy requirements for horizontal position xy. Typical
uncertainties for xy position of an antenna of a location-specific GNSS global navigation satellite system
would be 1 m to 2 m for DGPS differential global positioning system, 0.2 m for high-accuracy
DGPS and 0.05 m for real time kinematic RTK DGPS. Additional uncertainty will apply because of
uncertainty in antenna offset to the vertical axis of the thrust machine and because of directional drift of
the cone penetrometer. An upper limit for horizontal drift may be derived from measurement of omni-
directional inclination and assuming that the cone penetrometer is following a single horizontal direc-
tion. A multi-direction trajectory would imply a penetrometer position closer to the axis of the thrust
machine than calculated.

25
9.2 Penetration depth

Users of CPT data typically have more interest in penetration depth z than horizontal position. This is
because of many soils showing small correlation distances for the vertical direction compared to hori-
zontal directions. ISO 22476-1 defines penetration depth as vertical depth of the base of the cone, rela-
tive to a fixed point. ISO 22476-1 provides requirements for penetration length l defined as sum of the
lengths of the push rods and the cone penetrometer, reduced by the height of the conical part, relative to
a fixed horizontal plane. The accuracy requirements are 0.1 m or 1% of penetration length and 0.2 m or
2% (Fig. 18). This implies transition points from a fixed value to a percentage value at penetration
lengths of 10 m and 20 m respectively. It may be noted that the combined ISO 22476-1 requirements for
accuracy of penetration length and inclination imply a permissible uncertainty for penetration depth that
increases with penetrometer inclination. ASTM D5778 requires a depth accuracy of 0.1 m, where
depth probably has the ISO 22476-1 meaning of penetration length and accuracy probably has the
meaning of bias. Note that ASTM D5778 also uses the terms sounding depth, vertical depth and penetra-
tion depth.
Accuracy estimates for penetration depth depend on tool-specific deployment features, which typical-
ly focus on achieving maximum penetration at lowest cost (Table 3).

Table 3. Achieving penetration.

Practice Freq. Notes
Increasing thrust range
Allowable push from thrust machine ***** Increased push requires additional considerations for anchoring/ ballast-
ing/ suction resistance (offshore) of thrust machine; buckling of push rods
may become limiting factor particularly for soft-over-hard ground condi-
tions
Measuring range for penetrometer ***** Cone penetrometers used in practice typically match push capacity of
force and pressure sensors common thrust machines and allow penetration into a wide range of soil
conditions
Reducing friction along push rods
Casing installation (drilling, free fall *** Good track record; frequently applied in offshore cone penetration testing
etc.) and short-push penetration from
bottom of casing or drill pipe
Drill-out or full displacement hole for- ** Residual friction on push rods remains; use of low-friction spacers in cas-
mation, followed by installation of cas- ing can largely eliminate casing-push rod friction; external soil-casing re-
ing and push from ground surface sistance may reduce ballasting or anchoring requirements for thrust ma-
chine
Continuous drill-out with fixed-length * Difficult to achieve in practice because of control required on penetration
push rod latched in drill bit rate
Ground loosening or full displacement * Residual friction on push rods remains; percussion penetration of pene-
hole formation, followed by push from trometer can allow by-passing of competent zones at expense of no CPT
ground surface data
Cyclic displacement of push rods * Residual friction on push rods remains; can be effective in combination
with friction breaker
Friction breaker at short distance above * Effectiveness approximates that of 1500 mm2 cone penetrometer with
cone penetrometer 1000 mm2 push rods
1500 mm2 cone penetrometer with *** Routinely used in practice for deep surface-push deployment; 1500 mm2
1000 mm2 push rods penetrometer acts as friction breaker for 1000 mm2 push rods
500 mm2 cone penetrometer ** Routinely used in practice for short or shallow penetration; push rod con-
siderations dominate for deep surface-push deployment
Low-friction fluid or gas injection into * Some reported successes; appears to have some potential for narrow range
soil above cone penetrometer of soil conditions; increased deployment complexity
Change of in situ soil conditions above * For example change of in situ soil by high pressure water jetting, osmosis,
cone penetrometer thermal processes, vibratory liquefaction; the authors are not aware of use
in practice

26
Practice Freq. Notes
Temporary and local change of laws of * Imagination
physics
Frequency of use: * = less than 0.01% of CPTs, ** = between 0.01% and 0.1% of CPTs, *** = between 0.1% and 1% of
CPTs, **** = between 1% and 10% of CPTs; ***** = more than 10% of CPTs.

Uncertainty assessments for penetration depth require consideration of uncertainty in establishing a

fixed horizontal plane and uncertainty in the actual trajectory of a cone penetrometer, relative to this hor-
izontal plane. The following section discusses onshore CPTs with direct push from ground surface. ISO
(2012b) presents guidance on penetration depths for CPTs deployed from a jack-up platform and from a
floating vessel.
Ideally, the horizontal plane is a geodetic geoid, accurately fixed to a reference point for the push
rods. Location-specific GNSS can provide information in real time to an uncertainty of about 10 mm
to 20 mm (in real time) or better (static measurements). In routine practice, a horizontal plane is fre-
quently determined relative to location-specific ground surface and local geodetic monuments. For many
countries, uncertainty for elevation relative to geodetic monuments is probably in the range 5 mm to
10 mm. Uncertainty of elevation measurement relative to location-specific ground surface depends on
factors such as ground surface slope, surface smoothness and ground conditions. Slope and smoothness
relate particularly to offsets in horizontal positioning: a slight offset in horizontal positioning relative to
a target position may render a higher or lower horizontal plane. Ground conditions can affect uncertainty
for the horizontal plane by, for example, vertical displacement of a ballasted thrust machine operating on
soft ground. Such displacement will take place during initial positioning of the thrust machine and dur-
ing the test phase with variable force transferred from the thrust machine to the push rods. Measurement
of and correction for such displacements are feasible, albeit infrequently performed in routine practice.
In practice, uncertainty for a fixed horizontal plane is assessed to be in the order of 20 mm for favour-
able conditions. Uncertainty values exceeding 100 mm may be expected for unfavourable conditions.
Measurement of penetration length the trajectory of a cone penetrometer typically includes an incli-
nometer in the cone penetrometer and one or more of the following tools:
Tape measure
Mechanical distance sensor
Optical distance sensor
Fluid-flow distance sensor
Acoustic distance sensor.

Length or distance measurement is typically between a push rod above ground surface and a fixed
point on the thrust machine. Adverse conditions may impact length measurement, particularly: a) verti-
cal movements of the thrust machine relative to a fixed horizontal plane, b) length variation and bending
of the push-rod string.
Vertical movements of the thrust machine depend on system stiffness and ground conditions. For ex-
ample, stiffness, including any slack, of a thrust machine can be important for an anchored system. Soft
ground can give an adverse setting for a thrust machine relying on ballast. The thrust machine may ini-
tially settle. Heave may subsequently occur upon transfer of force from the thrust machine to the push
rods. Settlement of the ground surface by force transfer to the push rods is assessed to be negligible for
common adverse conditions. Force transfer will predominantly relieve the initial pressure imposed by
the thrust machine on the ground surface. Equivalent stiffness may be in the order of 2.5 kN/m for ad-
verse conditions.
Length variation and bending-type deformation of the push-rod string may take place when a cone
penetrometer reaches a competent zone. Axial stiffness is assessed to be in the order of 1.6105 kN/m of
initial length of a push rod string consisting of 36 mm diameter, 1 m long thick-walled push rods. This
implies push rod shortening of 0.06 % for a uniform axial force of 100 kN. It is assessed that bending
27
can possibly result in similar error under adverse conditions. Examples of adverse conditions are a later-
ally unsupported section and soft soil with limited lateral support. Thermal effects are assessed to be
negligible.
It may be inferred that a depth sensor on a thrust machine may incorrectly assume some non-existent
penetration upon build-up of soil resistance during cone penetration. Similarly, a penetrometer that
moves from a competent zone to a weak zone may show (very) rapid (unrecorded) penetration due to
compressive strain release. Some of these effects may be captured during a penetration interruption, for
example if required for addition of a push rod. NNI (1996) describes a so-called electronic heave com-
pensator for penetration interruptions. The tool provides data for correction for relative movements be-
tween the push rod string and the thrust machine.
It is assessed that CPTs with direct push from ground surface would typically provide trajectory data
to an uncertainty of 50 mm plus <1% of penetration length. These values are for favourable condi-
tions. For penetration depth, the value of <1% may increase by 0.5% for an average inclination of
the cone penetrometer of 4o, measured to 1o uncertainty. Values may double for adverse conditions.
Note that common interpretation of ISO 22476-1 requirements will typically assume zero uncertainty for
determination of a fixed horizontal plane. The value of 50 mm would then reduce to about 30 mm.
Accuracy improvement may be feasible by displacement modelling of a complete CPT system or se-
lected parts. In most cases, thrust applied to the push rods would be an important input parameter. Thrust
may be correlated to actual CPT measurements.

10. PENETRATION RATE

Penetration rate v is penetration length l normalised to (clock)time. ISO 22476-1 and ASTM D5778
specify permissible limits of 5 mm/s for a nominal penetration rate of 20 mm/s.
ISO 22476-1 requires data recording at lengths of 20 mm for Application Classes 1 and 2 and
50 mm for Application Classes 3 and 4. ASTM D5778 requires 50 mm. Increased frequency of data
recording can provide better stratigraphic detail and may provide opportunities for improved processing
of recorded data.
A typical thrust machine provides a push speed with an uncertainty within 5 mm/s under favourable
conditions. Under adverse conditions, penetration rates may be outside ISO 22476-1 and ASTM D5778
limits. The push speed may vary or reduce under adverse conditions, for example with strongly varying
thrust and towards the thrust limit of a thrust machine. In addition, the penetration rate is not equal to the
push speed because of vertical movements of the thrust machine and length variation and bending of the
push-rod string. The above section on penetration depth provides comments.
Influence of penetration rate on CPT parameters is particularly important where soil behaviour during
penetration shows domain shifts: drained, partially drained and undrained (DeJong et al. 2012). Fig-
ure 15 presents an extreme example for loss of penetration rate v at 74.1 m below ground surface. The
parameters u2 and Ic are excess pore pressure (u2 - u0) and soil behaviour type index according to Rob-
ertson & Wride (1998). The presented penetration rates include no post-processing for influence of data
point intervals.

28
 qt [MP a] ft [MP a] u [MP a] Ic [-] v [mm/s]
2

0 20 40 60 0 0.5 1 0 5 10 2 2.5 3 3.5 0 10 20 30

73.6

73.8

74
Penetration Depth [m]

74.2

74.4

74.6

Figure 15. Example of penetration rate affecting CPT parameters for chalk (Peuchen & Balthes 2013).

11. DATA ACQUISITION

A sensor requires stable or known electric input. Sensor output typically consists of a ratio of electrical
output versus input signal, expressed as mV/V or mV/mA. This signal is transmitted by wiring to an
electronic system that converts this analogue signal to a digital value, i.e. A/D conversion. Apart from
averaging or filtering, measurements expressed in digital counts are definitive after A/D conversion.
Excitation voltage and sensor output can be affected by deficiencies in the transmission system. Older
data acquisition systems will typically include a relatively long cable transmitting signals between the
cone penetrometer and electronics positioned on or near the thrust machine. Transmission may then be
influenced by external electric noise, moisture, temperature, mechanical strain and insulation issues for
wiring/cable etc. A voltage stabilizer and a signal amplifier in the cone penetrometer can mitigate such
influences. On-site checks on data acquisition influence can be made by a signal simulator replacing a
cone penetrometer. Older A/D conversion systems may have limits on resolution, for example 10 kPa on
cone resistance. Workarounds in practice may include built-in amplification in combination with 16bit
or 12bit A/D converters.
Newer systems incorporate electronics for excitation voltage and A/D conversion at or very close to
the measuring sensors, i.e. within the cone penetrometer. Digital data transfer to a recording unit then
takes place without loss of data quality. As discussed above for cone resistance, resolution presents no
limiting factor for 24bit A/D conversion.
Temperature variations in a cone penetrometer affect built-in electronics, such as the A/D converter.
Fugro experiments for cone resistance show an influence of about 0.001%/C plus a zero drift of
0.35 kPa/C. These values are an order of magnitude less than temperature sensitivity of a load cell or
pressure sensor.

12. CORRECTED, INFERRED AND NORMALISED PARAMETERS

ISO 22476-1 considers accuracy for measured parameters: qc, fs, u, i, l and v. ISO 22476-1, ASTM
D5778 and others also describe data processing for corrected, inferred and normalised parameters such

29
as friction ratio and pore pressure ratio. Uncertainty estimates for such parameters can be inferred from
the elementary considerations described above. The following paragraphs summarise selected topics.
Note that some corrections and normalisations require external data and their corresponding uncertainty
estimates. For example, net cone resistance qn requires an estimate of in situ vertical stress. An addition-
al uncertainty estimate is then required for the in situ vertical stress.
Many CPTs are without measurement of pore pressure u. The Netherlands probably hold a record for
CPT metres per person per year, yet pore pressure measurement applies to about 5% of the CPTs (Long
2010). This low percentage may be attributed to perceived cost/benefit, bias to end-bearing piles and
well-known soils that typically provide adverse conditions for pore pressure measurement, notably u2.
Nevertheless, an absence of u measurements should still allow fair confidence for derivation of correct-
ed, inferred and normalised parameters, particularly for well-known soils.
ISO 22476-1 and ASTM D5778 require graphical presentation of qc, fs and u data. In addition, ISO
22476-1 requires presentation of tabular qc, fs and u data for Application Classes 1 and 2. In practice,
presentation of these data may be versus time, length or depth intervals. Presentation at a fixed interval
may consider a nearest data record. Fixed-interval reporting may also rely on interpolation and averag-
ing techniques. The actual procedure will have little impact if CPT data are relatively uniform with
depth. Differences should be expected for strongly layered soils. For example, a peak in pore pressure
captured by high frequency recording may be missed or averaged by reporting to 50 mm length inter-
vals.
ISO 22476-1 provides no specific requirements for vertical shift of measured sleeve friction to pene-
tration depth z defined relative to the base of the cone. ASTM D5778 suggests the midpoint of the fric-
tion sleeve and the point of the cone. Consideration may be given to a matching cone resistance with
measured sleeve friction at a point slightly below the midpoint, allowing for a non-uniform distribution
of friction along the axis of the friction sleeve. Weighted averaging of multiple values of measured
sleeve friction is also used in practice.
A penetration interruption will lead to loss or change of penetration data, compared to continuous
penetration. Penetration rate will inevitable vary during the deceleration and acceleration of the pene-
trometer. The penetration interruption itself can lead to local or persistent change in qc, fs and u, as dis-
cussed above. In practice, data clips and smoothing will provide look-alike continuous penetration. Fig-
ure 16 shows an example of data as measured (unclipped) and processed data (data clips added where
penetration rate is close to zero and re-interpolated for penetration depth at 20 mm spacing.
qc [MP a] qc [MP a]

0 2 4 6 8 10 12 14 16 0 2 4 6 8 10 12 14 16
6.2
415
As Measured
420 Clipped and Re-int erpolat ed 6.3

425
Penetration Depth [m]

6.4
430
Time [s]

6.5
435

440 6.6

445
6.7
450

6.8

Figure 16. Penetration interruption - cone resistance qc versus time (left) and qc versus penetration depth (right).

30
 Drift of a sensor for qc, fs or u will occur during a CPT. ISO 22476-1 and ASTM D5778 consider zero
drift as indicator for possible anomalies in performance of a data acquisition system. In most cases, zero
drift will probably be incurred as a result of temperature variations and creep. For example, increasing
the penetration depth could imply penetration of harsh ground, temperature increase in the cone pene-
trometer and consequent increase in zero drift. The incurred zero drift would not apply to any shallow
soft soil. The ISO 22476-1 connection between zero drift and application class is doubtful and may lead
to unnecessary effort, for example conducting a drill-out to limit zero drift for a soft soil zone requiring
Application Class 2. Zero values at the start of the test should normally be assigned to the full penetra-
tion phase. Also, it is incorrect to assume zero drift as proof of meeting a particular application class.
A thrust machine may apply significant and variable stress conditions to the soil to be tested. In prac-
tice, no corrections are applied to obtain equivalent free-field conditions. Influence of external stress
conditions may be estimated from continuum mechanics such as elastic theory, where horizontal stress
change is likely to dominate change of qc and fs. Use may also be made of correlations for correcting qc
for a change in ground surface level.
A non-vertical trajectory (i > 0o) of a penetrometer can influence measured soil resistance, particular-
ly where coefficient of earth pressure at rest K0 1. ISO 22476-1 and ASTM D5778 provide no re-
quirements for correction of measured soil resistance for inclination of a cone penetrometer. Figure 17
shows qc and u1 for non-vertical penetration in clay with K0 values ranging between 0.4 and 1. The re-
sults consider calibration chamber tests and finite element modelling.

Cone Resist ance qc P ore P ressure u 1

1 0 1 0

0.9 0.9
-0.05 -0.05

0.8 0.8
-0.1 -0.1
K0 [-]

K0 [-]

0.7 0.7

-0.15 -0.15
0.6 0.6

-0.2 -0.2
0.5 0.5

0.4 -0.25 0.4 -0.25

0 20 40 60 80 0 20 40 60 80
o o
i [ ] i [ ]

Figure 17. Influence of non-vertical trajectory on cone resistance qc (left) and pore pressure u1 (right), expressed
as fraction of K0 = 1 and i = 0 (scale for colour bars is 0 to -0.25), after Wei et al. (2005).

ISO 22476-1 and ASTM D5778 require u2 for calculation of corrected cone resistance qt. These stand-
ards provide no guidance on resolving any differences between u2 and u2g, discussed earlier. In practice,
this implies that qt will usually depend on cone penetrometer design. The presumed u2g = u2 requirement
is in conflict with needs from practice.

13. EXAMPLE ESTIMATION OF CPT ACCURACY

The current state of knowledge suggests that it will be difficult in practice to demonstrate compliance
with any of the application classes of ISO 22476-1, except under favourable conditions.

31
 Table 4 shows example input values for uncertainty estimates according to Type B evaluation (EA
1999). Actual values for a particular CPT system may be better or worse.
For favourable conditions, it may be inferred that the selected example values would result in meeting
Application Class 1 of ISO 22476-1. This confirms current understanding derived from excellent test re-
peatability under uniform, favourable soil conditions (Fig. 18). Figures 19 and 20 show results for ad-
verse conditions: layered sands and clays. The examples may possibly represent first-ever estimates that
comply with ISO 22476-1 and its normative reference to ISO (2003).
The accuracy calculations for the adverse case show that the example CPT, as a whole, would fail to
meet any of the application classes of ISO 22476-1. Significant epistemic uncertainty is the principal
reason for this situation. Particularly, the prediction models for transient temperature and measurement
error take account of lack of understanding and are necessarily conservative. The example test profile
considers a single push from ground surface. Penetration interruptions for transient heat dissipation
would improve results. Actual accuracy is probably significantly better than calculated. Further study is
recommended.

Table 4. Example input for CPT accuracy estimates.

Cone Resistance, qc
Cone Load Cell, qc,trans 5kPa + 5 10 -3 q c(i)
where qc(i) is measured cone resistance
Cone Load Cell, qc,hys 5 10 -3 q c,max
qc,max is the maximum encountered hysteresis value of qc prior to reaching penetration
depth
Cone Area, qAtip 0.02 qc(i)
Temperature, qtemp (
T 10kPa C -1 + 5 10 -4 C -1 q c(i) )
where T is change in temperature in degrees centigrade at load cell position; T may
be estimated from measured cone resistance1) or measured
Non-axial Loading, qbend M 10kPa N m-1
where M is bending moment at load cell position; M may be inferred from cone pene-
trometer trajectory2), or measured
Sleeve Friction, fs
Sleeve Friction Load Cell, fs,trans
0.067 qc,trans + qc+f,trans
2 2

where qc+f,trans is as for qc,trans, except with added frictional force, for a substraction-
type cone penetrometer
Friction Sleeve Area, fAsleeve 0.015 fs(i)
where fs(i) is measured sleeve friction
Temperature, ftemp (
T 1kPa C-1 + 7 10-4 C-1 fs(i) )
Non-axial Loading, fbend M 0.5kPa N m -1
Pore Pressure, u
Pressure Sensor, utrans 1 kPa + 2 10 3 u (i) + 2.5 10 3 u max
where u(i) is measured pore pressure and umax is the maximum encountered value of u
prior to reaching penetration depth
Temperature, utemp (
T 0.5 kPa C -1 + 5 10 -4 C -1 u (i) )
Inclination, i
Inclination Sensor, itrans i trans = 1 + 1 10 3 i (i)

32
Penetration Length, l
Penetration Length Sensor, ltrans 5 103 m + 5 103 l (i)
where l(i) is measured penetration length
Sampling Frequency and Penetra- 1
tion Rate, lsamp 0.5 v (i )
fsamp
were v(i) is penetration rate calculated from penetration length sensor
Push Rod Compression and 12 106 kN1 Qthrust l(i)
Bending, lcompression
where Qthrust is total thrust on the push rods; Qthrust may be measured or estimated3)
Thrust Machine Stiffness, 10 4 m kN 1 Q thrust
lthrust machine
Elastic Soil Relaxing, lrelax 4 104 m kN1 Qthrust
Fixed horizontal Plane, lplane 0.05 m
Notes:
1) Temperature T may be estimated from cone resistance according to T = Tsoil + (Tcone Tsoil ) er t and
Tcone = 1 C MPa 1 qc(i) . The factor r depends on local convection heat transfer coefficient, the area over which convec-
tion occurs, cone penetrometer density and heat capacity. A (forward) smoothing function may account for delay effects
in generation of frictional heat and transition to a load cell location.
2) Bending moment may be inferred from Youngs modulus, second moment of area and curvature of a cone penetrometer
trajectory according to M = E I
l
1
Thrust on push rods may be estimated by Q thrust = q c(i) A tip +
3)
R
fs(i) d rod dl , where R typically ranges from 3 to 7.
0

qc [MP a] fs [kP a] u2 [kP a]

0 0.2 0.4 0.6 0.8 1 0 20 40 60 80 0 200 400 600 800 1000

0
Measured Value
Uncertainty Band
5

10
Penetration Depth [m]

15

20

25

30

35

40

Figure 18. Uncertainty estimates according to ISO 22476-1 for CPT values measured under favourable conditions
of homogeneous slightly overconsolidated clay.

33
 Cone Resistance Sleeve Friction
1 .10 1 .10
5 3

Measured
Application Class 1
Application Class 2
1 .10
4
Application Class 3 100

Uncertainty [kPa]
Uncertainty [kPa]

Application Class 4

1 .10
3
10

100 1

10 0.1
1 .10 1 .10 1 .10 1 .10 1 .10 1 .10
3 4 5 6 3 4
10 100 1 10 100

Measured [kPa] Measured [kPa]

Pore Pressure Penetration Length

1 .10 1 .10
3 4

1 .10
3
Uncertainty [kPa]

Uncertainty [mm]
100

10 100

1 10
1 .10 1 .10 0.1 1 10 100
3 4
1 10 100

Measured [kPa] Measured [m]

Figure 19. Uncertainty estimates according to ISO 22476-1 for CPT values measured under adverse conditions of
layered sands and clays.

qc [MP a] fs [kP a] u2 [kP a] Result ing Class

0 20 40 60 0 100 200 300 400 500 0 500 1000

0
Application Class 1
5 Application Class 2
Application Class 3
Application Class 4
10 Application Class >4

15
Penetration Depth [m]

20

25

30

35

40

45

Figure 20. Demonstrating compliance to ISO application classes, for adverse conditions of layered sands and
clays.

34
14. CHALLENGES IN MEASURING PRACTICE

This section presents examples of challenges faced in measuring practice.

Figure 21 illustrates difficulty in conclusively distinguishing in situ soil behaviour from CPT meas-
urement error. Values for qt and ft show unusual reductions at a penetration depth of about 17 m. Cor-
rected friction ratio Rft and pore pressure ratio Bq also show significant variations. Pore pressure u2
shows an approximately linear increase from ground surface to 40 m penetration depth.
qc,qt [MP a] ft [kP a] Rft [%] u2, uo [MP a] Bq [-] Ic [-] v [mm/s]

0 1 2 0 20 40 0 5 10 0 1 2 -1 0 1 0 2 4 0 20 40
0

10
Penetration Depth [m]

15

20

25

30

35

40

Figure 21. Unusual reduction in qt and ft at 17 m penetration depth.

Comments are as follows.

CPT deployment was by single push from ground surface, with no interruptions for push rods.
Quality monitoring records were according to expectations. Zero drift values for qc, fs and u2 were
+15 kPa, +2 kPa and +4 kPa respectively.
The qt trends with depth have negative intercepts at ground surface for penetration depths between
about 12 m and 17 m and between 17 m and 24 m. This is unusual for normally consolidated to
slightly overconsolidated clay, but may possibly reflect depositional regimes.
A similar profile was observed for a nearby location.
The 12 m level shows evidence for a depositional discontinuity: locally increased cone resistance and
reduced pore pressure. This zone may have triggered build-up of clay adhering to the cone penetrom-
eter. However, the plausibility of this hypothesis seems low.

Figure 22 illustrates challenges with measuring pore pressures at the u2 location, instead of u1. Partic-
ularly, u2 shows negative values as low as -1800 kPa upon penetration of soil depositional discontinui-
ties.

35
 qc,qt [MP a] ft [kP a] Rft [%] u2, uo [MP a] Bq [-] Ic [-] v [mm/s]

0 2 4 0 20 40 0 5 10 -2 0 2 -1 0 1 0 2 4 0 20 40
0

5
Penetration Depth [m]

10

15

20

25

30

Figure 22. Negative u2 values (-1800 kPa).

Comments are as follows.

CPT deployment was offshore and included a single push from seafloor, with no interruptions for
push rods.
Quality monitoring records were according to expectations. Zero drift values for qc, fs and u2 were
+16 kPa, 0 kPa and -1 kPa respectively.
A negative u2 value is followed by pore pressure recovery to positive values. The recovery distance is
at least 0.5 m or about 25 seconds. The delay in pore pressure recovery is probably a function of low
permeability of soil that inhibits fluid movement required for pore pressure measurement.
The u2 values for the recovery period appear to give a reasonable representation of pore pressure u2g
in the gap between the cone and the friction sleeve. This may be inferred from comparison of qc and
qt values.
An onshore test would have been limited in measuring negative pore pressures to around -100 kPa.
Pore pressure recovery would have been uncertain and estimation of u2g would have been difficult.

Figure 23 illustrates deployment challenges for a cone penetrometer, particularly very soft clay over-
lying claystone.
qc,qt [MP a] ft [kP a] Rft [%] u2, uo [MP a] Bq [-] Ic [-] v [mm/s]

0 20 40 0 200 400 0 5 10 0 10 20 -1 0 1 0 2 4 0 20 40
0

5
Penetration Depth [m]

10

15

20

25

Figure 23. Deployment challenges for a cone penetrometer.

36
Comments are as follows:
Discontinuous CPT profiles were acquired by multiple short-push penetrations from the bottom of a
drill pipe.
Quality monitoring records were according to expectations. Zero drift values for qs, fs and u2 for tests
in very soft clay were typically +50 kPa, +1 kPa and +1 kPa respectively. For the underlying clay-
stone, these values were typically -50 kPa, +1 kPa and +1 kPa. The claystone shows a soil behaviour
type index Ic of about 2, equivalent to sandy very clayey silt according to Robertson (1990).
Values for qc and u2 differ by factors of 300 and 420 respectively between 1 m and 22 m penetration
depths.
The data acquisition system for pore pressure fails between 22 m and 23 m penetration depth. The
measured pore pressure then exceeds 12,000 kPa, which is the equivalent of a water column with a
height of 1200 m.
Duration of a short-push test is less than 3 minutes. Temperature response to build-up of frictional
heat will be complex for the load cells for qc and fs. Simple prediction models for transient tempera-
ture may considerably overestimate measurement error for this situation.

Figure 24 illustrates an extreme case of transfer of axial force on the cone to the friction sleeve.
qc,qt [MP a] ft [kP a] Rft [%] u2, uo [MP a] Bq [-] Ic [-] v [mm/s]

0 20 40 0 1000 2000 0 5 10 -0.5 0 0.5 -1 0 1 0 2 4 0 20 40

0

4
Penetration Depth [m]

10

12

14

16

18

20

Figure 24. Transfer of axial force on the cone to the friction sleeve.

Comments are as follows:

CPT deployment was by single push from ground surface, with no interruptions for push rods.
Zero drift values for qc, fs and u2 were -31 kPa, +3 kPa and -4 kPa respectively. These values show no
evidence for anomalies. Force transfer was confirmed by post-project laboratory checks.
It is difficult to detect minor cases of force transfer. Extreme force transfer is evident from credibility
checks on ft signature replicating qc signature, with supplementary confirmation from friction ratio
and soil behaviour type index Ic. Note that the force transfer causes Ic to be mostly outside of classifi-
cation limits indicated by Robertson & Wride (1998).
The sum of the axial forces on the cone and friction sleeve should be approximately correct. For ex-
ample, it may be inferred that a presumed friction ratio of 1% at 10 m penetration depth would in-
crease qc from about 19 MPa to about 24 MPa and decrease ft from about 600 kPa to 240 kPa.
Measured pore pressure u2 shows robust response for testing in very dense sands. The difference be-
tween measured pore pressure and expected hydrostatic pore pressure is <50 kPa at 20 m penetration
depth.

37
15. CONCLUDING REMARKS

An ideal in situ penetration tool would provide:

zero measurement uncertainty,
unambiguous soil behaviour identification,
closed-form theoretical interpretation directly linked to fundamental soil mechanics,
rapid penetration to desirable depths into any type of ground at wide range of temperatures,
inexpensive operation from a small autonomous vehicle operating on any terrain, above and below
water, and,
unique standardisation.

Although no such tool exists, the CPT is closer to this wish list than any of its in situ rivals.
CPT practice has an admirable history on benchmarking of accuracy of parameter values. Such ambi-
tions are not yet seriously contemplated for parameter values inferred from soil sampling, sample han-
dling and geotechnical laboratory testing.
Estimation of CPT accuracy may distinguish favourable and adverse settings. The proposed fa-
vourable setting is comparable to a calibration laboratory. Homogeneous, slightly overconsolidated clay
would be an example of favourable conditions. The proposed adverse setting represents common hostile
site conditions and/or a setting with real-world limits on tool control. A setting of strongly layered dense
sands and soft clays with ground water at depth would be an example of adverse conditions. Particularly,
adverse conditions recognise the complexity of accurate measurement of resistance values differing by
typically four orders of magnitude, with no control on measuring environment. Furthermore, it may oc-
casionally be difficult to distinguish CPT measurement error from in situ soil behaviour.
Usually, discussions on CPT accuracy take place without a satisfactory definition for accuracy. Some
use accuracy in metrological terms of resolution or repeatability. Others presumably refer to accuracy
under calibration laboratory conditions. Many are ambiguous about: (1) coupling of qc, fs and u with
spatial position xyz and (2) true value related to permissible equipment-specific and procedure-specific
features.
A proposed Analytic CPT can serve as benchmark for true values, independent of equipment-specific
and procedure-specific CPT features. The proposed Analytic CPT features zero measurement uncertain-
ty and an imaginary cone penetrometer with a solid body of uniform geometry with no gaps.
Widely used CPT standards are ISO 22476-1:2012 (2012a) and ASTM D5778-12 (2012). ISO 22476-
1 defines accuracy, where (1) qc, fs and u are uncoupled from spatial position xyz and (2) true values de-
pend on permissible equipment-specific and procedure-specific features. This means that compliance
with an application class of ISO 22476-1 only provides a first indication of accuracy required for ge-
otechnical practice. Further processing will be necessary to obtain fit-for-purpose accuracy, i.e. with
benchmarking against the Analytic CPT or equivalent. ASTM D5778 has no aspirations on defining ac-
curacy. It considers precision and bias.
The ISO 22476-1 connection between zero drift and application class is doubtful and may lead to un-
necessary effort in practice. Typically, zero drift reflects a wide range of test conditions. It cannot be iso-
lated for the penetration phase applicable to a particular soil stratum.
The current state of knowledge suggests that it will be difficult in practice to demonstrate compliance
with the application classes of ISO 22476-1, except under favourable conditions. This is illustrated by
two examples of uncertainty estimates. The examples may possibly represent first-ever estimates that
comply with ISO 22476-1 and its normative reference to ISO (2003). The first example (Fig. 18) con-
siders favourable conditions, i.e. homogeneous slightly overconsolidated clay. The uncertainty estimate
indicates feasibility of meeting Application Class 1 of ISO 22476-1 and confirms current understanding
derived from excellent test repeatability under uniform, favourable soil conditions. The second example
(Fig. 20) considers routine adverse conditions: layered sands and clays. The accuracy calculations for

38
the adverse case show that the example CPT, as a whole, would fail to meet any of the application clas-
ses of ISO 22476-1. Significant epistemic uncertainty is the principal reason for this situation. Particu-
larly, the prediction models for transient temperature and measurement error take account of lack of un-
derstanding and are necessarily conservative. Actual accuracy is probably significantly better than
calculated. Further study is recommended.

ACKNOWLEDGEMENTS

The authors are indebted to many dedicated colleagues and gratefully acknowledge Fugros persistent
commitment to excellence in cone penetration testing. The opinions expressed in this paper are those of
the authors. They are not necessarily shared by Fugro.

GLOSSARY OF METROLOGICAL TERMS

Table 5. Glossary of metrological terms.

Term Definition Notes
Accuracy, closeness of agreement between a measured
measurement quantity value and a true quantity value of a
accuracy, measurand
accuracy of
measurement
Accuracy Class class of measuring instruments or measuring - An accuracy class is usually denoted by a number or
systems that meet stated metrological re- symbol adopted by convention.
quirements that are intended to keep meas- - Accuracy class applies to material measures.
urement errors or instrumental uncertainties
within specified limits under specified operat-
ing conditions
Bias, estimate of a systematic measurement error
Measurement
Bias ASTM2: bias, [Unit of Measure]in meas-
urements, a systematic error that contributes
to the difference between the mean of a large
number of test results and an accepted refer-
ence value
Drift, continuous or incremental change over time in Instrumental drift is related neither to a change in a
Instrumental indication, due to changes in metrological quantity being measured nor to a change of any recog-
drift properties of a measuring instrument nized influence quantity.
Error, measured quantity value minus a reference - The concept of measurement error can be used both
Error of meas- quantity value when there is a single reference quantity value to re-
urement, fer to, which occurs if a calibration is made by means
Measurement of a measurement standard with a measured quantity
error value having a negligible measurement uncertainty or
if a conventional quantity value is given, in which
case the measurement error is known, and b) if a
measurand is supposed to be represented by a unique
true quantity value or a set of true quantity values of
negligible range, in which case the measurement er-
ror is not known.
- Measurement error should not be confused with pro-
duction error or mistake.
Full scale out- the specified maximum output value for
put, FSO which the stated accuracy condition applies
Hysteresis ISRM4: incomplete recovery of strain during ASTM2: hysteresis is the dependence of a system not on-
unloading cycle due to energy consumption ly on its current environment but also on its past envi-

39
Term Definition Notes
ronment.
Measurement process of experimentally obtaining one or
more quantity values that can reasonably be
attributed to a quantity
Measurand quantity intended to be measured
Metrology, met- science of measurement and its application - Metrology includes all theoretical and practical as-
rological con- pects of measurement, whatever the measurement
firmation uncertainty and field of application
- Metrological confirmation generally includes calibra-
tion and verification, any necessary adjustment or re-
pair, and subsequent recalibration, comparison with
the metrological requirements for the intended use of
the equipment, as well as any required sealing and
labelling.
- Metrological confirmation is not achieved until and
unless the fitness of the measuring equipment for the
intended use has been demonstrated and documented.
- The requirements for intended use include such con-
siderations as range, resolution and maximum per-
missible errors.
- Metrological requirements are usually distinct from,
and are not specified in, product requirements.
Precision, closeness of agreement between indications or - Measurement precision is usually expressed numeri-
measurement measured quantity values obtained by repli- cally by measures of imprecision, such as standard
precision cate measurements on the same or similar ob- deviation, variance, or coefficient of variation under
jects under specified conditions the specified conditions of measurement.
- The specified conditions can be, for example, re-
ASTM2: precision, [Unit of Measure]in peatability conditions of measurement, intermediate
measurements, the closeness of agreement be- precision conditions of measurement, or reproduci-
tween test results obtained under prescribed bility conditions of measurement (see ISO 5725-
conditions. 3:1994/Cor.1:2001).
- Measurement precision is used to define measure-
ment repeatability, intermediate measurement preci-
sion, and measurement reproducibility.
- Sometimes measurement precision is erroneously
used to mean measurement accuracy.
Quantity property of a phenomenon, body, or sub-
stance, where the property has a magnitude
that can be expressed as a number and a refer-
ence
quantity value value of a quantity value number and refer-
ence together expressing magnitude of a quan-
tity
Random Meas- component of measurement error that in repli- - A reference quantity value for a random measure-
urement Error, cate measurements varies in an unpredictable ment error is the average that would ensue from an
Random error, manner infinite number of replicate measurements of the
Random error same measurand.
of measurement - Random measurement errors of a set of replicate
measurements form a distribution that can be summa-
rized by its expectation, which is generally assumed
to be zero, and its variance.
- Random measurement error equals measurement er-
ror minus systematic measurement error.

40
Term Definition Notes
Range ASTM2: range (of a deformation-measuring
instrument) the amount between the maxi-
mum and minimum quantity an instrument
can measure without resetting/ overloading. In
some instances provision can be made for in-
cremental extension of the range.
Repeatability, measurement precision under a set of repeata-
Measurement bility conditions of measurement
repeatability
Resolution smallest change in a quantity being measured Resolution can depend on, for example, noise (internal
that causes a perceptible change in the corre- or external) or friction. It may also depend on the value
sponding indication of a quantity being measured.
Stability, property of a measuring instrument, whereby Stability may be quantified in several ways.
stability of a its metrological properties remain constant in EXAMPLE 1 In terms of the duration of a time interval
measuring in- time over which a metrological property changes by a stated
strument amount.
EXAMPLE 2 In terms of the change of a property over a
stated time interval.
Systematic Er- component of measurement error that in repli- - A reference quantity value for a systematic meas-
ror, cate measurements remains constant or varies urement error is a true quantity value, or a measured
Systematic in a predictable manner quantity value of a measurement standard of negligi-
measurement ble measurement uncertainty, or a conventional quan-
error, tity value.
Systematic error - Systematic measurement error, and its causes, can be
of measurement known or unknown. A correction can be applied to
compensate for a known systematic measurement er-
ror.
- Systematic measurement error equals measurement
error minus random measurement error.
Type A evalua- evaluation of a component of measurement
tion of meas- uncertainty by a statistical analysis of meas-
urement ured quantity values obtained under defined
uncertainty measurement conditions
Type B evalua- evaluation of a component of measurement EXAMPLES
tion of meas- uncertainty determined by means other than a Evaluation based on information:
urement Type A evaluation of measurement uncertain- - associated with authoritative published quantity val-
uncertainty ty ues,
- associated with the quantity value of a certified refer-
ence material,
- obtained from a calibration certificate,
- about drift,
- obtained from the accuracy class of a verified meas-
uring instrument,
- obtained from limits deduced through personal expe-
rience.
Trueness, closeness of agreement between the average - Measurement trueness is not a quantity and thus can-
measurement of an infinite number of replicate measured not be expressed numerically, but measures for
trueness, true- quantity values and a reference quantity value closeness of agreement are given in ISO 5725/
ness of meas- Cor.1:2001)
urement - Measurement trueness is inversely related to system-
atic measurement error, but is not related to random
measurement error.
- Measurement accuracy should not be used for meas-
urement trueness and vice versa.

41
Term Definition Notes
Uncertainty, non-negative parameter characterizing the - Measurement uncertainty includes components aris-
Uncertainty of dispersion of the quantity values being at- ing from systematic effects, such as components as-
measurement, tributed to a measurand, based on the infor- sociated with corrections and the assigned quantity
Measurement mation used values of measurement standards, as well as the defi-
uncertainty nitional uncertainty. Sometimes estimated systematic
effects are not corrected for but, instead, associated
measurement uncertainty components are incorpo-
rated.
- The parameter may be, for example, a standard devi-
ation called standard measurement uncertainty (or a
specified multiple of it), or the half-width of an inter-
val, having a stated coverage probability.
- Measurement uncertainty comprises, in general,
many components. Some of these may be evaluated
by Type A evaluation of measurement uncertainty
from the statistical distribution of the quantity values
from series of measurements and can be character-
ized by standard deviations. The other components,
which may be evaluated by Type B evaluation of
measurement uncertainty, can also be characterized
by standard deviations, evaluated from probability
density functions based on experience or other in-
formation. - In general, for a given set of infor-
mation, it is understood that the measurement uncer-
tainty is associated with a stated quantity value
attributed to the measurand. A modification of this
value results in a modification of the associated un-
certainty.
Zero Error datum measurement error where the specified Zero error should not be confused with absence of meas-
measured quantity value is zero urement error.
Zero Adjust- adjustment of a measuring system so that it
ment of a meas- provides a null indication corresponding to a
uring system zero value of a quantity to be measured

Notes:
1. Definitions are according to JCGM (2012), except where shown
2. ASTM (2011)
3. ISO (2012)
4. ISRM (1972)
5. EA (1999)

REFERENCES

ASTM. 1979. Standard method for deep, quasi-static, cone and friction-cone penetration tests of soil, ASTM
D3441-79. West Conshohocken: ASTM International.
ASTM. 2011. Standard terminology relating to soil, rock, and contained fluids, ASTM D653-11. West Con-
shohocken: ASTM International.
ASTM. 2012. Standard test method for electronic friction cone and piezocone penetration testing of soils, ASTM
D5778-12. West Conshohocken: ASTM International.
ASTM. 2013. Form and style for ASTM standards. West Conshohocken: ASTM International.
Beer De, E.E. 1963. Scale Effects. Gotechnique, Vol. 13, pp. 39-75.
Berg, P. van den 1991. Numerical model for cone penetration. Seventh International Conference on Computer
Methods and Advances in Geomechanics, Cairns, Proceedings, Vol. 3, pp. 1777-1782.
Berg, P. van den, Borst, R. de, Huetink, J. (1992). Numerical model for penetration in frictional materials. Third
International Conference on Computational Plasticity, Barcelona, Proceedings, Vol. 1, pp. 971-982.

42
Berg, P. van den 1994. Analysis of soil penetration. Delft University Press 1994. Thesis. 180 pp.
Boggess, R. & Robertson, P.K. 2010. CPT for soft sediments and deepwater applications. CPT10: 2nd Interna-
tional Symposium on Cone Penetration Testing, Huntington Beach, CA: Conference Proceedings.
Boylan, N., Mathijssen, F., Long, M. and Molenkamp, F. 2008. Cone penetration testing of organic soils. Pro-
ceedings, Baltic Geotechnical Conference, September. Gdansk, Poland.
Commission of the European Communities CEC 1986. Nuclear science and technology, geotechnical deep ocean
research apparatus (DORA). Report EUR 10601 EN.
DeJong, J.T & Frost, J.D. 2002, "A Multisleeve Friction Attachment for the Cone Penetrometer"; American Socie-
ty for Testing and Materials, Journal of Geotechnical Testing, Vol. 25, No. 2, pp. 111-127, 2002.
DeJong, J.T., Jaeger, R.A., Boulanger, R.W., Randolph, M.F., Wahl, D.A.J. (2012), Variable penetration rate
cone testing for characterisation of intermediate soils, Geotechnical and Geophysical Site Characterization 4
Proceedings of the 4th International Conference on Site Characterization 4, ISC-4 1, 25-42
Diepstraten, E. 2003. Comparison of cone resistance of mini cone and Fugro standard cone in sand. MSc Disser-
tation Technical University of Delft, 113 pp. (in Dutch).
Elmgren, K. 1995. Slot-type pore pressure CPT-u filters behaviour of different filling media. In International
symposium on cone penetration testing CPT95, Linkping, Sweden, October 4-5, 1995, Vol. 2: 73-78. Link-
ping: Swedish Geotechnical Society. European Group of the Sub-committee for Static and Dynamic Penetra-
tion Test Methods, 1968
EA. 1999. Expression of the uncertainty of measurement in calibration, EA-4/02. European Co-operation for Ac-
creditation of Laboratories, 79 p.
Hird, C.C. & Springman, S.M. 2006. Comparative performance of 5 cm2 and 10 cm2 piezocones in a lacustrine
clay. Gotechnique, Vol. 56, No. 6, pp. 427-438.
ISO, 1994/Cor.1:2001, Accuracy (trueness and precision) of measurement methods and results Part 3: Inter-
mediate measures of the precision of a standard measurement method, International Standard ISO 5725-
3:1994/Cor.1:2001. Geneva: International Organization for Standardization.
ISO. 1998. General Principles on Reliability for Structures, International Standard ISO 2394:1998. Geneva: In-
ternational Organization for Standardization.
ISO. 2003. Measurement management systems - requirements for measurement processes and measuring equip-
ment, International Standard ISO 10012. Geneva: International Organization for Standardization.
ISO. 2012a. Geotechnical investigation and testing Field testing Part 1: Electrical cone and piezocone pene-
tration tests, International Standard ISO 22476-1:2012. (With Technical Corrigendum 1, January 2013). Ge-
neva: International Organization for Standardization.
ISO. 2012b. Petroleum and natural gas industries - specific requirements for offshore structures - Part 8: marine
soil investigations, Draft International Standard ISO/DIS 19901-8. Geneva: International Organization for
Standardization.
ISO/IEC. 2010. Conformity assessment general requirements for proficiency testing. Geneva: International Or-
ganization for Standardization & International Electrotechnical Commission.
ISO/IEC. 2011. ISO/IEC directives, Part 2: Rules for the structure and drafting of international standards. Gene-
va: International Organization for Standardization & International Electrotechnical Commission.
ISO/IEC. 2013. ISO/IEC directives, Part 1: consolidated ISO supplement procedures specific to ISO. Geneva:
International Organization for Standardization & International Electrotechnical Commission.
ISSMFE International Society for Soil Mechanics and Foundation Engineering. 1977. Report of the subcommittee
on standardization of penetration testing in Europe, Appendix A, recommended standard for the cone penetra-
tion test (CPT). Proceedings of the ninth international conference on soil mechanics and foundation engineer-
ing, Tokyo.
ISSMFE International Society for Soil Mechanics and Foundation Engineering. 1989. Report of the ISSMFE
Technical Committee on penetration testing of soils TC 16 with reference test procedures, CPT SPT DP
WST. Linkping: Swedish Geotechnical Institute.
ISSMGE International Society for Soil Mechanics and Geo-technical Engineering. 1999. International reference
test procedure for the cone penetration test (CPT) and the cone penetration test with pore pressure (CPTU):
Report of the ISSMGE Technical Committee 16 on Ground Property Characterisation from In situ Testing. In
F.B.J. Barends et al. (eds.), Geotechnical engineering for transportation infrastructure; Proceedings of the
twelfth European conference on soil mechanics and geotechnical engineering, Amsterdam, Netherlands, 7-10
June 1999, Vol. 3: 2195-2222. Rotterdam: A.A. Balkema.

43
JCGM. 2012. International vocabulary of metrology basic and general concepts and associated terms (VIM),
3rd edition, 2008 version with minor corrections, JCGM 200:2012. Joint Committee for Guides in Metrology.
Jekel, J.W.A. 1988. Wear of the friction sleeve and its effect on the measured local friction. In J. De Ruiter (ed.),
Penetration testing 1988; proceedings of the first international symposium on penetration testing, ISOPT-1,
Orlando, 20-24 March 1988, Vol. 2: 805-808. Rotterdam: A.A. Balkema.
Lima De, D.C. & Tumay, M.T. 1990. Development of the Louisiana State University calibration chamber system
(LSU/CALCHAS), 4th Int. Seminar on Calibration Chamber Research, Grenoble France.
Loon van J.C. & Schaareman A. 1991. Penetrometers, an investigation of internal friction effects. Thesis Tech-
nical Institute Rijswijk, The Netherlands (in Dutch).
Lunne, T., Eidsmoen, T., Gillespie, D. & Howland, J.D. 1986. Laboratory and field evaluation of cone penetrome-
ters. In S.P. Clemence (ed.), Use of in situ tests in geotechnical engineering; proceedings of In Situ 86, a spe-
cialty conference, Blacksburg, Virginia, June 23-25, 1986: 714-729. New York: American Society of Civil
Engineers.
Long, M. 2010. Regional report for northern Europe. CPT10: 2nd International Symposium on Cone Penetration
Testing, Huntington Beach, CA: Conference Proceedings.
Meave Silva O.R. 1999. Shallow cone penetration test / shallow penetration tests in dense saturated sand. MSc
Thesis Delft: International Institute for Infrastructural, Hydraulic and Environmental Engineering.
Nageswaran S. 1983. Effect of gas bubbles on the seabed behaviour. PhD Thesis for the University of Oxford.
NNI. 1996. Geotechnics. Determination of the cone resistance and the sleeve friction of soil. Electric cone pene-
tration test. Dutch Standard NEN 5140. Delft: Nederlands Normalisatie-instituut. (in Dutch).
NIST/SEMATECH. 2013. e-Handbook of Statistical Methods.
http://www.itl.nist.gov/div898/handbook/mpc/section1/mpc113.htm), last accessed November 2013.
Ooi, P.S.K. & Ramsey, T.L. 2003. Curvature and bending moments from inclinometer data. International Journal
of Geomechanics, Vol. 3, No. 1, pp. 64-74.
Peuchen, J. 1998. Commercial CPT profiling in soft rocks and hard soils. Geotechnical Site Characterisation,
Robertson & Mayne (eds.), 1998 Balkema, Rotterdam.
Peuchen, J. 2000. Deepwater cone penetration tests. In 32nd Annual Offshore Technology Conference, 2000: Pro-
ceedings, Vol. 1, OTC Paper 12094.
Peuchen, J., Adrichem, J. & Hefer, P.A. 2005. Practice notes on push-in penetrometers for offshore geotechnical
investigation. In S. Gourvenec & M. Cassidy, M. (eds.), Frontiers in offshore geotechnics ISFOG 2005; Pro-
ceedings of the first international symposium on frontiers in offshore geotechnics, University of Western Aus-
tralia, Perth, 19-21 September 2005: 973-979. London: Taylor & Francis.
Peuchen, J., Vanden Berghe, J.F. & Coulais, C. 2010. Estimation of u1/u2 conversion factor for piezocone.
CPT10: 2nd International Symposium on Cone Penetration Testing, Huntington Beach, CA: Conference Pro-
ceedings.
Peuchen, J. 2012. Site characterization in nearshore and offshore geotechnical projects. In Coutinho, R.Q. and
Mayne, P.W. (Eds.), Geotechnical and Geophysical Site Characterization 4: Proceedings of the Fourth Inter-
national Conference on Site Characterization ISC-4, Porto de Galinhas-Pernambuco, Brazil, 17-21 September
2012, Vol. I, CRC Press, Boca Raton, pp. 83-111.
Peuchen, J. & Balthes, R. 2013 Strategies for offshore geotechnical site investigations in chalk. In European Off-
shore Wind- and Infrastructure Developments, pp. 185-197.
Post, M.L. & Nebbeling, H. 1995. Uncertainties in cone penetration testing. In International symposium on cone
penetration testing CPT95, Linkping, Sweden, October 4-5, 1995, Vol. 2: 73-78. Linkping: Swedish Ge-
otechnical Society.
Powell, J.J.M. & Lunne, T. 2005a. A comparison of different sized piezocones in UK clays. In Proceedings of the
16th International Conference on Soil Mechanics and Geotechnical Engineering: 16ICSMGE, 2005, Osaka,
Vol. 2, Millpress, Rotterdam, pp. 729-734.
Powell, J.J.M. & Lunne, T. 2005b. Use of CPTU data in clays/fine grained soils. Studia Geotechnica et Mechani-
ca, Vol. 27, No. 3-4, pp. 29-66.
Rad, N.S. 1983. Modification and calibration of the piezocone penetrometer. Final report submitted to Fugro
Consultants International. Louisiana State University. Department of Civil Engineering.
Randolph, M.F. 2004. Characterisation of soft sediments for offshore applications. Proceedings of the Second In-
ternational Conference on Site Characterisation, Porto, Portugal, Vol. 1, pp. 209232.

44
Robertson, P.K. 1990. Soil classification using the cone penetration test. Canadian Geotechnical Journal, Vol. 27,
No. 1, pp. 151-158.
Robertson, P.K., Campanella, R.G., Gillespie, D. & Greig, J. 1986. Use of piezometer cone data. In Clemence,
S.P. (Ed.), Use of in situ tests in geotechnical engineering: proceedings of In Situ 86, Blacksburg, Virginia,
June 23-25, 1986, Geotechnical Special Publication, No. 6, American Society of Civil Engineers, New York,
pp. 1263-1280.
Robertson, P.K. & Wride (Fear), C.E. 1998. Evaluating cyclic liquefaction potential using the cone penetration
test. Canadian Geotechnical Journal, Vol. 35, No. 3, pp. 442-459.
Ruiter de, J. 1975, The use of in-situ testing for North Sea soil studies. In Offshore Europe Conference, Aberdeen,
Paper OE-75 219.1, Spearhead Publications.
Schaap, L.H.J. & Zuidberg, H.M. 1982. Mechanical and electrical aspects of the electric cone penetrometer tip. In
A. Verruijt, F.L. Beringen & E.H. De Leeuw (eds.), Penetration testing; proceedings of the second European
Symposium on Penetration Testing, ESOPT II, Amsterdam, 24-27 May 1982, Vol. 2: 841-851. Rotterdam:
A.A. Balkema.
Titi, H.H., Mohammad, L.N. & Tumay, M.T. 2000. Miniature cone penetration tests in soft and stiff clays, ASTM
Geotechnical Testing Journal, Vol. 23, No. 4, pp. 432-443.
Trevor, F.A., Paisley, J.M. & Mayne, P.W. 2010. Cone penetration tests at active earthquake sites. In CPT10: 2nd
International Symposium on Cone Penetration Testing, Huntington Beach, CA: Conference Proceedings.
Tufenkjian, M.R., Yee, E. & Thompson D.J. 2010. Comparison of cone and minicone penetration resistance for
sand at shallow depth. In CPT10: 2nd International Symposium on Cone Penetration Testing, Huntington
Beach, CA: Conference Proceedings.
Uesugi. M. & Kishida. H. 1986. Frictional resistance at yield between dry sand and mild steel. Soils and Founda-
tions, Vol. 26, No. 4, pp. 139-149.
VIM International Vocabulary of Metrology (1993) International vocabulary of basic and general terms used in
metrology 2nd Edition. International Organization for Standardization. Replaced by JCGM. 2012
Vreugdenhil, R., Davis, R. & Berrill, J. 1994. Interpretation of cone penetration results in multilayered soils, Int.
J. for Numerical and Analytical Methods in Geomechanics, Vol. 18, No. 9, pp. 585599.
Wei, L., Abu-Farsakh, M.Y. & Tumay, M.T. 2005. Finite-element analysis of inclined piezocone penetration test
in clays, International Journal of Geomechanics, Vol. 5, No. 3, pp. 167-178.

45