10
TECHNICAL NOTES CPT-DMT Correlations P. K. Robertson 1 Abstract: Although the cone penetration test CPT and flat-plat dilatometer test DMT have been used for over 30 years, relatively little has been published regarding comprehensive correlations between the two in situ tests. This paper presents preliminary correlations between the main parameters of the CPT and DMT. The key to the proposed correlations is the recognition that the main DMT parameters are normalized and hence, should be correlated with normalized CPT parameters. The suggested correlations are developed and evaluated using published records and existing links to various other parameters as well as comparison profiles. The suggested correlations may guide future more detailed correlations between these two in situ tests. DOI: 10.1061/ASCEGT.1943-5606.0000119 CE Database subject headings: In situ tests; Correlation; Cone penetration tests. Introduction The current electric cone penetration test CPT was developed in the Netherlands in the 1960s and has a strong theoretical back- ground as well as the advantages of being fast, near continuous, repeatable, and economical. These advantages have lead to a steady increase in the use and application of the CPT in many other places around the world. The flat-plat dilatometer test DMT was developed in Italy by Professor Silvano Marchetti in the 1980s and has become popular in some parts of the world. The DMT is simple, robust, repeatable and economical. However, the DMT is harder to push in very stiff ground compared to the CPT and the DMT is carried out every 20 cm whereas CPT readings are taken every 2–5 cm. The DMT requires a pause in the penetration to perform the test. Hence, the DMT produces less data than the CPT and is also slower than the CPT. Both tests do not include a soil sample, although it is pos- sible to take small diameter soil samples using the same pushing equipment used to insert either the CPT or DMT. Each test appears to correlate well with particular geotechnical parameters. For example, the CPT provides correlations with und- rained shear strength and overconsolidation ratio OCR in fine- grained soils and peak friction angle in coarse-grained soils. The CPT is commonly used for pile design and to evaluate the poten- tial for soil liquefaction. The DMT also provides correlations with undrained shear strength and OCR in fine-grained soils and cor- relations with one-dimensional constrained modulus for a wide range of soils. Both tests can be used to estimate consolidation/ drainage parameters such as the coefficient of consolidation and permeability from dissipation tests. However, in the past 30 years relatively little has been published regarding comprehensive links between the CPT and DMT parameters. The objective of this paper is to review published records of nearby CPT and DMT soundings as well as existing correlations for geotechnical parameters in an effort to identify possible cor- relations between normalized in situ test parameters. The key in the approach is the recognition that DMT interpretation param- eters are normalized and will likely correlate with normalized CPT parameters. Flat-Plat Dilatometer Test The DMT was developed in Italy by Professor Silvano Marchetti. It was initially introduced in 1980 and is currently used in over 40 countries. Marchetti 1980 provided a detailed description of the DMT equipment, the test method, and the original correlations. Subsequently, the DMT has been used and calibrated in soil de- posits all over the world. Various international standards and manuals are available for the DMT. Marchetti et al. 2001 pre- pared a comprehensive report on the DMT for Technical Commit- tee 16, ISSMGE. The flat dilatometer is a stainless steel blade with a flat circular steel membrane mounted flush on one side. The test involves two readings A and B that are corrected for membrane stiffness, gauge zero offset, and feeler pin elevation in order to determine the pressures p 0 and p 1 . Readings are taken every 20 cm during a pause in the penetration and the corrected pressures p 0 and p 1 are subsequently used for interpretation. The original correlations Marchetti 1980 were obtained by calibrating DMT results with high quality soil parameters from several test sites in Europe. Many of these correlations form the basis of current interpreta- tion, having been generally confirmed by subsequent research. The interpretation evolved by first identifying three “intermedi- ate” DMT parameters Marchetti 1980 Material index, I D = p 1 - p 0 / p 0 - u 0 1 Horizontal stress index, K D = p 0 - u 0 / vo 2 1 Professor Emeritus, Univ. of Alberta, and Technical Director, Gregg Drilling and Testing Inc., 2726 Walnut Avenue, Signal Hill, CA 90755. E-mail: [email protected] Note. This manuscript was submitted on August 4, 2008; approved on March 31, 2009; published online on May 2, 2009. Discussion period open until April 1, 2010; separate discussions must be submitted for individual papers. This technical note is part of the Journal of Geotech- nical and Geoenvironmental Engineering, Vol. 135, No. 11, November 1, 2009. ©ASCE, ISSN 1090-0241/2009/11-1762–1771/$25.00. 1762 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / NOVEMBER 2009 Downloaded 07 Jan 2010 to 66.237.65.34. Redistribution subject to ASCE license or copyright; see http://pubs.asce.org/copyright

CPT DMT Correlations

Embed Size (px)

DESCRIPTION

Technical Notes - CPT-DMT Correlations P.K. Robertson

Citation preview

  • TECHNICAL NOTES

    Coobe

    ilatomen thepropo

    PT paeterstests.

    ne pe

    equipment used to insert either the CPT or DMT. posits all over the world. Various international standards and

    Material index, ID = p1 p0/p0 u0 1March 31, 2009; published online on May 2, 2009. Discussion periodopen until April 1, 2010; separate discussions must be submitted forindividual papers. This technical note is part of the Journal of Geotech-Horizontal stress index, KD = p0 u0/vo 2nical and Geoenvironmental Engineering, Vol. 135, No. 11, November1, 2009. ASCE, ISSN 1090-0241/2009/11-17621771/$25.00.Each test appears to correlate well with particular geotechnicalparameters. For example, the CPT provides correlations with und-rained shear strength and overconsolidation ratio OCR in fine-grained soils and peak friction angle in coarse-grained soils. TheCPT is commonly used for pile design and to evaluate the poten-tial for soil liquefaction. The DMT also provides correlations withundrained shear strength and OCR in fine-grained soils and cor-relations with one-dimensional constrained modulus for a widerange of soils. Both tests can be used to estimate consolidation/drainage parameters such as the coefficient of consolidation andpermeability from dissipation tests. However, in the past 30 years

    manuals are available for the DMT. Marchetti et al. 2001 pre-pared a comprehensive report on the DMT for Technical Commit-tee 16, ISSMGE.

    The flat dilatometer is a stainless steel blade with a flat circularsteel membrane mounted flush on one side. The test involves tworeadings A and B that are corrected for membrane stiffness, gaugezero offset, and feeler pin elevation in order to determine thepressures p0 and p1. Readings are taken every 20 cm during apause in the penetration and the corrected pressures p0 and p1 aresubsequently used for interpretation. The original correlationsMarchetti 1980 were obtained by calibrating DMT results withhigh quality soil parameters from several test sites in Europe.Many of these correlations form the basis of current interpreta-tion, having been generally confirmed by subsequent research.The interpretation evolved by first identifying three intermedi-ate DMT parameters Marchetti 1980

    1Professor Emeritus, Univ. of Alberta, and Technical Director, GreggDrilling and Testing Inc., 2726 Walnut Avenue, Signal Hill, CA 90755.E-mail: [email protected]

    Note. This manuscript was submitted on August 4, 2008; approved onCPT-DMTP. K. R

    Abstract: Although the cone penetration test CPT and flat-plat dhas been published regarding comprehensive correlations betwebetween the main parameters of the CPT and DMT. The key to theare normalized and hence, should be correlated with normalized Cusing published records and existing links to various other paramguide future more detailed correlations between these two in situ

    DOI: 10.1061/ASCEGT.1943-5606.0000119

    CE Database subject headings: In situ tests; Correlation; Co

    Introduction

    The current electric cone penetration test CPT was developed inthe Netherlands in the 1960s and has a strong theoretical back-ground as well as the advantages of being fast, near continuous,repeatable, and economical. These advantages have lead to asteady increase in the use and application of the CPT in manyother places around the world.

    The flat-plat dilatometer test DMT was developed in Italy byProfessor Silvano Marchetti in the 1980s and has become popularin some parts of the world. The DMT is simple, robust, repeatableand economical. However, the DMT is harder to push in very stiffground compared to the CPT and the DMT is carried out every 20cm whereas CPT readings are taken every 25 cm. The DMTrequires a pause in the penetration to perform the test. Hence, theDMT produces less data than the CPT and is also slower than theCPT. Both tests do not include a soil sample, although it is pos-sible to take small diameter soil samples using the same pushing1762 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGIN

    Downloaded 07 Jan 2010 to 66.237.65.34. Redistribution subject to Arrelationsrtson1

    eter test DMT have been used for over 30 years, relatively littletwo in situ tests. This paper presents preliminary correlations

    sed correlations is the recognition that the main DMT parametersrameters. The suggested correlations are developed and evaluatedas well as comparison profiles. The suggested correlations may

    netration tests.

    relatively little has been published regarding comprehensive linksbetween the CPT and DMT parameters.

    The objective of this paper is to review published records ofnearby CPT and DMT soundings as well as existing correlationsfor geotechnical parameters in an effort to identify possible cor-relations between normalized in situ test parameters. The key inthe approach is the recognition that DMT interpretation param-eters are normalized and will likely correlate with normalizedCPT parameters.

    Flat-Plat Dilatometer Test

    The DMT was developed in Italy by Professor Silvano Marchetti.It was initially introduced in 1980 and is currently used in over 40countries. Marchetti 1980 provided a detailed description of theDMT equipment, the test method, and the original correlations.Subsequently, the DMT has been used and calibrated in soil de-EERING ASCE / NOVEMBER 2009

    SCE license or copyright; see http://pubs.asce.org/copyright

  • gested that KD could be regarded as the in situ horizontal stressratio, K , amplified by the DMT penetration. In genuinely nor-Dilatometer modulus, ED = 34.7p1 p0 3

    where u0=preinsertion in situ equilibrium water pressure andvo =preinsertion in situ vertical effective stress.

    The dilatometer modulus ED can also be expressed as a com-bination of ID and KD in the form

    ED/vo = 34.7IDKD 4

    The key DMT design parameters are ID and KD. Both parametersare normalized and dimensionless. ID is the difference betweenthe corrected lift-off pressure p0 and the corrected deflectionpressure p1 normalized by the effective lift-off pressure p0u0. KD is the effective lift-off pressure normalized by the in situvertical effective stress. Although alternate methods have beensuggested to normalize KD, the original normalization suggestedby Marchetti 1980 using the in situ vertical effective stress isstill the most common and is used in this paper. It is likely that amore complex normalization for KD would be more appropriate,especially in sands, but most of the available published records ofKD use the original normalization suggested by Marchetti 1980.

    According to Marchetti 1980, the soil type can be identifiedas follows: Clays ID0.6 Silt mixtures 0.6 ID1.8 Sands ID1.8Marchetti 1980 suggested that ID is a parameter reflecting themechanical behavior of the soil and not a soil classification basedon grain size distribution and plasticity. The link between ID andsoil type is shown in Fig. 1, which shows that ID can range from0.1 to 10 and is often presented on a log scale.

    KD provides the basis for several soil parameter correlationsand is a key parameter from the DMT. Marchetti 1980 sug-

    Fig. 1. Chart for estimating soil type and unit weight using theDMT normalized to w= water; modified from Marchetti et al.2001. Note that 1 bar=100 kPa.JOURNAL OF GEOTECHNICAL AND GEOE

    Downloaded 07 Jan 2010 to 66.237.65.34. Redistribution subject to A0mally consolidated clays i.e., no aging, structure, cementationthe value of KD is KD,NC2. The KD profile is similar in shape tothe OCR profile and hence, is generally helpful for understandingthe soil deposit and its stress history in clays Marchetti 1980.

    Cone Penetration Test

    The CPT was first introduced in The Netherlands in the 1930s asa mechanical test and in the 1960s the cone was updated to in-corporate electric strain-gauged load cells. Various internationalstandards and manuals are available for the CPT and Lunne et al.1997 presented a comprehensive book on the CPT.

    The CPT is a cylindrical probe pushed into the ground at2 cm/sec with essentially continuous readings of the tip stress, qc,sleeve friction stress, fs and sometimes the penetration pore pres-sure, u2, typically measured behind the cone. The tip stress, qc iscorrected for unequal end area effects to a total cone stress of qt.Campanella and Robertson 1982. Although a similar correctioncan be made to the sleeve stress, fs, the correction is rarely madewhen the cone has an equal end-area sleeve Lunne et al. 1997.

    Robertson 1990, based on the work of Wroth 1984, sug-gested using the following normalized CPT parameters to identifysoil behavior type SBT

    Qt1 = qt vo/vo 5

    Fr = fs/qt vo100% 6

    Bq = u2 u0/qt vo = u/qt vo 7

    where vo=preinsertion in situ total vertical stress; vo=preinsertion in situ effective vertical stress; u0=preinsertion insitu equilibrium water pressure; u2=measured pore pressure be-hind the cone; and u= u2u0=excess penetration porepressure.

    In the original paper by Robertson 1990 the normalized coneresistance was defined using the term Qt. The term Qt1 is usedhere to show that the cone resistance is the corrected cone resis-tance, qt and the stress exponent for stress normalization is 1.0.Although alternate methods to normalize CPT results have beensuggested e.g., Olsen and Malone 1988; Jefferies and Davies1991; Robertson and Wride 1998; Moss et al. 2006; Cetin andIsik 2007, especially in sands, and are more appropriate for awide range of soils, the original normalization suggested byWroth 1984, and shown in Eq. 5, will be used here to beconsistent with the simple normalization used by Marchetti1980 for DMT KD results. Hence, DMT KD and CPT Qt1 pa-rameters are normalized in a consistent manner using the verticaleffective stress. In the fullness of time, it is likely that DMT datawill become normalized using more complex techniques and thatfuture CPT-DMT correlations can use more appropriate normal-ized parameters. However, for typical stress levels in geotechnicalengineering of about 65200 kPa i.e., about 420 m, the nor-malization method has little influence on the normalizedparameters.

    Similar to Marchetti 1980, Robertson 1990 suggested thatthe CPT parameters reflect the mechanical behavior of the soiland not a soil classification based on grain size distribution andplasticity. Robertson 1990 suggested the term SBT to reflect themechanical characteristics of the soil measured using the CPT.NVIRONMENTAL ENGINEERING ASCE / NOVEMBER 2009 / 1763

    SCE license or copyright; see http://pubs.asce.org/copyright

  • Robertson 1990 suggested two charts based on either Qt1Fr orQt1Bq, but recommended that the Qt1Fr chart was generallymore reliable, as shown in Fig. 2.

    Jefferies and Davies 1993 identified that a SBT index, Ic,could represent the SBT zones in the Qt1Fr chart where Ic isessentially the radius of concentric circles that define the bound-aries of soil type. Robertson and Wride 1998 modified the defi-nition of Ic to apply to the Qt1Fr chart, as defined by

    Ic = 3.47 log Qt12 + log Fr + 1.2220.5 8Contours of the SBT index Ic, are shown on Fig. 3, where Qt1 isbased on Eq. 5.

    The CPT SBT index Ic can be used to represent the boundariesbetween different soil types, where Robertson and Wride 1998 Clays Ic2.95. Silt mixtures 2.05 Ic2.95. Sands Ic2.05.In general terms, the CPT SBT Ic can vary from 1 to 4.

    Cone Penetration TestDilatometer TestCorrelations

    Relatively few comprehensive correlations have been publishedbetween the DMT and normalized CPT parameters. Campanellaand Robertson 1991 suggested a link between normalized coneresistance, qt /vo and KD, in sands. Marchetti et al. 2001 sug-gested there was a link between the DMT constrained modulusMDMT and cone resistance, qt. Mayne and Liao 2004 sug-gested a link between ID and friction ratio Fr and between EDand qt, based on DMT and CPT data in Piedmont residuum soils.Mayne 2006 suggested interrelationships between the basicDMT measurements p0 and p1 and the CPTu measurements qtand u2 in soft clays.

    Many indirect correlations exist between DMT and CPT re-sults, since both tests are used to estimate various geotechnicalparameters. The main correlations used for both CPT and DMT

    Fig. 2. Normalized SBT charts for CPT after Robertson 19901764 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGIN

    Downloaded 07 Jan 2010 to 66.237.65.34. Redistribution subject to Aresults to estimate geotechnical soil parameters are for OCR, und-rained shear strength, peak friction angle and soil modulus. Cor-relations also exist for hydraulic permeability and shear wavevelocity.

    The literature has been reviewed for published records ofdocumented sites where adjacent CPT and DMT results are avail-able. Table 1 shows a summary of the published records fromadjacent CPT and DMT profiles in a wide range of soils. Thedepth range for the data shows that the methods used to normal-ized the results generally had little influence on the parameters.Unfortunately, some of the published records do not include ac-cess to the digital records of either CPT or DMT results andtherefore, estimates were made of the range in parameters fromthe published plots. Fortunately, digital records were available forsome of the published sites, and these are identified in Table 1.

    Soil TypeSince DMT ID and CPT Ic are both used to identify soil type,there is a strong possibility that a link exists between these nor-malized parameters. It is recognized that both parameters have thefollowing range: DMT 0.1 ID10. CPT 1.0 Ic4.0.Fig. 4 presents a summary of the published records in terms oflog ID versus Ic. Included on Fig. 4 are the common soil typeregions that overlap for both the CPT and DMT. Although indi-vidual values at each depth within a profile could be presented,the plots become crowded and confusing with many data points.Comparison between individual values from nearby in situ testprofiles at the same depth often show considerable scatter due tovariations in soil stratigraphy and consistency since many sitesare not uniform. Hence, adjacent in situ test data from the samedepth may not always represent the same soil. Any comparisonbetween in situ tests should be done in terms of the near continu-ous profiles with depth so that any variation in soil stratigraphycan be identified from the profiles. However, when there are alarge number of sites for comparison it is common to compare

    Fig. 3. Contours of SBT index, Ic on CPT normalized SBT Qt1FrchartEERING ASCE / NOVEMBER 2009

    SCE license or copyright; see http://pubs.asce.org/copyright

  • Table 1. Published Records from Adjacent DMT-CPT ProfilesDMTrangeED /v

    CPTrangeQt1

    CPTrange

    Fr %

    CPTrange

    Ic200600 40120 0.30.6 1.61.91430 24 1.52.5 3.33.61535 4.56 1.02.0 2.93.22040 46 1.02.5 3.13.31040 46 1.53.0 3.13.3

    400600 80100 0.40.6 1.61.82050 57 2.03.0 3.03.3

    130200 4080 0.50.9 1.82.11430 58 1.53.0 2.93.380175 1020 2.53.0 2.83.0

    110300 2555 1.42.2 2.32.5

    150250 3545 4.05.0 2.52.7

    70180 1230 7.09.0 2.93.240140 1520 0.91.2 2.52.7

    300500 80150 0.91.0 1.72.2

    3050 812 23 2.93.1100150 2060 1.52.5 2.52.7100200 2045 2.03.5 2.62.8100300 3050 3.56.0 2.62.9

    3050 57 0.41.0 2.93.1300800 80150 0.41.0 1.51.8

    JOURNAL

    OF

    GEO

    TECHNICALAN

    DG

    EOENVIRO

    NMENTAL

    ENGINEERING

    ASCE

    /NOVEM

    BER2009

    /1765

    rightNo. Site Soil Reference

    Depthrangem

    DMTrange

    ID

    DMTrange

    KD1a McDonalds Farm, BC, Canada Deltaic sand Campanella and Robertson 1991 512 3.08.0 261b McDonalds Farm, BC, Canada Soft silty clay Campanella and Robertson 1991 1730 0.20.3 232a Bothkennar, U.K. Soft clay Mayne 2006 315 0.30.4 233a Amherst, MA, U.S.A. Soft varved sensitive clay Mayne 2006 610 0.20.3 3.554a Ford Center, IL, U.S.A. Soft glacial clay Mayne 2006 716 0.10.3 355a Venice Lagoon, Italy Medium dense sand Marchetti et al. 2006 45 4.06.0 365b Venice Lagoon, Italy Soft clayey silt Marchetti et al. 2006 2930 0.30.5 236 Zelezny Mine, Poland Loose silty sandtailing Mlynarek et al. 2006 520 2.04.0 1.22.57 Hydraulic Fill, Brazil Loose silt and fine sandfill Penna 2006 48 0.20.3 238a Baton Rouge, LA, U.S.A. Stiff fissured clay Mayne 2006 1030 0.50.8 410

    9a Georgia Piedmont, U.S.A.Stiff silty sand to sandysiltresidual soil Mayne and Liao 2004 412 1.21.8 2.75.0

    10a Alabama Piedmont, U.S.A.Stiff silty sand, sandysiltresidual soil Mayne and Liao 2004 210 1.11.6 45

    11a North Carolina Piedmont, U.S.A.Stiff silty sand to clayeysiltresidual soil Mayne and Liao 2004 212 0.70.9 36

    12a Cooper Marl, SC, U.S.A. Stiff cemented silt Meng et al. 2006 2030 0.20.4 610

    13a Tainan, Taiwan Silty sandC. H. Juang and D.-H. Lee,personal communication, 2008 612 1.52.5 48

    14a Tainan, Taiwan Silty clayC. H. Juang and D.-H. Lee,personal communication, 2008 48 0.30.6 24

    15 Cowden, U.K. Very stiff clay Powell and Uglow 1988 410 0.50.7 51016 Brent Cross, U.K. Very stiff clay Powell and Uglow 1988 210 0.40.8 51517 Madingley, U.K. Very stiff clay Powell and Uglow 1988 212 0.50.8 816

    18 Pisa Clay Soft sensitive clayM. Jamiolkowski,personal communication, 2008 1220 0.20.3 34

    19 Univ of Central Florida, U.S.A. Sand to silty sand Anderson et al. 2007 35 2.05.0 48aSites where digital data for both CPT and DMT were available.

    Downloaded 07 Jan 2010 to 66.237.65.34. Redistribution subject to ASCE license or copyright; see http://pubs.asce.org/copy

  • Clay-Like Soils Ic>2.60, ID
  • u2/vo = Qt10.95 + 1.05 17The constant varies between 0.20.5 as the assumed val-ues for the undrained shear strength ratio for normally consoli-dated soils varies between 0.20 su /vo NC0.30 and soilrigidity index varies between 30 IR200, with an average=0.3 that represents

    su/vo NC = 0.25 and rigidity index, IR = 200 18

    Based on the observation that the corrected lift-off pressure p0 isessentially equal to the excess pore pressures u2 around theprobe in clays, it follows that

    KD = u2 u0/vo = u2/vo = Qt10.95 + 1.05 19where, on average, =0.3.

    Hence, KD should have similar values as the CPT parameteru2 /vo in soft clays. The relationship in Eq. 19 produces val-ues for KD that are remarkably similar to and are essentiallybounded by values from Eqs. 14 and 16, when 0.50.2,respectively. Eq. 19 was developed using a hybrid critical state-cavity expansion model for the CPT and assuming that the cor-rected DMT lift-off pressure p0 is essentially equal to the CPTpore pressures u2 around the probe. Eqs. 14 and 16 weredeveloped empirically based on case history observations withOCR.

    Fig. 5 presents a summary of the published records in terms oflog KD versus log Qt1, when Ic2.60. Included on Fig. 5 are therelationships represented by Eqs. 14, 16, and 19. Eq. 19with =0.3 appears to provide the best overall fit to the fieldcomparison data over the full range of values. Eq. 19 also showsthe correct trend in very soft clays i.e., at low values of Qt1,where 1.0KD2.0. The slightly higher measured values for KDin the region where Qt110, may be due to high penetration porepressures around the DMT probe in sensitive soils, since moresensitive soils generate higher pore pressures during probe pen-etration Schneider et al. 2008. Schneider et al. 2008 also de-veloped a relationship for excess CPT pore pressures in sensitiveclays of the form

    u2/vo = 0.67Qt10.91 + 1.1 20Eq. 20 can be related to KD assuming KD=u2 /vo and is alsoshown on Fig. 5 in the region where Qt110. Eq. 20 representsan approximate upper bound to the measured values. The clays

    Fig. 5. Summary of published average values from adjacent CPT andDMT profiles of Qt1 versus KD in fine-grained soils where Ic2.60see Table 1 for site detailsJOURNAL OF GEOTECHNICAL AND GEOE

    Downloaded 07 Jan 2010 to 66.237.65.34. Redistribution subject to Afrom sites 1b, 3, and 4 are somewhat sensitive and plot closer toEq. 20, as predicted.

    Sand-Like Soils Ic2.60, ID>1.0The DMT parameter KD is used extensively in many correlationsfor the DMT. Marchetti 1980 noted that for most normally con-solidated, uncemented, young soils, KD2.0. Robertson 1990noted that most normally consolidated, uncemented, young soilsplot down the center of the Qt1Fr chart, as shown in Fig. 2.Hence, there is a possibility that, in coarse-grained soils, KD var-ies with both Qt1 and Fr and likely follows approximately theregion marked normally consolidated on Fig. 2.

    Possible EDQt1 RelationshipMayne and Liao 2004 presented CPT and DMT data from threesites with Piedmont residual soils that are silty sands to sandy siltswith very high small strain stiffness that is possibly associatedwith cementation and suggested a correlation between ED and qt,as follows:

    ED = 5 qt 21

    Professor Mayne kindly made the digital CPT and DMT recordsavailable for these and other published sites. The data from thesethree sites fit equally well in terms of net cone resistance,qnet= qtvo, since qtvo at these sites. Hence

    ED = 5qt vo 22

    The normalized form then becomes

    ED/vo = 5 Qt1 23Fig. 6 presents a summary of the published records for all thesoils in terms of ED /vo versus Qt1 both on log scales, andshows that Eq. 23 provides a reasonable average fit to the data.The range of values presented in Fig. 6 suggests that the relation-ship can be represented in a more general manner by

    ED/vo = Qt1 24Based on Fig. 6, it would appear that 210, where mayvary with relative density, age, and stress history in a mannersimilar to the variation of the CPT modulus factor, E, with thesefactors Baldi et al. 1989; Lunne et al. 1997. Based on the pub-

    Fig. 6. Summary of published average values from adjacent CPT andDMT profiles of Qt1 versus ED /vo see Table 1 for site detailsNVIRONMENTAL ENGINEERING ASCE / NOVEMBER 2009 / 1767

    SCE license or copyright; see http://pubs.asce.org/copyright

  • lished field records shown in Fig. 6, it would appear that a valueof =5 is a reasonable average for a wide range of soils where

    1000

    155Qt1200.The data presented on Fig. 6 span a wide range of soils, from

    coarse-grained soils where the CPT and DMT results are essen-tially drained, to fine-grained soils where the test results are es-sentially undrained. Fig. 6 also includes some unusual soilsSchnaid et al. 2004 where the small strain stiffness Go is sig-nificantly higher than would be expected from the cone resistancei.e., high Go /qt ratios. This confirms the observation made byCampanella and Robertson 1991 that the DMT modulus ED isessentially a large strain response.

    Since ED /vo is also a function of ID and KD Eq. 4, itfollows that:

    34.7IDKD = 5 Qt1 25hence

    KD = 0.144 Qt1/ID 26Using the link between ID and Ic Eq. 10, this becomes

    KD = 0.144 Qt1/101.670.67Ic 27hence, there appears to be an approximate relationship betweenDMT KD and CPT Qt1 for different values of Ic in a wide range ofsoils where 5Qt1200. Since Eq. 27 is based on an averagevalue of =5, the relationship between KD and Qt1 will not beunique for all soils, since may vary with soil type, relativedensity, age, and stress history. However, Eq. 27 may representa framework for future refinements, as more well-documentedcomparison data becomes available. Although alternate normal-ization techniques could be used to normalize both KD and Qt1,any influence on the correlation in Eq. 27 will likely be small,provided consistent normalization is applied to both parameters.Also for typical stress levels in geotechnical engineering of about65200 kPa about 420 m, the normalization has little influenceon the CPT SBT index, Ic. When Ic2.60 ID1.0 the relation-ship between KD and Qt1 may be better captured by Eq. 19 seeFig. 5.

    Proposed Cone Penetration TestDilatometer TestCorrelations

    With the above observations as a framework i.e., Eqs. 10, 19,and 23, approximate contours of DMT ID and KD were devel-oped on the normalized CPT soil behavior chart Qt1Fr, asshown in Fig. 7. Fig. 7 shows that KD varies with both normalizedQt1 and Fr, except in the region defined by Ic2.60 ID0.85,where KD likely becomes essentially independent of Fr.

    The proposed correlations can be summarized, as follows:

    ID = 101.670.67Ic 28

    KD = 0.3Qt10.95 + 1.05 when Ic 2.60 29

    ED/vo = 5 Qt1 30The correlation for KD can be sensitive to the cutoff for Icwhen CPT data fall on or close to the boundary between fine-grained and coarse-grained soils. The suggested cutoff in Eq. 29is Ic2.60. However, some soils can straddle this value whichcan result in a rapid variation in estimated DMT KD values de-pending on the values for Qt1 and Fr. If this occurs the cut-off1768 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGIN

    Downloaded 07 Jan 2010 to 66.237.65.34. Redistribution subject to Avalue for Ic can be modified slightly to obtain a more smoothprofile of predicted KD. The suggested correlations for KD shownin Fig. 7 identifies a possible transition zone in the region com-prised of silt mixture soils where 2.40 Ic2.90 i.e., 1.2 ID0.60. This region represents a transition from primarily drainedCPT and DMT in sands Ic2.40 and ID1.2 to primarily und-rained CPT and DMT in clays Ic2.90 and ID0.60. DMTresults in this transition region of silt-mixture soils can be furtherinfluenced by possible drainage during the pause between pen-etration and testing.

    The contours for KD shown in Fig. 7 where Ic2.60 ID0.85 were based on ED /vo =5 Qt1. This relationship could beextended to stiff soils, where Ic2.60 ID0.85 and Qt120,since the suggested contours change little in form in this region.This is confirmed by the results presented by Mayne and Liao2004 for the stiff Piedmont soils that included one site site 11where Ic2.60.

    The suggested contours for KD in Fig. 7 may partly explainthe somewhat poor published correlations between KD and rela-tive density Dr and peak friction angle p in coarse-grainedsoils.

    The proposed correlations between normalized CPT param-eters Qt1 and Fr and DMT parameters ID, KD, and ED areapproximate and will likely be influenced by variations in in situstress state i.e., Ko, soil density, stress history, age, cementation,and soil sensitivity. The general relationship for KD in Eq. 29 forfine-grained soils where Qt110 will tend to under predict KD insensitive fine-grained soils, as illustrated in Fig. 5. The proposedcorrelations are unlikely to be unique for all soils but the contoursshown in Fig. 7 may form a framework for future refinements.

    The profiles of measured cone resistance qt may also differslightly from those of adjacent DMT since the CPT senses soilslightly ahead and behind the cone tip due to the size of the zoneof influence. Ahmadi and Robertson 2005 showed that the conecan sense a soil interface up to 15 cone diameters ahead andbehind, depending on the strength/stiffness of the soil and the insitu effective stresses. The DMT appears to be less influenced bysoil layers ahead and behind since the probe is stopped and themembrane expanded in a horizontal direction. Hence, in interbed-

    NORM

    ALIZE

    DCO

    NERE

    SISTANCE

    ,Qt1

    NORMALIZED FRICTION RATIO, Fr

    0.1 1 101

    10

    100

    1

    2

    3

    4

    5

    6

    7 8

    9

    KD

    6.2

    1.98

    0.85

    0.50

    0.18

    ID =

    ID =

    ID =

    ID =

    ID =

    10

    5

    2

    ID1510

    5

    2

    Fig. 7. Proposed contours of DMT KD and ID on the CPT normalizedSBT Qt1Fr chartEERING ASCE / NOVEMBER 2009

    SCE license or copyright; see http://pubs.asce.org/copyright

  • ing sitded soils the CPT may be influenced by adjacent soil layers some-what more than the DMT.

    Recently, two CPTs and one DMT were carried out at adjacentlocations only 1 m apart at the Moss Landing site WoodwardMarine in California. Details about the Moss Landing site areprovided by Boulanger et al. 1997. The results from two adja-cent CPT soundings at Moss Landing are shown in Fig. 8. Thesite is composed of about 2.6 m of silty sand to silt over about 4.4m of sand. Below the sand is a deposit of firm plastic clay ex-tending to a depth of 13.4 m. A thin soft clay layer is at a depth of6 m within the sand deposit. The ground water level is at a depthof about 2.2 m below ground surface but fluctuates somewhatwith the tide. The two CPT profiles show good repeatability interms of tip resistance qt and friction ratio Rf down to about 11m, after which the sand layers show considerable differences interms of depth, thickness, and density, especially the sand layersbetween 13 and 15 m. Some soil liquefaction was observed atshallow depth at the site during the 1989 Loma Prieta earthquakeBoulanger et al., 1997. A comparison between measured andpredicted DMT parameters ID, KD, and ED with depth based onone of the CPT profiles CPT-04, using Eqs. 2830, is shownin Fig. 9. Note that the Moss landing data were not used in de-veloping the correlations shown in Figs. 46. The comparisonplots for ID and ED are presented on log scales to present therange of values more clearly. In general, the comparison betweenmeasured DMT parameters and those predicted from the CPTusing the proposed correlations show reasonable trends. The rapidand large variations in both ID and ED are fully captured. It isinteresting to note the small scale effect where the CPT appearsto sense the soft clay layer at a depth of about 6 m, which isearlier than the DMT responds to the same layer. Eq. 19 under-predicts KD in the clay layer between 7.5 and 13 m where Icvalues are close to or slightly larger than 2.60. In places, e.g., at8.7 and 12 m the CPT Ic values fall just below the cutoff of 2.60which produces a sudden change in predicted KD. This illustratesthe sensitivity of the proposed correlations for KD in transitionsilt-mixture soils where 2.40 Ic2.90. The predicted DMTvalues are significantly different from the measured values be-

    Fig. 8. CPT profile at Moss LandJOURNAL OF GEOTECHNICAL AND GEOE

    Downloaded 07 Jan 2010 to 66.237.65.34. Redistribution subject to Atween 12.5 and 14 m, which may, in part, be due to rapid varia-tions in these sand layers, as indicated from the two CPT profilesshown in Fig. 8. In general, the Moss Landing site provides agood test for the proposed correlations since the soils range fromsoft to firm clay and loose to dense sand.

    Marchetti 1997 has suggested that the DMT is more sensitiveto a surface crust than the CPT. However, previous publishedcomparisons were carried out using non-normalized CPT param-eters. It is possible that the CPT may also be sensitive to a surfacecrust when the comparison is carried out using normalized values,since the normalized cone resistance Qt1 increases close to theground surface due to the low vertical effective stresses.

    Conclusions

    The CPT and DMT have been used worldwide for over 30 years.Each test has certain advantages and limitations Mayne et al.2002. A preliminary set of correlations is proposed that links thekey DMT parameters ID, KD, and ED to normalized CPT param-eters Qt1 and Fr. The proposed correlations are approximate andwill likely be influenced by variations in in situ stress state, soildensity, stress history, age, cementation, and soil sensitivity. Theproposed correlations are unlikely to be unique for all soils butthe suggested relationships may form a framework for future re-finements. The proposed general relationship for KD in fine-grained soils Ic2.60 will tend to under predict KD in sensitivesoils. The trends in the proposed correlations illustrated in Fig. 7may provide further insight into possible future correlations forthe DMT with other geotechnical parameters and design applica-tions since the CPT has a somewhat more extensive theoreticalbackground compared to the DMT as well as a larger database ofdocumented case histories for certain applications e.g., liquefac-tion evaluation.

    The correlations presented are based on a simple but consis-tent normalization of the key parameters KD and Qt1. The result-ing correlations between normalized CPT and DMT parameters

    e Woodward Marine, CaliforniaNVIRONMENTAL ENGINEERING ASCE / NOVEMBER 2009 / 1769

    SCE license or copyright; see http://pubs.asce.org/copyright

  • may be somewhat influenced by the normalization technique.However, provided consistent normalization methods are appliedto each in situ test the correlations may not change significantly,although further research will be required to verify thisassumption.

    Acknowledgments

    This research could not have been carried out without the support,encouragement and input from John Gregg, Kelly Cabal, andChris Atwell, and other staff at Gregg Drilling and Testing Inc.The sharing of data by Professors Paul Mayne, Silvano Marchetti,Hsein Juang is also appreciated.

    References

    Ahmadi, M. M., and Robertson, P. K. 2005. Thin layer effects on theCPT qc measurement. Can. Geotech. J., 429, 13021317.

    Anderson, J. B., Townsend, F. C., and Rahelison, L.2007. Load testingand settlement prediction of shallow foundation. J. Geotech. Geoen-viron. Eng., 13312, 14941502.

    Baldi, G., Bellotti, R., Ghionna, V. N., Jamiolkowski, M., and Lo Presti,D. F. C. 1989. Modulus of sands from CPTs and DMTs. Proc.,12th Int. Conf. on Soil Mechanics and Foundation Engineering,Balkema, Rotterdam, The Netherlands, 165170.

    Boulanger, R. W., Mejia, L. H., and Idriss, I. M. 1997. Liquefaction at

    Fig. 9. Comparison between measured DMT parameters ID, KD, andsite, California. Note that ID and ED are presented on log scale.1770 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGIN

    Downloaded 07 Jan 2010 to 66.237.65.34. Redistribution subject to AMoss Landing during Loma Prieta earthquake. J. Geotech. Geoenvi-ron. Eng., 1235, 453468.

    Campanella, R. G., and Robertson, P. K. 1982. State-of-the-art in in-situ testing of soils: Developments since 1978. Proc., EngineeringFoundation Conf. on Updating Subsurface Sampling of Soils andRocks and Their In-Situ Testing, Santa Barbara, Calif., 245267.

    Campanella, R. G., and Robertson, P. K. 1991. Use and interpretationof a research dilatometer. Can. Geotech. J., 28, 113126.

    Cetin, K. O., and Isik, N. S. 2007. Probabilistic assessment of stressnormalization for CPT data. J. Geotech. Geoenviron. Eng., 1337,887897.

    Douglas, B. J., and Olsen, R. S. 1981. Soil classification using electriccone penetrometer. Proc., Symp. on Cone Penetration Testing andExperience, ASCE, New York, 209227.

    Jefferies, M. G., and Davies, M. P. 1991. Soil classification by the conepenetration test. Can. Geotech. J., 281, 173176.

    Jefferies, M. G., and Davies, M. P. 1993. Use of CPTU to estimateequivalent SPT N60. Geotech. Test. J., 164, 458468.

    Kulhawy, F. H., and Mayne, P. H. 1990. Manual on estimating soilproperties for foundation design. Rep. No. EL-6800, Electric PowerResearch Institute, EPRI, Palo Alto, Calif.

    Ladd, C. C. 1991. Stability evaluation during staged construction22ndTerzaghi Lecture. J. Geotech. Eng., 1174, 540615.

    Lunne, T., Robertson, P. K., and Powell, J. J. M. 1997. Cone penetrationtesting in geotechnical practice, Routledge, New York.

    Marchetti, S. 1980. In-situ tests by flat dilatometer. J. Geotech. Eng.,1063, 299321.

    Marchetti S., 1997. The flat dilatometer: Design applications. 3rdGeotechnical Engineering Conf., Keynote lecture, Cairo Univ., Cairo,

    ith depth and predicted parameters using the CPT at Moss LandingED wEERING ASCE / NOVEMBER 2009

    SCE license or copyright; see http://pubs.asce.org/copyright

  • Egypt, 421448.Marchetti, S., Monaco, P., Calabrese, M., and Totani, G. 2006. Com-

    parison of moduli determined by DMT and backfigured from localstrain measurements under a 40 m diameter circular test load in Ven-ice area. Proc., 2nd Int. Conf. on the Flat Dilatometer, Washington,D.C., In-Situ Soil, Virginia, 220231.

    Marchetti, S., Monaco, P., Totani, G., and Calabrese, M. 2001. TheDMT in soil investigations. A report by the ISSMGE TC 16. Proc.,Int. Conf. on In Situ Measurement of Soil Properties and Case Histo-ries, Bali, Indonesia, Parahyangan Catholic Univ., Bandung, Indone-sia, 95132.

    Mayne, P. W. 2001. Stress-strain-strength-flow parameters from en-hanced in-situ tests. Proc., Int. Conf. on In-Situ Measurement of SoilProperties and Case Histories, Bali, Indonesia, Parahyangan CatholicUniv., Bandung, Indonesia, 2748.

    Mayne, P. W. 2006. Interrelationships of DMT and CPT readings insoft clays. Proc., 2nd Int. Conf. on the Flat Dilatometer, Washington,D.C., 231236.

    Mayne, P. W., and Bachus, R. C. 1989. Penetration pore pressure inclay by CPTU, DMT, and SBP. Proc., 12th Int. Conf. on Soil Me-chanics and Foundation Engineering, Rio de Janeiro, Balkema, Rot-terdam, The Netherlands, 291294.

    Mayne, P. W., Christopher, B. R., and DeJong, J. 2002. Subsurfaceinvestigations: Geotechnical site characterization. Pub. No. FHwANHI-01031, National Highway Institute, Arlington, Va.

    Mayne, P. W., and Liao, T. 2004. CPT-DMT interrelationships in pied-

    Olsen, R. S., and Malone, P. G. 1988. Soil classification and site char-acterization using the cone penetrometer test. Proc., PenetrationTesting 1988, ISOPT-1, Vol. 2, J. De Ruiter, ed. Balkema, Rotterdam,The Netherlands, 887893.

    Penna, A. 2006. Some recent experience obtained with DMT in Brazil-ian soils. Proc., 2nd Int. Conf. on the Flat Dilatometer, Washington,D.C., In-Situ Soil, Virginia, 170177.

    Powell, J. J. M., and Uglow, I. M., 1988. The interpretation of theMarchetti dilatometer test in UK clays. Proc., Penetration Testing inthe UK. Thomas Telford, London, 269273.

    Robertson, P. K. 1990. Soil classification using the cone penetrationtest. Can. Geotech. J., 271, 151158.

    Robertson, P. K. 2009. Interpretation of cone penetration testsA uni-fied approach. Can. Geotech. J. in press.

    Robertson, P. K., Campanella, R. G., Gillespie, D., and By, T. 1988.Excess pore pressures and the dilatometer. Proc., 1st Int. Symp. ofPenetration Testing, ISOPT-1, Balkema, Rotterdam, The Netherlands,567576.

    Robertson, P. K., Campanella, R. G., Gillespie, D., and Greig, J. 1986.Use of piezometer cone data. Proc., In-Situ 86 Use of In-Situ Test-ing in Geotechnical Engineering, GSP 6, ASCE, Reston, Va., 12631280.

    Robertson, P. K., and Wride, C. E. 1998. Evaluating cyclic liquefactionpotential using the cone penetration test. Can. Geotech. J., 353,442459.

    Schnaid, F., Lehane, B. M., and Fahey, M. 2004. In situ characteriza-

    mont residuum. Proc., 2nd Int. Conf. on Geophysical and Geotech-nical Site Characterization, ISC-2, Porto, Portugal, Millpress,Rotterdam, The Netherlands, 345350.

    Mayne, P. W., and Martin, G. K. 1998. Commentary on Marchetti flatdilatometer correlations in soils. Geotech. Test. J., 213, 222239.

    Meng, J., Hajduk, E. L., Casey, T. J., and Wright, W. B. 2006. Obser-vations from in-situ testing within a calcareous soil. Proc., 2nd Int.Conf. on the Flat Dilatometer, Washington, D.C., In-Situ Soil, Vir-ginia, 237243.

    Mlynarek, Z., Gogolik, S., Marchetti, S., and Marchetti, D. 2006. Suit-ability of the SDMT method to assess geotechnical parameters ofpost-floatation sediment. Proc., 2nd Int. Conf. on the Flat Dilatom-eter, Washington, D.C., In-Situ Soil, Virginia, 148153.

    Moss, R. E. S., Seed, R. B., and Olsen, R. S. 2006. Normalizing theCPT for overburden stress. J. Geotech. Geoenviron. Eng., 1323,378387.JOURNAL OF GEOTECHNICAL AND GEOE

    Downloaded 07 Jan 2010 to 66.237.65.34. Redistribution subject to Ation of unusual geomaterials. Proc., 2nd Int. Conf. on Site Charac-terization, ISC2, Millpress, Rotterdam, The Netherlands, 4974.

    Schneider, J. A., Randolph, M. F., Mayne, P. W., and Ramsey, N. 2008.Influence of partial consolidation during penetration on normalizedsoil classification by piezocone. Proc., 3rd Int. Conf. on Site Char-acterization, Geotechnical and Geophysical Site Characterization, A.B. Huang and P. W. Mayne, eds., Taylor and Francis, London, 11591165.

    Smith, M. G., and Houlsby, G. T. 1995. Interpretation of the Marchettidilatometer in clay. Proc., 11th European Conf. on Soil Mechanicsand Foundation Engineering, Copenhagen, Denmark, 1.2471.253.

    Wroth, C. P. 1984. The interpretation of in-situ soil tests. RankineLecture. Geotechnique, 4, 449489.

    Yu, H. S. 2004. In-situ soil testing: From mechanics to interpretation.Proc., 2nd Int. Conf. on Site Characterization, ISC2, Millpress, Rot-terdam, The Netherlands, 338.NVIRONMENTAL ENGINEERING ASCE / NOVEMBER 2009 / 1771

    SCE license or copyright; see http://pubs.asce.org/copyright