reporter: s. amoroso 6th february 2013 analysis of results and parameters derived from sdmt banbury...

Reporter: S. Amoroso 6th February 2013 www.marchetti-dmt.it Analysis of results and parameters derived from SDMT Banbury United Kingdom

SDMT Test Layout Measurements performed after penetration independant from insertion method DMT (static) SDMT (dynamic)

DMT test: P 0 & P 1 every 20cm Z (m) P 0 (kPa) P 1 (kPa) 0.20 0.40 0.60 0.80 1.00 1.20 220 210 305 310 285 290 300 310 420 450 380 390

DMT Intermediate parameters Intermediate Parameters Id: Material Index DMT Readings P0P0 P1P1 Kd: Horizontal Stress Index Ed: Dilatometer Modulus

DMT Formulae Interpreted parameters Intermediate Parameters Id Kd Ed Interpreted Parameters Cu: Undrained Shear Strength Ko: Earth Pressure Coeff (clay) OCR: Overconsolidation ratio (clay) : Safe floor friction angle (sand) : Unit weight and description M: Constrained Modulus

DMT Formulae (1980) Po and P1 Intermediate parameters Interpreted parameters

I D contains information on soil type Performing DMT, immediate notice that: p 1 CLAY p p 0 SAND p 0 p 1 p SILT falls in between

I D contains information on soil type SAND CLAY

Obviously I D is not a sieve analysis. Eg. a mixture sand-clay would probably be "wrongly" interpreted as silt. On the other hand such mixture could perhaps behave mechanically as a silt. The engineer is often interested to the grain size distribution not "per se", but just to infer mechanical properties, PERHAPS, in SOME cases, it could be better to have the I D interpretation than the sieve analysis results and to infer from them the mechanical behaviour. A mechanical information (a sort of Soil Type Behaviour Index) that, in design, might be even more important than the granulometric composition. Reliability of material index I D

K D contains information on stress history K D is an amplified K 0, because p 0 is an amplified h due to penetration K D = vv (p 0 - u 0 ) K D well correlated to OCR and K 0 (clay) p0p0 DMTDMT formula similar to Ko: (p 0 u 0 ) h

Depth Z Kd K D contains information on stress history 2 K D = 2 in NC clay (OCR = 1) NC OC K D > 2 in OC clay (OCR > 1)

K D contains information on stress history OC Kd > 2 NC Kd 2 Taranto 1987

Experimental Kamei & Iwasaki 1995 Theoretical Finno 1993 Theoretical Yu 2004 OCR=KdKd 1.56 Marchetti 1980 (experimental) 0.5 K D correlated to OCR (clay)

Theoretical 2004 Yu Experimental Marchetti (1980) K0K0 = KdKd 0.47 Marchetti 1980 (experimental) 1.5 0.6 K D well correlated to K 0 (clay)

E D contains information on deformation Theory of elasticity: E D = elastic modulus of the horizontal load test performed by the DMT membrane (D=60mm, 1.1 mm expansion) 1.1 mm DMTDMT ED =ED = 34.7 (P 1 - P 0 ) Gravesen S. "Elastic Semi-Infinite Medium bounded by a Rigid Wall with a Circular Hole", Danmarks Tekniske Hjskole, No. 11, Copenhagen, 1960, p. 110. E D not directly usable corrections (penetration,etc)

M obtained from Ed using information on stress history (Kd) and soil type (Id) EdEd KdKd IdId M Constrained Modulus

How can be an undrained modulus Ed related to a drained modulus M In the early days of DMT (1980) the initial idea was obviously to correlate Ed-Eu, not Ed-M. But Ed-Eu impossible. Lab Eu values too dispersed. M values from oedometers less dispersed. Though the link Ed-M is presumably weaker, at least it can be tested. A correlation Ed-M must be a complex function of many variables, among them the Skempton pp parameters A & B and anisotropy (horiz. to vert. modulus), which in turn depend on soil type (to some extent represented by Id) and on OCR (to some extent represented by Kd). These considerations encouraged investigating Ed-M using Id,Kd as parameters. (Prof. Lambe of MIT (Jnl ASCE March 1977 Foundation Performance of Tower of Pisa p.246) wrote: E typically 1/3-1/4 Eu. Thus a connection E-Eu already invoked in the past). Final word goes to real world. Several decades of observations appear to confirm that Mdmt is a reasonable estimate of the operative M.

Ed must be corrected to obtain M M=Rm Ed with Rm=f(Kd,Id) Dont use Ed as Youngs Rm has various correction tasks Distortion Horiz to vertical Drained Undrained Once Ed is converted to M, Youngs E 0.8-0.9 M (elasticity)

Vertical drained confined tangent modulus (at ' vo ) Same as E oed, traditionally measured by oedometer Usual range M 0.4 - 400 Mpa Except highly structured clays (sharp break), M variation across pc is moderate Error in assuming M ~ constant : often acceptable (other methods for M : not infrequent error factors 2-3) May use M = constant if 'v large ? M = Eoed=1/mv= ' v / v (at ' vo )

Ladd 1971 Terzaghi 1967 Compressibility of even good samples

M Comparison from DMT and from Oedometer Norwegian Geotechnical Institute (1986). "In Situ Site Investigation Techniques and interpretation for offshore practice". Report 40019-28 by S. Lacasse, Fig. 16a, 8 Sept 86 ONSOY Clay NORWAY Constrained Modulus M (Mpa) Tokyo Bay Clay - JAPAN Iwasaki K, Tsuchiya H., Sakai Y., Yamamoto Y. (1991) "Applicability of the Marchetti Dilatometer Test to Soft Ground in Japan", GEOCOAST '91, Sept. 1991, Yokohama 1/6 Virginia - U.S.A. Failmezger, 1999

Cu correlation from OCR Ladd SHANSEP 77 (SOA TOKYO) Ladd: best Cu measurement not from TRX UU !! Using m 0.8 (Ladd 1977) and (Cu/ v ) NC 0.22 (Mesri 1975) Cu vv OC = Cu vv NC OCR m OCR=0.5KdKd 1.56 Cu = vv 0.5 KdKd 1.25 0.22 best Cu from oed OCR Shansep

Cu comparisons from DMT and from other tests Mekechuk J. (1983). "DMT Use on C.N. Rail Line British Columbia", First Int.Conf. on the Flat Dilatometer, Edmonton, Canada, Feb 83, 50 Skeena Ontario Canada Tokyo Bay Clay - JAPAN Iwasaki K, Tsuchiya H., Sakai Y., Yamamoto Y. (1991) "Applicability of the Marchetti Dilatometer Test to Soft Ground in Japan", GEOCOAST '91, Sept. 1991, Yokohama 1/6 Recife - Brazil Coutinho et al., Atlanta ISC'98

Complaint : Cu field vane > Cu dmt (in very plastic clay) From Book Soil Mechanics in Eng. Practice (by Terzaghi, Peck, Mesri) Cu field vane needs a correction factor before it can be used in stability analysys. The Bjerrum correction is eg 0.70 when PI = 70. Cu field vane reduced by Bjerrums correction is often considered the best available Cu for stability analysis. The DMT 1980 correlations for Cu were developed using for calibration such operative Bjerrums-corrected Cu values. It is therefore Cu field vane uncorrected which is too high - in plastic clays.

Summary of DMT 2 step data processing Id(soil type) Kd(stress history) Ed(elastic modulus) M Cu Ko OCR DMT Readings Intermediate Parameters Geotechnical Parameters P0P1P0P1

Main SDMT applications Settlements of shallow foundations Compaction control Slip surface detection in OC clay Quantify ' h relaxation behind a landslide Laterally loaded piles Diaphragm walls FEM input parameters Liquefability evaluation Seismic design (NTC08, Eurocode 8) In situ G- g decay curves

Main Application: Settlement prediction 1-D approach (classic Terzaghi) Primary settlement at working loads (Fs 2.5-3 to b.c.) M must be treated as if by oedometer LOADSOIL DMT Boussinesq S = z M zz v M Z z

Circular area: Settlement Prediction from DMT M is evaluated for each Dz (0.20 m) by DMT Ds is evaluated for each Dz (0.20 m) by Poulus & Davis q = load Primary settlement is evaluated by Poulus & Davis (1974)

Rectangular area: Settlement Prediction from DMT D Fadum Abacus Rectangular area D M is evaluated for each Dz (0.20 m) by DMT Ds is evaluated for each Dz (0.20 m) by Fadum Abacus q = load I = influence value Primary settlement is evaluated by

Numerous case histories of favourable comparisons measured vs DMT-predicted settlements (or moduli): Lacasse & Lunne (1986) Dilatometer Tests in Sand. Proc. In Situ '86 ASCE Spec. Conf. Virginia Tech, Blacksburg.Very good agreement between DMT-predicted and measured settlements under a silos at a sandy site Steiner W. (1994) Settlement Behaviour of an Avalanche Protection Gallery Founded on Loose Sandy Silt. Settlement '94 ASCE Conf. at Texas A&M.The DMT- predicted settlements agreed well with observed settlements Mayne & Liao Tianfei (2004) CPT-DMT interrelationship in Piedmont residuum. Proc. In ISC2 Porto. Over two decades of calibration between the DMT and measured foundation performance records have shown its value & reliability in settlements computation Vargas (2009), Bullock (2008), Monaco (2006), Lehane & Fahey (2004), Mayne (2001, 2004), Failmezger (1999, 2000, 2001), Crapps & Law Engineering (2001), Tice & Knott (2000), Woodward (1993), Iwasaki et al. (1991), Hayes (1990), Mayne & Frost (1988), Schmertmann (1986,1988), Steiner (1994), Leonards (1988), Lacasse (1986). Settlement Prediction from DMT

Summary of comparisons DMT- predicted vs. observed settlements Monaco et al. (2006) Large No. of case histories good agreement for wide range of soil types, settlements, footing sizes Average ratio DMT-calculated/observed settlement 1.3 Band amplitude (ratio max/min) < 2 i.e. observed settlement within 50 % from DMT-predicted

M observed vs. predicted by DMT M by DMT vs. M back-calculated from LOCAL vertical strains measured under Treporti full- scale test embankment (Italy) Marchetti et al. (2006) Sliding Micrometers installed every meter

Treporti Test Embankment (Venezia) Conclusion: OC increases stiffness especially at operative modulus (working strain) Before embankment constructionAfter embankment removal

Applicability of Oedometer, SPT, CPT, PMT, DMT, to predict settlements of shallow foundations (GeoRisck ASCE Failmezger & Bullock 2011) Oedometer:...Testing is time- consuming and is typically performed at depth intervals exceeding 3 m...Sampling and handling disturbance... SPT:...The hammer type is often omitted... Extrapolation from a failure strain to an intermediate strain CPT:...Extrapolation from a failure strain to an intermediate strain PMT:... Static deformation to strain the soil to intermediate strains... Relatively slow test... Drillers skill and experience DMT:...Static deformation to strain the soil to intermediate strains... The dilatometer test is therefore the best choice of in-situ tests for settlement prediction of shallow foundations... Mayne (2001)

Settlement Prediction DMT vs SPT

Possible reasons DMT good settlement predictions 1.Wedges deform soil less than cones Baligh & Scott (1975) measure zone 2. Modulus by mini load test relates better to modulus than penetration resistance Stiffness Strength 3.Availability of Stress History parameter Kd Jamiolkowski (1988) Without Stress History, impossible to select reliable E (or M) from Qc Robertson et al. (1986) Prediction of soil stiffness from cone resistance can be rather poor, especially for OC soils Leonards (Asce 88) Calculating settlements on granular soils based on correlations [Penetr. Resistance Soil Modulus] will seriously overestimate settlements if deposit has been prestressed. Similar statements by Schmertmann 70, Terzaghi 67

RATIO = E/Qc OCR??? CC Jamiolkowski: = 2.5 to 25. Factor 10 ! Depends on OCR(?) Jamiolkowski concludes (Isopt-1, '88, Vol. 1, p.263) : "without Stress History it is impossible to select reliable E (or M) from Qc"

Lee 2011, Eng. Geology CC in sand Effects of Stress History on CPT and DMT Effect of stress history on norm. Qc (x 1.10-1.15) Effect of stress history on Kd (x 1.30-2.50)

PRESTRAINING CYCLES simulated AGING (similar mechanism: grain slippage) CC TEST N. 216 IN TICINO SAND Jamiolkowski (ISC'98 Atlanta) applied prestraining cycles in calibration chamber. Found : K D (DMT) 3 to 7 times more sensitive to AGING than penetration resistance K D + 20 % q D + 3 % K D ++ sensitive to Stress History and aging than penetration resistance

Stress History also fundamental for liquefiability (e.g. Jamiolkowski 1985) Lack of SH : probably reason high scatter in the CPT- liquefaction correlations, possibly reduced with the SH info from Kd or use directly Kd-CRR for liquefaction (eg Fig.14 Rob 2012). Kd : thanks for existing a formidable parameter for settlem. & liquef. Appears the only parameter readily available today reflecting clearly SH not many SH tools Yet for decades Terzaghi, Skempton, Leonards, Schmertmann, Jam have been preaching (in essence) : without SH go nowhere. Kd is a bargain.

Example: Settlement Prediction from DMT

DMT for Compaction Control The high sensitivity to changes of stresses and density make the DMT particularly suitable for detecting benefits of SOIL IMPROVEMENT Compaction of a loose sandfill Resonant vibrocompaction technique Van Impe, De Cock, Massarsch, Meng New Delhi (1994) Depth (m)

DMT vs CPT sensitivity to Compaction Schmertmann (1986) DYNAMIC COMPACTION of sand site. M DMT % increase twice % increase in q c. Jendeby (1992) monitored DEEP COMPACTION in a sand fill by VIBROWING. M DMT increase twice increase in q c. Pasqualini & Rosi (1993) VIBROFLOTATION job : "DMT clearly detected improvement even in layers where benefits were undetected by CPT". Ghent group (1993) before after CPTs DMTs to evaluate effects ( h, Dr) by PILE (Atlas) INSTALLATION "DMTs before-after installation demonstrate more clearly [than CPT] beneficial effects of Atlas installation".

Compaction Control DMT vs CPT Jendeby (1992): Qc & Mdmt before & after compaction of a loose sandfill Before compaction After compaction

Subgrade Compaction Control M DMT acceptance profile (max always found at 25-26 cm) Bangladesh Subgrade Compaction Case History 90 km Road Rehabilitation Project Acceptance M DMT profile fixed and used as alternative/fast acceptance tool for quality control of subgrade compaction, with only occasional verifications by originally specified methods (Proctor, CBR, plate), (Marchetti, 1994)

Slip surface detection in OC clay slopes DMT-K D method Verify if an OC clay slope contains active (or old quiescent) slip surfaces (Totani et al. 1997)

Slip surface detection in clay slopes SS. N. 83 Marsicana Gioia dei Marsi (2006) blocked Mine of lignite S. Barbara (San Giovanni Valdarno)

Validation of DMT-K D method LANDSLIDE "FILIPPONE" (Chieti) LANDSLIDE "CAVE VECCHIE" (S. Barbara) DOCUMENTED SLIP SURFACE (inclinometers)

healthy clay Reconstruction of multiple slip surfaces active: Kd=2 quiescent: Kd=2 qualitative recontruction infected clay (K D 2 due to active/quiescent slip surfaces)

Quantify ' h relaxation behind a landslide Case History: Landslide in Milazzo, Sicily Horizontal Stress h Z (m) above sea level 1 2 3 h obtained using K 0 from DMT RAILWAY 12 3 clay

Design of laterally loaded piles (Winkler) Three different methods using DMT results for evaluating P-y curves for laterally loaded piles: E s = constant Deflection y Soil reaction, p Linear P-y curve E s = f (y) Deflection y Soil reaction, p Non Linear P-y curve Recommended methods Gabr & Borden (1988) Robertson et al. (1989) Marchetti et al. (1991)

Observed vs. DMT predicted pile deflections single pile, 1st time monotonic loading In clay Validation of: 2 independent methods (Robertson 1989 and Marchetti 1991) provide similar predictions, in very good agreement with measured full-scale pile behaviour (1989)

DMT for DESIGN of DIAPHRAGM WALLS Tentative correlation for deriving the Winkler model springs for design of multi- propped diaphragm walls from M DMT Indications on input moduli for FEM analyses (PLAXIS Hardening Soil model) based on M DMT Monaco & Marchetti (2004 ISC'2 Porto)

Linear elastic model: E 0.8 M DMT (Hamza & Richards, 1995) DMT aims to calibrate FEM parameters PLAXIS hardening soil model: E 50,ref is correlated to M DMT (Schanz, 1997) FEM input parameters Monaco & Marchetti (2004)

LIQUEFACTION RISK ASSESSMENT very cautious recommendations using SPT and CPT Robertson & Wride (1998) CRR by CPT adequate for low-risk projects. For high-risk: estimate CRR by more than one method Youd & Idriss (NCEER Workshops 2001) use 2 or more tests for a more reliable evaluation of CRR Idriss & Boulanger (2004) the allure of relying on a single approach (e.g. CPT-only) should be avoided Jamiolkowski (1985, 11 ICSMFE) reliable predictions of CRR require the development of some new in situ device [other than CPT or SPT] much more sensitive to the effects of past STRESS AND STRAIN HISTORIES Leon et al. (ASCE GGE 2006) South Carolina sands. Ignoring AGING and evaluating CRR from in situ tests insensitive to aging (SPT, CPT, VS) underestimated CRR by a large 60 % Monaco & Schmertmann (ASCE GGE 2007) Disregarding AGING omitting a primary parameter in the correlation predicting CRR

Liquefaction: CRR from DMT Correlations for evaluating Cyclic Resistance Ratio (CRR) from K D developed in past 2 decades, stimulated by: 1.Sensitivity of K D to factors known to increase liquefaction resistance: stress history prestraining/aging cementation structure 2.Correlation K D Relative Density D r (Reyna & Chameau, 1991, Tanaka & Tanaka,1998) 3.Correlation K D In situ State Parameter (relative density + stress level) (Yu, 2004) Intuitively K D expresses propensity/reluctance of sand to decrease in volume ( !! )

Liquefaction: K D related to D r K D - D r correlation Reyna & Chameau (1991) Tanaka & Tanaka (1998)

Liquefaction: K D related to CRR from Mayne 2009 LIQUEFACTION NO LIQUEFACTION K d - correlation Yu (2004) theoretical = vertical distance between the current state and the critical state line in the usual v - ln p' plot

alone: incomplete indicator of Liquefaction resistance lacks structure, stress history, aging: applying / removing load causes only small e ( small ), but big CRR It does not appear illogical to expect that Kd, being related to , but at the same time incorporating stress history and aging, could be uniquely well correlated with CRR governs the attitude of a sand to increase or decrease in volume when sheared, hence it is strongly related to liquefaction resistance

Have seen various reasons for expecting good Kd-CRR. But how to translate the large experimental base behind Qc1-CRR? ( e.g. Youd & Idriss 2001). Translation done by Tsai (2009). He first determined a Kd-Qc1 correlation by running side-by-side CPT-DMT in loose saturated clean sand. Then he used said Kd-Qc1 correlation to replace Qc1 with Kd in Youd & Idriss, thereby obtaining a correlation CRR-Kd.

Youd & Idriss 2001 Tsai translated the CRR-Qc database into CRR-Kd Kd Side-by-side CPT-DMT parallel profiles of Qc1-Kd Qc1=f(Kd) CRR=f(Qc1) CRR=f(Kd) (scatter)

At first sight one might consider doubtful the resulting Kd-CRR correlation, being based on the highly dispersed Qc1-Kd correlation. Not so. The scatter is just natural, is the consquence of Kd reacting to factors unfelt by Qc1. E.g. for a same Qc1, there can be many Kd - depending if the site has had Stress History. Scatter is healthy. If there was no scatter : Qc1 and Kd contain the same information, i.e. Qc1 reactive to SH as Kd. Not so. Dispersion of the Qc1-Kd relation

Reason of the dispersion of the Qc1-Kd curve The fact that the translation occurs via the average eliminates that part of scatter due to the insensitivity of Qc1 to stress history. Hence expectable Kd-CRR less scatter.

Dispersion of intercorrelations Qc1-Kd-CRR

SDMT for LIQUEFACTION Monaco et al. (2005) ICSMGE Osaka SDMT 2 independent evaluations of CRR from K D and V S (Seed & Idriss 1971 simplified procedure) Andrus & Stokoe (2000) Andrus et al. (2004) CRR from VsCRR from K D

Jamiolkowski 1992 Soils & Foundations SDMT provides two independent CRR estimates From Kd From Vs Sometimes different CRR. We consider more reliable CRR(Kd) Vs insensitive to STRESS HISTORY Waves produce strains far too small to initiate trend to dilate/contract (essence of liquefaction) Vs measured on sand specimen in the calibration chamber during loading and unloading (Jamiolkowski and Lo Presti, 1992)

Vittorito LAquila (Earthquake, 6th April 2009) Kd Vs Moment magnitude MW: 6.3 Distance from the epicentre: 45 km Peak ground acceleration PGA: 0.065 g CSR Liquefaction case history in Italy LAquila

Monaco et al. (2009, 2010) Liquefaction depth from Vs: 1-2.5 m 0 0.1 0.2 0.3 0.4 0.5 0.6 050100150200250 Normalized shear wave velocity, Vs1 (m/s) Cyclic Stress Ratio, CSR or Cyclic Resistance Ratio, CRR Fc = 35% LIQUEFACTION NO LIQUEFACTION Liquefaction depth from K D : 2-6 m 0 0.1 0.2 0.3 0.4 0.5 0246810 Cyclic Stress R a tio CSR or Cyclic Resistance Ratio CRR K D Proposed CRR-K D curve (Monaco et al. 2005) LIQUEFACTION NO LIQUEFACTION Both Kd and Vs indicated Liquefaction (red points) Liquefaction case history in Italy LAquila

Liquefaction case history in Costarica Just a few weeks after the SDMT execution, the cyclic wave action due to a storm induced liquefaction of the soil deposit.. (Vargas & Coto 2012) cofferdam Design Earthquake (M Richter = 7,5 and PGA = 0,25 g) LIQUEFACTION NO LIQUEFACTION LIQUEFACTION NO LIQUEFACTION

Diagram groups results of 34 international sites in various soils & geography SDMT same depth values for: Id, Kd, M, Go (Vs) M, Id, Kd may provide rough Vs in previous DMT sites G 0 (Vs) M = G0G0 constant G0G0 M 0.5 - 20 Correlation to estimate Vs (G 0 ) from mechanical DMT data (I D, K D, E D ) Marchetti et al. (2008) Would it be possible predict Go from one-number test (no stress history)?

G 0 /M DMT for detecting cementation as a consequence of these data analysis, it becomes clear that both [G 0 /E D vs. I D ] and [G 0 /M DMT vs. K D ] can be used to detect the presence of cementation.. (Cruz, 2010)

measured by SDMT estimated from "mechanical" DMT data Vs profiles Earthquake in LAquila, 6 April 2009 Monaco et al. (2009)

Simplified use of Vs for Seismic Design Vs profile Vs 30 Soil category (NTC08, Eurocode 8)

Vs for Seismic Design Go profile (Vs) EERA, ProShake (or similar software) auxiliary input soil surface behaviour Input motion Output motion Bedrock Soil Period, T AGI (2005)

SDMT small strain modulus G 0 from Vs working strain modulus G DMT from M DMT (Marchetti et al. 2008) Tentative methods to derive in situ G- curves by SDMT Two points help in selecting the G- curve In situ G- decay curves by SDMT 0.05 0.1 % Mayne (2001) 0.01 1 % Ishihara (2001)

Treporti Venice (Italy), Texas: SDMT vs observed settlements LAquila (Italy): SDMT vs dynanic laboratory tests Western Australia: SDMT vs SBP, SDMT vs triaxial tests In situ G- decay curves by SDMT Amoroso et al. (2012) 2%

Conclusions 1/2 A CPT investigation costs less, but remains orphan of capability of providing SOA predictions of settlements 220 papers ISC4: no one Qc for settlements Countless researchers: Qc insensitive to Stress History W/o info Stress History, impossible predict well settlements Robertson (1986) Prediction of soil modulus from Qc can be rather poor, with a large potential error This well known Qc weakness is no little thing. Often settlement prediction is a > 50% task in a geo-report (Similar considerations for liquefiability, where Stress History also fundamental) pay less, get less - unless uninterested in settlements

Conclusions 2/2 SDMT is simple, accurate, cost-effective, repeatable and supplies results real time Practical : Any operator gets same results. No need highly skilled workers. Short training time Robust correlations to design parameters based on intermediate parameters: Id, Ed and Kd Key parameter is Kd, which captures stress history and is sensitive to aging, prestraining, cementation and structure (fundamental for settlements and liquefaction) SDMT provides two moduli in situ: G 0 (Vs) - low strain M DMT operative (settlement prediction) Used in many everyday applications

Thank you for your attention

reporter: s. amoroso 6th february 2013 analysis of results and parameters derived from sdmt banbury...

Documents

d slide

clay p p

ocr clay slide

dmt membrane d

sand p

h slide

p silt

constrained modulus