Keyword: PIRT, KPS, stabilisation, clay/silt, strength Draft · verification is of paramount...

Draft

Strength Verification of Stabilised Soil-Cement Columns: A

Laboratory Investigation of the Push-In Resistance Test (PIRT)

Journal: Canadian Geotechnical Journal

Manuscript ID cgj-2016-0230.R1

Manuscript Type: Article

Date Submitted by the Author: 15-Dec-2016

Complete List of Authors: Timoney, Martin; Arup Consulting Engineers, formerly of National

University of Ireland, Galway McCabe, Bryan; National University of Ireland, Galway, Civil Engineering

Keyword: PIRT, KPS, stabilisation, clay/silt, strength

https://mc06.manuscriptcentral.com/cgj-pubs

Canadian Geotechnical Journal

Draft

1

Strength Verification of Stabilised Soil-Cement Columns:

A Laboratory Investigation of the Push-In Resistance Test (PIRT)

Martin J. Timoney, PhD, MIEI ([email protected])

Project Engineer, Arup, 50 Ringsend Road, Dublin 4, Ireland, formerly of the College of Engineering & Informatics, National University of Ireland, Galway, Ireland.

Bryan A. McCabe, PhD, CEng, MIEI, Eur Ing ([email protected])

Lecturer, College of Engineering & Informatics, National University of Ireland, Galway, Ireland.

Corresponding author: +353 91 492021 [email protected]

Word Count: 6797 (Introduction to Acknowledgements)

Page 1 of 50



Draft

2

Abstract: The Push-In Resistance Test (PIRT) is an in situ means of verifying the

strength of stabilised soil columns. During the test, a winged penetrometer penetrates the

stabilised column at a constant rate and the column strength is estimated from the

recorded probing resistance using a bearing capacity factor, N. While N values between 8

and 20 (although typically between 8 and 15) are quoted in the literature (based almost

exclusively on empirical and Scandinavian experience), there have been few field tests

and no laboratory investigations aimed at investigating the value of N and the factors

upon which it depends, thereby limiting international confidence in the method.

This paper presents the findings of a unique laboratory-scale PIRT series conducted at

NUI Galway. The development of appropriate column construction and testing methods is

discussed. The results of 11 no. PIRT tests on pre-drilled stabilised columns with

unconfined compression strengths (UCS) from 150 kPa to over 800 kPa are reported, as

well as those of a complementary cone-only series of tests to assess additional frictional

forces acting on the penetrometer. Appropriate corrections to the data for temperature and

time-consistency between the probing forces and UCS values are discussed. Test results

indicate that the strength of the column has a mild influence on the N value. Further

investigation of this finding is recommended at both laboratory and field scales.

Keywords: PIRT, KPS, cement, stabilisation, clay, silt, column, laboratory testing

Page 2 of 50



Draft

3

Introduction

Soil stabilisation (or soil mixing) is a method of ground improvement in which

cementitious and/or pozzolanic binder materials (such as lime, Ordinary Portland Cement

(OPC), Ground Granulated Blast Furnace Slag (GGBS), Pulverised Fuel Ashes (PFA) or

combinations of these) are mixed with a soil in situ to enhance its strength and stiffness

characteristics (EuroSoilStab 2002). Dry soil mixing (DSM) is one such method in which

the binder is injected as a dry powder and is appropriate for soils with natural moisture

contents in excess of approximately 40% (Topolnicki 2012) as this moisture content is

sufficient to instigate the binder hydration reactions.

Developed in Scandinavia in the 1970s (Bredenberg 1999), Deep Dry Soil Mixing

(DDSM) provides a method with which stabilised columns can be created in soft

clays/silts, peats and other weak soils. A mixing tool is inserted into the ground, rotating

at a specific rate to break up the soil. Once the design depth is reached, the mixing tool is

retracted as the binder is injected and mixed with the parent soil. DDSM columns can be

constructed as either single units, rows or interlocking panels, to depths of up to 30 m

(Topolnicki 2012), although depths are typically less than 10 m. Column diameters

between 0.5 m and 1.0 m are typical in European practice (Dahlström 2012), with

diameters up to 1.6 m possible using Japanese equipment (Kitazume and Terashi 2013).

Applications of DDSM include foundations for roads, railways and light commercial and

residential structures, slope and embankment stability solutions, confinement and

remediation of contaminated soils, vibration reduction solutions for rail applications and

temporary works (Dahlström 2012). In recent years DSM processes have been deployed

beyond the traditional markets of Europe (principally Scandinavia) and Japan, with

projects carried out in the USA (Burke et al. 2007; Filz 2009; Hussin and Garbin 2012),

Australia (Liyanapathirana and Kelly 2011) and Asia (Horpibulsuk et al. 2012), and

therefore it is important that there is international confidence in the methods used to

establish in situ stabilised field strengths.

The factors which influence the effectiveness of the stabilisation process have been

generalised into four categories by a number of authors (e.g., Babaski and Suzuki 1996;

Kitazume 2005) as follows:

(i) Soil characteristics: moisture content (Babaski et al. 1996; Porbaha et al.

2000; Jacobson 2002; Åhnberg et al. 2003; Marzano et al. 2009), organic

Page 3 of 50



Draft

4

content (Tremblay et al. 2002; Timoney et al. 2012a), pH (Kitazume 2005)

and the particle size distribution. Many soft soils are inherently variable in

nature and this variability can be reflected in the characteristics of the

stabilised soil.

(ii) Binder characteristics: the type of binder added to the soil (Janz and Johansson

2002; Axelsson et al. 2002; EuroSoilStab 2002; Timoney et al. 2012a).

(iii) Mixing conditions: the quantity of binder added to the soil (Babaski et al.

1996; Porbaha 2000; Axelsson et al. 2002; Kitazume 2005; Horpibulsuk et al.

2011) and the degree and duration of mixing (Larsson 2003).

(iv) Curing conditions: the time (Nagaraj et al. 1996; Jacobson 2002; Horpibulsuk

et al. 2003; Åhnberg 2006; Filz 2009), temperature (Marzano et al. 2009;

Åhnberg and Holm 2009) and confining pressure (Åhnberg 2007) under which

the stabilised soil cures.

Site-specific binder trials can help estimate achievable column strengths provided the

laboratory conditions are representative of the field conditions under categories (i)-(iv).

This can be very difficult to achieve in practice, as evidenced by variable field-to-

laboratory strength ratios reported by Larsson (2003); therefore in situ strength

verification is of paramount importance. Although rotary coring, in situ column shear

vane (Larsson 2006) and column extraction (Holm et al. 1999) methods are used to assess

in situ column strengths directly, the most commonly-used method remains the Push-In

Resistance Test (PIRT), also known in Scandinavia as Kalk-Pelar-Sondering (KPS) or the

Conventional Column Penetration Test (CCP).

The probing resistance recorded during the PIRT is related to the column strength using a

bearing capacity factor, N. Some guidance exists in the literature on appropriate values of

N (Åhnberg and Holm 1986; Carlsten and Ekström 1996; EuroSoilStab 2002; Larsson

2006; Trafikverket 2011), mainly based on Scandinavian experience, but with limited

supporting data.

In this paper, a reduced-scale PIRT facility developed at NUI Galway is presented. These

experiments, believed to be the first of their kind, were conducted under controlled

laboratory conditions, an advantage over previous PIRT studies. The research is intended

to provide some new insights into the N value applicable to stabilised clays/silts spanning

a wide range of stabilised strengths.

Page 4 of 50



Draft

5

Push-In Resistance Test (PIRT)

PIRT Details

During a PIRT, a winged penetrometer (Figure 1) is advanced into the column at a

constant rate and the force-depth profile recorded. Details and guidance on in situ

penetrometer dimensions are provided by Carlsten and Ekström (1996), Larsson (2006)

and TK Geo 11 (Trafikverket 2011), see Figure 1a. The width of the penetrometer (B) is

typically 75% of the diameter of the column to be tested or 100 mm less than the column

diameter. The leading edge of the penetrometer has a circular bulb shape to reduce

friction along the vertical plate section of the wing; this is typically either 15 mm or 20

mm in thickness (d), while the wing thickness is 5 mm less than that of the leading edge.

The cone tip of the penetrometer is typically 50 mm in diameter and sounding rod

diameters are between 36 mm and 50 mm. The larger diameter bars are used in high

strength columns where probing resistances are high and there is a risk of the bars

bending during the PIRT (Carlsten and Ekström 1996).

PIRT offers a number of conceptual advantages over the popular Cone Penetration Test

(CPT). CPT cones only probe at a point location within the column cross-section and tend

to follow the weakest path, i.e., that previously followed by the kelly bar of the mixing

tool, whereas in the case of PIRT, a chord of the column is probed. In addition, CPT

cones are more prone to deviation from vertical than the winged penetrometer,

particularly in strong columns.

Equipment used for carrying out a PIRT can be either truck-mounted (as used for CPT),

excavator/track-mounted or wheel-loader-mounted (Hussin and Garbin 2012). The

penetrometer is pushed into the column at a constant rate of 20±4 mm/sec (Carlsten and

Ekström 1996; EuroSoilStab 2002; Larsson 2006; Trafikverket 2011), similar to that used

in the CPT (Lunne et al. 1997), until it has reached a hard stratum or, in the case of

floating columns, has passed 2 m beneath the column base. Column lengths of 6 m to 8 m

can normally be tested, but in long and/or strong columns, the penetrometer can deviate

off vertical and exit the column (Halkola 1999). Current Swedish guidelines suggest that

columns longer than 6 m to 8 m (Larsson 2006; Trafikverket 2011) or columns with

undrained shear strengths greater than 300 kPa should be pre-drilled. This has been

shown to prevent the penetrometer deviating from the centre of the column as the hole,

typically 50 mm to 65 mm in diameter, provides a guide for the penetrometer tip to

Page 5 of 50



Draft

6

follow (Axelsson and Larsson 2003) and testing of columns up to 15 m long has been

achieved in this way (Bergman and Larsson 2014).

Interpretation of PIRT

From the results of a PIRT, the undrained shear strength (cu) of a stabilised column is

estimated using Equation 1, where N is an overall dimensionless bearing capacity factor

for the device, P is the probing force (kN) (after correction for sounding bar friction) and

A is the projected plan area of the wings and cone of the penetrometer (m2). For pre-

drilled columns, the area of the drill hole is deducted from the projected area of the

penetrometer. The relationship in Equation 1, proposed by Boman (1979), is a

simplification of the semi-empirical equation for the Iskymeter penetrometer, a device

used for sounding in soft ground (Kallstenius 1961).

A

P

Ncu ×=

1 1

Early guidance (Åhnberg and Holm 1986) recommended the use of an N value of 8 and

10 for 400 mm wide penetrometers with thicknesses of 15 mm and 20 mm, respectively,

while more recent guidance documents (Carlsten and Ekström 1996; EuroSoilStab 2002;

Larsson 2006; Trafikverket 2011) recommend the use of an N value of 10. N values

reported for PIRT in the literature from field studies are compiled in Table 1, the majority

of which originate from Scandinavian experience.

While an N value of 10 appears to have a broad (although not complete) acceptance, it is

difficult to draw a firm conclusion from the data in Table 1 due to:

(i) Variation in the stabilised materials tested

(ii) Variation in penetrometer dimensions

(iii) Differences in the test methods by which the column strength was measured

or inferred, and

(iv) Limits of strength for which the quoted N value has been derived.

Furthermore, the variation in N observed in the literature has prompted some authors

(Axelsson and Rehnman 1999; Bergman et al. 2013) to suggest that N may be a function

of the strength and stiffness properties. Axelsson and Rehnman (1999) found good

agreement between strengths (estimated using N = 10) from PIRT data and UCS tests on

Page 6 of 50



Draft

7

column samples, and suggest that a somewhat higher value may be required for regions

with shear strengths less than 100 kPa. Finite element modelling reported by

Liyanapathirana and Kelly (2011) indicated a dependence of N on both the rigidity index

and the probe-soil interface roughness, concluding that an N value of 12 to 14 is

reasonable.

Other relevant penetrating devices

As mentioned, the convention for interpreting PIRT data is to use a single N value, and

there is no reference in the literature, including the numerical work of Liyanapathirana

and Kelly (2011), to separate bearing factors for the wings and cone. However, T-bar

penetrometers and Deep Penetrating Anchors (DPA) provide some insight into the

relative contributions of the PIRT penetrometer components. This is a relevant frame of

reference given that the columns were pre-drilled for the tests reported here, as there was

no cone-bearing contribution to the total force.

T-bar Penetrometer:

Stewart and Randolph (1994) have shown that for a 50 mm dia. and 200 mm long T-bar

penetrometer, the relationship between undrained shear strength and T-bar resistance is

dependent on the surface roughness of the T-bar with bearing factors (Nb) of between 9

and 12 suggested for smooth and rough probes, respectively. Using a bearing factor of

10.5, recorded resistances from field T-bar tests in a soft clay deposit were used to

estimate the undrained shear strength and the results were found to be in agreement with

shear vane and triaxial test results. The authors also noted the similarities in plan area and

bearing factor with that of the Iskymeter probe.

Deep Penetrating Anchors (DPA):

Used in mooring offshore structures, DPAs consist of a cylindrical shaft with a number of

flukes or wings attached to the upper end and a rounded conical tip (O’Loughlin et al.

2004). While obviously not used for in situ testing, the similarities between DPAs and the

PIRT penetrometer are noteworthy. O’Loughlin et al. (2004) describes an equation to

calculate the pull-out capacity of a DPA to include (i) the weight of the device

(significant for DPAs), (ii) frictional resistance along the wings and shaft, and (iii)

bearing resistance along the wings and end of the shaft. Published bearing N values are

between 9 and 12 for the conical tip of the shaft and 7.5 for the flukes.

Page 7 of 50



Draft

8

Laboratory Experimental Work

Soil and Stabilisation Properties

An organic clayey silt from Kinnegar, Co. Down, Northern Ireland, known locally in the

Belfast area as sleech, was used as the parent soil for construction of the stabilised

columns. Extensive classification and testing of the sleech has been reported by Crooks

and Graham (1976), Bell (1977), McCabe (2002) and McCabe and Lehane (2006) and

details are provided in Table 2.

Approximately 5 m3 of sleech was obtained from between 3.0 m to 4.5 m below ground

level and transported to the laboratory in 1 m3 bulk containers. Swedish fall cone tests on

undisturbed samples, taken during sampling, indicated an undrained shear strength of

approximately 15 kPa and a remoulded strength of 4.5 kPa. Moisture contents (wi) at the

time of stabilisation ranged from 42% to 56% and in some cases are below the in situ

ranges reported by the aforementioned studies (see Table 2) due to drying out that

occurred due to reworking of the sleech during the construction process and during

storage between tests. The sleech was found to have a plastic limit of approximately 28%

and a liquid limit of approximately 75%. Results from loss on ignition tests on the sleech

used for the columns, carried out at 440 °C, ranged from 2.2% to 4.6%.

The Unconfined Compression Strength (UCS, q) is the typical measure of strength used

in DSM applications, and is equivalent to twice the undrained shear strength, i.e., q = 2cu.

It was envisaged that columns with various UCS values would be constructed, covering at

least the range of strengths (approximately 100 kPa to 600 kPa) represented in Table 1.

To assess achievable strengths, an initial series of stabilisation trials were performed in

which OPC, GGBS and lime binders were used to stabilise the sleech. From the results of

these trials, OPC at binder contents of 75, 100 and 150 kg/m3 were deemed most

appropriate for the testing programme (Timoney 2015).

Curing Basin and the Reduced-scale PIRT Penetrometer

Columns with a diameter of 200 mm were constructed, cured and tested in a 750 mm

diameter, 1.0 m high High-Density Polyethylene (HDPE) pipe sealed at the base with a 6

mm HDPE base plate. The column diameter was chosen following consideration of

possible methods to form the column and transpired to be one-quarter that of the

commonly-used diameter of 800 mm. The basin diameter was chosen based on a

Page 8 of 50



Draft

9

simplified axisymmetric finite element model, ensuring that the horizontal stress increase

at the basin boundary due to the expansion caused by the insertion of the penetrometer

was no greater than 5% of the increase experienced at the edge of the column (Timoney

2015).

Using guidance provided by Carlsten and Ekström (1996), two one-quarter scale PIRT

penetrometers were designed and manufactured. The stainless steel penetrometers (Figure

2) had a width (B) of 150 mm, leading edge thickness (d) of 6 mm and a cone diameter of

14 mm; the penetrometer shaft and the sounding bars were 10 mm in diameter. A bespoke

cone-only penetrometer (Figure 2a) was also developed, originally intended to assess

cone bearing resistance, but ultimately used in pre-drilled columns for a series of

sounding bar friction tests to isolate the force on the penetrometer wings only.

PIRT Column Construction

A total of 14 no. stabilised columns were constructed and subjected to PIRT testing. The

first three (PI-1 to PI-3) were preliminary trials to devise a robust test methodology;

details of the remaining 11 column tests (PI-4 to PI-14) are provided in Table 3 and

details of an additional 4 PIRTs carried out in the unstabilised sleech surrounding the

columns are provided in Table 4.

The use of a reduced-scale mixing tool was considered for column production. However,

based on verticality, binder delivery and mixing challenges encountered by Larsson

(1999), Al-Tabbaa and Evans (1999) and Kosche (2004), it was deemed that higher

quality columns could be produced using casting techniques. Due to significant deviations

of the PIRT penetrometer during initial trials, the aforementioned pre-drilling stage was

replicated to enable the penetrometer remain within the column. This was achieved by

pre-forming a hole in the column using a 13 mm diameter metal bar during the

construction process, and extracting it immediately prior to the test.

Using two semi-circular pieces of 200 mm diameter pipe, bound together with adhesive

tape, columns were constructed in four lifts within the curing basin. The stages below

correspond to those depicted in Figure 3:

(i) The form pipe was placed on a 200 mm thick layer of sleech, which represents

unstabilised soil beneath a floating column; for construction purposes the sleech

directly beneath the column was held in a 225 mm diameter, 200 mm high pipe.

Page 9 of 50



Draft

10

(ii) EuroSoilStab (2002) laboratory procedures for column applications (applicable

to soft cohesive soils) were followed to create each stabilised mixture, using

either 75, 100 or 150 kg/m3 of OPC. The first stabilised mix was compacted in

40mm layers within the form pipe.

(iii) A 13 mm diameter hole-form bar was inserted into the centre of the column to

create the guide hole. The form pipe was removed (Figure 4) and unstabilised

sleech was carefully moulded by hand around the column.

(iv) The form pipe was replaced around the top of the column and bound closed with

adhesive tape.

(v) A second stabilised mix was created and compacted into the form pipe as

described in step (ii).

(vi) Steps (iii) to (v) were repeated until the desired column height was reached, with

an overall construction time of 6 to 7 hours. The hole form bar was left in place

during curing and removed just prior to the PIRT.

To simulate columns at greater depths and to increase confinement around the top of the

column, PI-9S and PI-14S were cured and tested under a surcharge. This surcharge was

applied over the entire top surface of the 750 mm dia. basin, loading both the column and

the surrounding sleech. The surcharge magnitude of 13.7 kPa, limited by testing setup

constraints, was nevertheless of the same order as surcharge loading applied in practice.

PIRT Column Testing

Columns were allowed to cure for between 0.9 and 11.9 days (Table 3), with the

maximum strengths achieved at the latter time adjudged to be approaching the limit of the

testing equipment. The reduced-scale PIRT penetrometer was then pushed into the

column using the NUI Galway CPT rig, mounted on an independent reaction frame over

the test basin (Figure 5a). Column probing was carried out in two pushes, with rates

designated δ1 (first push) and δ2 (second push), with a short pause in between required to

connect additional sounding bars. In keeping with the full-scale PIRT, a target push-in

rate of 20±4 mm/sec was sought. While some of the rates shown in Table 3 are slightly

above this range, consideration of research on rate effects, including that by Brown and

Hyde (2008) and Robinson and Brown (2013), indicates that such relatively small rate

differences will have a negligible effect on strength. The push-in force and penetrometer

displacement were recorded by a 5 kN load cell and a draw wire gauge, respectively

Page 10 of 50



Draft

11

(Figure 5b). During testing, minor buckling of the sounding bars was observed due to

their slenderness, resulting in contact with the column, the extent of which cannot be

quantified.

A total of four reference PIRTs were also carried out in the unstabilised sleech that

surrounded a number of columns (Table 4). These were carried out in the same manner as

those in the stabilised column with the second PIRT penetrometer. In two of these, a 13

mm diameter hole was formed in the sleech immediately prior to the test, while in the

other two, pre-drilling was not replicated.

Once tested, each column was carefully exhumed from the basin and the location of any

cracks was noted. Cylindrical samples, 50 mm in diameter, were obtained from the

column where possible (i.e., from intact sections of column) and their UCS determined in

accordance with EuroSoilStab (2002) procedures. For quality assurance purposes, the

uniformity of the surrounding unstabilised sleech was assessed during the exhumation

process through its moisture content and bulk density (determined using 50 mm diameter,

50 mm long core samples) and its shear strength (determined using a 49.4 mm by 33 mm

shear vane).

Cone and Sounding Bar Tests

The cone-only penetrometer (shown in Figure 2a) was originally envisaged as a means of

ascertaining the bearing resistance of the cone element of the PIRT penetrometer in

isolation. However, as occurred in the trial PIRT tests, the cone-only penetrometer was

found to deviate excessively from verticality. As the PIRT columns were pre-drilled, the

cone-only tests enabled quantification of the frictional contribution and therefore allowed

the force on the penetrometer wings to be isolated. An additional series of 7 such tests

was carried out on 104 mm diameter pre-drilled stabilised columns. Each column was

constructed, cured and tested in a 104 mm diameter pipe without any surrounding sleech,

from a single stabilised mix, with construction taking approximately 1 hour. Pre-drilling

was replicated using the same 13 mm diameter bar as used in the PIRT series and the hole

was offset from the centre to allow for recovery of 50 mm diameter samples for UCS

testing (Figure 6). A summary of the test details is provided in Table 5. For comparison,

Page 11 of 50



Draft

12

a cone-only penetrometer test (see Table 4) was also carried out in the unstabilised sleech

surrounding column PI-14S.

Experimental Results and Interpretation

PIRT Probing Force Data

The variation in the uncorrected probing force profile (Pu) with height from the basin base

(h, relative to the leading edge of the penetrometer wings) is shown in Figure 7 for tests

PI-4 to PI-14S. An expanded version of the notation PI-X (Yd-Zc) is used in this and

subsequent figures, to aid in their interpretation, where X is the test number, Y is the

curing time (rounded to the nearest day) and Z is the cement content (kg/m3). The data in

Figure 7 are grouped according to curing times: (a) 1-4 days, (b) 6 days and (c) 12 days.

The following general observations can be made about the probing force profiles at the

top, middle and bottom of each column:

(i) Top of the column: At the beginning of each test, the force increased instantly as

the penetrometer began to press into the column. The delayed increase in the case

of PI-4, PI-9S and PI-14S is due to a thin layer of sleech at the top of those

columns. The force peaked and subsequently dropped as the column cracked

ahead of the penetrometer. This drop is less significant in the case of the

surcharged columns; for example, the force drops from a peak of 2.8 kN to 1.1 kN

for PI-10, but from 2.8 kN to 1.9 kN for PI-14S, even though UCS values in the

top 200 mm, corrected to the time of PIRT are similar at 688 kPa (PI-10) and 634

kPa (PI-14S). After this drop the force rises again before becoming constant or

gradually increasing with depth, albeit with fluctuations.

(ii) Middle of column: A temporary reduction in force identifies where the test was

paused to insert additional sounding bars. Interestingly, the recommencement of

probing after the addition of sounding bars does not show similar peaks and drops

in force to those occurring at the top of the column. It appears that the increased

confining stresses around the column at this level limit the amount by which the

column can crack open. In PI-12*, the varied binder content column, a reduction

in force was recorded between h = 615 mm and h = 420 mm, i.e., the vertical

extents of the 100 kg/m3 Mix 2, and showed a notable drop at h ≈ 450 mm.

(iii) Bottom of column: Within the last 50 mm of the column, the force dropped,

believed to be due to a crack forming ahead of the penetrometer and extending to

Page 12 of 50



Draft

13

the base of the column. In all cases, the magnitude of the probing force further

reduced upon entering the unstabilised sleech underneath the column.

Upon exhumation, the softer columns showed little evidence of crack development due to

PIRT probing, although a weak plane was noted along boundaries between individual

mixes (Figure 8a). However, the stronger columns behaved in a more brittle fashion; with

cracks typically running from the centre of the column diagonally downwards and

outwards to the face of the column at approximately 45° (Figure 8b). In places, especially

in the mid-sections of columns whose q values lay approximately between 300 kPa and

450 kPa, cracking was significant and limited the amounted of samples that could be

retrieved in these regions. From a comparison of the Pu profiles in Figure 7, it is likely

that the jaggedness in the profiles is related to the increased material brittleness and the

localised effects of cracking. The soft columns, in which few cracks were observed, show

a smoother profile.

Two of the PIRTs in unstabilised sleech were performed alongside columns also

subjected to PIRTs, i.e., PI-13-U and PI-14S-U, where U denotes that the material was

unstabilised. PI-13-U and PI-14S-U were pre-drilled. The other two tests (PI-T-U1 and

PI-T-U2) were performed in the sleech alongside an instrumented column that was

constructed solely to assess the impact of curing temperature, i.e. a PIRT was not

performed on the column. PI-T-U1 and PI-T-U2 were not pre-drilled. The probing force

profile with depth for the PIRTs carried out in the unstabilised sleech surrounding the

columns typically increased from 0.1 kN at the top to 0.25 kN at the bottom. This

increase is believed to be due the mild increase in strength (cu = 7-12 kPa) with depth and

friction due to the adhesive nature of the sleech.

PIRT Column Strength Data

In Figure 9, q values at the time of the UCS test (qmeas) are plotted for each individual

column sample against the depth of the midpoint of each sample, from which it can be

seen that a wide range of qmeas values (between 148 kPa and 816 kPa) has been obtained,

as intended, by varying the binder content and curing time.

The number of samples obtained from each column ranged from 7 to 22 with column

cracking (caused by the PIRT itself) affecting the number of samples retrievable and their

heights (with samples measuring between 75 mm and 108 mm in height).

Page 13 of 50



Draft

14

Column secant stiffnesses at 50% of the failure strength (E50) were found to range from 5

MPa to 110 MPa (Figure 10) and increased with both time and binder content. E50/qmeas

were found to lie between 45 and 130, comparing well with the range 50-150 found in

literature. The value of E50/qmeas reduced with increasing strain to failure; further details

are available in Timoney (2015).

Column Strength Correction

Given the inevitable delay between the PIRT on the column and UCS testing of the

column’s samples (between 4 and 10 hours), it is inappropriate to relate qmeas and Pu

values in the context of the subsequent calculation of an N value. A framework was

developed to this end to allow the q value at the time of PIRT to be estimated; this value

is referred to as the corrected UCS (qcor). The framework unites all PIRT data by

accounting for the principal variables of curing time, curing temperature, moisture

content and binder content (represented in Table 3).

Curing Temperature: Column temperatures during curing were measured as part of this

research through a series of thermistors embedded within a column constructed

specifically for this purpose (Timoney 2015). Column temperatures were compared to the

average ambient room temperature (T) recorded by the Building Management System

(see Table 3) and it was found that the column temperature equalised to ambient

temperature within 12 hours of stabilisation. Therefore in this study, the average ambient

temperature of the laboratory during curing (see Table 3) provides a reasonable measure

of the curing temperature of the column. Temperature variations were accounted for by

adjusting the curing time using Equation 2, a formula used for adjusting concrete curing

times for temperature (BSI 2008):

[ ]( )tet

n

i

T

adj ×=∑=

−+−

1

65.13273/4000 2

where Temperature-Adjusted Time (tadj) is the adjusted curing time of the column

samples and T is the average ambient temperature (°C) over a curing period of t (days).

Moisture Content and Binder Content: It has been shown that the effect of variations in

moisture content and binder content on stabilised strengths can be assessed using a single

Page 14 of 50



Draft

15

parameter such as water-to-binder ratio (WTBR, η) or a similar parameter (Åhnberg et al.

1995; Jacobson et al. 2005; Filz 2009; Timoney et al. 2012a, 2012b). η is defined as the

mass per unit volume of water (mw) divided by the mass per unit volume of (active)

binder (mb), and can be calculated using Equation 3 (Timoney et al. 2012a):

+

==

i

b

soil

b

w

wm

m

m

11

ρη

3

where ρsoil is the bulk density of the soil (kg/m3), mb is the binder content (kg/m3) and wi

is the initial moisture content of the unstabilised soil. Binder trials reported by Timoney

(2015) have shown that q varies approximately linearly with 1/η for cement-only binders,

rendering 1/η a suitable parameter by which to normalise q (i.e. the normalised parameter

is qη).

Values of (qmeasη)avg are plotted against their corresponding tadj values for all PIRTs in

Figure 11. This correlation is stronger than that found by plotting q against time. In

keeping with the mathematical form of the normalised strength versus time relationship

proposed by other authors for cement-stabilised soils (Nagaraj et al. (1996), Horpibulsuk

et al. (2003; 2011; 2012), Åhnberg (2006) and Filz (2009)), a best-fit natural log

relationship was arrived at and is shown in Equation 4 (R2 = 0.787). The standard

deviation in (qmeasη)avg is also shown.

(qmeasη)avg = 531.78 ln(tadj) +1010 (1 ≤ tadj ≤ 12) 4

The aforementioned qcor is ascertained for each individual sample from Equation 5, using

(i) the qmeas value, (ii) the η of the actual sample and (iii) the term ∆(qmeasη)avg, determined

as the difference between (qmeasη)avg at tadj relevant to the UCS test and (qmeasη)avg back

calculated using tadj relevant to the time at which the PIRT was carried out; both of these

values are determined from Equation 4. Corrections to the strength ranged from 2.5 kPa

to 25 kPa and, as expected, were most significant for early-age columns where the rate of

strength gain is at its highest.

Page 15 of 50



Draft

16

( )η

ηavgmeas

meascor

qqq

∆−=

5

PIRT Cone-Only Results

Probing force profiles for the cone-only tests (Pc) in columns are shown in Figure 12,

using the notation C-X (Yd-Zc), where X is the test number, Y is the curing time (to the

nearest 0.1 day) and Z is the cement content (kg/m3). Pc values were near-constant with

depth (with values up to 0.25 kN) indicating that minimal friction occurred in the cone-

only tests between the sounding bars and the column; bending of the sounding bars did

not occur due to the lower probing forces. Values of qcol ranged from 195 kPa to 839 kPa

and showed similar strengths with depth in any one column. E50 values of 8 MPa to 108

MPa were observed and E50/qcol was found to range from 60 to 125.

The same correction strategy described previously was applied based on data in Table 5

but with an equivalent of Equation 4 specific to the cone-only column results. The qcor

values are plotted against the average push-in force over the respective sample length

(Pc,avg) in Figure 13. A linear trend is apparent and Equation 6 (R2 = 0.926) can be used to

estimate the PIRT cone-only friction correction (kN):

Pc,avg = 0.0003qcor 6

Probing forces for the cone-only test in unstabilised sleech (C-14S-U), also shown in

Figure 12, showed a relatively linear increase with depth to 0.07 kN at 800 mm. In

keeping with the column tests, the cone was guided by a hole had been formed with the

13 mm diameter bar prior to the test.

Discussion and N value calculation

Statistical Analysis

The variability of the stabilised soil used for the columns is now discussed, which has

implications for the correlations derived from the test results (such as Equation 4) and in

turn, the N values. Coefficients of variation (CoV) were used to assess variability within

each given column, while bivariate correlation analysis and linear mixed model

regression analysis were carried out, using the software package SPSS, to statistically

determine the impact of a number of variables on the measured column strengths.

Page 16 of 50



Draft

17

Coefficient of Variation:

CoVs were calculated for the initial sleech moisture content (wi), density of the stabilised

column samples (ρcol), qmeas and E50 within each column and compared with values quoted

in published literature. All CoV values for wi were found to be less than 6.2%, which is

below the 8% lower limit quoted by Phoon and Kulhawy (1999) for in situ clay and silt

samples. CoV values for ρcol were less than 3.5%, within the limit of 10% quoted by the

same authors. CoV values for qmeas were between 5.6% and 17.5%, generally below the

15% to 60% values quoted in the literature (Larsson 2005) for field mixing. E50 values

fell between 19.7% and 40.5% (no CoV ranges were found in the literature with which to

compare). Similar CoV results were observed for the cone-only columns.

Bivariate Correlation:

A bivariate correlation analysis using Pearson Correlation was performed to indicate

covariates (t, mb, wi, T, ρcol, depth d, loss on ignition LOI) that may have an effect on the

dependant variable (qmeas), and the significances of each. Highly significant correlations

were observed between qmeas and curing time, binder content and temperature only; an

interaction between wi and ρcol was also noted. The effects of other variables were not

significant. An interaction between wi and ρcol was also noted, as might be expected.

Linear Mixed Model Regression:

A number of linear mixed model regression analysis were carried out, increasing in

complexity with the addition of variables. qmeas was the dependent variable and the

covariates and fixed effects considered were tcol, mb, wi, T, ρcol, sample depth (d) and LOI,

depending on the analysis being carried out. A summary of the results is provided in

Table 6.

In all models for the PIRT columns, curing time and binder content were found to have a

highly significant effect on the qmeas value; values less than 0.05 are considered

significant and values less than 0.01 are considered highly significant. The natural log of

the qmeas and the curing time was taken to remove flaring noted in the residual versus

predicted graphs. Normal distribution of the residuals was observed. Addition of wi and T

to the models showed both properties not to be significant. In the PIRT models, following

inclusion of the natural log data, all models show similar R2 values of approximately

92%.

Summary:

Page 17 of 50



Draft

18

The statistical analyses performed have clearly illustrated: (i) the quality/repeatability of

the column construction process and (ii) that all of the variables upon which the strength

of the column depends have been captured in the single normalised framework in Figure

11.

N Value Calculation

The value of N can be deduced by rearranging Equation 1. In order to determine the force

P due to the penetrometer wings only, the cone-only force contribution (Pc,avg) must be

deducted from the PIRT force (Pu, avg), where Pu, avg is the average probing force over the

depth range originally occupied by the sample. An exception to this is for samples at the

bottom of the column where a reduction in probing force occurred; for these samples the

reducing Pu is not considered in arriving at the Pu,avg value. Therefore Equation 1 can be

revised, in terms of UCS, as:

( )Aq

PPN

cor

avgcavgu ,,2 −= 7

The PIRT penetrometer plan area (A) of 0.00082 m2 was determined in this case as the

penetrometer width minus the diameter of the cone-only penetrometer multiplied by the

wing thickness, as any contribution of the cone has been removed by subtracting Pc,avg. In

a similar fashion, N values for the tests in unstabilised sleech (Nunstab) were back-

calculated using the cu values measured using the hand shear vane. It should be noted that

Nunstab values are very sensitive to the value of cu used in the calculation, given the low

undrained strengths of the unstabilised sleech and the small probing forces. As such, the

Nunstab values should be considered only as an approximate frame of reference for the

column N values.

N Value Discussion

In an initial plot of N value with depth, low N values (less than 7) were observed in the

top 200 mm of some unsurcharged columns. These values are due to low confining

stresses around the top of the column. This effect is depicted in Figure 14 where N values

for two unsurcharged columns (PI-6 and PI-10) are compared with those of two

equivalent surcharged columns (PI-9S and PI-14S); the higher N values shown in the

latter case are more consistent with those over the full column length.

Page 18 of 50



Draft

19

Average values of N for each column are plotted against corresponding average corrected

undrained strengths cu,cor (= qcor/2) in Figure 15; error bars illustrate the standard

deviation in both N and cu,cor for each test. Based on the foregoing argument, only the

surcharged values are included at shallow depths (for cu,cor > 150kPa) and a small number

of points were omitted which were adversely affected by cracking during the PIRT or

show discrepancies in their UCS test data. The average Nunstab values, also shown in

Figure 15, were found to lie between 15 and 22 (cu = 7-15kPa); however it is strictly only

the Nunstab for those two tests that also incorporated pre-drilling (i.e. PI-13-U and PI-14S-

U) that provide a reference for the N values for the stabilised sleech. The comment in the

N value calculation section regarding probing forces in low strength materials should also

be noted.

The data indicate a mild dependence of N on undrained strength (for cu,cor values up to

400 kPa), in keeping with that postulated by a some authors (Axelsson and Rehnman

1999; Bergman et al. 2013). The higher N values, recorded at lower strengths, imply an

additional contribution to the probing force occurred. In Figure 16, stabilised material can

be seen adhered to the penetrometer wings in test PI-5 and is believed to be evidence of

additional friction occurring between the penetrometer and the stabilised sleech,

particularly for a low strength column. Interestingly, the modelling of Liyanapathirana

and Kelly (2011) also identify increased adhesion between column and tool as a source of

higher N values.

It is recommended that the potential dependence of the value of N on column strength is

investigated further at laboratory and field scale, under realistic stress conditions, as this

may provide a possible explanation for the spread of values reported in Table 1.

Conclusions

The research presented in this paper is the first attempt (known to the authors) to carry out

reduced-scale PIRT, in pre-drilled columns, under laboratory conditions and the results

have helped to identify the factors which influence the relationship between stabilised

column strength and probing resistance. The following points summarise the research:

(i) A range of stabilised column UCS values (up to approximately 800 kPa, typical of

the range encountered in the field) were achieved by varying the binder content

Page 19 of 50



Draft

20

(75 kg/m3, 100 kg/m3 or 150 kg/m3 of OPC) and curing time (1 to 12 days).

Strengths were also influenced by moisture content and curing temperature.

(ii) The quality/repeatability of the columns produced was found to be consistently

high, with CoV values typically below ranges seen in the literature.

(iii) A new framework was successfully developed to unify all of the data in terms of

time, temperature and moisture content, thereby allowing the measured UCS

strengths to be corrected to correspond to those at the time of the PIRT. Statistical

analyses have confirmed that there are no other significant variables affecting the

column strengths besides those captured in this framework.

(iv) Low column confining stresses can significantly influence the N value as splitting

of the column resulted in lower than representative probing forces relative to the

strength of the column.

(v) The data obtained from this study indicate a mild inverse dependence of the N

value on the undrained strength of the stabilised columns; N was found to vary

from approximately 13 at low undrained strengths to approximately 8 at the

highest strengths considered in this study. This may form an explanation for the

range of N values reported from various previous studies.

While it is acknowledged that there is scope for further work to investigate other factors

that may influence N, such as scale, soil type and confinement, an appropriate set of

laboratory procedures and an interpretation framework has now been developed to do so.

While the influence of low confinement is not expected to be an issue in the field, the

potential dependence of the value of N on column strength is worthy of further

consideration at both laboratory-scale and field-scale, under realistic stress conditions, as

it may provide a possible explanation for the spread of values reported in Table 1.

Acknowledgements

The first author acknowledges the financial support provided by the Irish Research

Council for Science, Engineering and Technology under the EMBARK Initiative. The

funding provided by Keller Group to carry out the PIRT series is greatly appreciated, as is

guidance on the statistics provided by Dr. Jerome Sheahan and assistance from the civil

engineering technical staff at NUI Galway and visiting student Mr. Antoine Boutin.

Page 20 of 50



Draft

21

Finally, the authors are extremely grateful to Dr. Alan L. Bell for his technical insights

and support throughout this study.

References

Åhnberg, H. 2006. Strength of stabilised soils - A laboratory study on clays and organic

soils stabilised with different types of binder. PhD Thesis. Lunds Tekniska

Högskola, Lund University.

Åhnberg, H. 2007. On yield stresses and the influence of curing stresses on stress paths

and strength measured in triaxial testing of stabilized soils. Canadian Geotechnical

Journal 44(1), pp. 54–66.

Åhnberg, H. and Holm, G. 1986. Kalkpelarmetoden. Resultat av 10 års forskning och

praktisk användning samt framtida utveckling. Report No 31. (The lime column

method. Results from 10 years of research and practical use and future development;

in Swedish). Swedish Geotechical Institute.

Åhnberg, H. and Holm, G. 2009. Influence of laboratory procedures on properties of

stabilised soil specimens. In: Kitazume, M. and Terashi, M. eds. Deep Mixing 2009

Okinawa Symposium, International Symposium on Deep Mixing & Admixture

Stabilization. Okinawa, Japan.

Åhnberg, H., Johansson, S.-E., Pihl, H. and Carlsson, T. 2003. Stabilising effects of

different binders in some Swedish soils. Proceedings of the ICE - Ground

Improvement 7(1), p. 9–23(14).

Åhnberg, H., Johansson, S.-E., Retelius, A., Ljunkrantz, C., Holmqvist, L. and Holm, G.

1995. Cement och Kalk för Djupstabilisering av Jord - En Kemisk Fysikalisk Studie

av Stabiliseringseffekter (Cement and Lime for Deep Soil Stabilisation - a Chemical

Physical Investigation of Soil Improvement Effects; in Swedish), Report 48.

Linköping, Sweden.

Al-Tabbaa, A. and Evans, C.W. 1999. Laboratory-scale soil mixing of a contaminated

site. Proceedings of the ICE - Ground Improvement 3(3), pp. 119–134.

Axelsson, K., Johansson, S.-E. and Andersson, R. 2002. Stabilization of Organic Soils by

Cement and Puzzolanic Reactions - Feasibility Study; Report 3. Swedish Deep

Stabilization Research Centre c/o Swedish Geotechnical Institute, Linkoping.

Axelsson, M. 2001. Djupstabilisering med kalkcementpelare – metoder för

produktionsmässig kvalitetskontroll i fält (Deep stabilization with lime cement

columns - methods for production quality control in the field; in Swedish). Royal

Institute of Technology.

Axelsson, M. and Larsson, S. 2003. Column penetration tests for lime-cement columns in

deep soil mixing - experiences in Sweden. In: Grouting and Ground Treatment 2003

- Geotechnical Grouting and Deep Mixing. New Orleans, LO, USA, pp. 681–694.

Page 21 of 50



Draft

22

Axelsson, M. and Rehnman, S.E. 1999. Field methods for quality control at the dry jet

mixing method. In: Bredenberg, H., Holm, G., and Broms, B. B. eds. Dry Mix

Methods for Deep Soil Stabilization. Stockholm, Sweden, pp. 303–310.

Babaski, R., Maekawa, A., Terashi, M., Kawamura, M., Suzuki, T. and Fukazawa, E.

1996. JGS TC Report: Factors influencing the strength of improved soil. In:

Yonekura, R., Terashi, M., and Shibazaki, M. eds. Grouting and Deep Mixing:

Proceedings of IS-TOKYO’96, the Second International Conference on Ground

Improvement Geosystems. Balkema, Tokyo, pp. 913–919.

Babaski, R. and Suzuki, K. 1996. Open cut excavation of soft ground using the DCM

method. In: Yonekura, R., Terashi, M., and Shibazaki, M. eds. Grouting and Deep

Mixing: Proceedings of IS-TOKYO’96, the Second International Conference on

Ground Improvement Geosystems. Balkema, Tokyo, pp. 467–473.

Bell, A. 1977. Laboratory Studies of the Belfast Estuarine Deposits. Queen’s University

of Belfast.

Bergman, N., Al-Naqshabandy, M.S. and Larsson, S. 2013. Variability of strength and

deformation properties in lime–cement columns evaluated from CPT and KPS

measurements. Georisk: Assessment and Management of Risk for Engineered

Systems and Geohazards 7(1), pp. 21–36.

Bergman, N. and Larsson, S. 2014. Comparing column penetration and total-sounding

data for lime-cement columns. Proceedings of the ICE - Ground Improvement

167(4), pp. 249–259.

Boman, P. 1979. Kontroll av kalkpelare; Del II resultat av en undersökning i

Sollentunaholm (Quality control of lime columns; Part 2 Results from an

investigation at Sollentunaholm; in Swedish). Stockholm, Sweden.

Bredenberg, H. 1999. Keynote lecture: Equipment for deep soil mixing with the dry jet

mix method. In: Bredenberg, H., Holm, G., and Broms, B. B. eds. Dry Mix Methods

for Deep Soil Stabilization. Stockholm, Sweden, pp. 323–331.

Brown, M. and Hyde, A.F.L. 2008. Rate effects from pile shaft resistance measurements.

Canadian Geotechnical Journal (45), pp. 425–431.

Robinson, S. and Brown, M. 2013. Rate effects at varying strain levels in fine grained

soils. In: Proceedings of the 18th International Conference on Soil Mechanics and

Geotechnical Engineering. ICSMGE, Paris, pp. 263–266.

BSI 2008. BS EN 1992-1:2004 Eurocode 2- Design of Concrete Structures.

Burke, G., Sehn, A., Hussin, J., Hull, V. and Mann, J. 2007. Dry Soil Mixing at Jewfish

Creek. In: Soil Improvement. American Society of Civil Engineers, pp. 1–11.

Carlsten, P. and Ekström, J. 1996. Lime and Lime Cement Columns: Guide for Design,

Construction and Control, Report 4:95. Swedish Geotechnical Society.

Page 22 of 50



Draft

23

Crooks, J.H.A. and Graham, J. 1976. Geotechnical properties of the Belfast estuarine

deposits. Géotechnique 26(2), pp. 293–315.

Dahlström, M. 2012. Dry soil mixing. In: Bell, A. and Kirsch, K. eds. Ground

Improvement, Third Edition. 3rd ed. Taylor & Francis, pp. 435–494.

Edstam, T., Ekström, J., Hallingberg, A. and Nilsson, L. 2004. Provning av

kalkcementpelare i Göta älvdalen (Testing of lime-cement columns in Göta River

Valley; in Swedish). In: NGM 2004 XIV Nordic Geotechnical Meeting; Foundations,

piles and soil improvements. Swedish Geotechnical Society, Linköping, p. D 43-54.

EuroSoilStab 2002. Development of design and construction methods to stabilise soft

organic soils. Design Guide: Soft Soil Stabilisation.

Filz, G.M. 2009. Design of Deep Mixing Support for Embankment and Levees. In:

Kitazume, M. and Terashi, M. eds. Deep Mixing 2009 Okinawa Symposium,

International Symposium on Deep Mixing & Admixture Stabilization. Okinawa,

Japan, p. KL-2.

Halkola, H. 1983. In-situ Investigations of Deep Stabilized Soils. In: Rathmayer, H. and

Saari, K. H. O. eds. Proceedings of the Eight European Conference on Soil

Mechanics and Foundation Engineering - Improvement of Ground. Balkema,

Helsinki, pp. 33–36.

Halkola, H. 1999. Keynote lecture: Quality control for dry mix methods. In: Bredenberg,

H., Holm, G., and Broms, B. B. eds. Dry Mix Methods for Deep Soil Stabilization.

Stockholm, Sweden, pp. 285–294.

Holm, G., Ruin, M. and Håkansson, S. 1999. Column penetration tests and extraction of

lime/cement columns. In: Bredenberg, H., Holm, G., and Broms, B. B. eds. Dry Mix


Horpibulsuk, S., Miura, N. and Nagaraj, T.S. 2003. Assessment of strength development

in cement-admixed high water content clays with Abrams’ law as a basis.

Géotechnique 53(4), p. 439–444(5).

Horpibulsuk, S., Rachan, R. and Suddeepong, A. 2012. State of the art in strength

development of soil–cement columns. Proceedings of the ICE - Ground

Improvement 165(4), pp. 201–215.

Horpibulsuk, S., Rachan, R., Suddeepong, A. and Chinkulkijniwat, A. 2011. Strength

development in cement admixed bangkok clay: Laboratory and field investigations.

Soils and Foundations 51(2), pp. 239–251.

Hussin, J. and Garbin, E. 2012. Recent Advancements in the use of Dry Mass

Stabilization for Structural Support in the United States. In: Johnsen, L., Bruce, D.,

and Byle, M. eds. Grouting and Deep Mixing 2012. American Society of Civil

Engineers, pp. 728–737.

Page 23 of 50



Draft

24

Jacobson, J.R. 2002. Factors Affecting Strength Gain in Lime-Cement Columns and

Development of a Laboratory Testing Procedure. Virginia Polytechnic Institute and

State University.

Jacobson, J.R., Filz, G.M. and Mitchell, J.K. 2005. Factors Affecting Strength of Lime-

Cement Columns Based on a Laboratory Study of Three Organic Soils. In: Rydell,

B., Westberg, G., and Marrarsch, K. R. eds. Deep Mixing - Best Practice and Recent

Advances, Deep Mixing ’05. Swedish Deep Stabilization Research Centre,

Stockholm, Sweden, pp. 87–94.

Janz, M. and Johansson, S.-E. 2002. The Function of Different Binding Agents in Deep

Stabilisation: Report 9. Swedish Deep Stabilization Research Centre c/o Swedish

Geotechnical Institute, Linkoping.

Kallstenius, T. 1961. Development of two modern continuous sounding methods. In:

Proceedings of the 5th International Conference on Soil Mechanics and Foundation

Engineering, (ICSMFE). Paris, pp. 475–480.

Kitazume, M. 2005. State of Practice Report – Field and laboratory investigations,

properties of binders and stabilized soil. In: Rydell, B., Westberg, G., and Marrarsch,

K. R. eds. Deep Mixing - Best Practice and Recent Advances, Deep Mixing ’05.

Swedish Deep Stabilization Research Centre, Stockholm, Sweden, pp. 660–684.

Kitazume, M. and Terashi, M. 2013. The Deep Mixing Method. CRC Press.

Kosche, M. 2004. A laboratory model study on the transition zone and the boundary layer

around lime-cement columns in kaolin clay. Royal Institute of Technology (KTH),

Stockholm.

Larsson, R. 2006. Djupstabilisering med bindemedelsstabiliserade pelare och

masstabilisering – En vägledning Rapport 17 (Deep stabilization of binder stabilized

columns and mass stabilization - a guide, Report 17; in Swedish). Linköping.

Larsson, S. 1999. Shear box apparatus for modelling chemical stabilised soil -

Introductory tests. In: Bredenberg, H., Holm, G., and Broms, B. B. eds. Dry Mix


Larsson, S. 2003. Mixing Processes for Ground Improvement by Deep Mixing. Royal

Institute of Technology, Stockholm.

Larsson, S. 2005. State of Practice Report – Execution, monitoring and quality control.

In: Rydell, B., Westberg, G., and Marrarsch, K. R. eds. Deep Mixing - Best Practice

and Recent Advances, Deep Mixing ’05. Swedish Deep Stabilization Research

Centre, Stockholm, Sweden, pp. 732–786.

Liyanapathirana, D.S. and Kelly, R.B. 2011. Interpretation of the lime column penetration

test. Computers and Geotechnics 38(1), pp. 69–79.

Lunne, T., Robertson, P.K. and Powell, J.J. 1997. Cone Penetration Testing in

Page 24 of 50



Draft

25

Geotechnical Practice. Blackie Academic/Routledge Publishing, New York.

Marzano, I.P., Al-Tabbaa, A. and Grisolia, M. 2009. Influence of Sample Preparation on

the Strength of Cement-Stabilized Clays. In: Kitazume, M. and Terashi, M. eds.

Deep Mixing 2009 Okinawa Symposium, International Symposium on Deep Mixing

& Admixture Stabilization. Okinawa, Japan.

McCabe, B.A. 2002. Experimental Investigation of Driven Pile Group Behaviour in

Belfast Soft Clay. Trinity College Dublin.

McCabe, B.A. and Lehane, B.M. 2006. Behaviour of Axially Loaded Pile Groups Driven

in Clayey Silt. Journal of Geotechnical and Geoenvironmental Engineering, ASCE

132(3), pp. 401–410.

Nagaraj, T.S., Yaligar, P., Miura, N. and Yamadera, A. 1996. Predicting strength

development by cement admixture based on water content. In: Yonekura, R.,

Terashi, M., and Shibazaki, M. eds. Grouting and Deep Mixing: Proceedings of IS-

TOKYO’96, the Second International Conference on Ground Improvement

Geosystems. Balkema, Tokyo, pp. 431–436.

O’Loughlin, C., Randolph, M. and Richardson, M. 2004. Experimental and Theoretical

Studies of Deep Penetrating Anchors. 36th Annual Offshore Technology Conference,

p. Paper No. OTC 16841.

Phoon, K.-K. and Kulhawy, F.H. 1999. Characterization of geotechnical variability.

Canadian Geotechnical Journal 36(4), pp. 612–624.

Porbaha, A. 2000. State of the art in deep mixing technology. Part IV: design

considerations. Proceedings of the ICE - Ground Improvement 4(3), pp. 111–125.

Porbaha, A., Shibuya, S. and Kishida, T. 2000. State of the art in deep mixing technology.

Part III: geomaterial characterisation. Proceedings of the ICE - Ground Improvement

4(3), pp. 91–110.

Rogbeck, Y., Hansson, K. and Säfström, L. 2000. Deep soil stabilisation in organic soil

and clay - field experiments. In: Rathmayer, H. ed. Grouting soil improvement,

geosystems including reinforcement: proceedings of the 4th International

Conference on Ground Improvement Geosystems, Helsinki, Finland. Building

Information Ltd, Helsinki, pp. 117–126.

Stewart, D. and Randolph, M. 1994. T-Bar Penetration Testing in Soft Clay. Journal of

Geotechnical Engineering 120(12), pp. 2230–2235.

Timoney, M.J. 2015. Strength Verification Methods for Stabilised Soil-Cement Columns:

A Laboratory Investigation of PORT and PIRT. PhD Thesis. NUI Galway.

Timoney, M.J., McCabe, B.A. and Bell, A. 2012a. Experiences of Dry Soil Mixing in

Highly Organic Soils. Ground Improvement 165(1), pp. 3–14.

Page 25 of 50



Draft

26

Timoney, M.J., Quigley, P. and McCabe, B.A. 2012b. Some laboratory soil mixing trials

of Irish peats. In: Denies, N. and Huybrechts, N. eds. ISSMGE - TC 211

International Symposium & Short Courses; Recent Research, Advances & Execution

Aspects of Ground Improvement Works. Brussels, pp. 511–520.

Topolnicki, M. 2012. In-situ soil mixing. In: Bell, A. and Kirsch, K. eds. Ground

Improvement, Third Edition. 3rd ed. Taylor & Francis, pp. 329–434.

Trafikverket 2011. TK Geo 11; Trafikverkets tekniska krav för geokonstruktioner

(Trafikverket technical requirements for geotechnical structures; available in

Swedish and English).

Tremblay, H., Duchesne, J., Locat, J. and Leroueil, S. 2002. Influence of the nature of

organic compounds on fine soil stabilization with cement. Canadian Geotechnical

Journal 39, pp. 535–546.

Wiggers, A.G. and Perzon, J. 2005. The Lekkerkerk trial: Mixed-in-place dike

improvement in Netherlands. In: Rydell, B., Westberg, G., and Marrarsch, K. R. eds.

Deep Mixing - Best Practice and Recent Advances, Deep Mixing ’05. Swedish Deep

Stabilization Research Centre, Stockholm, Sweden, pp. 179–183.

1. LIST OF SYMBOLS

A Penetrometer plan area Avg Average cu,cor Stabilised column undrained shear strength corrected to the time of the column

test (= qcor/2) cu Undrained shear strength (= q/2) CCP Conventional column penetration test, alternative name for PIRT C-X Cone-only penetrometer experiment on cone-only column no. X d Depth from top of column dcol Depth from the top of column to centre of the sample DDSM Deep dry soil mixing dia. Diameter DSM Dry soil mixing E50 Secant stiffness at 50% of the failure stress h Height from basin base KPS Kalk-Pelar-Sondering (Lime-Column Probing), alternative name for PIRT LOI Loss on ignition

mb Binder content (kg/m3) N Bearing capacity factor relating undrained shear strength to penetrometer probing

resistance relationship NUI Galway National University of Ireland, Galway Pu PIRT Penetrometer push-in probing force (uncorrected) Pu,avg Average PIRT penetrometer push-in probing force over the depths occupied by a

column sample Pc Cone-only penetrometer push-in force Pc,avg Average cone-only penetrometer push-in probing force over the depths occupied

by a column sample PIRT Push-in resistance test PI-X PIRT experiment on PIRT column no. X

Page 26 of 50



Draft

27

PI-X-Y-Z PIRT experiment on PIRT column no. X, where Y is the curing time (days) and Z is the binder content (kg/m3)

q Unconfined compression strength qcor Stabilised column UCS corrected to the time of the column test qmeas Stabilised column UCS from UCS testing qmeasη Product of the stabilised column UCS (from UCS testing) and its WTBR Sleech Soft dark grey organic clayey silt soil local to Belfast, Northern Ireland St. Dev Standard deviation T Average ambient laboratory temperature during the curing period tadj Temperature-Adjusted Time tcol Time from mixing to UCS testing for column samples t Time from mixing to the PIRT UCS Unconfined compression strength wi Initial soil moisture content WTBR Water-to-binder ratio δ Push-in rate η Water to binder ratio ρcol Stabilised column sample density ρsoil Soil bulk density ∆(qmeasη)avg Change in qmeasη of a sample occurring in the time between PIRT of the column

and the UCS testing of the column’s samples References from Table 1. These are required in the document for the reference manager to compile the bibliography but are not needed for the final paper. (Halkola 1983) (Axelsson and Rehnman 1999) (Halkola 1999) (Rogbeck et al. 2000) (Axelsson 2001)

(Axelsson 2001) (Axelsson and Larsson 2003) (2001) (Edstam et al. 2004) (Wiggers and Perzon 2005)

(Burke et al. 2007) List of Figure captions

Figure 1: Typical PIRT penetrometer: a) guideline dimensions (Trafikverket 2011) & b) a 400 mm PIRT penetrometer

Page 27 of 50



Draft

28

Figure 1: a) 150 mm scaled PIRT penetrometer and cone-only penetrometer, b) wing profile

Figure 2: PIRT column construction process (all dimensions in mm)

Figure 3: PIRT column during construction with 13 mm dia. hole-form bar in place

Figure 4: PIRT Experimental Setup; a) test frame with CPT rig & b) load cell and draw wire gauge instrumentation

Figure 5: 104 mm dia. cone-only test column showing the pre-formed 13 mm dia. hole

Figure 6: Recorded PIRT probing force with height from basin base: a) 1-4 day columns, b) 6 day columns and c) 12 day columns

Figure 7: PIRT columns following testing; a) Extracted PI-13 column with few cracks & b) PI-10 during extraction with highlighted diagonal cracks in the column mid-section

Figure 8: Uncorrected PIRT column UCS with depth

Figure 10: Column UCS with column stiffness

Figure 11: Average column qmeasη (with St. Dev error bars) against temperature-adjusted time

Figure 12: Cone-only probing force with depth for pre-drilled stabilised columns and unstabilised sleech

Figure 13: Corrected column UCS with cone-only probing force

Figure 14: PIRT column N value at shallow depth: unsurcharged and surcharged

Figure 15: PIRT column N value with corrected shear strength

Figure 16: PIRT penetrometer at final location in PI-5 with evidence of stabilised sleech adhered to the penetrometer wings

Page 28 of 50



Draft

1

Figure 1: Typical PIRT penetrometer: a) guideline dimensions (Trafikverket 2011) & b) a 400

mm PIRT penetrometer

(b) (a)

Page 29 of 50



Draft

1

Figure 2: a) 150 mm scaled PIRT penetrometer and cone-only penetrometer, b) wing profile

(a) (b)

6 mm

150 mm PIRT

Penetrometer

14 mm Cone-only

Penetrometer

Page 30 of 50



Draft

1

Figure 3: PIRT column construction process (all dimensions in mm)

Page 31 of 50



Draft

1

Figure 4: PIRT column during construction with 13 mm dia. hole-form bar in place

Page 32 of 50



Draft

1

Figure 5: PIRT Experimental Setup; a) test frame with CPT rig & b) load cell and draw wire gauge

instrumentation

(a) (b)

Load cell

Draw-wire

gauge

CPT rig

Page 33 of 50



Draft

1

Figure 6: 104 mm dia. cone-only test column showing the pre-formed 13 mm dia. hole

Page 34 of 50



Draft

1

Figure 7: Recorded PIRT probing force with height from basin base: a) 1-4 day columns, b) 6 day columns and c) 12 day columns

0

100

200

300

400

500

600

700

800

900

1,000

1,100

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5

He

igh

t fr

om

ba

sin

ba

se, h(m

m)

Uncorrected Probing Force, Pu (kN)

PI-4 (2d-150c) PI-5 (1d-100c)

PI-8 (4d-150c) PI-13 (1d-75c)

Column Base0

100

200

300

400

500

600

700

800

900

1,000

1,100

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5


PI-6 (6d-150c) PI-7 (6d-100c)

PI-9S (6d-150c) Column Base

0

100

200

300

400

500

600

700

800

900

1,000

1,100

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5


PI-10 (12-150c) PI-11 (12d-100c)

PI-12 (12-150*c) PI-14S (12d-150c)

Column Base

(a) (b) (c)

Page 35 of 50



Draft

1

Figure 8: PIRT columns following testing; a) Extracted PI-13 column with few cracks & b) PI-10

during extraction with highlighted diagonal cracks in the column mid-section

(b) (a)

Page 36 of 50



Draft

1

Figure 9: Uncorrected PIRT column UCS with depth

0

100

200

300

400

500

600

700

800

900

0 100 200 300 400 500 600 700 800 900

Co

lum

n D

ep

th, d

(mm

)

PIRT Column UCS, qmeas (kPa)

PI-4 (2d-150c)

PI-5 (1d-100c)

PI-6 (6d-150c)

PI-7 (6d-100c)

PI-8 (4d-150c)

PI-9S (6d-150c)

PI-10 (12d-150c)

PI-11 (12-100c)

PI-12 (12d-150*c)

PI-13 (1d-75c)

PI-14S (12d-150c)

Page 37 of 50



Draft

1

Figure 10: Column UCS with column stiffness

0

20

40

60

80

100

120

140

0 100 200 300 400 500 600 700 800 900

Co

lum

n S

tiff

ne

ss, E50

(MPa

)

Column UCS, qmeas (kPa)

PI-4 (2d-150c) PI-5 (1d-100c)

PI-6 (6d-150c) PI-7 (6d-100c)

PI-8 (4d-150c) PI-9S (6d-150c)

PI-10 (12d-150c) PI-11 (12-100c)

PI-12 (12d-150*c) PI-13 (1d-75c)

PI-14S (12d-150c)

Page 38 of 50



Draft

1

Figure 11: Average column qmeasη (with St. Dev error bars) against temperature-adjusted time

PI-4PI-5

PI-6

PI-7

PI-8PI-9S

PI-10PI-11

PI-12*

PI-13

PI-14S

y = 531.78ln(x) + 1010

R² = 0.7872

0

500

1,000

1,500

2,000

2,500

3,000

0 2 4 6 8 10 12 14

Av

era

ge

qmeasη

(kP

a)

Temperature Corrected Time, tadj (days)

Page 39 of 50



Draft

1

Figure 12: Cone-only probing force with depth for pre-drilled stabilised columns and unstabilised

sleech

0

100

200

300

400

500

600

700

800

900

0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35

Co

lum

n D

ep

th, d

(mm

)

Cone-Only Probing Force, Pc

(kN)

C-1 (1d-100c) C-2 (2d-150c)

C-3 (6d-100c) C-4 (6d-150c)

C-5 (12d-150c) C-6 (1.6d-150c)

C-7 (12.7d-100c) C-14S-U

Page 40 of 50



Draft

1

Figure 13: Corrected column UCS with cone-only probing force

y = 0.0003x

R² = 0.92560.00

0.05

0.10

0.15

0.20

0.25

0.30

0 100 200 300 400 500 600 700 800 900

Av

era

ge

Co

ne

-On

ly P

ush

-In

Fo

rce

. Pc,avg

(kN

)

Corrected Column UCS, qcor (kPa)

C-1 (1d-100c)

C-2 (2d-150c)

C-3 *(6d-100c)

C-4 (6d-150c)

C-5 (12d-150c)

C-6 (1.6d-150c)

C-7 (12.7d-100c)

Page 41 of 50



Draft

1

Figure 14: PIRT column N value at shallow depth: unsurcharged and surcharged

0

20

40

60

80

100

120

140

160

180

200

2 4 6 8 10 12

Co

lum

n D

ep

th, d

(mm

)

PIRT Column N Value

PI-6 (6d-150c)

PI-9S (6d-150c)

PI-10 (12d-150c)

PI-14S (12d-150c)

Page 42 of 50



Draft

1

Figure 15: PIRT column N value with corrected shear strength

Sur

Sur

0

50

100

150

200

250

300

350

400

4 6 8 10 12 14 16 18 20 22 24

Co

rre

cte

d S

he

ar

Str

en

gth

, ccor

(kPa

)

N Value

Page 43 of 50



Draft

1

Figure 16: PIRT penetrometer at final location in PI-5 with evidence of stabilised sleech adhered to

the penetrometer wings

Page 44 of 50



Draft

1

Table 1: Published field observed N values for PIRT penetrometers

N

Value:

Soil

Type:

Binder Achieved

Strengths:

Test Details: Reference:

10 Clay Lime cu < 160 kPa 3 month curing period.

400 mm probe (Boman 1979)

11 * Lime

cu ≈ 255 kPa,

(190-320

kPa)

Menard pressuremeter

tests on columns

compared with 400mm

PIRT

(Holm et al. 1981)

12.5-

16.7 * * cu < 255 kPa

375mm PIRT,

comparison with column

vane (85 mm high by 132

mm dia.). 0-320 day

curing periods

(Halkola 1983)

10

(8-11) * * *

Experience based value

Stiffness related

(Axelsson &

Rehnman 1999)

10

10-15

* * *

Sweden

Finland, defined by site

specific column vane

tests

(Halkola 1999)

10 Clay &

gyttja

Cement-Lime

Cement-

GGBS

cu <185 kPa

Field column tests

400mm PIRT

28-134 day curing

periods

(Rogbeck et al.

2000)

10 Clay * Not provided PIRT in unstabilised clay (Axelsson 2001) (in

Swedish)

10-15 Very soft

clay Lime-cement cu < 600 kPa PIRT

(Axelsson 2001) (in

Swedish)

10

15

* * *

600/800mm column,

350-700mm PIRT

Columns under active

loading conditions

Conservative N, for use

with columns in direct

shear with low

confinement

(Axelsson & Larsson

2003). See Axelsson

(2001) for further

details on the tests.

10 Clay Lime-cement cu < 250 kPa. Pre-drilled columns

3-4 day curing period

(Edstam et al. 2004)

(in Swedish)

20 Clayey-

peat *

cu ≈ 100-300

kPa

N = 20 derived from field

PIRT in pre-drilled mass

stabilised clayey peat

(Wiggers & Perzon

2005)

10 Bentonite Cement *

Field calibration tests on

cement-bentonite mixture

with shear vane and

laboratory tests

(Burke et al. 2007)

* No specific data provided in the source literature

Page 45 of 50



Draft

1

Table 2: Published and recorded results for and for Kinnegar sleech

Parameter Crooks &

Graham (1976)

Bell (1977) McCabe &

Lehane (2006)1

Classification

Testing1,2

Depth 3.0-8.0 3.0-6.0 1.7-6.5 3.0-4.5 m

Moisture Content 60-80 54-80 60±10 48-71 %

Organic Content 3.0-5.0 2.5-5.0 11±1 2.6-5.6 %

Bulk Density 1,500-1,750 1,500-1,780 - 1,620 kg/m3

Peak Vane Strength 3-5 - 22±2 - kPa

Fall Cone Strength - 10-17 - 15 kPa

pH 7.5-8.2 7.5-8.5 - 8.0-8.5 - 1This site comprised approximately 1.0 m of fill not present at the other sites. 2Results from classification tests carried out at time of sampling.

Page 46 of 50



Draft

1

Table 3: Constructed PIRT column details

PIRT No. PI-4 PI-5 PI-6 PI-7 PI-8 PI-9S PI-10 PI-11 PI-12 PI-13 PI-14S

Test Time, t: 1.9 0.9 5.9 5.9 3.9 5.9 11.9 11.9 11.9 0.9 11.9 days

Binder Content: 150 100 150 100 150 150 150 100 100/ 150* 75 150 kg/m3

Column Height, h: 955 1,025 1,020 1,020 1,000 1,000 1,010 1,010 1,000 1,005 980 mm

Sleech Moisture

Content, wi: 42-48 45-47 46 44-47 48-52 48-54 53-56 48-53 47-53 48-49 48-51 %

Loss on Ignition: 2.6-2.4 2.4-4.6 2.2-3.2 2.8-2.9 2.6-3.6 2.6-3.4 2.4-3.8 3.1-4.1 2.5-4.4 2.6 -3.9 2.9-4.6 %

No. of Samples: 14 16 13 7 13 16 14 10 15 16 22 -

Average Column

UCS, qmeas: 347.5 221.6 474.4 397.4 459.4 463.2 712.5 479.4 539.8 180.5 632.7 kPa

qmeas Standard

Deviation: 60.9 24.8 59 55.5 37.7 78.7 51.8 54.3 54.4 22.5 35.4 kPa

Corrected Average

Column UCS, qcor: 327.8 199.1 469.7 393.8 425.2 458.5 708.6 476.9 535.8 155.7 625.4 kPa

Average Ambient

Temperature, T: 19.2 18.5 18.4 18.7 17.9 18.3 17.8 18.7 17.6 17.5 16.9 °C

Average Push-In

Rate: δ1 & δ2

18.3 31.2 24.9 23.5 27.4 22 34.5 19 19.8 22.1 18.2 mm/sec

21.1 21.3 26.1 29.9 25.3 30.7 26.3 28.9 19.9 20.5 21.6 mm/sec

S designated columns which cured under a surcharge loading

* Column created with 3 no. 150kg/m3 stabilised mixes and 1 no. 100kg/m

3 stabilised mix

Page 47 of 50



Draft

1

Table 4: PIRT and cone-only tests in unstabilised sleech

Test Ref: PI-13-U PI-14S-U PI-T-U1 PI-T-U2 C-14S-U

Penetrometer: PIRT

No. 2

PIRT

No. 2

PIRT

No. 1

PIRT

No. 2

Cone-

Only

Probing Distance: 820 823 832 858 822 mm

Sleech Moisture

Content, wi: 43-51 43-49 43-53 43-53 43-49 %

Average Shear

Strength cu (Range):

8

(6-12)

9

(7-14)

11.5

(8-16)

10.8

(8.5-13)

9

(7-14) kPa

No. Strength Tests: 12 18 12 12 18 -

cu Standard

Deviation: 1.5 1.7 2.5 1.5 1.7 kPa

Pre-drilled: Yes Yes No No Yes -

Average Push-In

Rate δ1 & δ2:

22.1 20.4 15.1 15.1 16.9 mm/sec

20.5 17.9 14.2 18.7 15.8 mm/sec

Page 48 of 50



Draft

1

Table 5: Constructed cone-only & sounding bar column details

PIRT Cone Test No. C-1 C-2 C-3 C-4 C-5 C-6 C-7

Test Time, t: 1 2 6 6 12 1.6 12.7 days

Binder Content, mb: 100 150 100 150 150 150 100 kg/m3

Column Length: 750 745 730 747 730 740 740 mm

Sleech Moisture Content,

wi: 45.8 46.0 49.1 49.2 47.0 47.2 45.5 %

Loss on Ignition: 4.1 4.1 4.4 4.8 4.3 4.1 3.1 %

No. Column Samples: 6 7 7 7 6 7 7 -

Average Column UCS,

qmeas: 226.6 389.8 390.9 506.9 788.2 301.0 562.3 kPa

qmeas Standard Deviation: 22.9 49.1 31.3 41.0 29.0 23.7 40.4 kPa

Corrected Average Column

UCS, qcor: 208.0 373.7 387.7 501.3 785.5 285.9 561.4 kPa

Average Ambient

Temperature, T: 16.7 15.7 16.2 16.6 17.1 16.0 14.5 °C

Average Push-In Rate: δ1

& δ2

17.8 19.6 12.9 13.1 16.5 14.2 15.7 mm/sec

19.3 22.6 15.0 15.0 19.4 19.1 18.1 mm/sec

Page 49 of 50



Draft

1

Table 6: Mixed Model Linear Regression output for PIRT column data

Analysis: 1 2 3 4 5 6 7 8 9

R2 0.868 0.919 0.919 0.92 0.92 0.921 0.921 0.923 0.921

Output q Ln(q) Ln(q) Ln(q) Ln(q) Ln(q) Ln(q) Ln(q) Ln(q)

Covariates: Significances:

Intercept 0.076 0 0 0 0 0.26 0 0.289 0.216

Time 0 - - - - - - - -

Ln(Time) - 0 0 0 0 0 0 0 0

Binder 0 0 0 0 0 0 0 0 0

wi - - 0.914 - 0.869 0.066 0.931 0.08 0.044

Temperature - - - 0.308 0.306 0.64 0.278 0.534 -

Organics - - - - - - 0.81 0.664 -

Density - - - - - 0.069 0.819 0.084 0.046

Depth - - - - - - 0.326 0.464 -

wi - Density - - - - - 0.066 - 0.081 0.043

Page 50 of 50



Keyword: PIRT, KPS, stabilisation, clay/silt, strength Draft · verification is of paramount...

Documents

Transcript of Keyword: PIRT, KPS, stabilisation, clay/silt, strength Draft · verification is of paramount...