Pinheiro, M.; Proskin, S. & Li, B. (2017) Laboratory plate load testing of non-segregating tailings. In: Proc.

21st Intl.
Conf. Tailings & Mine Waste, Eds. G.W. Wilson, D.C. Sego & N.A. Beier, 5-8 November 2017, Banff, AB, Canada.

Laboratory Plate Load Testing of Non-Segregating Tailings

M. Pinheiro & S. Proskin

Thurber Engineering Ltd., Calgary, Alberta, Canada
B. Li
Canadian Natural Resources Ltd., Calgary, Alberta, Canada

ABSTRACT: Non-segregating tailings (NST) are tailings produced by mixing mature fine tail-
ings or thickened tailings with cyclone underflow coarse tails, and a coagulant to accelerate sol-
ids settling and densification. NST is prepared at a target sand-to-fines ratio of about 4 or more,
thereby producing a predominantly frictional material. A research project was setup to measure
the bearing capacity of NST, in order to assess its readiness for capping or landform construc-
tion activities. Five plate load tests were carried out on NST dewatered to different solids con-
tent. Index testing, and strength testing (using pocket penetrometer, handheld field vane and a
Rimik® cone device) were also conducted to further characterize the tailings. The results re-
vealed that the ultimate bearing capacity of NST increases with plate settlement, as expected for
a frictional material whose shear strength depends on stress and confinement levels. For a small
settlement (20 mm), the ultimate bearing capacity varied from 3–16 kPa. For larger settlements
(60 mm), ultimate bearing capacities reached nearly 30 kPa for the NST at 80.7% solids content.

1 INTRODUCTION

In 2009, the Alberta Energy Regulator (AER), then known as Energy Resources Conservation
Board (ERCB), issued “Directive 074: Tailings Performance Criteria & Requirements for Oil
Sands Mining Schemes.” Under this directive, oil sands operators were required to come up
with strategies to meet specific targets, such as having a trafficable tailings deposit with a mini-
mum shear strength of 10 kPa within five years after tailings has been deposited in a dedicated
disposal area. Directive 074 proved itself to be difficult to comply with, and has since been re-
placed in 2016 by “Directive 085: Fluid Tailings Management for Oil Sands Mining Projects”
(AER, 2009, 2016). Directive 085, which is currently being completed and enhanced, does not
set out a particular strength target for trafficability. Instead, it sets out a tailings management
framework in which performance criteria are first developed by the operator for each tailings
deposit. The proposed performance criteria are then assessed by the regulator for approval. Each
criterion must identify indicators (e.g. material properties, residual settlement, trajectory to traf-
ficability) and measures (e.g. solids content, sand-to-fines ratio, cone penetration testing) that
will be used to assess and track progress towards the various stages of reclamation – trafficabil-
ity being one of them.
Non-segregating tailings (NST) are tailings that have been significantly dewatered by using
thickeners and cyclones. They are produced by mixing either mature fine tailings or thickened
tailings with sand derived from cyclone underflow coarse tails. Coagulants (e.g. CO2, CaO) are
added to accelerate solids settling and densification. NST is typically prepared at a target sand-
to-fines (SFR) ratio of about 4 or more, thereby producing a predominantly frictional material.
Plate load testing and ultimate bearing capacity are here investigated as an alternative gauge
to assess the shear strength of NST. The underlining concept would be to conduct plate load
tests on NST samples prepared at different solids contents by weight (SBW), and estimate the
ultimate bearing capacity of each sample. Solids content would then be used as a measure to as-
sess the deposit’s trafficability and readiness for capping or landform construction activities and
monitor progress towards reclamation.
A research project was setup, and the scope of work entailed the following tasks:
 Develop procedures for the index testing, strength characterization, and plate load testing.
 Design and fabricate a plate load assembly for testing the NST under laboratory conditions.
 Conduct plate load tests.
 Conduct strength characterization tests on the NST samples.
a. Cone penetration testing using a hand-held electronic cone penetrometer.
b. Shear vane testing using a hand-held shear vane device.
c. Pocket penetration testing using a hand-held pocket penetrometer.
 Collect samples for geotechnical index testing, including Atterberg limits, solids content, bi-
tumen content, and particle size distribution.
 Analyze results from index, strength characterization and plate load testing programs.
 Determine the ultimate bearing capacity of the NST material by means of analytical methods.

2 METHODOLOGIES

The following sections briefly describe the methodologies employed to prepare the NST and to
characterize it in a laboratory setting. NST production and testing, including plate load testing
were all carried out at Canadian Natural’s “Bitumen Production / Applied Process Innovation
Centre,” located at Horizon Oil Sands mine, just north of Fort McMurray, Alberta, Canada.

2.1 NST Material Preparation

The NST material was produced at a SBW of 64.4% and at a SFR of about 5. The tailings mate-
rial was then discharged into four 1.0 m × 1.0 m × 0.7 m (length × width × height) aluminum
pans: Pans #12, #11, #09 and #08 (Figure 1). The initial plan was to prepare batches of NST at
solids contents ranging from around 65% to 85%, and discharge each batch into separate pans.
However, this plan was found to be impractical – it would take more time and resources to pre-
pare several batches at distinct solids content than to prepare a single homogeneous material.
Moreover, NST produced at SBWs above 80% would not be representative of field conditions.
The tailings in each pan was then subject to different dewatering conditions so that by the end
of two weeks, they were expected to have reached target solids contents of 65%, 70%, 75% and
85%. The dewatering strategies involved:
 Laying a 0.10 m thick layer of dry and loose beach sand at the bottom of the pan to provide
bottom drainage, at least to the maximum storage capacity of this sand layer.
 Installing a French drain at the sand layer to allow the under-seepage water to drain out, and
enhance the bottom drainage.
 Decanting the “free” water accumulated on the tailings surface.
 Mounting lights, and subjecting the surface of the tailings material to varying light schemes
and intensity (i.e. number of light bulbs on), to enhance evaporation.
 Setting up fans to blow air moisture away from the tailings surface, and enhance evaporation.

Table 1 summarizes the various dewatering conditions applied to the NST in each pan.

Table 1. Pan setup and dewatering conditions.

Pan #12 Pan #11 Pan #09 Pan #08
Average target SBW 65% 70% 75% 80%
Sand beach layer at the bottom of pan Yes Yes Yes Yes
French drain No No Yes Yes
Removal of decanted water No Yes Yes Yes
Lights / Light intensity No No Yes / Medium Yes / High
Fan No No No Yes
Figure 1. (a) Overview of pans and lab setup. (b) Overall plate load test setup showing pan, reaction
frame, hydraulic ram, load cell, LVDT, data acquisition equipment and laptop.

All four pans were initially placed on top of Desna scales to monitor the change in weight (in
kilograms) of the tailings material with time prior to plate load testing. Figure 1a shows the
pans, light stands and Desna scales used in this research project. Figure 1b depicts the plate load
testing setup, which will be described in the next subsection.
NST dewatering occurred over a two-week period. Weights were recorded at irregular inter-
vals, usually during weekdays, as no one worked full time in the lab. The results of the dewater-
ing program are presented in Figure 2, in terms of solids content. The NST material dewatered
rapidly after deposition as indicated by the sharp jump in solids content from 64.4% to almost
75% in Pans #11, #09 and #08. The NST material in Pan #12 also underwent similar rapid de-
watering; however, this is not readily observed in Figure 2 because the decant water was not
pumped out from this pan. Figure 2 also shows target (from Table 1) and actual average SBW
values achieved prior to plate load testing. The solids content of the tailings in Pan #08 was on
target, but the others were not. Anyway, an NST at a solids content below 75% would have been
too soft/weak to be tested, as discussed in Section 2.2.

90 0.29
Pan #12 [Target 65%] plate load test
Pan #11 [Target 70%] on Pan #08
Average Solids Content by Weight (%)

85 0.46
Pan #09 [Target 75%] plate load test
Pan #08 [Target 80%] on Pan #09 80.7, 0.63
80 0.65
Average Void Ratio

80.3, 0.64
76.3, 0.81
75 0.87
plate load test 75.4, 0.85
decant water pumped out on Pan #11
70 from Pans #11, #09, #08 1.11

65 1.40

60 1.73
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
Days of Measurement
Figure 2. Average solids by weight of the NST in each pan over an approximately two-week period.
 The NST in Pans #09 and #08 developed millimetre-to-centimetre thick cracks. Desiccation
was not uniform because of uneven wind and light exposure (some of the lightbulbs fixed to the
stand were burnt out, and air was blown from a fan located at one of the corners of the pan). Un-
saturation in the driest areas may have reached down a couple centimetres. The topmost surface
of the NST was notably glossy, likely due to bitumen migration during deposition and sedimen-
tation. The topmost layer was also visibly richer in fine particles than the underlying one.

2.2 Plate Load Testing

The plate load testing equipment included the following:
 Custom-made load frame.
 Custom-made 203.2 mm (8-inch) diameter steel plate.
 Custom-made adapters and extension rods.
 4-ton capacity hydraulic Porta-Power ram assembly.
 1-ton capacity electronic load cell.
 Linear Variable Differential Transformer (LVDT).
 Data acquisition system and laptop.

The load frame was especially designed for this project to accommodate the dimensions of
the pans, the expected maximum loads, as well as transportation requirements. The load frame,
steel plate, adapters and extension rods were all manufactured in Calgary. The hydraulic ram,
the disassembled frame and companion parts were shipped to Horizon mine, and reassembled on
site, as shown in Figure 1b.
The plate load test consisted of pushing the steel plate onto the tailings material, and record-
ing both plate displacement (in millimetre) and load (in kilogram). Any desiccated/crusted tail-
ings material within the testing area was removed, and the exposed surface evened out with a
spatula before laying down the steel plate.
The data acquisition system was set to automatically collect and store readings of displace-
ment and load every second. These readings were plotted on a laptop in real-time for quick in-
terpretation of the tailings behaviour. The first displacement increments were very small (0.01 –
0.05 mm) because the initial NST response was relatively stiff. As the NST deformed and sof-
tened, the imposed displacement increments were increased in steps up to a maximum value of
about 1 – 3 mm. The recorded total displacement and load values were transferred to an Excel
spreadsheet to calculate the relative total displacement of and applied pressure under the plate.
These were then plotted to help assess the results and determine when to terminate the test.
Plate load tests were completed on Pans #11, #09 and #08. No test was carried out on Pan #12
due to time constraints and because its NST was still very soft. Moreover, the bearing capacity
of the tailings material in Pan #12 might have been below the sensitivity of the load cell. The re-
sults of the plate load testing are discussed in Section 3.3.

2.3 NST Strength Characterization

Strength properties of the NST material were also indirectly assessed by other means based on
pocket penetrometer test, field shear vane test and cone penetration test. The methodology of
each test is briefly described below. Desiccated and cracked material was removed and, the ex-
posed surface levelled before testing.
The pocket penetrometer test consisted of pushing the piston of a Humboldt pocket pene-
trometer device onto the NST, and recording the unconfined compressive strength, Sq (in tons/ft2
or kg/cm2) directly from a scale indicator. An adaptor foot was attached to the piston to increase
the device’s accuracy in the lower range of measurement.
The field shear vane test consisted of pushing a Humboldt vane into the NST material to the
desired depth, slowly turning the handle clockwise at constant speed until failure, and recording
the value, Sv, registered on the graduated scale of the shear vane device. A shaft rod was used to
prevent friction between NST and extension rod. A large 50.8 mm × 101.6 mm four-bladed vane
size was used to measure shear strengths in the range of 0 to 8.125 kPa.
The cone penetration test consisted of pushing a Rimik® CP140II cone vertically into the
NST material at a constant speed until the tip of the cone reached the bottom of the pan. The
cone was pushed at three locations. The Rimik assembly comes with two size cones (areas: 130
mm2 and 323 mm2). The larger size cone was used to improve the device’s accuracy in the lower
range of measurement. The Rimik device automatically tracks insertion depth, speed and load. It
then averages the recorded data over a 25-mm range, and converts the load into a cone re-
sistance index, qc (in kPa).
The limitations of these tests and their applicability to NST-like material is well understood.
Here, they were used as qualitative tools to improve our understanding of the NST material, to
help interpret the plate load test results, and to compare the NSTs prepared at different solids
contents – they provide a relative measure of resistance when comparing these materials.

2.4 NST Index Characterization

Two 4-litre pails (pails #1 and #2) were filled with the NST prepared on site at 64.4% SBW, and
shipped to Calgary for index characterization, including Atterberg limit tests, solids content and
bitumen content determination, and sieve-hydrometer analyses. These tests were conducted ac-
cording to the following methodologies:
 Atterberg limit tests were completed according to ASTM D4318 standards.
 Solids contents were measured using the conventional oven technique, as per ASTM D2216
standards.
 Bitumen content was determined by means of Dean-Stark extraction method.
 Sieve-hydrometer analyses were carried out on non-bitumen extracted samples according to
ASTM E11-09 and D422-63 standards, respectively. Mining fines content, defined as the
mass of fines (< 44 μm) divided by the mass of mineral solids and bitumen, was obtained di-
rectly from the particle size distribution curves.

The solids contents of the NST in Pans #11, #09 and #08 were also measured on site using the
quick oven technique, as per ASTM D4959 standards. NST samples were collected at various
depths by pushing a 30.5 cm (1-foot) long, 2.5 cm (1-inch) diameter cylinder sampler into the
tailings, then digging around the cylinder, and placing a cap at the top of the cylinder to mini-
mize water loss when carrying the sample for quick oven testing.

3 RESULTS

3.1 NST Index Characterization

The results of the index testing completed on the fresh NST tailings (before any dewatering took
place) are summarized below:
 Atterberg limits could not be measured because of the high sand content (about 80%). The
NST is classified as a cohesionless non-plastic material.
 The solids contents of the NST in pails #1 and #2 were 66.3% and 65.6%, respectively. These
values were determined using the conventional oven drying method, and are slightly higher
than 64.4%, measured on site using the quick oven method.
 Bitumen content, determined by Dean-Stark, ranged from 0.12% to 0.15%.
 Particle size distribution curves are shown in Figure 3 for pails #1 and #2. Mining fines con-
tent is about 16%, and SFR is about 5.2.

The solids content of the NST material in Pans #11, #09 and #08 was again determined im-
mediately prior to plate load testing, as described in the previous section. The measured solids
contents agree with the average values estimated based on the weights of the pans. Averaged
void ratio values were around 0.65 or higher. Interestingly, Robertson et al. (2011) pointed out,
based on field measurements on Suncor’s composite tailings in Pond 5, that effective stress
starts to develop at a critical void ratio of approximately 0.65. The strength characterization re-
sults presented in the following section corroborate their findings.
 100
clay silt sand
90

geotechnical fines < 75 μm

80 pail #1
Percent Finer (by Weight)

mining fines < 44 μm

70 pail #2
60
50
40
30
20
10
0
0.001 0.01
0.1 1
Particle Size (mm)
Figure 4. Particle size distribution curves of NST from pails #1 and #2

3.2 NST Strength Characterization

Both pocket pen and field shear vane tests generate readings that are directly correlated with
shear strengths. The correlations depend on the testing device employed. For the Humboldt
pocket pen, the shear strength, Sp (in kPa) was calculated using the following equation:
1 Sq 
Sp   95.76   (1)
16  2
where Sq is the unconfined compressive strength; “95.76” is a factor that converts tons/ft2 into
kPa; and “16” is to account for the foot adaptor, which has an effective area sixteen times great-
er than the piston (Humboldt, 2010). For the Humboldt shear vane, the shear strength, Sf (in
kPa), was calculated using the equation below:
Sf  0.625  Sv (2)
where Sv is the value registered on the graduated scale of the vane device (Humboldt, 2011).
Shear strengths obtained from field shear vanes are normally corrected to account for the plas-
ticity and liquidity of the material being tested (Bjerrum, 1972). Such correction is not required
here, as the NST is non-plastic.
The cone resistance index (qc) has been correlated with several soil properties, including shear
strength and friction angle (Kulhawy & Mayne, 1990). For the hand-held Rimik cone, the fol-
lowing equations were used for the shear strength, Sc (in kPa), and friction angle, ϕ (in degrees),
respectively:
qc  σv q 
Sc  and tan ϕ  0.10  0.38  log  c  (3)
N kt  σv 
where σv is the total vertical stress; σv' is the effective vertical stress; Nkt is the cone bearing fac-
tor. Nkt is an empirical factor that ranges from 10 to 20 with an average of 15, the value assumed
here (Eid & Stark, 1998).
Figure 4a shows the shear strengths calculated from the pocket pen and field shear vane
measurements. We refrained from using the term “undrained” as we are not sure of the actual
drainage conditions of the tests carried out for the project. For Pan #11, shear strengths increase
gradually with depth, whereas for Pans #09 and #08, much higher shear strengths are encoun-
tered at the surface due to desiccation; however, at about 25 cm from tailings surface, shear
strength values are almost as low as those measured in Pan #11.
Figures 4b,c display the shear strengths and friction angles estimated from the Rimik cone.
For Pan #11, shear strengths are relatively constant. For Pans #09 and #08, much higher shear
strengths are encountered at the surface of the NST due to desiccation – and near the contact be-
tween NST and beach sand layer due to under drainage. Inferred friction angles vary from about
18º–24º for the very loose NST in Pan #11 to about 27º–30º for the loose to medium dense NST
in Pans #09 and #08. The desiccated NST, as expected, shows higher friction angles, above 40º
in average. The results from the Rimik cone, pocket pen and vane shear tests are consistent with
one another.

0.6
(a) (b) Rimik (c) Rimik

pocket pen NST (top)

0.5
Distance from Bottom of Pan (m)

0.4

vane shear
0.3
Pan #11
Pan #09
0.2 Pan #08

NST (bottom)
0.1

beach sand

0.0
0 3 6 9 12 0 3 6 9 12 10 20 30 40 50
Shear Strength (kPa) Shear Strength (kPa) Fricion Angle (deg)

Figure 4. Profiles of (a) shear strength from pocket pen and vane shear tests, (b) shear strength from
Rimik tests and (c) friction angle from Rimik tests for Pans #11, #09 and #08.

3.3 Ultimate Bearing Capacity

Five plate load tests were completed within two days: one in Pan #11, two in Pan #09 and two in
Pan #08. The results of the plate load testing are presented in Figure 5, in terms of averaged
pressure (force divided by plate area) versus displacement curves. These curves portray the typi-
cal plate load test response of loose to medium cohesionless soils (Lambe & Whitman, 1969).
That is, for the loose to medium NST in Pans #09 and #08, the curves show an initial stiffer me-
chanical response followed by a softer behaviour after a sharp bend in the curve, corresponding
to a local shear failure. The initial stiffer response is likely governed by the presence of desic-
cated material at the surface of the NST. For the very loose NST in Pan #11, the shear zones at
the sides of the footing are not well defined and no surface heave was observed. This is termed
“punching failure”, much like that observed in Pans #09 and #08 (Figure 6).
Typically, the “bearing capacity” (ql) of the soil is defined as the pressure at which there is a
pronounced change in slope, corresponding to the “knee” or bend in the pressure-displacement
curve. The “ultimate bearing capacity” (qult) is the bearing pressure that causes a sudden cata-
strophic settlement of the foundation (Lambe & Whitman, 1969). This catastrophic settlement
was not observed in our tests for two reasons: (a) the NST was loaded at a displacement-
controlled mode, and (b) the shear strength of sand-dominated materials depends on stress and
confinement levels; that is, when the load increases, so does the stress level and therefore the
shear strength. As such, the ultimate bearing capacity of sands increases as more load is applied,
and failure is not clearly defined. For that reason, some authors have defined the ultimate bear-
ing capacity of coarse-grained soils as the bearing pressure for a settlement equivalent to 10% of
the plate diameter (Briaud, 2013).
 Vertical Pressure (kPa)
0 10 20 30 40
0
(a)
10
displacement at 1m
10% plate diam. (b) Pan
20
Pan #11
30 Pan #09
Vertical Displacement (mm)

Pan #08 Test Test

Location Location
40
[2] [1]

1m
Load
50 [2-1] Frame
[1]

60

0.4 m
diam. 0.2 m
[2]
70
[1]

80 [2]
0.3 m 0.3 m
test
90 [2-2] locations

100
Figure 5. Plate load testing (a) results and (b) locations.

Figure 6. Punching failure observed on the NST in (a) Pan #11 and (b) Pan #08

In light of the above discussion, the bearing capacity of the NST material varied from about
1 kPa for the 76.3% SBW material to around 9–12 kPa for the 80.3–80.7% SBW material. If
one resorts to Briaud’s definition of ultimate bearing capacity, then qult varied from about 3 kPa
for Pan #11 to around 12–16 kPa for Pans #09 and #08. Evidently, if larger settlements can be
tolerated (e.g. in case of construction equipment), then higher ultimate bearing capacities can be
expected. For example, for a plate displacement of 60 mm, an ultimate bearing capacity of 20-30
kPa is obtained.
Results of field plate load tests can be used to estimate the ultimate bearing capacity of actual
footings (track or wheel footprint). For cohesionless soils, the ultimate bearing capacity of actu-
al footings is proportional to the ultimate bearing capacity obtained from a plate load test by the
ratio of footing size to plate size (Das, 2010).
4 SIMPLIFIED BEARING CAPACITY ANALYSIS

Several formulations are available to calculate the ultimate bearing capacity (qult) of soils. Ter-
zaghi’s, Meyerhof’s, Hansen’s and Vesic’s methods are among the most widely used ones
(Bowles, 1988). Meyerhof’s seems to be the best method for frictional soils, and will be applied
here. For concentrically loaded circular footings, Meyerhof’s formulation (in SI units) reduces
to the following equation:
q ult  c  N c  s c  d c  q  N q  s q  d q  0.5 γ  B  N γ  s γ  d γ (5)
where:
 c (in kPa) is the cohesion (or shear strength) of the soil.
 ϕ (in degrees) is the friction angle of the soil.
 γ' (in kN/m3) is the effective unit weight of the soil.
 q (in kPa) is an external load (equal to γ'D for loading plate embedded into the soil).
 B (in m) is the plate diameter; D (in m) is the embedment depth of the plate.
 Kp is the coefficient of passive resistance (= tan2(45º + ϕ/2)).
 Nc, Nq and Nγ are dimensionless bearing capacity factors; Nc = (Nq – 1) cot(ϕ), Nγ = (Nq – 1)
tan(1.4ϕ), Nq = eπ tanϕ Kp.
 sc, sq and sγ are dimensionless shape factor; sc = 1 + 0.2Kp; sq = sγ = 1 + 0.1Kp for ϕ > 10º.
 dc, dq and dγ are dimensionless depth factor; dc = 1 + 0.2√Kp (B/D); sq = sγ = 1 + 0.1√Kp (B/D)
for ϕ > 10º.

Table 2 compares measured and calculated ultimate bearing capacities under various model-
ling assumptions. The mechanical properties of the NST (i.e. shear strength and friction angle)
were obtained from the Rimik cone data. The saturated unit weights were calculated directly
from the solids contents in Figure 3. In all cases, Meyerhof’s method overestimates the ultimate
bearing capacities, with the best estimates being based on shear strength values. This suggests
that the friction angles in Figure 4 are slightly overestimated.
In Meyerhof’s formulation, the soil was assumed homogenous. This may not apply to Pans
#09 and #08 because of the upper 5–10 cm NST material. Employing Meyerhof’s method for a
strong soil layer over a weak soil layer profile (Hanna & Meyerhof, 1980) yields ultimate bear-
ing capacity estimates that are still comparable to the homogeneous case with the loading plate
at surface.

Table 2. Measured versus calculated ultimate bearing capacities based on Meyerhof’s method
Pan #11 Pan #08 Pan #09
SBWave (%) 76.3 80.3 80.7
γs(ave) (kN/m3) 18.6 19.8 19.8
Measured qult for plate 20 mm and 60 qult-20mm (kPa) 3 16 12–14
mm below NST surface qult-60mm (kPa) 8 28–32 20–24
Plate at NST surface and 60 mm below qult-surface (kPa) 4 17 30
surface (ϕ from Rimik data) qult-60mm (kPa) 10 32 51
ϕave (deg) 22º 29º 32º
Plate at NST surface / NST as a cohe- qult (kPa) 4 13 20
sive material (S from Rimik data) Save (kPa) 0.7 2.1 3.2
Plate at NST surface / Two-layered qult-surface (kPa) NST in Pan 21 30
frictional material (ϕ from Rimik data) ϕave-top10cm (deg) #11 is homo- 40º 41º
ϕave-elsewhere (deg) geneous 27º 30º

5 CONCLUSIONS

The bearing capacity of the NST material varied from as low as 1 kPa to about 9–12 kPa. The
ultimate bearing capacity was found to increase with plate settlement, as expected for a sand-
dominated frictional material, whose shear strength depends on stress and confinement levels.
Based on the plate load test results, it was also evident that the ultimate bearing capacity of NST
increases with solids content. However, because we tested three pans effectively at two solids
content (Pan #11 at about 76% SBW and Pans #09 and #08 at about 80-81% SBW), there is in-
sufficient data to derive a reliable relationship between bearing capacity and solids content. This
relationship is very likely to be non-linear, as is the case for undrained shear strength and void
ratio (Sobkowicz et al., 2013; Moore et al., 2014).
While considering the above, it is also important to note that we have presented a preliminary
evaluation based on laboratory measurements using about 0.50 m thick layers of laboratory pre-
pared and deposited NST with a loading surface of 0.20 m diameter plate. Inferring field esti-
mates of bearing capacity must consider the geotechnical issues associated with extrapolating
laboratory data and empirical relationships to field scale designs. This NST was prepared under
controlled lab conditions. The NST in the pans were exposed to dewatering conditions, and both
may differ significantly from field deposits. Furthermore, field bearing capacity must consider
the larger volume and depth of soil experiencing higher stresses imposed by larger scale loads.
The NST material was at a very loose state after deposition and initial dewatering. Liquefac-
tion of loose, saturated sands may be caused by either static or dynamic, undrained loading. The
assessment of static and seismic liquefaction potential is certainly advised as part of the design
of NST deposits. A similar investigation is advisable if there were potential for deposits to be
trafficked by equipment applying dynamic/cyclic loads.

6 ACKNOWLEDGEMENTS

The authors would like to acknowledge the contributions of Messrs. Vincent Gao, Chenxi
Zhang, Iain Gidley and Trempess Moore for this research project, as well as the support from a
number of staff within Canadian Natural.

7 REFERENCES

Alberta Energy Regulator (AER). 2009. Directive 074: Tailings Performance Criteria and Requirements
for Oil Sands Mining Schemes. February 3, 2009.
Alberta Energy Regulator (AER). 2016. Directive 085: Fluid Tailings Management for Oil Sands Mining
Projects. July 14, 2016.
Bjerrum, L. 1972. Embankments on soft ground. In Proc. Specialty Conf. of Earth and Earth Supported
Structures: 1-54, Vol. 2, ASCE, New York.
Bowles, J.E. 1988. Foundation Analysis and Design. 4th edition, McGraw-Hill.
Briaud, J.-L. 2013. Geotechnical Engineering: Unsaturated and Saturated Soils. Wiley.
Das, B.M. 2010. Principles of Geotechnical Engineering. 7th edition. Stamford: Cengage Learning.
Eid, H.T. & Stark, T.D. 1998. Undrained shear strength from cone penetration test. In: Geotechnical Site
Characterization, P.K. Robertson & P.W. Mayne (Eds.), Rotterdam: Balkema.
Hanna, A.M. & Meyerhof, G.G. 1980. Design charts for ultimate bearing capacity of foundations on sand
overlying soft clay. Canadian Geotechnical Journal, 17, 300-303.
Humboldt. 2010. H-4200 & H-4200F Product Manual. Soil Penetrometer. Illinois, US.
Humboldt. 2011. H-4227 Product Manual. Vane Inspection Kit. Illinois, US.
Kulhawy, F.H. & Mayne, P.W. 1990. Manual on Estimating Soil Properties for Foundation Design. Re-
port EL-68000, Electric Power Research Institute, EPRI, August 1990.
Lambe, T.W. & Whitman, R.V. 1969. Soil Mechanics. New York: John Wiley & Sons.
Moore, T., Zhang, C. & Pinheiro, M. 2014. Geotechnical performance of ATA™-treated fluid fine tail-
ings. In Paste 2014, Vancouver, BC.
Robertson, A.M., Caldwell, J., Wells, P.S. & Barnekow, U. 2011. Advances from the Wismut project on
soft tailings cover technologies. Retrieved from: http://www.infomine.com/publications/docs/Robert-
son2011.pdf
Sobkowicz, J.C., Pinheiro, M., Moore, T.W. & Moffett, R.H. 2013. Predictive model for thick lift place-
ment of silica-treated fluid fine tailings. In Tailings & Mine Waste 2013, Banff, AB, Canada.