See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/350836532

FIELD TEST VERFIFICATION OF GROUND BEARING PRESSURE UNDER RIG

TRACKS AND IMPLICATIONS FOR WORKING PLATFORM DESIGN

Article in Ground Engineering · April 2021

CITATIONS READS

0 364

6 authors, including:

Michał Topolnicki Grzegorz Soltys

Keller Holding GmbH
5 PUBLICATIONS 24 CITATIONS
30 PUBLICATIONS 258 CITATIONS
SEE PROFILE
SEE PROFILE

All content following this page was uploaded by Michał Topolnicki on 13 April 2021.

The user has requested enhancement of the downloaded file.

FIELD TEST VERFIFICATION OF GROUND BEARING PRESSURE UNDER RIG
TRACKS AND IMPLICATIONS FOR WORKING PLATFORM DESIGN

Michał Topolnicki1, Jürgen Wäger2, Steffen Schweizer2, Alfred Koller2, Tadeusz Brzozowski3 and
Grzegorz Sołtys3
1
Keller Holding GmbH, 2 Liebherr-Werk Nenzing, GmbH, 3 Keller Polska Sp. z o.o.

1. INTRODUCTION
A realistic assessment of the safety of tracked plant operation on construction sites, such as piling or
drilling rigs, depends, inter alia, on the accuracy of determining the pressures exerted on the ground by
the tracks. If the imposed ground bearing pressure (GBP) is too high for the actual condition of the
working platform, excessive ground deformations occur and may, ultimately, cause rig overturning.
The GBP distributions along the left and right track depend largely on the position of the overall
centre of gravity (COG) of the rig with respect to the centre of the tracks' system. Traditionally, the
location of the COG has been calculated manually for critical rig orientations, taking into account the
weights and centres of gravity of the various components of the particular item of plant. With the sensor
technology installed in modern geotechnical rigs it is possible, however, to determine the overall COG
for any rig configuration in real time. A new assist system, developed by Liebherr [8], utilises the actual
position of the COG to calculate and display the GBP in real time, and to compare it with the safety
limit specified by the user for the actual working conditions. In this way dangerous situations can be
controlled or prevented before the load approaches a critical range, reducing the risk of a rig falling over.
In addition, the assist system also helps to optimize rig utilisation within its full range of capacity,
increasing work efficiency without compromising safety.
The complicated construction of plant tracks, and especially the flexible connection of track links
below the track rollers and the largely variable ground conditions in which the rigs operate, make the
interaction of the track and the ground very complex and difficult to capture in calculation models. This
includes the prediction of GBP. An example of a problem is the determination of the maximum load
under the track, which has not been satisfactorily solved so far. The maximum pressure is often estimated
on the basis of the average pressure, which is multiplied by an empirically determined amplifying factor
(in the range of 1.5 to 2.5, depending on the ratio of the average roadwheel spacing to track pitch, and
on the ground stiffness [1]). Similarly, for the purposes of designing working platforms, drastically
simplified GBP distributions are typically used in practice, calculated as for rigid shallow pad
foundations, cf. [2, 3]. While a full-scale experimental substantiation of these distributions is rare, it is
commonly believed that they represent a conservative and safe approach.
In this study the principles of calculation of the GBP under a track are discussed first. Subsequently,
a field test conducted in co-operation between Keller and Liebherr is presented, which involved the use
of a LRB 125XL drilling rig equipped with a triple-shaft soil mixing tool. The track loads and the GBPs,
determined with Liebherr's Ground Pressure Visualisation (GPV) assist system installed on the rig, were
recorded during soil mixing operations carried out at five leader positions relative to the tracks. Pressure
cells, embedded in the soil below the right track, were used to measure the vertical stress induced in the
ground. Based on the comparison of calculated and measured stresses, conclusions are drawn regarding
the reliability of the GBP determined. Plausible modifications of the calculation method are proposed,
and the implications of the modified GBP distributions for the design of working platforms are
discussed.

2. ANALYTICAL FRAMEWORK
A model track system and the parameters used are depicted in Figure 1. The left and right track are
represented by their footprint area in contact with the ground surface, B×L, and are rigidly connected by
the undercarriage assembly. The origin of the coordinate system XOY is located in the geometrical
centre of the model, and the X-axis is parallel to track length. The magnitude of the imposed total vertical
load, F, and the associated eccentricities, ex and ey, are known variables. They are determined depending
on the orientation of the body of the rig with respect to the orientation of the tracks, defined by the slew

1
Published in Ground Engineering, April, 2021, p. 26-38
angle θ. Pressure distribution along the width of the track is considered uniform given that the pressure
variation is minor, and the ground surface is assumed horizontal.

Fig. 1. Model track system and parameters used for ground bearing pressure (GBP) calculations

Ground bearing pressures under tracks of geotechnical rigs are conventionally calculated as for shallow
pad foundations (e.g. Tab. F1 in [3]). For this purpose the total vertical load of the rig, which
encompasses contributions from a number of load elements including plant weight and operational
loads, needs to be adequately assigned to the left and right track. The split of the vertical load between
both tracks, fulfilling the equilibrium condition
F = FL + FR (1)
is straightforward, and depends on the eccentricity ey and track spacing s only, namely
1 e  1 e 
FL = F  − y  , FR = F  + y  (2)
2 s  2 s 
The individual loads FL and FR generate either a trapezoidal or a triangular GBP distribution over the
length of the track that is in contact with the ground, depending on their distance from the Y-axis. A
proper assignment of the position of both loads is therefore essential. This important step is often
oversimplified assuming that load eccentricities adopted for the left and right tracks, exL and exR, are the
same as the eccentricity ex of the total load. Examples of this approach, hereinafter dubbed the 'equal ex'
method, can be found in [5, 7] and in the spreadsheet included in appendix 3 to the Guide to Working
Platforms, issued by EFFC/DFI [4]. However, it should be noted that in the case of a susceptible ground
the movement of the resultant vertical load toward the right track, for example, will cause the rig to tilt,
and the rear part of the left track may lose contact with the ground. In such a situation, the pressure
distribution may be triangular under the left track and trapezoidal under the right track. The
corresponding eccentricities of individual track loads are then different. Therefore, in general it holds
for ex ≠ 0, and except leader positions θ = 0 and θ = 180, that
exL  exR  ex (3)

as also observed in practice. A closed form analytical solution which accounts for (3) is not available
for a full range of possible double eccentricities. The negative GBPs which can occur computationally
under the tracks cannot be transmitted to the ground, and must be redistributed to compressive zones
until the equilibrium condition
F  ex = FL  exL + FR  exR (4)

is met. This is achievable with iterative computations, as shown in [6]. The Liebherr's GPV assistant
splits the full range of double eccentricities into the following load cases:
- trapezoidal pressure distribution under both tracks,
- triangular pressure distribution under both tracks,
- trapezoidal pressure distribution under one track and triangular under the opposite track,
2
Published in Ground Engineering, April, 2021, p. 26-38
and checks which of those is applicable for the actual loading utilising the equilibrium condition (4).
Within the detected load cases, the left and right GBP distributions, assumed to be uniform across the
track's width, are calculated based on the eccentricities of the individual track loads FL and FR. Both
methods are compared in Figure 2 for the rig parameters adopted in [6]. The authors considered a full
rotation of the rig body in relation to stationary tracks, i.e. θ = 0 to 360, and calculated the maximum
bearing pressure induced under a track corner (point M in Fig. 1). Variable positions of the vertical load,
F = 1, were defined by the radius r

r = ex2 + e y2 , ex = r  cos  , ey = r  sin  (5)

selected in the range of 0.2 to 1.2 m (maximum: ex/L = 0.3 for θ = 0, and ey/s = 0.46 for θ = 90).

Fig. 2. Maximum GBP under the right track for a full rotation of the leader, calculated with two
different methods (case analysed in [6])

The comparison shows that both methods give consistent results over the entire range of leader positions.
In fact, the coincidence is even better than displayed because the method [6] takes into account uneven
stress distribution across the track's width. Therefore, a more adequate comparison should involve the
arithmetic mean of stresses computed in both track corners along the same edge (however, the data for
the second track corner were not shown). In practice, the assumption of a uniform GBP distribution over
the width of the track is more justified because the rollers do not transfer bending moment on the track
pads in the same way as a rigid connection.
It is also interesting to compare the 'equal ex' method with the GPV method, as shown in Figure 3.
The comparison demonstrates that the discrepancies between the two methods increase significantly
with increasing eccentricity. Moreover, in the range of 0< θ <180 the GBP calculated with the 'equal
ex' method are always greater, which may lead to the conclusion that this method is conservative.
However, this doesn't apply to the range of 180< θ <360, where the relationship is reversed.
Consequently, when there is a need to check the stability of both the right and the left tracks, the use of
the 'equal ex' method can be unsafe.

In addition to the accurate GPV method, a closed form solution for the determination of GBP can be
proposed for a limited but practically useful range of double eccentricities. It is postulated that the
eccentricities exL and exR are inversely proportional to individual track loads FL and FR, namely
exL exR = FR FL (6)
3
Published in Ground Engineering, April, 2021, p. 26-38
 Fig. 3. Maximum GBP under the right track for a full rotation of the leader, calculated with the GPV
and the 'equal ex' method (case analysed in [6])

Equation (6) together with the equilibrium condition (4) leads to

F  ex e s F  ex e s
exL = = x , exR = = x (7)
2 FL s − 2e y 2 FR s + 2e y

Once the corresponding pairs FL, exL and FR, exR are determined based on the input variables F, ex, ey and
s, the resulting GBP distributions read:
- for |ex,L/R|  L/6
FL / R  6  ex , L / R  FL / R  6  ex , L / R 
p1, L / R = 1 +  , p2, L / R = 1 −  (8)
BL  L  BL  L 

- for L/6 < |ex,L/R|  L/3 and ex,L/R > 0

2 FL / R
p1, L / R = , p2,L/ R = 0 , cL/ R = L / 2 − ex,L/ R (9)
3 cL / R B

- for L/6 < |ex,L/R|  L/3 and ex,L/R < 0

2 FL / R
p1, L/ R = 0 , p2, L / R = , cL/ R = L / 2 + ex, L / R (10)
3 cL / R B

For the case analysed in Figure 3 a better agreement is achieved with the assumption (6) than with the
'equal ex' method, both in terms of the shapes of the curves within a full slewing range, as well as for
stress magnitudes (Fig. 4). As expected, the accuracy of the proposed modification, hereinafter dubbed
the 'divergent ex' method, gradually decreases with increasing eccentricity, excluding specific leader
positions, for which the results are accurate or the GBP is zero. Therefore limit of application must be
considered. Based on a more detailed comparison, it is estimated that the assumption (6) is adequate for
the following range of relative eccentricities of the total vertical load
4
Published in Ground Engineering, April, 2021, p. 26-38
 r
 0.22 for 0< θ <90, 90< θ <180 and 270< θ <360 (11)
L

Within the specified range, the accuracy of calculated GBP compared to the GPV method is

GBP ('divergent ex ')

 0.95 (12)
GBP (GPV)

noting that with increasing relative eccentricity the 'divergent ex' method will slightly underestimate the
GBPs calculated with the GPV method for leader positions 0< θ <90, and slightly overestimate for
leader positions 90< θ <180 and 270< θ <360. For specific leader positions θ = 0, θ = 90 and for
θ ∊ <180; 270> both methods give the same results for a full range of possible eccentricities or the
GBP is zero.

Fig. 4. Maximum GBP under the right track for a full rotation of the leader, calculated with the GPV
and the 'divergent ex' method (case analysed in [6])

3. FIELD TEST ARRANGEMENT AND DATA SELECTED FOR ANALYSES

To observe realistic conditions of rig operation, a soil mixing construction site in Lublin, Poland, was
chosen for test location (Fig. 5). The LRB 125XL used had a total weight of approximately 570 kN,
including the weight of a triple soil mixing shaft and three driving heads MA35. This rig is equipped
with a track system shown in Figure 6a. The ratio of average roadwheel spacing to track pitch is 1.71.
The flat part of the track in contact with a horizontal surface has a length of 3250 mm, and the chamfered
front and rear sections of the track are 440 and 520 mm long, respectively. Since the inclination angle,
β, of chamfered sections is less than 2, the track length of L= 4210 mm should be used for GBP
calculations according to EN 16228-1, app. F [3]. The ground pressure visualisation program was
implemented into the rig's software for the experiment, and the calculated GBPs were displayed directly
on-board and recorded for further analyses.
The ground conditions in the surface zone of the test area allowed for direct operation of the rig
without the need for an additional working platform. Therefore, site preparation was minimal, and
mainly involved placing six plate sensors at a depth of 0.3 m below the target operating level to measure
vertical stresses mobilised in the subsoil. The applied pressure cells, type Geosense TPC-4000, were
aligned coaxially with the planned position of the right track (Fig. 6b). The accuracy of stress
measurement was 0.1% of the full-scale range of 690 kPa. The excavated trench was subsequently filled
5
Published in Ground Engineering, April, 2021, p. 26-38
with sand and compacted with a plate vibrator, and the ground surface was surveyed and marked to
enable exact positioning of the rig. The pressure cells were zeroed before the rig entered the
instrumented area.

Fig. 5. Testing set up: (a) installation of plate sensors at the depth of 0.3 m, (b) positioning of the rig
for testing, (c) wet soil mixing operation carried out at five leader positions relative to the tracks

Fig. 6. Track geometry and the arrangement of plate stress sensors

The conducted test programme included multiple series of measurement and operating conditions, and
allowed close inspection of loads, eccentricities, and ground stresses corresponding to different leader
positions and phases of deep soil mixing. An example of a record of measurements is shown in Figure
7, which illustrates the variability of stress induced in the soil and of the resultant vertical load during
mixing operation for the slew angle of 0.

Only selected experimental data are utilised for the purpose of this study and are summarised in Table
1. They comprise five leader positions, defined by θ = 0, 45, 90, 270 and 315, and include values
of the total vertical load, load eccentricities and stresses measured on plate sensors in the soil,
corresponding to the selected times. The adopted times were chosen in the range of 'stable'
measurements to assure that the data recorded with two independent logging systems operating with
different frequency rates are compatible with each other.

6
Published in Ground Engineering, April, 2021, p. 26-38
 Fig. 7. An example of measured data selection for comparison with calculations

Table 1. Field test data selected for verification analyses

Slew angle θ 1 0 45 90 270 315
F [kN] 572.75 557.49 563.09 549.53 560.34
ex [m] 0.682 0.435 -0.058 -0.110 0.402
ey [m] 0.038 0.539 0.846 -0.586 -0.457
r [m] 0.683 0.693 0.848 0.596 0.609
Plate sensors:
S1: σ [kPa] 17.0 36.1 29.3 12.9 22.4
S2: σ [kPa] 132.0 147.5 110.7 56.0 90.5
S3: σ [kPa] 174.4 162.5 129.5 59.7 97.6
S4: σ [kPa] 156.3 167.4 173.3 83.1 102.8
S5: σ [kPa] 160.4 189.6 201.8 100.0 131.4
S6: σ [kPa] 68.2 108.0 137.4 58.3 69.0
1
Slew angle values are indicative and are used to distinguish leader positions only; the actual
eccentricities ex and ey, determined by Liebherr's Asist System, represent the true leader positions
and are used in calculations.

The maximum eccentricity in Table 1 for leader positions other than θ = 0, 90 and 270 corresponds
to the relative eccentricity of r/L= 0.693/4.21= 0.165 < 0.22, and is comfortably within the validity range
of equation (6). Consequently, the GPV and the 'divergent ex' method can alternatively be used in the
analyses presented in the following.

4. DETERMINANTION OF GBP AND EXPERIMENTAL SUBSTANTIATION

4.1 Manual calculations

The purpose of the analysis is to determine such a GBP distribution under the track that is able to induce,
as closely as possible, the same stresses on the plate sensors as are actually measured. Accordingly, the
right track and the plate sensors are represented by a simplified model shown in Figure 8. The track
contact area, B × L, is divided into nine disconnected rectangular sections, T1 to T9. Six circular stress
7
Published in Ground Engineering, April, 2021, p. 26-38
sensors are replaced by equivalent rectangular plates, S1 to S6. The applied track loads, FR, and the
corresponding eccentricities, exR, for the investigated leader positions are determined based on the input
data presented in Table 1. The subgrade extending below the track represents a linear-elastic and
isotropic half space.

Fig. 8. Adopted model of the right track and plate sensors

In this model the vertical stress at a selected location inside the elastic half space can be calculated based
on the theory of Boussinesq (1885) and its derivative. The solution of Steinbrenner, which determines
the vertical stress, σ, at the depth, z, below a corner point of a rectangular area, Bp × Lp, loaded with a
uniform load, p, reads

p  B L B L z 1 1 
=  arctan p p + p p  2 +  (13)
2  4 zW 4W  Bp / 4 + z 2 L2p / 4 + z 2  

where W = Bp2 / 4 + L2p / 4 + z 2 .

Equation (13), together with the superposition principle valid for linear-elastic systems, allows
calculation of the vertical stress at a chosen point on the plate sensor. The imposed stress results from a
combined action of patch loads adopted along the track (pTi pressures). Because of a relatively shallow
embedment of plate sensors below the track, it is sufficient to superimpose stresses generated by the
patch load positioned above a selected sensor, pi, and by no more than two adjacent patch loads on both
sides, i.e. by pTi-1, pTi-2 and pTi+1, pTi+2 (see point B in Figure 8). The resultant stress on each sensor was
determined as a mean value of stresses calculated in points A to D.

Individual track sections B × LTi , where i = 1 to 9, are loaded with uniform patch loads, pTi. The
magnitude of pTi can be determined by iterative trials until the best possible agreement between the
calculated and measured stresses on the plate sensors is obtained, and the following compatibility
conditions are simultaneously satisfied with the assumed accuracy:

8
Published in Ground Engineering, April, 2021, p. 26-38
 - for the track load
i =9
FR −  pTi BLTi  0.05 kN (14)
i =1

- for the eccentricity

i =9 i =9
exR −  ( pTi LTi  xi ) p Ti LTi  0.005 m , (15)
i =1 i =1

where FR and exR are determined by equations (2) and (7), respectively, depending on the investigated
leader position. Furthermore, the lengths of the front and rear sections of the track may be shortened, if
needed to reach a better coincidence between the calculated and measured stresses (cf. Fig. 8)

LT 1 := LT 1 − a f , LT 9 := LT 9 − ar (16)

If a reduced track length is adopted in the analysis due to the need to consider the effect of inclined
sections of the crawler in the front and rear parts of the track, and af ≠ ar is selected, the centre of the
shortened track will no longer coincide with the Y-axis, and the corresponding shift along the X-axis
equals (cf. Fig. 8)
ex = (ar − a f ) / 2 (17)

Consequently, adjusted eccentricity should be used in the equation (4), (5) and (7)
ex := ex − ex (18)

together with the reduced track length in contact with the ground in equations (8) to (11)
L := L − a f − ar (19)

The results of analyses are shown in Figure 9, where column (a) presents calculated and measured
stresses on plate sensors, and column (b) presents back calculated distributions of patch loads in
comparison with two theoretical GBP distributions. The best fit between the measured and calculated
stresses at the depth of 0.3 m below the track was achieved adopting af = 0.2 m and ar = 0.28, that is for
the track length of 3.73 m. Since the track length in contact with the ground depends on track settlement,
s, and for the utilised rig equals (cf. Fig. 6a)
2s
L = 3250 + = 3250 + 80  s [mm] for 0  s  11 [mm] (20)
tan 
the resulting length of 3.73 m corresponds, in an idealised situation, to a uniform track settlement of 6
mm (the inclination of chamfered track sections is not constant and slightly varies depending on the
position of the crawler chain). It also represents a 'mean' track length when compared with the regulation
included in EN 16228-1 [3], which makes the length of the track dependent on the angle of inclination
of its front and rear parts.
Taking into account that soil stresses measured under rig tracks often exhibit irregular patterns and
inevitable peaks, it can be assumed that the obtained compliance is satisfactory (slightly better for leader
positions determined by θ = 0, 45, 90 and 315, and less accurate for θ = 270). The good
compatibility between the measured and calculated stresses on the plate sensors allows consideration of
the corresponding GBPs, represented by the patch loads, as reference distributions for further
comparisons. The graphs presented in column (b) demonstrate that the distributions calculated with the
'divergent ex' (or equivalently with the GPV) method using the 'mean' track length of 3.73 m are in better
agreement with the patch load distributions than results using the 'equal ex' method and the track length
of 4.21 m, compliant with [3]. The improvement is primarily due to the reduction of the effective track
length and, to a lesser extent, due to the corrected position of the resultant track load FR, determined by
the eccentricity exR.

9
Published in Ground Engineering, April, 2021, p. 26-38
 Fig. 9. (a) Best fit between the measured and calculated stresses on plate sensors, (b) back calculated
patch load distributions in comparison with two theoretical GBP distributions under the right track.

10
Published in Ground Engineering, April, 2021, p. 26-38
4.2 Finite Element calculations
A limited FE study using a simple 3D model of the track assembly and the subsoil was conducted to
eliminate the need for an a priori split of the vertical load into the left and right track and the assumption
about their positions, and to investigate - in a narrow range - the effect of ground and track stiffness on
GBP distributions under the tracks. The utilised model, built in PLAXIS 3D code, is presented in Figure
10.

Fig. 10. Simple FE model of the track assembly and a homogeneous subsoil layer.

The crawlers in contact with the ground were modelled by 22 platens with dimensions of 180×700 mm
(blue), placed on the ground at a distance of 10 mm from each other. The platens were assumed rigid,
and their stiffness corresponded to a steel section of 60×700 mm (arbitrary thickness). The stiffness of
the joints between the platens, also modelled with a plate element, was 1000 times lower. To achieve
the required interaction, the platens of both tracks were connected by means of structural plate elements
with zero thickness (green), placed along the longitudinal axis of both tracks. The left and right tracks
were connected with a transverse rigid plate (orange), on which the total vertical load was applied
according to the analysed leader position. A contact element was also adopted under the tracks to
disconnect the mesh nodes in the event of tensile stresses.

To comply with the manual analysis, a linear elastic constitutive model was adopted. Since the subsoil
at the field test location was not investigated in detail, arbitrarily selected lower and upper bound values
of the Young's modulus E, equal to 40 and 120 MPa, were adopted to study the influence of ground
stiffness on the mobilised stresses. Based on observation of the earth and compaction works conducted
during test site preparation it is believed, however, that the actual soil stiffness on site was closer to that
represented by the upper bound value of the Young's modulus.

Similar to the analyses presented above, the calculated and measured stresses on the plate sensors were
inspected first, and the FE model was purposely 'calibrated' to reach the best possible compatibility. The
process of model adjustment mainly included changes of the stiffness of the plate elements connecting
the platens, as well as changes of the width of the transverse plate connecting both tracks. The achieved
results are presented in Figure 11 for five leader positions and the adopted values of the Young's
modulus. The dotted curves shown on the graphs are composed of segmented least-square trend lines
determined for short sections of the track. The use of average values allows a more unambiguous
comparison of the calculation results with the stress measurements. A satisfactory agreement was
obtained in all cases investigated, and the measured stresses lay closer to FE curves computed with E
=120 MPa, as anticipated.

11
Published in Ground Engineering, April, 2021, p. 26-38
 Fig. 11. Measured and calculated stresses on plate sensors, determined with the FE model for lower
and upper bound values of the elastic modulus of the subsoil

An example of the bearing pressure mobilised below both tracks is presented in Figure 12 for the leader
position defined by θ = 45. By comparing the two stress maps it can be seen that the tension cut-off
zones formed in the rear parts of both tracks depend on soil interaction, and are larger in the case of a
stiffer ground. The resultant vertical force acting on the left track, FL, moves forward, and the
corresponding eccentricity, exL, increases and exceeds the eccentricity, ex, of the total load. At the same
time the situation on the right track is the opposite. The resultant vertical force, FR, moves backwards,
and the eccentricity, exR, decreases below the ex-value. In terms of a quantitative evaluation, reference is
made to Table 2, which includes the eccentricities obtained from the computed stress maps, as well as
the eccentricities calculated with the proposed equations (7). The achieved compatibility leads to the
conclusion that the introduced modification, i.e. the inverse proportionality (6), has been also
corroborated by the FE calculations.

Tab. 2. Track load eccentricities for leader position θ = 45

Calculation method exL [m] exR [m]

Equations (7) and L= 4.21m 0,599 0,317

FE with E = 40 MPa 0,572 0,333
FE with E = 120 MPa 0,593 0,357
Note: ex = 0.435 (cf. Tab. 1)

12
Published in Ground Engineering, April, 2021, p. 26-38
 Fig. 12. GBP in dependence of soil stiffness for leader position θ = 45

In the second step of the analyses, the GBP distributions computed with the FE model under the right
track were compared with the back calculated patch load distributions, and with the GBP distributions
calculated using the 'divergent ex' (or equivalently the GPV) method. A rather satisfactory compatibility
in all cases examined, as shown in Figure 13, leads to the conclusion that the use of the reduced track
length of 3.73 m, together with the 'divergent ex' method or - in general - the GPV method, significantly
improves the accuracy of predicted GBPs.

Fig. 13. GBPs computed with the FE model in comparison with the patch load distributions and the
distributions calculated with the 'divergent ex' (or equivalently with the GPV) method
13
Published in Ground Engineering, April, 2021, p. 26-38
FE analyses also allow examination of the maximum stresses predicted under the tracks. For
comparative purposes, the case presented in Figure 12 was supplemented with calculations in which the
linear elastic perfectly plastic constitutive model of Mohr-Coulomb was used to describe the subsoil
(Fig. 14). The stress map obtained illustrates that along track edges the ultimate shear strength of the
soil is quickly mobilised and localised plastic deformations occur, triggering stress redistribution.
Consequently, the peak stress is also reduced compared to values predicted with the linear elastic model.
The discrepancies between the edge stresses calculated numerically and analytically, visible in Figure
13, result from the assumption of a perfectly rigid track ('foundation') adopted in the analytical methods.
Consequently, in the case of the considered analytical methods the maximum GBP always appears on
the track's edge, while in the analyses involving advanced soil-structure interaction models it usually
occurs along the track, depending on load position and the relative stiffness of the track and the ground.
Similar results have been obtained with the computer-aided method known as RTVPM. This model
takes into account all major rig design parameters, including weight, COG location, number, location
and diameter of roadwheels, location of sprocket and idlers, supporting roller arrangements, track
dimensions and geometry, initial track tension and drawbar hitch location, as well as pertinent ground
characteristics, such as the pressure-settlement and shearing characteristics and the response to repetitive
loading [9].

Fig. 14. GBP distributions computed with the FE model for two different constitutive models of the
soil

5. IMPLICATIONS FOR THE DESIGN OF WORKING PLATFORMS

GBP applied for the design of working platform should generally be determined by the plant supplier
for the specific task. The use of dedicated systems for this purpose, such as the GPV assistant, reduces
the risk of making a mistake. However, there will be instances when the designer needs to calculate
imposed track loads from first principles (e.g. older plant, to check imposed loads supplied by others).
In such cases, the outlined 'divergent ex' method can be used within the limit of its validity (11). Along
with the adoption of a mean track length, corresponding to the results of the analysis performed, the
impact of modified GBP distributions on the ultimate bearing capacity of a simple working platform
consisting of a granular subgrade is analysed below for illustration.
For the design of working platforms, the trapezoidal or triangular GBP distributions need to be converted
to an equivalent uniformly distributed loading (UDL), p', equal

FL / R
pL / R = (21)
B L
where L' represents the converted track length, corresponding to a centrally loaded track section.
Restricting the analyses to the right track and leader positions corresponding to θ = 0, 45, 90 and
315, the converted track length can be calculated as follows

14
Published in Ground Engineering, April, 2021, p. 26-38
 2 L( p1R + 2 p2 R )
- for 0 < exR  L/6 L = (22)
3 ( p1R + p2 R )
2 ex s
- for L/6 < exR  L/3 L = L − (23)
s + 2e y
If a reduced track length is considered in the analysis, the ex and L values in equations (22) and (23)
should be adjusted using formulas (18) and (19), respectively.

The ultimate bearing capacity, V, of a granular subgrade is typically calculated with a classical formula
normally used for shallow pad foundations
1 B
V =   (1 − 0.3 ) N B 2 L (24)
2 L
where γ' represents effective bulk unit weight of granular soil, and Nγ a bearing capacity factor
depending on the angle of internal friction. For the same conditions on site and the same track width,
the ultimate bearing capacity is a function of the converted track length only, and for a direct comparison
of two calculation approaches, denoted as 1 and 2, it holds

V (2) L(2) − 0.3B

= (25)
V (1) L(1) − 0.3B

The results of calculations for four leader positions exercised during the field test are presented in Table
3. The introduced modifications, comprising the position of the individual track loads along the X-axis
and the mean track length, lead to a reduction in the platform's ultimate bearing capacity when compared
with the 'equal ex' method and the track length defined by EN 16288-1 [3]. The reduction is -14.1%
for θ = 0, -12.8% for θ = 90 and θ = 315, and -6.8% for θ = 45. In the latter case, however, the 'equal
ex' method overestimates the maximum stress by about 10% (compare with Fig. 3), and this effect
decreases the resulting difference. In conclusion, the use of the 'equal ex' method together with the
current regulation for the track length according to [3] may lead to an overprediction of the calculated
safety factor of the designed working platform.

Tab. 3. Impact of two approaches on the ultimate bearing capacity of a granular platform (right track)

Slew angle θ = 0 θ = 45 θ = 90 θ = 315

Approach (*) 1 2 1 2 1 2 1 2
FR [kN] 292.54 364.64 417.73 207.02
p1R [kPa] 195.7 225.3 200.4 207.4 129.9 142.9 110.5 141.7
p2R [kPa] 2.8 0 47.1 71.9 153.6 177.1 30.0 16.8
p' [kPa] 146.8 169.0 155.9 166.6 145.8 165.9 94.7 107.5
L' [m] 2.846 2.473 3.341 3.127 4.093 3.597 3.122 2.751
V(2)/V(1) [%] -14.1 -6.8 -12.8 -12.8
(*)
Approach 1: 'equal ex' method and L = 4.21 m according to EN 16228 [3]
Approach 2: 'divergent ex' (equivalently the GPV) method with the track length L = 3.73 m

6. CONCLUSIONS
The conducted research allows for the formulation of the following conclusions, which are important
for the safety of tracked plant operation on working platforms.

Considering the methodology of calculating the GBP caused by the vertical load, which in general can
act on a double eccentricity with respect to the centre of the track system, it was found that assuming
15
Published in Ground Engineering, April, 2021, p. 26-38
 for the load on the left and right track the same eccentricity ex on which the resultant load acts, leads -
with increasing eccentricities and depending on the position of the leader - to a significant
overestimation or underestimation of the GBP (except leader positions θ = 0, 90 and 270, for which
the results are accurate). While overestimation of ground pressure may be considered conservative,
underestimation is unsafe. Consequently, it should be considered good practice to use the GBP
distributions specified by the plant manufacturer, and computed with adequate calculation methods. In
particular, the use of a dedicated assist system installed on the rig is recommended. The GPV assist
system utilised in this study, available for Liebherr's plant, proved not only convenient during site
operations, but also reliable in terms of evaluating the position of the COG and bearing pressures under
the tracks in real time. As an alternative, the GBPs can also be determined using the simplified 'divergent
ex' method, provided that the eccentricity of the resultant vertical load is within the validity range defined
by the equation (11).
It is also important how the length of the track in contact with the ground is determined. On the basis of
the conducted analyses, it was found that the current recommendation of EN 16228-1 [3], which
provides for a step change in the track length depending on the value of the β-angle below or above 2
degrees, is oversimplified. It seems that it would be more appropriate to make the contact length
dependent on the track settlement under a centrally acting load, i.e. dependent on the susceptibility of
the working platform. The conducted study does not yet allow the formulation of a binding
recommendation in this matter. However, it was found that adopting the average track length, which in
the studied case is less than the length resulting from [3], ensures better compliance of the calculated
GBP distributions with the measured ones. Further investigation is advisable to develop precise
recommendations on this matter.
It has been also demonstrated that the corrected pressure distributions under the tracks, which better
correspond to the results of measurements, influence the dimensioning of the working platforms.
Compared to the conventional approach, i.e. pressures determined by the 'equal ex' method and the track
length in accordance with [3], lower ultimate bearing capacity of the subsoil was obtained. This means
that the designed working platform may actually be less secure than is shown by conventional
calculation.
Finally, it should be emphasized that the continued research on the GBPs under the tracks of
geotechnical plants should be aimed at departing from the rigid track scheme. The cited example of the
RTVPM method [9] shows that it is already possible to adequately take into account the flexible
construction of the track and the susceptibility of the subsoil. Along with the progressive equipping of
foundation rigs with modern control systems, such solutions will become part of the assist systems
dedicated to control GBPs. This will also eliminate the need for separate regulations regarding the
variable track length in contact with the ground.

REFERENCES

[1] Bodziony P., Kasztelewicz Z, Patyk M., Characteristics of the dozer chassis and track-type
undercarriage used in surface mining (in Polish), Przegląd Górniczy, 10/2016.
[2] BR 470 (2004): Working platforms for tracked plant: good practice guide to the design, installation,
maintenance and repair of ground-supported working platforms, BRE Bookshop.
[3] EN 16228-1 (2014). Drilling and foundation equipment – Safety – Part 1: Common requirements.
Beuth Publishing Company.
[4] Guide to Working Platforms (2020), EFFC and DFI Working Platforms Task Group.
[5] Kummeter B. (2001), Bodenpressung unter Raupenketten, Tiefbau 6/2001.
[6] Meißner H., Becker A., Geib S. (1992), Zur Bodendruckverteilung unter Raupenketten hoher
Baumaschinen bei Schwerpunktwanderungen, Tiefbau 8/1992.
[7] TWf (2019), Working Platforms Design of granular working platforms for construction plant, A
guide to good practice, Temporary Works Forum, www.twforum.org.uk, document: TWf2019: 02
[8] Zörner S., Dona M., Bacher T. (2019), Bodendruckanzeige für Spezialtiefbaumaschinen. In Theorie
& Praxis des Spezialtiefbaus. Proc. 12th Austrian Geotechniktagung in Vienna, p. 323–336.
[9] Wong J. Y., Theory of ground vehicles, John Wiley & Sons, 3rd ed., 2001.

16
Published in Ground Engineering, April, 2021, p. 26-38

View publication stats