46

How the Pressure Build-Up Affects the Penetration Length of

Grout: New Formulation of Radial Flow of Grout Incorporating
Variable Pressure

Authors:
Funehag, Johan
Claesson, Johan
Presentation Preference: Oral
Track: Track 2/ Drilling and Grouting
Session: 2A31/ Design and Assessment, Part I
Is this a student presentation?: None Specified
Bio: None Specified
Submission type: Technical Papers (6-10 pages in length)
Date Uploaded: 3/7/2017 4:26:38 AM
____________________________________________________________

Project Manager Submission separator pages

Grouting 2017, Honolulu
Theme: Innovation and Developments in Ground Treatment Technology

How the pressure build-up affects the penetration length of grout- New formulation of
radial flow of grout incorporating variable pressure

Johan Funehag, Ph.D.,1 and Johan Claesson, Prof.2

1
Chalmers University of Technology, Gteborg, Sweden, johan.funehag@chalmers.se
2
Chalmers University of Technology, Gteborg, Sweden, Johan.claesson@chalmers.se

ABSTRACT

For around two decades of research and development in the field of grouting in hard jointed
rock, the design process has taken some leaps forward. Stille and Gustafson, 2005 and
Funehag and Gustafson 2008, shows how a grouting design can be computed. A grouting
design in hard rock can based on the penetration length of grout in rock fractures. The design
comprises considerations of the fracture apertures in the rock mass, the type of grout and its
rheological properties and how the grout is injected i.e pressure and grouting times. When
knowing these parameters an optimized geometry fitting the design is made. Thrn, et al,
2014 describes a fundamental analysis with a comprehensive tool to retrieve the fracture
distribution and aperture distribution of the fractures crossing a cored borehole. The data
needed about the core is geological mapping and hydraulic section tests.
In Gustafson, Claesson and Fransson, (2013) a full derivation of a radial Bingham flow
in a slit is described for constant pressure. By optimizing with a specific pressure and an
efficient grouting time (efficient time means the time when the pressure has reached the
designed pressure) a prognosis a more realistic time consumption for grouting can be
computed. However, the time it takes to reach a certain pressure is dependent on the capacity
of the pump and the how large the fractures widths are. For poorly chosen pumps together
with large fractures the time to reach the design pressure can be significant.
The overall objective for this new formulation was to involve the grouting pressure as a
variable rather than constant. A pressure build-up mimic more a realistic pumping scenario
which enables better prognosis of grouting works. This paper brings up this new formulation
of the radial Bingham flow with variable injection pressure in slit. The benefits of this new
formulation is that it can easily be integrated in other computer programs. One program that
uses this new formulation is a grouting simulator owned and developed by Edvirt AB. The
simulator has been used to pedagogically demonstrate how a variable pressure and restrictions
in grout flow (the pump capacity) affect the penetration length. Further, the results show that
it can be used to predict suitable pump capacity to fit the coming grouting works.
INTRODUCTION

Grouting involves several fields of knowledge: hydrogeology, geology, material science and
machine knowledge. For several decades the research of grouting has been intense and has
presented suggestions on design tools for grouting in hard rock. One tool which seems to
work well in design is the penetration length of grout in deterministic fractures. In analytical
calculations it is difficult to incorporate the pressure build up. Therefore, the overpressure
used in the calculations can often be set as the average pressure over the total grouting time,
i.e. a little bit lower than the presumed pressure to be reached. For design, the calculation of
the penetration length has been incorporated with the time spent on grouting (see Gustafson
and Claesson, 2005 and Claesson, Gustafson and Fransson 2013).
The new formulation made it possible to interpret the penetration length reached for a
certain time. It means in reality that when the design pressure is reached the pressure should
be kept for a certain time to assure that the desired penetration length is reached. One
understands that even if the pressure is lower than the design pressure, the penetration is still
developing. A large wide fracture will have a large grout take and with a low capacity of the
pump the pressure is low or/and will take time to reach a certain design pressure. This also
affects one common stop criteria which is the stop pressure. It means that when the certain
stop pressure is reached the grouting can end. For large fracture apertures and/or low
capacity of the pumps means that it will take significant time to reach the this stop pressure.
This paper presents a new formulation where this pressure build-up is incorporated into
the earlier known equations. It is the authors intention to show that for different situations
with low and high flowing boreholes, the pump capacity has a large effect on the spent
grouting time. The results could be used to choose a suitable pump and a more efficient way
to set the stop criterias for the grouting.

THEORY AND METHOD

Basic considerations
One way of design grouting work is based on the penetration length of grout in fractures. The
main point is that the maximum grout spread depends on the applied pressure, p [Pa],
fracture aperture, b [m] and the yield stress of the cement grout, 0 [Pa] (Lombardi, 1985,
Hssler, 1991):
p b
I max,cement =
2 0
Recently Gustafson and Claesson developed the expressions further to also incorporate the
grouting time. The formulation was utilized by Stille and Gustafson, 2005, where it is
suggested on how to use the formulation in design works for grouting.
The expressions of penetration length consider the pressure applied as static pressure
that is acting constantly through the whole grouting process. Depending on the efficiency of
the pump (capacity) and the fracture aperture, the grouting pressure take some time to be built
up.
The conceptual model for fractures in hard rock is that the fractures are seen as
individual features, the hydraulic properties are statistically independent and the flow is radial
(2-D). The total transmissivity is equal to the sum of the individual transmissivities (Fransson,
2002). The link between the out-flow from a borehole to a certain fracture aperture is done
using the cubic law.
Water bearing fractures and boreholes and the pressure build up
Large fracture apertures and the capacity of the grouting pumps has a great influence on the
time spent for grouting. The basic consideration suggests that for one borehole it is likely that
there are many fractures crossing the borehole, and that the sum of these fractures yields a
large aperture that has effects on the grouting time. A simple sketch of the problem is shown
in Figure 1.

Figure 1. Conceptual model for fractures crossing a borehole and how this can be
interpreted as a single fracture with a larger aperture.

New mathematical formulation

The differential equation for the radial Bingham flow in a crack for any grout flux Q may be
formulated in dimensionless form using a grout flux Qb and a pressure pb.

Here, rb is the borehole radius, b the fracture aperture, g the yield stress of grout, and g the
viscosity of grout.

The pressure p(r,Q) is obtained from the general dimensionless solution by:

The dimensionless solution , which is zero at the injection radius r=1, is given by
explicit formulas for two functions. The details are given in Claesson, 2016. The pressure
solution p(r, Q) for any Q is then quite easy to implement in any computer code.

The prescribed injection pressure p(Q) for any particular Q shall be equal to the pressure
drop from the injection radius r = rb to the radius of the grout front r = rb + I :

Combining the two above equations, a relation between I and Q is obtained:

The front position I as a function of grout flux Q is readily obtained from a root solver in
any mathematical computer program. The motion of the grout front I(t) is then obtained from
the mass balance of grout at the grout front as described for example in (Lombardi, 1985, and
Hssler, 1991).
 The function p(Q) describes the pump characteristic. In this paper a simple pump-
curve is used. An example of how the pump characteristic is modelled in the new
mathematical formulation is shown in Figure 2.

Figure 2. An example of a pump characteristic. The line is the limiting values of flows
and pressures. When the pump works with the lower flows (0-15 l/min) it can pump with
maximum pressure (30 bar). For larger flows the pressure gradually decreases. For the
largest flows (>30 l/min) the maximum pressure it can pump is 18 bar.

A commonly used pump for grouting is the piston pump. In short, it works using the law of
mass conservation. A large volume of grout in the cylinder is pushed by the piston through a
smaller hole in the cylinder which raises the pressure. The velocity and number of the strokes
of the piston is adjusted to fit the desired pressure. The faster the strokes, the larger is the flow
and a higher pressure can be reached.
The pressure in the system pump-borehole-fracture is governed by the withstanding
forces (water pressure and friction forces build up in the grout). Since the size of the hoses,
packers and borehole is much larger than the fracture aperture, the pressure is constant up to
the fracture. In the grout front the pressure is equal to the withstanding water pressure, or
equal to zero overpressure. While the penetration goes on the resistance of moving increases
and the pressure in the system is also increased. If flow is low and the fractures are wide, then
the pressure will not come in balance until a long penetration is reached.

RESULTS

The pressure build-up in different fracture apertures.

The above method involving the general dimensionless solution is incorporated in a newly
developed grouting simulator owned by Edvirt AB. The simulator is based on the theories
of grout spread published in scientific papers such as Gustafson and Stille (2005) and
Funehag and Gustafson, (2008). In figure 3 below it is shown how pressure is built up to the
desired 3 MPa (30 bar) dependent on the fracture aperture. The pump capacity curve, as
defined in Figure 2 is determined by QA1, PA1= 15 l/min, 30 bar and QA2, PA2= 30 l/min, 18
bar.
Figure 3. Four pump curves that show pressure (in bar, 10 bars= 1 MPa) and flow
against time. The pressure is the top line (black) and the flow is the lower one (green).
The top left picture is for a total aperture of 250 m. The top right figure is for the
aperture of 300 m and the bottom left is 320 m and the bottom right is 350 m.

For the fracture aperture 250 m in Figure 3 the pressure of 3 MPa is reached almost
instantly. For the case of 300 m the pressure build-up takes around 2,5 minutes to reach the
3 MPa. For 320 m, it takes 10 minutes, and for 350 m in fracture aperture the pressure is
only reached to 2,5 MPa after 10 minutes of grouting, but still rises. One should keep in mind
that the simulations is an ongoing research/development.
For the fracture aperture of 350 m and a desired pressure of 3 MPa, the capacity of the
pump needs to be increased to QA1=20 l/min and QA2=50 l/min, respectively (Figure 4).

Figure 4. Pump curve for fracture aperture of 350 m. The pump capacity is here
increased to 20 and 50 l/min respectively.
DISCUSSION

That the pump capacity plays a large role for the pressure build-up in the whole system is
shown Figure 3 and 4. At a slow pressure build-up it is suggested that the penetration shall
move slower. In figure 5 it shown how the penetration is affected for the 350 m fracture
aperture where the pressure has not reached the desired 3 MPa. The capacity is hence the
lower ones previously used (30 and 15 l/min respectively). The penetration is shown with
pressure build-up compared with a constant pressure of 30 bar.

Figure 5. An example showing how the penetration length over time in a fracture
aperture of 350 m is affected by the pressure build up in the borehole compared with a
constant pressure. The bottom green line is the penetration with pressure build-up and
the top line is the penetration with a constant pressure of 30 bar.

One can reason that the penetration is not remarkably affected for a pumping that has a slow
pressure build-up compared with one having a constant pressure. In the example it is only
around 10% shorter penetration length for the penetration having a slow pressure build-up.
The case compared is the fracture having an aperture of 350 m where the pressure only
reached 25 bar after 10 minutes of grouting (compare with Figure 4). This means that the
difference in pressure is 0,5 MPa for the largest part of the penetration. A lowered constant
pressure of 0,5 MPa affects the penetration of 10%. The computed pressure build-up has
hence little effect on the penetration length. Note that the pressure increases rather steeply and
the time where the pressure is lower than the desired 3 MPa is short. The mathematical
formulation assumes that the borehole is filled with grout instantly which in reality is not the
case.
To summarise, the time it takes to reach design pressures for different fracture apertures
is shown in Figure 6. For the smallest pump capacity (QA1 and QA2, 10 and 20 l/min
respectively) and a design pressure of 3 MPa the pressure build-up is highly affected already
at a fracture aperture of 250 m. For medium capacity (QA1 and QA2, 20 and 50 l/min
respectively) and a design pressure of 3 MPa the pressure build up takes 10 minutes for a
fracture aperture of 350 m. For the most powerful pump (QA1 and QA2, 40 and 80 l/min
respectively) the fracture aperture needs to be larger than 450 m of the pressure build-up
should take 10 minutes or longer.
Figure 6. Time for pressure build up for different fracture apertures and pump
capacities. The design pressure is 30 bar for all cases.

Choice of pump
Pre-grouting (permeation grouting) of hard crystalline rock usually involves drilling 20-25
long boreholes ahead of the tunnel front in fan-like formation. The amount of water
transmitted by fractures can vary several magnitudes, from completely dry holes to very large
flows, an example is shown in figure 7. In the figure the flows are also converted to hydraulic
apertures, using the specific capacity and the cubic law, for one single fracture crossing the
borehole.

Figure 7. Example of inflows (dots) in boreholes in a rock mass. The flows are rather
well described by a log-normal distribution (line). For the inflows, a corresponding
hydraulic aperture at a groundwater pressure of 0,1 MPa (10 meter water head) is
computed and illustrated.
Knowing the water flows and hydraulic apertures enables to choose a pump suitable for the
job. 80% of the boreholes have smaller apertures than 265 m. One choice of pump could be
a pump that reach a pressure build-up for fractures up to 265 m, and then accept longer
pumping times for the larger fractures. This is valid for when the criteria on the grouting
works are stipulated by a stop pressure. Other criteria can be used, for instance volume or a
combination of pressure and time. Having an oversized pump for the smaller apertures would
result in large variations of pressure and would not be cost-effective.
The design methodology described in Gustafson and Stille (2005) suggests that the
penetration length is calculated for the minimum fracture aperture required to cope with
stipulations on ingress of water. If the pressure build-up affects the penetration length the
more realistic pressure could be chosen in the design or to cope with observations of pressure
build during grouting by adjusting the grouting time. Until a refined formulation is valid
where more than one fracture is grouted. Table 1 can give guidance on how the grouting time
can be changed to reach the same penetration length.

Table 1. Suggestion on how the grouting time can be adjusted when the design pressure
is not reached.
Over pressure, p Grouting time
Original design 3 MPa 15 min
Pressure reached 2,5 MPa 15+5 min
Pressure reached 2 MPa 15+10 min
Pressure is hardly reached <2 MPa in 10 min Make stiffer grout

The content of Table 1 is that if pressure build up is not reached within 5-10 minutes the
grouting time should be increased to reach the same penetration length. Typically, for every
lowering of pressure of 0,5 MPa, the grouting time should be increased by 5 minutes. If the
reached pressure is far from the designed the grout properties should be changed to stiffer
ones to avoid unnecessary spill and loss of grout.

CONCLUSIONS

The new formulation of grout spread is incorporated with the pressure build up based on the
chosen pump characteristics. This is a step for theories around grouting to close in on more
practical issues like choice of pump and observations done at the grout pump. When the
design of grouting does not fully work on site the new formulation can bring clarity to
problems faced. Further, estimates of the real time spent on grouting can be more accurate
when the pressure build-up is incorporated.
The calculations presented are based on radial Bingham flow in one plane parallel plate.
With the new formulation the time to reach design pressure can be estimated if the fracture
aperture is known. This is a step forward in modelling more realistic grouting scenarios.
However, the pump characteristics modelled in this paper needs further development and a
calibration to real pump curves is planned.
It is shown that the penetration length in one fracture is not affected to a large extent
when the pressure build-up takes approximately 10 minutes. The frame work for multiple
fractures remains to be developed. Experience shows that when grouting a borehole, it can
take some time to reach the designed pressure. One effect is the capacity. Another is that the
fractures do not behave with radial flow, and can even be three dimensional. A three
dimensional network requires large volumes of grout to reach a pressure balance and can be
an explanation of long pressure build-up times. The results show that for large fracture
apertures, > 250 m is where the pressure build-up time is significant. In shallow
infrastructural tunnels in Scandinavia where the grouting works have been followed for
research purposes, large hydraulic apertures like 250 m have seldom been reported. This
requires that the relation between flow dimension-fracture aperture-pump capacity needs to be
analysed further.

ACKNOWLEDGEMENT

The financial support from the Swedish Rock Foundation (BeFo) is highly appreciated. The
simulations of the pressure build up were done using the grouting simulator owned by
Edvirt AB.

REFERENCES
Claesson, J. (2016). Radial Bingham Flow. Revised mathematical background report,
Chalmers University of Technology, Department of Civil- and Environmental
Engineering. In prep.
Fransson, . (2002). Characterisation of Fractured Rock for Grouting Using Hydro-
geological Methods. Doktorsavhandling, Chalmers tekniska hgskola, Gteborg.
Funehag, J. and Gustafson, G. (2008). Design of grouting with silica sol in hard rock - New
methods for calculation of penetration length, Part I. Tunneling and Underground
Space Technology, ISSN 0886-7798, Volume: 23.
Gustafson, G. and Stille, H. (2005). Stop criteria for cement grouting, Felsbau, vol. 23, no.
3, pp. 6268, 2005.
Gustafson, G. Claesson, J. and Fransson, . (2013). Steering parameters for rock grouting.
Journal of Applied Mathematics. Volume 2013 (2013), Article ID 269594, 9 pages.
Hssler, L. (1991). Grouting of rocksimulation and classification [Thesis]. Department of
Soil and Rock Mechanics, KTH, Stockholm, Sweden.
Lombardi, G. (1985). The Role of Cohesion in Cement Grouting of Rock. Proceedings of
the 15th International Commission on Large Dams (ICOLD) Congress, Lausanne,
Switzerland, June.
59
Practical Aspects of Water Pressure Testing for Rock
Grouting

Authors:
Paisley, Adam
Wullenwaber, Jesse
Bruce, Donald
Presentation Preference: Oral
Track: Track 2/ Drilling and Grouting
Session: 2A31/ Design and Assessment, Part I
Is this a student presentation?: None Specified
Bio: None Specified
Submission type: Technical Papers (6-10 pages in length)
Date Uploaded: 1/24/2017 3:35:31 PM
____________________________________________________________

Project Manager Submission separator pages

 Practical Aspects of Water Pressure Testing for Rock Grouting

Adam Paisley, P.E.1, Jesse Wullenwaber, P.E.2, and

Donald A. Bruce, Ph.D., C.Eng.3
1
Schnabel Engineering, Inc., 11 Oak Branch Drive, Suite A, Greensboro, NC 27407;
Email: apaisley@schnabel-eng.com
2
Schnabel Engineering, Inc., 1380 Wilmington Pike, Suite 100, West Chester, PA 19380;
Email: jwullenwaber@schnabel-eng.com
3
Geosystems, L.P., P.O. Box 237, Venetia, PA 15367;
Email: dabruce@geosystemsbruce.com

ABSTRACT

Rock mass permeability testing has been performed to characterize rock formations since the
early 20th century. It was further popularized by Maurice Lugeon (1933) after he defined a
standard unit for quantifying the transmissive ability of rock discontinuities. As the dam and
grouting industries developed and flourished in the latter half of the 20th century, more attention
was devoted to better understanding water pressure testing and its application to rock grouting
projects. Today, water pressure testing is a well-established practice and is one of many useful
tools for characterizing rock formations. However, it seems the benefits of water testing may not
be fully understood from a rock grouting perspective. In several recent grouting designs, the
authors have observed that water pressure testing has been incorporated for the sole purpose of
measuring the permeability of the untreated formation. Such a limited use of the test and results
inhibits important evaluation of project quality and production efficiency. In the last two
decades, the grouting industry has seemingly dedicated more effort toward technical
interpretation of the water pressure test while not promoting its many practical and technical
benefits. In this paper, the authors expound on numerous uses for water pressure testing which
add to the overall value of a rock grouting program and advocate for increasing quality and
production efficiency. To maintain such progress, geoprofessionals in these industries should
maximize the application of available tools and continue promoting innovation in this dynamic
field of work.

INTRODUCTION

The bulk of rock fissure grouting is conducted for seepage control, as opposed to some form of
mechanical improvement of the rock mass. Therefore, it is natural and correct that the main
acceptance criterion for a grouted cutoff should be some type of permeability test (to quantify the
residual permeability of the treated rock mass). This is commonly known as a water pressure

1
test. Such tests also have equal value in characterizing the site before construction begins and to
monitor and control the intensity of the work during its implementation. Prior to the work of
Maurice Lugeon, as described below, the intensity of grouting programs was principally dictated
by an analysis of grout (not water takes), or by the available budget. Neither path is acceptable.
However, the authors have begun to note a lack of awareness of the true needs for systematic
water pressure testing in certain quarters, and seek to reaffirm its essential nature in this paper.

WHAT IS WATER PRESSURE TESTING?

Water pressure testing in rock consists of pressurized injection of water into boreholes drilled
into the rock mass and recording the measured water flow rate under the applied pressure(s).
The borehole is discretized into intervals, or stages, which are isolated through the use of single
or double packers. Water pressure testing is typically performed in an upstage manner in a
hole that has been drilled to full design depth. In such cases, the bottom stage may be tested
using a single packer, and the stages above the bottom stage are tested using a double packer
assembly which consists of two packers that are connected by a central perforated pipe.

RELATION TO CEMENT GROUTING

The grouting industry generally consents that water pressure testing data and pressure grouting
data are generally not able to be directly correlated. The authors disagree, however, and think
that there is much that can be gained from a qualitative comparison of water testing and grouting
data. This paper describes multiple uses for water pressure testing results during the course of
rock grouting projects, including comparison with rock grouting results. The authors consider it
imperative to note that the qualitative comparison between water pressure testing and
cementitious rock grouting results relies largely on consistency of grout properties (e.g. density,
viscosity, etc.) during the grout injection process. Therefore, cement grouts mentioned in this
paper refer to stable grouts, also known as high mobility grouts (HMG), which have a superior
resistance to bleed and pressure filtration.

PERMEABILITY AND THE LUGEON VALUE

The measured data and observations recorded during water pressure tests are used to calculate
the permeability of the stage. It is important to qualify and clarify the use of the term
permeability. The terms permeability and hydraulic conductivity are more accurately
applied to soils which typically contain a relatively regular network of pore spaces which allows
fluid to be transmitted uniformly through the soil mass. As noted by Quiones-Rozo (2010), a
rock mass transmits seepage through discrete discontinuities. Therefore, it is more accurate to
consider that the data collected during water pressure testing reflect the ability of the
discontinuities in a rock mass to transmit water. While the authors note this distinction as an

2
important concept to be recognized, the term permeability is used herein to refer to the average
transmissive ability of rock mass discontinuities.

Maurice Lugeon (1933) developed a method for quantifying the permeability of rock based on a
water pressure test performed in a discrete interval of a borehole. The method produced a
permeability unit which became known as the Lugeon. The Lugeon (Lu) is defined as 1 liter per
minute (L/min) of water flow into a water test stage with a length of 1 meter (m) under an excess
pressure of 10 bars. This may be more clearly understood as follows:

1 /
1 = 10

Weaver and Bruce (2007) noted that the method was originally devised by Lugeon for measuring
water well inflow and that the applied pressure of 10 bars (about 145 psi) initially selected to
mimic heads created by the Alpine dams of the day is excessive for shallower grouting projects.
Adjustments for applied pressure (P), fluid flow (Q), and stage length (LS), which may differ
from the unit values in Lugeons method, are accounted for in the following equation for the
Modified Lugeon (LuMod):

/ /
= 1000 = 1803.2

It is important to note that the Lu and LuMod units only apply when injecting water. When a fluid
other than water is injected, LuMod must be adjusted to account for the difference in apparent
viscosity. In rock grouting practice, the apparent viscosity of stable, non-sanded grouts is
commonly represented by the marsh viscosity (vmarsh), which is defined as the amount of time, in
seconds, for 1 quart of a fluid to flow from a filled marsh cone. The marsh viscosity of water is
26 seconds. A LuMod value that is calculated when injecting a stable fluid other than water is
referred to as an Apparent Lugeon (LuApp) value and is calculated as:

,
=
26
/ , / ,
= 38.5 = 69.3

The LuApp equation is rearranged as such to emphasize that the calculated value is exclusively
based on the flow:pressure ratio when a single fluid is injected into a fixed stage.

Since LuMod is intended to represent the in-situ permeability at the depth of the stage, the
pressure value used in the LuMod calculation should represent the actual pressure applied in the
stage. In practice, the term effective pressure is used to describe the calculated net pressure

3
applied to the stage and it should be used in the Lugeon equation. Effective pressure (Peff) is a
function of the gauge pressure (Pgauge) applied by a pump as measured at surface, the net
hydrostatic pressure (Phydro) due to weight of fluid in packer pipe above the stage and the pore
pressure in the formation around the stage, and pressure loss, or line loss (PLL), in the packer
pipe due to friction, and is calculated as follows:

= +

Readers are encouraged to consider that the magnitude of the Lugeon value should not be
regarded in the same manner on all projects. As shown above, the Lugeon becomes dependent
on the relationship between pressure and flow if the fluid viscosity and stage length are fixed.
Using this convention to define maximum effective pressure for a deeper project as for a shallow
project, a relatively high flow (e.g. 10 gallons per minute) will produce a considerably lower
Lugeon at deeper depths as compared to the Lugeon produced from the same flow at shallower
depths. Therefore, testing data should be evaluated with respect to project-specific constraints.

The length of the water pressure test stage can also have a misleading effect on the Lugeon
value. The US Bureau of Reclamation (2001) noted that the measured flow during a water test
may be transmitted through the rock mass by only a few discontinuities, and that the resulting
calculated Lugeon may not accurately represent the permeability of the formation. Designers
may want to specify the option to perform tests with reduced stage length in boreholes where
drilling and testing data suggest the presence of large discontinuities.

WATER TESTING IN PRACTICE

Professionals in the rock grouting industry have endeavored in recent decades to converge in
opinion on historically variable water pressure testing conventions. The USACE (2014)
presented a comprehensive assembly of commonly accepted practices, but the authors have
noted that such practices may not be widely recognized and that reiteration is warranted. Several
notable concepts are discussed below.

Stepped Tests vs Single-Pressure Tests

As explained by Houlsby (1990), the behavior of fractures in a borehole interval can be

interpreted from a test where a series of effective pressures are applied to the borehole interval.
The LuMod from each pressure step in the series can collectively suggest the relative number and
size of the fractures, the suitability of the maximum design pressure, and the tendency for infilled
particles in the fracture to be dislodged and carried by the pressurized flow. An alternative to the
stepped test is the single-pressure test, where one pressure is applied for a brief period of time.
Single-pressure tests typically provide time and cost savings, but further interpretation of the test

4
results is limited. Weaver and Bruce (2007) recommended that every stage should be tested,
regardless of the type of test. The authors encourage designers to specify a combination of
stepped and single-pressure tests for production holes based on project constraints, and that
stepped pressure tests be specified as a minimum for Primary and Verification holes.

Consistency Between Water Testing and Grouting

The ability to identify trends between water test results and grout take relies greatly on
consistency between equipment and methods used to perform water testing and grouting. The
same types and sizes of equipment must be used to allow work crews to concentrate more on
performing the work and less on using the correct equipment for each respective activity. The
stage length and applied effective pressure range are two elements of methodology that have a
greater effect on the numerical results and the response of the formation. Though effective
pressure applied during water testing and grouting will vary with time during each respective
activity, the data collected during water testing provides grouters with insight about how the
formation responds to the range of specified pressures. Additionally, trends between water
testing formation response and grout performance, which are often identified after the
completion of many holes, allow grouters to make timely decisions about switching grout mixes.
Using different pressures for water testing as compared to grouting hinders the ability to identify
such trends and thus reduces the efficiency of the grouting program. The identification of trends
and differences between water and grout performance cannot be accomplished unless the
response of a discrete group of fractures to water testing and grouting can be evaluated. The
USACE (2014) recommended that stage intervals selected for water testing be identical in length
and position in a borehole to subsequent grout stages, and the authors concur with the
recommendation.

Water Testing in Weak Rock or Near Soil/Rock Interface

Weak, weathered rock and materials encountered near the soil/rock interface can be
unconsolidated and sensitive to low amounts of stress. Additionally, the susceptibility to
deformation of these types of materials tends to increase when a pressure gradient is applied. If
efficient reduction of permeability is the goal of a rock grouting project, then water testing in the
vicinity of such materials should be approached with extreme caution. This can be a
predicament in some cases where the upper portion of rock has a higher permeability and poses a
significant challenge for seepage control. The USACE (2014) advocated for more conservative
testing in weak zones, such as shorter stage lengths, reduced pressures, or elimination of water
testing altogether (specifically near the soil/rock interface). If the design team determines that
water testing of the upper portion of rock is not to be performed, the grouting permeability of and
closure through the upper rock zone may still be evaluated based on LuApp and grout take.

5
Communication

Real time monitoring has become the standard for larger grouting projects over the last two
decades. However, communication between experienced grouters observing the real time
monitoring and experienced field personnel at the hole remains vital. A sophisticated program
cannot replace experienced personnel or the ability to communicate observations at the subject
hole. Clear communication, whether about a leaking packer, faulty pressure gage, water flowing
from surface or adjacent holes, or any other issue, provides clarity to the task at hand and
promotes overall project quality and efficiency.

Advancements in Technology Downhole Pressure Transducers

On certain recent projects, downhole pressure transducers have been required to measure the
pressure in the stage. Downhole pressure is collected in real time and displayed by the computer
monitoring system. Based on recent results, the cost of outfitting a packer with a downhole
pressure transducer is easily outweighed by the benefit of being able to see what is happening
in the stage being tested or grouted. Since the variables used to calculate effective pressure are
based on the testing system (i.e. filled hoses, valves, gages, etc.), the downhole transducer allows
grouters to recognize flaws in the system upstream of the stage via comparison of the calculated
effective pressure against the measured downhole pressure. The downhole pressure record
additionally allows for communication of a more coherent explanation to designers or review
panels not present at the time of testing. The authors recommend use of downhole pressure
transducers and the measured downhole pressure to calculate Lugeon values on larger projects.

APPLICATIONS OF WATER PRESSURE TESTING DATA

The effectiveness of grouting designs and the efficiency of the grouting process are improved
through pragmatic use of water testing data and participation of experienced grouting
professionals. Accordingly, the authors recommend that designers and project owners employ
the expertise of experienced grouters during each phase of a grouting project. The following
paragraphs present applications of water pressure testing results that can add value to rock
grouting projects.

Site Exploration

Drilling, water pressure testing, televiewing, and grouting processes implemented during the
course of a grouting project provide a means to characterize the rock formation. Subsurface
explorations are typically performed during the design phase of rock grouting projects, but the
scale of the exploration scope is often limited by economics. As a result, exploratory borings are
often widely spaced and designers are forced to make conservative design assumptions. Ewert
(1985) recommended designers consider that local differences in permeability can affect a large

6
area. Ewert further explained that the macro-scale permeability of a rock mass can differ greatly
from the conductivity measured in a borehole. As a rock grouting project progresses, designers
and contractors are afforded the opportunity to confirm design assumptions, to further
characterize the rock formation, and to recognize various subsurface conditions. Enhanced
understanding of subsurface conditions is facilitated, in part, by an increased number of borings,
a decrease in distance between the borings, and water pressure testing procedures that promote
recognition and interpretation of the various manners in which rock discontinuities respond to
pressurized seepage flow.

Evaluation of the Work

The process of rock grouting consists of a continuous process of modification and evaluation.
Typically, an initial grout hole series (i.e. primary) is drilled, water tested, and grouted prior to
installing subsequent hole series (i.e. secondary, tertiary, etc.). Water testing is performed in the
initial hole series to measure unmodified permeability, in part, as a baseline for comparison to
future water test data. As each hole series is completed, the water testing data is evaluated to
confirm design assumptions and evaluate the effectiveness of the materials, equipment, and
methods which are used to perform the grouting work. In this way, water testing allows grouters
to control the quality and value of the work during the course of the project.

Planning During Grouting Construction

The performance of water testing as part of a rock grouting program can provide grouters the
opportunity to identify conditions that will affect the progress and nature of the work. Such
conditions can indicate potentially large grout takes, connections between grout holes, challenges
sealing stages due to leaks around the packer, and surficial leaks (when rock is near surface).
Grouters must appropriately plan an approach to these types of situations and coordinate efforts
accordingly to maintain the quality of the work, the progression of the project schedule, and the
safety of construction staff. Through use of water testing data, grouters can plan to have
sufficient materials for completing stages with a large LuMod and to begin grouting such stages
early in the day. Appropriate planning for large grout takes also promotes safety by reducing the
chances of working past normal working hours. Sections of a grout row where multiple holes
have produced low LuMod can be grouted quickly to more efficiently progress the program
sequence. Lastly, water test results can aid in avoidance of partially-grouted holes. When
grouting in a hole is paused at the end of a scheduled work interval, such as a daily shift, the hole
can become partially contaminated before grouting resumes at a later date. Such contamination
could inhibit grout take. Common causes include emptying of the grout line as it is removed
from the hole and grout rebound into the hole as injection pressures dissipate in the formation.
The authors recommend that water testing data be used to avoid any situation where a portion of
a grout hole could become compromised.

7
Amenability

The suitability of a stable particulate grout to move into fractures that are accessible to water is
termed amenability. The LuMod and LuApp values calculated during water testing and grouting,
respectively, are related to the aperture size of fractures in the subject stage. Amenability can be
quantified by comparing LuApp from grouting to LuMod from antecedent water testing. Naudts
(1995) presented the amenability coefficient (Ac) as a measure to evaluate the efficiency of the
fractures to accept grout.

Note that a reasonably accurate Ac can only be obtained if the LuApp and LuMod are produced
from the same borehole interval (i.e., same stage length and location in borehole). Additionally,
the LuApp value used for calculating Ac must be selected early during the grouting of a stage
when the fractures are as unmodified as possible.

The maximum particle size and the viscosity of a particulate grout affect the ability of the grout
to penetrate an aperture. Evaluation of amenability allows grouters to identify the need to
modify viscosity or change one or more components in the grout. For example, a coarse cement
may be selected during design and evaluation of amenability may lead to use of a finer cement
or, in certain cases, a non-particulate grout. Amenability should be constantly evaluated
throughout a grouting program, particularly in double or triple row grout curtains, because finer
fractures tend to play a larger role as the project progresses. Naudts (1995) recommended a
minimum Ac value of about 0.75. The authors note that amenability should only be evaluated for
stages that intersect apertures that are large enough to accept typical cement grouts but are small
enough to present a challenge to grout injection. Weaver and Bruce (2007) noted that stages
producing LuMod values of less than 10 are typically less likely to accept cement grouts and
stages which produce LuMod values of 100 or greater will almost certainly accept cement grouts
readily. While these LuMod values may vary from project to project, readers should be aware that
apertures of a more intermediate size are more likely to create an amenability issue.

A Case Study in Amenability Rio Verde Dam

Rio Verde Dam is a small concrete gravity structure used to store drinking water in south central
Kentucky, United States. A two-row grout curtain was installed in the dams right abutment in
late 2013 in response to reported vortices upstream of the dams right abutment and seepage
flows exiting the right stream bank downstream of the dam. A suite of Type III cement stable
grouts were used to construct the curtain, and the thinnest grout mix had a marsh viscosity of 40
seconds. Initial Primary (PP) holes were installed 20 feet apart, Primary (P) holes were installed
between PP holes, and Secondary (S) holes were installed between PP and P holes. Each hole

8
generation was drilled, water tested, and grouted before adjacent hole generations were installed,
and the downstream row was grouted to closure prior to the grouting of the upstream row. PP, P,
and S holes were required on the downstream row, and PP and P holes were required on the
upstream row. Additional holes were installed based on the results of water testing and grouting.
A few indications of poor amenability issues were observed during the grouting of the
downstream row. However, no changes to the mixes were made because observed LuMod values
and grout takes appeared to be more influenced by very small or very large apertures and
because a general decreasing trend in LuMod was observed. As the grouting of the upstream row
commenced, water testing and grouting data suggested that some intermediately-sized apertures
remained open and that the thinnest grout mix may not be able to penetrate such apertures.
Figure 1 shows a comparison between calculated LuMod and LuApp values for stages where water
take was observed.

Type III cement

Figure 1. Comparison of Water Test and Grouting Lugeon Values at Rio Verde Dam.

9
The amenability of two upstream PP stages with LuMod values over 100 was not concerning due
to large respective grout takes when using the Type III cement grouts. However, an impeding
amenability issue became apparent when one upstream PP stage and the first upstream P stage
produced LuMod of 91 and 73, respectively, and subsequently produced respective LuApp of 1 and
9 when using the Type III cement grouts. The project team discussed the observed discrepancies
between water testing and grouting and decided that the use of a microfine grout was warranted
to improve the quality of the work. With agreement from the project owner, a microfine grout
was developed with a marsh viscosity of 32 seconds and was used to grout the remaining
upstream holes. The upstream P and S stages shown near the perfect amenability line in Figure 1
were grouted immediately following the change to microfine grout. The data collected in the
remaining upstream P and S stages and verification holes indicated that the residual LuMod could
be reduced to no less than 20-30 Lu using the microfine grout. Exposed excavations around the
site revealed localized pockets of silty sand which exhibited the relict structure of the non-
soluble limestone formation upon which the dam was founded. The team concluded that similar
silty sand pockets, which only non-particulate fluids may be able to penetrate, may exist in the
right abutment. Despite higher localized residual LuMod values, the previously observed seepage
on the downstream right bank was no longer visible, no further vortices have since been
reported, and the project owner has been satisfied with the results.

REDUCING ERRORS IN WATER TESTING DATA

Water testing procedures, equipment, and input variables affect the quality of water testing
results. Errors in testing data can be reduced by experience observing real time data collection
and by having an operational knowledge of the equipment being used and the engineering
principles behind the test. Several scenarios that may contribute to error are discussed below.

Trapped air in the grouting/testing hoses can produce errors in instruments such as flow meters
and pressure transducers. Air exists in the system initially and can enter the system when the
packer is not submerged or if a leak occurs in the equipment. Water can be pumped through the
system to purge air before testing. At a minimum, the theoretical volume of the hoses should be
pumped. During testing, the measured downhole pressure can indicate if the packer line is full.

Water can flow around the packer and up the borehole if the packer is inflated in a fractured
zone. This can be identified by water flow from the hole or a rising water level. The packer can
be moved up or down in the borehole until a good seal is achieved.

Packers are pressurized through inflation lines that are connected to the bladders. Leaks at the
connection or in the line can result in partial inflation. This scenario can be identified by rising
water level or multiple consecutive large water takes (if bottom packer in double packer
assembly is partially inflated). However, a clear indication is an observed decrease in packer
pressure when shutting off the pressure source.

10
As stated previously, effective pressure is a function of the pore pressure in the formation, which
is represented by the water level in the borehole. The amount of time between borehole drilling
(using water) and water testing may not allow for the water level to settle to static level. This is
a common source of error in water testing and grouting, and is often not avoidable without
disrupting schedule. A reasonable representation of water level between holes can be developed
from a series of overnight water levels collected from multiple holes along a grout line.
However, ground water in the vicinity of a grout curtain should rise in elevation as the program
progresses, and as such water levels will need to be collected through the course of the program.

The application of very low pressures requires that pumping equipment operate in the lowest
range of design capacity. Additionally, instruments that measure pressure and flow may not
function as well near the lower boundary of their calibrated range. Thus, data collected at very
low pressures may be obscured by the mechanical limitations of testing equipment.

Higher LuMod values calculated from relatively high flows may be misinterpreted if the formation
is able to accept more water than can be supplied by the pump. In this case the resulting LuMod
value may represent pump capacity instead of formation permeability.

CLOSING REMARKS

Water testing is routinely performed as part of rock grouting programs, but may be underutilized.
The quality and efficiency of a grouting program can benefit from appropriately specified and
performed water testing, especially when performed consistently with grouting. As the grouting
industry evolves, grouting professionals should take advantage of advancements in technology
and methods while reinforcing the value of experienced personnel. Water testing is an important
component to a rock grouting program and should be given careful consideration for the purpose
of obtaining the most information and value from the test.

11
REFERENCES

Ewert, F.K. (1985). Rock Grouting with Emphasis on Dam Sites. Springer-Verlag, Berlin,
Heidelberg, Germany.
Houlsby, A.C. (1976). Routine Interpretation of the Lugeon Water Test. Quarterly Journal of
Engineering Geology, 9(4), 303-13.
Houlsby, A.C. (1990). Construction and Design of Cement Grouting: A Guide to Grouting in
Rock Foundations, John Wiley & Sons, Inc., New York, NY.
Lugeon, M. (1933). Barrages et Geologie. Dunod, Paris.
Naudts, A. (1995). Grouting to Improve Foundation Soil. Chapter 5B in Practical Foundation
Engineering Handbook. 1st Ed. McGraw Hill, New York.
Quiones-Rozo, C. (2010). Lugeon Test Interpretation, Revisited. Collaborative Management
of Integrated Watersheds, U.S. Society of Dams, 30th Annual Conference, 405414.
USACE (U.S. Army Corps of Engineers) (2014). Grouting Technology. Engineering Manual
1110-2-3506, July 3, 2014.
USBR (U.S. Department of Interior, Bureau of Reclamation) (2001). Engineering Geology
Field Manual. Second Edition.
Weaver, K.D. and D.A. Bruce (2007). Dam Foundation Grouting, Revised and Expanded
Edition, American Society of Civil Engineers, ASCE Press, New York.

12
96
Application of Low-Frequency Rectangular Pressure Impulse in
Rock Grouting

Authors:
Nejad Ghafar, Ali
Sadrizadeh, Sasan
Draganovic, Almir
Johansson, Fredrik
Hkansson, Ulf
Larsson, Stefan
Presentation Preference: Oral
Track: Track 2/ Drilling and Grouting
Session: 2A31/ Design and Assessment, Part I
Is this a student presentation?: None Specified
Bio: None Specified
Submission type: Technical Papers (6-10 pages in length)
Date Uploaded: 2/21/2017 9:27:56 AM
____________________________________________________________

Project Manager Submission separator pages

 Application of Low-Frequency Rectangular Pressure Impulse
in Rock Grouting
A. N. Ghafar*a, S. Sadrizadehb,c, A. Draganovica,
F. Johanssona, U. Hkanssona, S. Larssona

a
Division of Soil and Rock Mechanics, School of Architecture and the Built Environment,
KTH Royal Institute of Technology, SE-100 44 Stockholm, Sweden
b
Division of Fluid and Climate Technology, School of Architecture and the Built
Environment, KTH Royal Institute of Technology, SE-100 44 Stockholm, Sweden
c
Danish Building Research Institute, Aalborg University, DK-2450 Copenhagen, Denmark
Email address: alng@kth.se (A. N. Ghafar*)

Abstract
In order to sufficiently seal an underground facility in fractured rock, it is essential to obtain
adequate grout spread into the surrounding fractures. The grout spread itself depends on
parameters, the most significant of which are the filtration tendency and rheological
properties. These properties can be affected by the applied pressure. High-frequency
oscillating pressure has been shown to improve grout spread by virtue of reducing the grout
viscosity. However, this method has not yet been industrialized due to the quick dissipation of
the oscillation along a fracture. In a recent investigation, we examined a low-frequency
rectangular pressure-impulse using a short slot. The results showed significant improvements
in the injected grout volume in comparison to the static pressure results. In this paper, we
examine the method in a considerably longer artificial fracture in order to investigate the
dissipation of the pressure impulses. The study indicates the potential of the method to
improve the grout spread in rock grouting.

Keywords
Grout spread, filtration, erosion, cement-grout, dynamic grouting, rectangular pressure-
impulse, VALS
1. Introduction
In order to provide the sealing required in subsurface infrastructures, located in fractured hard
rock, a critical issue is achieving a sufficient grout spread into the surrounding fracture
network, especially fine fissures (such as those <70 m). This issue is more discernible in
areas that demand very high sealing, such as nuclear and/or toxic chemical waste repositories,
and oil/gas storage caverns (Houlsby 1990; Warner 2004; Fransson 2008; Gustafson et al.
2013; Stille 2015). The required/obtained grout spread depends on numerous parameters: the
grout properties, the geological and hydrogeological properties of rock mass, the objective of
the grouting operation, as well as the environmental, logistical, and economic circumstances
(Hkansson et al. 1992; Schwarz 1997; Eriksson et al. 2000; Banfill 2006). The most
important of these parameters are the grout properties, such as filtration tendency and
rheological properties. In particular the yield stress and apparent viscosity, are the most
influential factors (Lombardi 1985; Eklund and Stille 2008; Ewert 2012). Studies have shown
that these properties in turn can be affected by grouting pressure (Hjertstrm 2001;
Draganovic and Stille 2014; Ghafar et al. 2016a).
1.1 Background
According to Mller and Bruce (2000), the outflow pressures of all contemporary grouting
pumps that can be categorized as either of progressive cavity pumps, piston pumps, or
plunger pumps are more or less associated with some fluctuations. Each of these categories
has its own benefits and drawbacks as summarized in Table 1. The piston and plunger pumps,
and in particular the plunger pumps, have several advantages over the progressive cavity
pumps. However, by virtue of their inherent design, they are associated with partially
controlled fluctuations, whereas the progressive cavity pumps have shown the least
fluctuation in output pressure that can be considered as nearly constant (Bruce 1992; Mller
and Hny 2000).

Table 1. Benefits and drawbacks of three categories of contemporary grouting pumps

(Mller and Bruce 2000)

Pump Costs Pressure Clogging

Wear Pump ability
Type Investment Maintenance Labor range Risk

Progressive
Adjustable pressure
cavity Low Moderate High Limited High Moderate
and flow
pumps
Adjustable pressure
High with and flow, Applicable
Piston Relatively
High Lower Lower Full abrasive for sanded grout
pumps high
grout
(with small particles)
The most versatile
Plunger Extremely pumps, Applicable for
High Lowest Lowest Full Non
pumps low heavily sanded grout
(with large particles)

2
According to Bruce (1992), there has been on-going controversy among experts about
whether the grouting process should be conducted at either constant or dynamic pressure (that
is, whether the fluctuations in output pressure of the pumps can be accommodated). The old
North American grouting philosophy, which recommended the low-pressure injection to
maintain the original structure of the fracture network, preferred application of constant
pressure that can be principally provided by progressive cavity pumps (Weaver and Bruce
2007). However, the new North American and European grouting philosophies,
proposing the high-pressure injection to break through weak formations in order to improve
the penetration into adjacent fractures, encourage application of fluctuated/dynamic pressure
that can be inherently provided by the piston and plunger pumps (Weaver 1991; Bruce and
Dugnani 1994; Bruce 2013).
In progressive cavity pumps, the resulting constant pressure causes arching/bridging of the
cement particles and clusters at constrictions along a fracture (pressure filtration) that might
be eroded by applying stepwise pressure increments (Mller and Bruce 2000; Nobuto et al.
2008). This method (application of stepwise pressure increments) has been acknowledged as a
common practice with a satisfactory control on filtration via eroding partially built plug
formations and unstable filter cakes (Stille 2015). On the contrary in piston and plunger
pumps, the sudden pressure drop that occurs within each pump cycle might somehow be
advantageous in terms of transporting the cement particles and clusters into micro fractures
(Mller and Bruce 2000). The latter suggested that, over the short rest period of each pump
cycle, the cement particles are encouraged to reorient and conform to the constrictions within
the fracture. Further, application of high-frequency oscillating pressure has also been shown
to improve the grout spread in fractures, both in lab and field settings, by reducing the grouts
apparent viscosity (Pusch et al. 1985; Borgesson and Jansson 1990; Wakita et al. 2003;
Mohammed et al. 2015). However, this method has not yet been industrialized due to short
influence length of the oscillation (quick dissipation) along a fracture and limited efficiency.
To increase the potential length of influence of dynamic pressure along a fracture, Ghafar et
al. (2015, 2016a) developed and examined a low-frequency rectangular pressure impulse. The
aim was to improve the grout spread by successive erosion of the produced filter cakes at a
fracture constriction in consecutive cycles. The mechanism of action, as detailed by Ghafar et
al. (2015), was variation in flow pattern due to the change of pressure and thereupon flow
velocity. The results revealed a significant improvement (up to 11 times) in grout spread
volume compared to that in the experiments at static pressure condition using a short slot with
30 and 43 m apertures. This approach can be a new potential with significant improvement
of grout spread in rock grouting, provided that the dissipation length of the pressure impulses
along the fractures can be prolonged.
1.2 Scope
In this paper, we investigate the dissipation of the dynamic pressure impulses along a
considerably longer artificial fracture (with 4 m length and varying aperture of 230-10 m);
this is the so-called varying-aperture long slot (VALS) developed by Ghafar et al. (2016b).
The goal is also to monitor the filtration and erosion processes at two constrictions with 60
and 40 m apertures (at distances of 2.0 and 2.7 m from the slots beginning, respectively) for
a dynamic pressure setup with a maximum pressure of 15 bar and 2 s/2 s peak/rest periods to
examine its effect on grout penetrability. These constrictions were selected due to the fact that
penetrability of our examined grout was determined in terms of bmin and bcrit equal to 40 and
60 m as described in detail by Ghafar et al. (2016b). The values of bmin and bcrit according to
Eriksson and Stille (2003) have been defined as the minimum fracture aperture a grout can
3
penetrate at all and the minimum fracture aperture a grout can penetrate without filtration,
respectively.
2. Materials and Methods
2.1 Test plan, Materials, and Sample Preparation
Here we present the results of two experiments with tap water (20C) and a grout with one of
the most frequently used grout recipes in Swedish grouting industry. Both tests were
conducted at maximum pressure of 15 bar with durations of the peak/rest periods of 2 s/2 s,
due to maximum influence on improving the grout spread at the short slots with 30 and 43 m
apertures illustrated by Ghafar et al. (2016a).
The cement used was Injektering 30 from Sika Sverige AB with d95 of 30 m (that is, at least
95% by weight of the cement particles pass through a sieve of 30 m). A superplasticizer,
Sika iFlow-1 with 0.5 percent of concentration by cement weight, was used to disperse the
cement particles within the suspension with a water-to-cement ratio (w/c) of 0.8.
An electric hand-mixer was employed to premix the materials. Subsequently, a rotor-stator
dispersion system, operated at 10000 rpm, was utilized to mix the materials for a period of
4 min, where the superplasticizer was added after 2 min. In order to prevent any eventual
clogging of the cement particles and clusters, the grout tank was filled with the prepared grout
to run the experiment immediately. The viscosity and yield stress of the prepared grout were
measured (right after the mixing process) 7.5 mPas and 0.05 Pa, respectively using a
Brookfield LV-II+ Programmable rheometer with a SC4-31 spindle. The bleeding of the grout
sample was measured approximately 5% according to Swedish standard (SS 13 75 40) and
the values of bmin and bcrit were evaluated as 40 and 60 m using the so-called varying
aperture long slot (VALS).
2.2 Schematic of Experimental Setup
Fig. 1 presents a schematic of the test setup. It comprises three main components: a nitrogen
gas tank (200 bar), a grout tank with 2.6 l capacity, and the VALS to mimic grouting in hard
rock fractures. A pressure regulator adjusts the gas pressure to 15 bar. The grout tank is
suspended via an S-shaped load cell to register the weight of the injected grout over time. A
pressure transducer is attached to the grout tank to record the variation in tank pressure over
time. Valve Vp is to control (open/close) the pressurized gas inlet into the grout tank. Valves
Vi and Vf are to control the grout inlet and release the gas from the tank after each test,
respectively.
The VALS is equipped with three pressure transducers, P1P3, located before valves V1 and
V3 and at the beginning of the slot, to monitor the pressure variation at the constrictions with
40 and 60 m apertures as well as the entrance of the slot, respectively. In this way,
dissipation of the pressure along the slot and eventual filtration and erosion processes can be
studied. A three-way ball valve is connected to the VALS entrance. The third pressure sensor
P3 reads the pressure at the VALS entrance.
2.3 Experimental Procedure and Evaluation Methods
The grout tank was first filled with the prepared grout or water and pressurized at a constant
pressure of 15 bar. The three-way ball valve was frequently switched between two positions
to either open the grout flow from the tank toward the VALS or to drop the internal pressure
of the VALS down to the atmospheric pressure (Fig. 1). In the given case, the grout was
injected through valve V1.

4
The pressure variations registered by the pressure sensors P1P3 show the pressure impulses
at distances of 0.0, 2.0, and 2.7 m from the slots beginning and the corresponding
dissipations. The max-pressure envelope (that is, a polyline connecting the consecutive
vertices of the maximum pressures of the peak periods throughout the pressure-time
measurement) can also indicate eventual filtration and erosion processes as discussed by
Ghafar et al. (2016a). For example, if both pressure P1 and P2 decrease it indicates filtration
at the 70-m constriction. If both these pressures increase it indicates filtration at the 40-m
constriction. Filtration may occur at all these constrictions at the same time and the
interpretation may be more complex.

Fig. 1. Schematic depiction of the experimental setup: (1) gas tank, (2) pressure
regulator, (3) load cell, (4) grout tank, (5) pressure transducer, (6) DAQ (7) PID-control
unit (8) three-way pneumatic driven ball valve
3. Results and Discussion
3.1 Analysis of the pressure-impulse dissipation
Fig. 2 presents the results of the pressure-impulse dissipation of the water test. The light gray
line represents the initial pressure applied to the system at the slots beginning, with peak-to-
peak amplitude of 14 bar (between 15 and 1 bar pressures). The results, registered by P2
(medium gray line), show a dissipation of approximately 38% in the peak-to-peak amplitude
(reduction in the magnitude of the pressure at peak from 15 to 12.5 bar and increase in the
magnitude of the pressure at rest from 1 to 3.8 bar) after reaching the stabilized condition at
25 s. This dissipation has occurred throughout 2.0 m length of the slot with varying apertures
from 230 to 60 m. The results, registered by P1 (black line), show a dissipation of
approximately 70% in the peak-to-peak amplitude. This dissipation has occurred throughout
5
2.7 m length of the slot, with varying apertures from 230 to 40 m. This clearly indicates that
the remaining amplitude of the pressure impulses after 2.0 and 2.7 m lengths of VALS can be
as large as 60% and 30% of the initial/applied amplitude in water injection, respectively.
Fig. 3 presents the illustrative results of the pressure-impulse dissipation of the grout test. The
light gray line represents the initial pressure applied to the system at the slots beginning. The
pressures, registered by P2 (medium gray line), show a dissipation of 54% in the mean value
of the peak-to-peak amplitude of the pressure-impulses after 2 min of injection. This
dissipation has occurred throughout 2.0 m length of the slot (23060 m). The pressures,
registered by P1 (black line), show a dissipation of 75% in the mean value of the peak-to-peak
amplitude after 3 min of injection. This dissipation has occurred throughout 2.7 m length of
the slot (23040 m). The analysis shows that, after the 2.0 and 2.7 m lengths of the VALS,
the remaining amplitude of the pressure-impulses can be as large as 46% and 25% of the
initial/applied amplitude, respectively.
The results also reveal the potential for reducing the pressure-impulse dissipation by
increasing the durations of the peak and rest periods. This will be investigated in the later
stages of the project using longer peak/rest periods. It must be mentioned that during injection
and switching the three-way valve to control the P3-pressure at the VALS entrance, a
significant amount of grout passed out through the three-way valve. The passed amount of
grout through the VALS was significantly less than 2.6 l but still pressure dissipation and
indications of filtration and erosion could be studied.

P3
15
P2
P1
Pressure [bar]

10

0
0 10 20 30 40 50 60
Time [s]

Fig. 2. Depiction of the pressure-time measurements using P1-P3 in water test with 2s/2s
peak/rest period (P1 and P2 are located before 40 and 60 m constrictions, respectively)

6
 Pressure [bar] 15 P3
P2
10
P1
5
0
0 50 100 150 200 250
Time [s]

15 P3
Pressure [bar]

10 P2
5 P1
0
100 110 120 130 140 150
Time [s]

Fig. 3. Depiction of the pressure-time measurements using pressure sensors P1P3 in

grout test with 2 s/2 s peak/rest period
3.2 Analysis of the max-pressure envelope
Fig. 4 shows the max-pressure envelopes obtained from the pressure-time measurements
using P1 and P2 during the water test. The results show approximately stabilized envelopes
that represent no-filtration/no-erosion conditions after 25 s of injection. This is due to the fact
that water is a one-phase fluid (without any suspended solid particles) and there are no
filtration/erosion phenomena during a water injection.
Fig. 5 depicts the max-pressure envelopes obtained from the pressure-time measurements
using P1 and P2 during the grout test. These max-pressure envelopes can be divided in five
different periods: 0-20s; 20-40s; 40-120s; 120-140s; and 140-290s. The first period (0-20s)
shows the initial filling of the slot. In the second period (20-40s), both pressures decrease
which indicates filtration at the 70-m constriction. This is close to the evaluated bcrit of this
grout. In the third period (40-120s) both pressures increase linearly which indicates filtration
at the 40-m constriction. In the fourth period (120-140s), the pressure fluctuates somewhat
and the interpretation is difficult. In the last period (140-290s) the pressures are relatively
stable. Since the pressures P1 and P2 do not increase to 15 bar during the period it indicates
that the 40-m constriction was still partially open. Furthermore, the value of the max-
pressure envelope registered by P1 in grout test was approximately 10.5 bar, compared to
nearly 8.5 bar for the water test (Figs. 4, 5). This pressure jump (from 8.5 to 10.5 bar) is most
probably caused by the filter cakes produced at the constriction that reduced the slots
opening size.

7
 14
12

Pressure [bar]
10
8
6
4 P2
P1
2
10 15 20 25 30 35 40 45 50 55
Time [s]
Fig. 4. Depiction of the max-pressure envelope using P1 and P2 in water test with 2 s/2 s
peak/rest period

12
10
Pressure [bar]

8
6
4 P2
P1
2
0 50 100 150 200 250
Time [s]
Fig. 5. Depiction of the max-pressure envelopes using P1 and P2 in grout test with 2 s/2 s
peak/rest period
4. Final Remarks
By applying a low-frequency rectangular pressure-impulse with a maximum pressure of 15
bar and 2 s/2 s peak/rest period, the dissipation of the pressure impulses, even after 2.0 and
2.7 m lengths of an artificial fracture with varying apertures (23040 m), was not critical.
The remaining amplitude of the pressure-impulses was sufficient to extend the grout
penetration due to counterbalancing the filtration and erosion of the cement particles at the
constrictions. Even though the experiments conducted in this study were limited, the results
showed the potential of low-frequency rectangular pressure-impulse on improvement of grout
penetrability in fractured hard rock. The study will be continued with further experiments on
grouts with different recipes and pressure setups.
5. Acknowledgments
The authors would like to acknowledge the Rock Engineering Research Foundation (BeFo),
the Swedish Transport Administration (Trafikverket), the Development Fund of the Swedish
Construction Industry (SBUF), and the Transparent Underground Structure (TRUST) project
for their technical and financial support.

8
6. References
Banfill, P. F. G. 2006. Rheology of Fresh Cement and Concrete. Rheology Review 2006.
British Society of Rheology: pp. 61130. doi:10.4324/9780203473290.
Borgesson, L., and L. Jansson. 1990. Grouting of Fractures Using Oscillating Pressure. In
Proceeding of the International Conference on Mechanics of Jointed and Faulted Rock,
pp. 87582. Vienna: A. A. Balkeme, Rotterdam.
Bruce, D. A. 1992. Current Technologies in Ground Treatment and In Situ Reinforcement.
In Proceedings, Second Interagency Symposium on Stabilization of Soils and Other
Materials, pp. 27195. Metairie, LA.
Bruce, D. A. 2013. Contemporary Rock Drilling and Grouting Practices. In New Zealand
Society of Large Dams (NZSOLD) and the Australian National Committee on Large
Dams (ANCOLD) Conference, 12 p. Rotura, New Zealand.
Bruce, D. A., and G. Dugnani. 1994. Innovations in American Grouting Practice. In
Proceeding of International Symposium on Anchoring and Grouting Techniques, 21 p.
Guangzhou, China.
Draganovic, A., and H. Stille. 2014. Filtration of Cement-Based Grouts Measured Using a
Long Slot. Tunnelling and Underground Space Technology 43: pp. 10112.
doi:10.1016/j.tust.2014.04.010.
Eklund, D., and H. Stille. 2008. Penetrability due to Filtration Tendency of Cement-Based
Grouts. Tunnelling and Underground Space Technology 23 (4): pp. 38998.
doi:10.1016/j.tust.2007.06.011.
Eriksson, M, and H Stille. 2003. A Method for Measuring and Evaluating the Penetrability
of Grouts. In The 3rd International Conference on Grouting and Ground Treatment,
132637. New Orleans, Louisiana, United States: ASCE. doi:10.1061/40663(2003)79.
Eriksson, M., H. Stille, and J. Andersson. 2000. Numerical Calculations for Prediction of
Grout Spread with Account for Filtration and Varying Aperture. Tunnelling and
Underground Space Technology 15 (4): pp. 35364. doi:10.1016/S0886-7798(01)00004-
9.
Ewert, F. K. 2012. Rock Grouting: With Emphasis on Dam Sites. Illustrate. Springer Science
& Business Media.
Fransson, . 2008. Grouting Design Based on Characterization of the Fractured Rock,
Presentation and Demonstration of a Methodology (Technical Report R-08-127).
Swedish Nuclear Fuel and Waste Management AB (SKB) Stockholm, Sweden.
Ghafar, A. N., A. Draganovic, and S. Larsson. 2015. An Experimental Study of the Influence
of Dynamic Pressure on Improving Grout Penetrability (BeFo Report 149). Rock
Engineering Research Foundation-BeFo, Stockholm, Sweden.
Ghafar, A. N., A. Mentesidis, A. Draganovic, and S. Larsson. 2016a. An Experimental
Approach to the Development of Dynamic Pressure to Improve Grout Spread. Rock
Mechanics and Rock Engineering. 49: pp. 3709-3721. doi:10.1007/s00603-016-1020-2.
Ghafar, A. N., S. Sadrizadeh, K. Magakis, A. Draganovic, and S. Larsson. 2016b. A New
Laboratory Apparatus, Varying Aperture Long Slot (VALS), for Studying Grout
Penetrability into Fractured Hard Rock. ASTM International - Geotechnical Testing
Journal, (In review).
9
Gustafson, G., J. Claesson, and . Fransson. 2013. Steering Parameters for Rock Grouting.
Journal of Applied Mathematics 2013: pp. 19. doi:10.1155/2013/269594.
Hjertstrm, S. 2001. Microcement-Pentration Versus Particle Size and Time Control. In
Proceeding of the Fourth Nordic Rock Grouting Symposium, SveBeFo Rapport 55,
pp.6171. Stockholm, Sweden.
Houlsby, A. C. 1990. Construction and Design of Cement Grouting: A Guide to Grouting in
Rock Foundations (Vol. 67). John Wiley & Sons.
Hkansson, U., L. Hssler, and H. Stille. 1992. Rheological Properties of Microfine Cement
Grouts. Tunnelling and Underground Space Technology 7 (4): pp. 45358.
doi:10.1016/0886-7798(92)90076-T.
Lombardi, G. 1985. Some Theoretical Considerations on Cement Rock Grouting. Lombardi
Engineering Ltd.
Mohammed, M. H., R. Pusch, and S. Knutsson. 2015. Study of Cement-Grout Penetration
into Fractures under Static and Oscillatory Conditions. Tunnelling and Underground
Space Technology 45: pp. 1019. doi:10.1016/j.tust.2014.08.003.
Mller, R., and D. A. Bruce. 2000. Equipment for Cement Grouting: An Overview. In
Proceeding of Advances in Grouting and Ground Modification, pp. 15572. ASCE.
doi:10.1061/40516(292)11.
Mller, R., and A. G. Hny. 2000. Contemporary Grouting Equipment. In Proceeding of
the CUC Course, Waterproofing of Tunnels. Switzerland.
Nobuto, J., M. Nishigaki, S. Mikake, S. Kobayashi, and T. Sato. 2008. Prevention of
Clogging Phenomenon with High-Grouting Pressure. Doboku Gakkai Ronbunshuu C.
64 (4): pp. 81332 (in Japanese with English abstract). doi:10.2208/jscejc.64.813.
Pusch, R., M. Erlstrm, and L. Brgesson. 1985. Sealing of Rock Fractures A Survey of
Potentially Useful Methods and Substances (Technical Report 85-17). Swedish Nuclear
Fuel and Waste Management Co (SKB), Stockholm, Sweden.
Schwarz, L. G. 1997. Roles of Rheology and Chemical Filtration on Injectability of Micro
Fine Cement Grouts. Northwestern University.
Stille, H. 2015. Rock Grouting-Theories and Applications. Rock Engineering Research
Foundation-BeFo, Stockholm, Sweden.
Wakita, S., K. Aoki, Y. Mito, Y. Kurokawa, T. Yamamoto, and K. Date. 2003. Development
of Dynamic Grouting Technique for the Improvement of Low-Permeable Rock Masses.
In Proceeding of the First Kyoto International Symposium on Underground
Environment, pp. 34148. doi:10.1201/NOE9058095565.ch44.
Warner, J. 2004. Practical Handbook of Grouting: Soil, Rock and Structures. John Wiley &
Sons.
Weaver, K. D., and D. A. Bruce. 2007. Grout Injection Pressure. In Dam Foundation
Grouting, Revised Ed, pp. 18592. ASCE. doi:10.1061/9780784407646.ch06.
Weaver, K.D. 1991. Dam Foundation Grouting. ASCE.

10
163
Drilling and Grouting Challenges in Broadly Graded Soils

Authors: Byle, Michael

Presentation Preference: Oral
Track: Track 2/ Drilling and Grouting
Session: 2A31/ Design and Assessment, Part I
Is this a student presentation?: None Specified
Bio: None Specified
Submission type: Technical Papers (6-10 pages in length)
Date Uploaded: 1/31/2017 8:34:22 AM
____________________________________________________________

Project Manager Submission separator pages

 Drilling and Grouting Challenges in Broadly Graded Soils

Michael J Byle1 P.E., D.GE, F.ASCE

Tetra Tech, Inc. One Oxford Valley, suite 200, Langhorne, PA 19047; e-mail:
1

michael.byle@tetratech.com

ABSTRACT

Broadly graded soils are characterized by having high coefficient of uniformity and typically are
deposited in fluvial and alluvial environments. Galciofluvial soils are common broadly graded
soils found in glaciated areas in the northern Continental United States and other locations
around the world. Particle sizes in glaciofluvial soils range from fine sand to boulder sizes and
typically have very little fines. The coarse fraction poses some problems for drilling and some
methods of grouting, but the greater problems occur due to internal instability. Internal
instability results from the inability of the coarse fraction of the soil to act as a filter to prevent
movement of the finer fraction. This causes water based drilling methods and grouting to wash
the finer fraction out leaving the open graded coarse material. This can induce settlement of the
ground surface during drilling and grouting. Drilling and grouting in these soils can also
increase the permeability of the ground, require much larger volumes of grout to complete the
project than initially anticipated, and limit the effectiveness of some grouting methods. While
some practitioners are knowledgeable about these conditions and their impacts, the impacts to
drilling and grouting have not been well documented in the literature. This paper presents a
summary of the nature of these soils including appropriate references, a discussion of the specific
impacts to drilling and grouting, and summaries of example cases where these soils have
impacted construction cost schedule, and performance.

INTRODUCTION

In his State of the Practice paper, George Burke (2012) stated Gravelly and Cobbly ground can
pose unexpected problems, including loss of return spoil, air escape into the formation,
unraveling of the formation and consequential lock-up of jet grout tooling. The reason for this
has not been explained, but where experienced in the field, conditions are explained as due to a
variety of causes including refusal on cobbles, borehole erosion, and nested boulders. The
results of such projects are typically not published, since engineers and contractors tend to
publish only their successes and failures that are well explained (Burke, 2012). On closer
examination, there is a greater commonality of conditions among the sites in question. While the
observations bringing this condition into the light do not constitute a rigorous scientific
investigation, a common thread becomes apparent.

INTERNALLY UNSTABLE SOILS

Internal instability is a well-known concept among dam safety engineers, but seems less
well known outside of dams practice. The term internal instability refers to soils susceptible to
internal erosion through the migration of finer particles through the matrix of coarser particles.
This most frequently occurs in glacial and fluvial soils but can occur in a wide range of soil
types. In these cases, the soils are susceptible to suffusion where the finer fraction is enclosed
within and does not completely fill the voids between coarser particles and the finer fraction is
transported through the void spaces within the coarser fraction. Suffusion can also occur in soils

Material
Susceptibility
Internal Instability
Filter Incompatibility
Voids

Fines Soil
Migration Distress

Concentrated
Seepage Velocity Erosion Arching
Hydraulic Gradient Vibration
Heave
Pore Pressures Hydraulic Low Stress
Hydraulic Fracture Stress
Load Condition

Figure 1 - Factors Affecting Initiation of Internal Erosion

(After USBR, Garner and Fanin, 2010)
where the finer grains fill the openings, but a only a fraction of the finer grains are stressed as
part of the structural fabric of the soil.
The United States Department of the Interior Bureau of Reclamation (USBR) has
assembled a guidance document that summarizes the factors related to the causation of internal
erosion and internal instability. The guide includes a figure similar to the one below that places
the causative factors in three categories: Material Susceptibility, Hydraulic Load, and Stress
Condition. The material susceptibility is related to the gradation and of the soil and is generally
limited to non-cohesive soils. The stress condition is related to the overburden, but also to the
distribution of stress within the soil structure. In the case of drilling and grouting, the Hydraulic
load is provided by the drilling fluid or grout injection.

MATERIAL SUSCEPTIBILITY

Cohesionless soils that are gap graded or broadly graded can be susceptible to internal instability.
Gap graded soils generally have a steep portion of the gradation curve separating flatter portions
as shown in Figure 2. Broadly graded soils are those that have a relatively flat gradation curve.
In coarse cobbly and boulder soils one must consider the limitations of the sampling tools where
the gradation is truncated at the maximum size that is sampled. In such cases, the laboratory
gradation will exclude cobbles and boulders that may affect the true gradation.
In either case, the finer fraction can be eroded through the coarser matrix. In the case of
gap graded soils, defining the finer fraction is simply identifying the gap location from the
gradation curve as indicated in Figure 2. For broadly graded soils, this can be less obvious.

U.S. Standard Sieve

12-in. 3-in. #4 #200
100

90
Gap
80 Graded
70 Soil Finer
Fractio
60
50
Broadly
Percent

40 Graded
30 Soil
20
10
0
Figure 2 - Potentially Internally Unstable Soils
 A common approach is the Kezdi (1979) method. In this approach the soil is arbitrarily
divided into two fractions and the filter criteria are checked between the two fractions. To meet
the filter criteria, the size corresponding to 15% of the grains being (D15) of the coarser fraction
must be at least 4 times the size corresponding to 85% finer (D85) of the finer fraction. For
broadly graded soils, a spreadsheet can be used to assess this criteria at different division points.
A method that is simpler for assessing broadly graded soils was developed by Sherard
(1979). In this approach one simply plots the gradation to assess whether it falls into the
Potentially Unstable Band and checks to see that the gradation does not lie below a 4% slope line
drawn on the gradation chart.

Figure 3 Sherard Approach - Potentially Unstable Soils (USBR 2011)

There are a number of other methods to evaluate the susceptibility. The capillary tube
model (Kovacs, 1981) determines instability by comparing the 85th percentile of the finer
fraction (d85) gradation to the 50th percentile opening size in the coarser fraction (O50). O50 is
computed based on the porosity, grain shape, and grain size distribution. Li (2008) expanded on
this concept producing the following criterion for instability.
Soil is potentially unstable where:
d85/ O50 > 0.42 (1)

Li (2008) also discusses the impact of gradation on the critical gradient to induce piping
erosion and found that when D15 > 12 * d15 that piping occurred spontaneously. This means for
these soils the critical gradient is essentially zero and that any injected fluid will transport the
finer fraction even at very low pressures.

STRESS CONDITION

The stress condition includes assessment of the micro and macro stress distributions. Macro
stress distribution is based on the total load being carried by the soil consistent with conventional
computation using unit weight, overburden depth and earth pressure coefficients. The micro
stress distribution can vary significantly, since not all particles carry the same load. In most non-
cohesive soils, the stresses are shared among the soil grains. In broadly graded and gap graded
soils, a significant fraction of the soil grains may be unstressed, since the finer fraction that does
not completely fill the pore space of coarse fraction will not be loaded. In this case the coarse
fraction is supported by grain-to-grain contacts between the coarse particles and perhaps only a
fraction of the finer particles. Shire, et al. (2014) completed discrete element modeling to
quantify the effective stress distribution and considered the stress reduction factor as the ratio of
the stress on the fine fraction to that on the overall soil. This factor is used to estimate the change
in critical gradient needed to mobilize the fine particles and induce internal erosion. The results
of this study identifies what are termed Case (i) and (iii) materials, in which the coarse
particles dominate stress transfer and the finer particles sit loose within the voids between, and
can be eroded with no loss of matrix integrity. These materials are highly susceptible to
suffusion.

Stress produces interlocking and frictional force, which act to prevent migration of
particles. Also, soil arching can produce zones of low stress within the larger soil mass. In
general, for soils that Shire, et al. show that this stress distribution is important for soils that have
D15 of the coarse fraction 7 to 10 times the d15 of the fine fraction. The process of assessing this
on the micro scale is complex and not suited to practice, but reveals a relationship that indicates
the role of stress distribution in conjunction with gradation.

HYDRAULIC LOAD

The hydraulic load is created by imbalance in pore water pressures that induces flow that
can erode and transport the soil or its components. In any case, the parameter of interest is the
critical gradient, which in these soils, is dependent upon a number of factors as noted above.
Research by a number of investigators have yielded a number of relationships for computation of
the critical gradient. Internally unstable soils can have critical gradients 1/3 to 1/5 of those for
other soils (Skempton and Brogan, 1994). Wan And Fell (2004) noted that internally unstable
soils began to erode at gradients of 0.5 to 0.8.
In dams, the gradients are easily computed or generated from models to assess the flow.
In drilling and grouting, more localized gradients are created that can be complex to analyze.
Since the intent of drilling fluid circulation and jet grouting is to have the returns flow up the
hole to the surface and to isolate the pressures in the drilling and grouting fluid from the
surrounding groundwater regime, the gradients will be steep. At a minimum, the head acting at
any point along the drill stem will be the difference between the discharge elevation and the
hydrostatic level in the ground. However, this head difference can be substantially higher
considering head losses occurring in the hole.

Fluid Injection and Internal Erosion. Where jetting is used, either to advance the hole or to
expand it (as in jet grouting) the intent is to erode the soil. In such cases, the critical gradient
must exist at the point of impingement in order to erode and transport particles up the hole.
These methods generally rely on the viscosity of the grout to prevent migration into the soil. The
problem occurs when the permeability of the stratum is high enough that the pressure induces
flow carrying soil particles outward or downward into the formation instead of up the hole.
The process initiates during the drilling, or pre-drilling of holes. If air is used, the air
bubbles expand as they rise in the annulus around the drill stem creating an air-lift. This creates
a reduced pressure at the bottom of the hole, which draws ground water into the hole. Where this
inflow exceeds the critical gradient, fines are eroded from the soil surround the hole. This
enlarges the opening sizes in the soils that can lead to progressive suffusion and suffusion.

EXAMPLE CASES

Shaft Drilling. The author witnessed a driller attempting to install a 10-inch diameter shaft in
alluvial sand and gravel with cobbles using a down-the-hole hammer and air as the drilling fluid.
Drilling progressed well to a depth of about 15 feet, where progress slowed. The drilling
continued for 2 days without appreciable progress, when the ground collapsed under the rig.
This is the result of eroding the fines and causing the cobbles to nest and the creation of a
continuous erosion as further drilling was attempted while encountering increased resistance in
now nested cobbles.

Foundation Excavation and Dewatering. Excavation of foundations was underway at a site

that had been filled with broadly graded cobbles and boulders. The contractor encountered
boulders and sand and groundwater within 4-6 feet of the ground surface. The fill had been
previously tested to be 95% of the maximum Standard Proctor Density. As the contractor
attempted to dewater the excavations, the excavation sidewalls collapsed, the excavations rapidly
became wider than deep. It was suspected that internal instability might have been a factor, but
gradation testing of small samples was inconclusive. Accordingly, a full gradation of the soil
was undertaken to better assess the susceptibility to internal instability. To do this, a large
sample was needed to assess the full gradation because of the significant large size fraction. This
was accomplished using a backhoe to excavate a sample into a tare weighed truck, dumping the
sample onto a tarp (Figure 4), manually sorting and measuring the large fraction greater than 100
mm (4 inches) and performing laboratory testing on the balance of the sample. The resulting
gradation curve was found to fall within the Sherard susceptibility zone.

Figure 4 Full sample of broadly graded soil on tarp for gradation testing

Jet Grouting Case. Jet grouting was undertaken at a site in New York in an area of glaciofluvial
soils. The gradations for the soils indicated broadly graded soils containing cobbles and
boulders. Despite an initial concern, the jet grouting contractor considered that jet grouting
should work based on another site previously completed in cobbly glacial soils. Jet grouting
encountered numerous problems including, difficulty drilling, caving boreholes, loss of
circulation, cross communication between distant boreholes, and incomplete columns.
Figure 5 shows a typical gradation with the Sherard Band added. The sample represented
came from a 37mm (1.5 in) sampler, so the gradation does not include the larger fraction. Even
so, the gradation clearly falls within the Sherard Unstable Band and the full range of actual soils
would be even flatter and broader with the inclusion particles larger than 37 mm. The potential
for interal instability was not noted by any of the parties to the work.
Jet grouting was completed after extensive delays and cost overruns using polymer mud
to stabilize the boreholes and increased quantities of grout. The methodology used was defined
in the test grouting, and the contractor initially felt that any problems would be worked out
during the course of the work. This proved not to be the case and led to large claims late in the
project.
 Such cases are not unusual. A similar situation of problems related to jet grouting in
glacial sands and gravels was reported by Walker (1997), but the explanation of the problems
was speculated as due to factors including the underlying geology, and open gravels not
identified during the investigation. Walker (1997) indicates that use of pre-grouting with tremied
cement, higher viscosity grout, and sodium silicate accelerator were used to mitigate the

Figure 5: Typical gradation for jet grouting case

difficulties in the mixed overburden. Gradations of the materials were not presented, though the
conditions might be related to suffusion of broadly graded glacial soils.
MITIGATION MEASURES
Drilling and grouting difficulties can be anticipated and mitigated with increased awareness of
these conditions. Mitigation could occur at any number of points including those listed below:
1. Geotechnical Baseline Report Identification of the soils as internally unstable provides
insight to the project designers to consider alternatives to jet grouting, and puts the
contractor on warning. Further, discussion of the impact of oversized materials that are
not included in the collected samples and laboratory gradation results should be included
to accurately assess the full gradation and behavior of the in situ soils.
2. Design Phase Evaluation of the soils for internal instability issues can lead to different
weighting of grouting in the alternatives analysis phase and may lead to a different means
and methods or a completely different recommended solution. Awareness of the internal
stability issues in these soils would enable communication of this condition to the
contractor prior to bidding.
3. Bidding Phase If the grouting contractor is alert for internal instability issues, review of
the gradations will reveal the potential problems for drilling and jet grouting and enable
improved scheduling and budgeting, or consideration of an alternative approach.
4. Test Grouting Phase When problems are encountered in the test grouting, a review of
the gradation of the material is appropriate to provide the reason for the problems.
 Awareness of internal instability impacts could and should provide a basis for a remedy
that could be clearly shared with the owner to revise budget and schedule expectations.
5. Production Grouting Phase As the drilling and grouting progress and continued
problems occur, it should become obvious very quickly that the problems are due to
internal instability of the soils and would not be going away.

Baseline Report. The key factor in mitigation is to identify the condition. Where granular soils
containing gravels larger than 37 mm (1-1/2 in), conventional sampling will not recover them.
Geotechnical engineers must consider the potential impact of oversized material. Where such
material is present, geotechnical engineers should not ignore it. In glacial and alluvial deposits
of mixed sizes, geotechnical engineering reports should indicate the potential for internal
instability for any project where drilling with a fluid may be needed, or grout injection. The
extent of the problem may not be apparent in auger drilling where there is no unbalanced head
during drilling, but any problems of running sand or similar conditions should be reported. The
geotechnical engineer should provide information on the presence of oversized gravels, cobbles,
and boulders to the extent indicated by drilling resistance and observation of cuttings returned,
even where large fraction is not recovered. It is not enough to just provide a gradation of the
collected samples. The soil should be described and evaluated as a single composite material
spanning the full range of its in situ gradation. Where the soil is broadly graded, the potential for
internal instability should be discussed in the report.
Design Phase. In the early phases of design, particularly during alternatives analysis, design
engineers should consider the impacts of potential internal instability on the methods being
considered. A specialist should be consulted who is knowledgeable about the methods being
considered, and who has no financial or contractual interest in the project, to assess the
suitability of the methods to the site and site conditions. This specialist must include internal
instability together with other factors in assessing the viability and risks associated with
alternative approaches. Keep in mind that internal instability will be a factor into difficulties and
costs of other construction aside from drilling and grouting; such as dewatering, excavation and
excavation support. Owners need to be made aware of the project risks and be prepared to deal
with the consequences if risks are taken.

Bid Phase. Awareness is also a primary precursor to mitigation in construction. Contractors

have to read the baseline report and understand the impact of conditions on the methods
proposed, even where not explicitly stated in the report. Specialty contractors are experts at the
methods they are using and will have much greater knowledge of the impacts of conditions on
the performance of drilling and grouting methods than most engineers.
Accordingly in reviewing the bid documents, contractors should examine all of the data
provided on a more than cursory basis, including:
 Boring Logs
o Notes on running sands or hole collapse
o Notes on bit grinding or presence of oversized material
o Low recovery in drive samples
o Soil classifications of SP, SW, GP, GW.
Gradation Curves
o Flat linear gradation curves
o Stepped gradation curves
Geologic description
o Glacial Glaciofluvial
o Fluvial
o Alluvial
Where conditions are suspect, specific questions should be raised during the bid process to raise
awareness of all bidders to the condition. This levels the playing field and avoids the potential
for the winning bidder to be the one that leaves this issue out of his bid.
Test Grouting Phase. Completing test grouting in advance of production work is intended to
provide proof of viability of the techniques used, tune in parameters to achieve the project
objectives, set the process for production work, and verify performance of the selected
production parameters. Accordingly, it is essential that the test grouting be fully documented
and difficulties be addressed and resolved prior to proceeding with production drilling and
grouting. While it may add cost, and lengthen the schedule to permit time to digest and report
the test grouting information, it is essential to do so.
In drilling and grouting test holes, problems should be identified and fully explained. Oft times
it is tempting to explain problems as due to variations in subsurface conditions that may differ in
the test area, but this should not be concluded without paired testing that is done with the same
equipment and methods within the production area and the test area to assure that the perceived
variation is real. Also, fluvial and alluvial soils naturally vary within the deposits, so variations
should be anticipated and may have more to do with the intervals sampled than variations across
the site.
Where hole collapse, running sand, perceived nesting of cobbles and boulders, grout loss, cross
communication between distant holes, or loss of circulation during drilling are encountered, the
potential for internal stability should be seriously considered and ruled out before proceeding. If
necessary, the test grouting should be modified to better evaluate the cause and provide a
solution. All parties must be willing to accept that the methods originally proposed may not be
suitable for the conditions and it may be necessary to employ other methods. Where drilling and
grouting difficulties are encountered, it is tempting to proceed on the hope that the process will
evolve to solve them, but this is not good practice, and can lead to enormous cost overruns and
delays later in the project. The time to address this is before production grouting starts.
Production Phase. In production is the latest and most expensive time to deal with internal
instability issues. Conditions that may have appeared manageable in the test grouting may be
found to be more severe than they appeared in the test grouting and become unmanageable.
Owners must be advised of conditions and measures taken. Internal instability issues that are
discovered in the production phase are typically resolved through change orders, claims, and
litigation. It is therefore the least desirable time to address these issues.
CONCLUSION
Internal instability of soils is a real and critical factor in drilling and grouting, It can render
ineffective, methods that may have worked exceedingly well in soils with similar USCS
classifications. This can be true even in apparently similar geologic settings, since geologic
mapping as glacial or alluvial soil, is not sufficient in itself to make a conclusion as to the
internal instability of the soils under the conditions of the proposed methods. Alluvial granular
soils, especially glaciofluvial soils, should be specifically assessed for internal erosion potential
for any drilling and grouting project.
REFERENCES
George Burke State of the Practice in Jet Grouting, Conference Proceedings from Grouting and
Deep Mixing 2012 ASCE GSP No.228. ASCE, Reston, VA.
Burke, George (2004) Jet Grouting Systems: Advantages and Disadvantage GeoSupport 2004,
ASCE, Reston, VA.
United States Bureau of Reclamation (USBR) (2015) Dam Safety Best Practices Chapter IV-4
Internal Erosion Risks for Embankments and Foundations.
Walker, A.D. (1997) An Example of the use of Jet Grouting to permit Tunneling in Chemically
Weathered Limestone. Proceedings of 6th Multidisciplinary Conference on Sinkholes and
the Engineering and Environmental Impacts of Karst, Springfield, Missouri.
Li, Maoxin (2008). Seepage Induced Instability in Widely Graded Soils. Ph.D. Thesis.
University of British Columbia (Vancouver)
Shire, T.; OSullivan, C.; Hanley, K.J.; Fannin, R.J. (2014). Fabric and Effective Stress
Distribution in Internally Unstable Soils. Journal of Geotechnical and Geoenvironmental
Engineering, vol. 140, no.12, 04014072, ASCE GT.1943-5605.001184 ASCE, Reston,
VA.
190
Statistical Evaluation of Groutability with Data from
Hydraulic Tests and Fracture Mapping: Case Studies from
Sweden

Authors:
Runsltt, Edward
Thrn, Johan
Kvartsberg, Sara
Fransson, sa
Presentation Preference: Either Oral or Poster
Track: Track 2/ Drilling and Grouting
Session: 2A31/ Design and Assessment, Part I
Is this a student presentation?: None Specified
Bio: None Specified
Submission type: Technical Papers (6-10 pages in length)
Date Uploaded: 1/17/2017 11:32:59 AM
____________________________________________________________

Project Manager Submission separator pages

 Statistical evaluation of groutability using data from hydraulic tests and
fracture mapping
Case studies from Sweden

Edward Runsltt, M.Sc.,1 Johan Thrn, Ph.D,2 sa Fransson, Professor3, Sara Kvartsberg,
Ph. Lic. 4

1
Golder Associates AB, Division of Geotechnical Engineering, Stockholm, Sweden; e-mail:
Edward_runslatt@golder.se
2
Chalmers Univeristy of Technology, Division of GeoEngineering, Gothenburg, Sweden;
Bergab, Gothenburg; e-mail: Johan.thorn@chalmers.se
3
Chalmers Univeristy of Technology, Division of GeoEngineering, Gothenburg, Sweden; Golder
Associates AB, Gothenburg, Sweden; e-mail: Asa_fransson@golder.se
4
Norconsult AB, Division of Rock and Hydrogeology, Gothenburg, Sweden; e-mail:
Sara.kvartsberg@norconsult.com

ABSTRACT

Sweden has a long history of research within the field of rock fissure grouting in hard crystalline
rock mass due to strict environmental requirements regarding allowable ground water draw
down. These requirements normally implies that fractures down to aperture size between 50 to
100 m needs to be sealed and within these ranges the size of the particles for cementitious
grouting agents becomes a limiting factor. For a grouting design it is therefore of importance to
consider the aperture size distribution of the rock mass in order to predict the groutability for
both cementitious and non-cementitious grouting agents. Transmissivity data from hydraulic
tests (water pressure tests) and number of fractures along a borehole can be assessed from core
logging for further use as input for a statistical interpretation of fracture data to simulate an
aperture size distribution. A methodology developed at Chalmers University of Technology in
Gothenburg, Sweden, is proposed. The method is a statistical evaluation of groutability (SEG)
and is based on the Pareto distribution. A computational design tool has been developed to
simplify the use of the statistical evaluation and to make the research more accessible to end
users, designers, in the grouting industry. The aim of this article is to present two case studies
where the statistical interpretation of fracture data is performed by using the computational
design tool and how the outcome can be of great use in finding a more accurate grouting design.
The case studies include fracture data sets from two large infrastructure rock tunnel projects in
Sweden; a road tunnel in Stockholm and a railroad tunnel in Gothenburg.

1
INTRODUCTION

Sweden has a long history of research within the field of rock fracture grouting in fractured
crystalline rock mass. This is due to strict environmental requirements in urban environments
regarding groundwater ingress in rock tunnels and other underground structures below the
groundwater table. If the groundwater table is lowered settlements will occur due to the
geological conditions with soft clays which will cause damages to the structures. To receive a
building permit the owner needs to present a sustainable sealing strategy in order to minimize the
environmental impact on the surroundings approved from the Swedish Land and Environment
Court. In the approval process various types of measurable sealing demands need to be specified
such as maximum allowed groundwater ingress, groundwater drawdown, risk assessment etc.
Monitoring programs are issued to assure that the inflow requirements are fulfilled. The County
Administrative Board do regular follow up, during construction, with the owner to verify that the
grouting works is in line with the approved permit for water works which include the specific
sealing demands and restrictions for allowable ground water lowering. These requirements
normally implies that fractures down to aperture size between 50 to 100 m needs to be sealed
and within these ranges the size of the particles in a cementitious grouting agent becomes a
limiting factor.

Depending of the magnitude of the project various extensive pre-investigations are executed to
characterize the hydraulic properties of the rock mass. When cities grow and create a denser and
more compact urban environment, authorities introduce tougher requirements to protect existing
interests (buildings, structures, environment etc.). This results in a need to make better
predictions and designs in terms of sealing against water ingress during construction.

For grouting design it is of great importance to be able to describe the fracture aperture
distribution in the rock mass in order to make predictions of groutability for both cementitious
and non-cementitious grouts. Factors that affects the groutability is fracture aperture distribution,
fracture network connectivity, fracture intensity and rheology of the grouting agent.

A method for Statistical Evaluation of Groutability (SEG) developed at Chalmers University of

Technology suggests the use of the Pareto (Power Law) distribution and combinatorics to
describe a fracture transmissivity distribution and further a fracture aperture distribution
(Fransson 2002, Gustafson et al. 2004). The SEG method is based on transmissivity data from
short duration hydraulic tests conducted in test sections along a borehole. These are linked to
fracture position data from core mapping and/or optic televiewer from the corresponding section
in the borehole.

The fracture aperture distribution can be used to design a preferable type of grouting agent to use
for an investigated rock mass to be grouted. It can also be used to assess whether cementitious or

2
non-cementitious grout is needed, where the first is normally preferred and the latter is normally
only used in situations where the penetrability of cementitious grout is poor, i.e. fractures smaller
than 50 to 100 m. The outcome of the fracture aperture distribution can also be of use for other
applications such as evaluating grout penetrability and grout spread (Gustafson and Stille 1996).
Fransson et al. (2016) presents examples of hydraulic testing, grout selection and estimate of
grout spread (penetration length).

Runsltt et al. (2013) highlighted that in order to make SEG method more available, user friendly
and transparent there is a need for a more extensive description and better understanding of how
to deal with model uncertainties regarding quality of input data gathered from borehole
investigations. Furthermore a more uniform approach of how to set up an analysis was requested.
In response, Thrn et al. (2015) developed a computational tool, coded with Visual Basics for
Applications (VBA) in excel, that in a uniform way simplifies the use and makes the SEG
method more available to the designers in the Swedish grouting industry. This tool also makes
the analysis more transparent and easier to review. The computational design tool comes with a
report that gives an in depth description of the theoretical background of the SEG method,
guidance for data collection in field and a user manual in how to perform the data input.

It should be kept in mind that fairly simple models are used for the SEG method and that the
result should be considered a useful guidance in grouting design and not as an exact answer.

The aim of this article is to present two case studies where the statistical interpretation of fracture
data is done by using the computational design tool and show how the outcome can be of use in
finding an accurate grouting design. The case studies include fracture data sets from two large
infrastructure rock tunnel projects in Sweden; a road tunnel in Stockholm and a railroad tunnel in
Gothenburg. This article will present a summary of the theoretical background while Fransson
(2002), Gustafson & Fransson (2005), Thrn et al (2015) and Thrn et al (2016) gives the reader
a more in depth description.

THEORY

Input data/ Data Collection

The SEG method use two input parameters: hydraulic transmissivity measurements and fracture
data collected from water pressure tests and mapping of rock cores respectively. A suggested
approach to reach sufficient resolution is to perform water pressure tests (WPT) in three meter
sections. Longer sections are used, 5-10 m, but will result in a loss of resolution for the analysis.
Fracture data mapping should be performed by a person with good skills to judge whether
fractures are open or closed. The open fractures along the borehole should be correlated to the
corresponding WPT section as a first step before the analysis is performed.

3
Evaluation of the transmissivity, T [m2/s], from WPTs can for example be performed with
Moyes equation (Gustafson 2012):

= 1+ Eq. 1
2 2

where Q [m3/s] is the stationary steady state water flow logged at the end of the test period, dp
[Pa] is the constant water injection pressure, w [kg/m3] is the density of water, g [m/s2] is the
constant of gravity, rw is the borehole radius and L [m] is the length of water test section.

To perform a good quality statistical analysis it is of great importance to collect fracture data
with high quality and good accuracy. High quality in this instance is achieved through data
collected by experienced field personnel being aware of the purpose of the investigation, using
calibrated equipment relevant for the task and record any event or anomaly that could be of
interest for a designer to interpret the data in a correct way. Good accuracy is acquired with
calibrated flowmeter and piezometers with appropriate detection ranges and small error. Listed
below are a number of factors that needs to be considered during data collection.

Core drilling, minimum double tubing, is strongly recommended since it result in a less
rough borehole wall in comparison to a percussion drilled hole. This will allow a double
packer that is used during WPTs to better seal against the borehole wall in each
individual test section. The risk of leakage around the packers will also be minimized.
Core drilling also produce a rock core which can be used for the fracture mapping which
is needed for the analysis. An optic televiewer can be used as a substitute or
complementary method.
Measurement limits and accuracy for the flowmeter and piezometers should be based
upon the smallest fractures that need to be grouted to meet the specific sealing
requirements, i.e. inflow, environmental impact etc. It is especially important that the
lower detection limit for the flowmeter is small enough and corresponds to the inflow
requirements.
A maximum section length of three meter is recommended for WPT. A longer section
length will result in fewer sections and a loss in resolution for the analysis.
The water pressure and the water flow should be logged in real time. It is recommended
to install a piezometer in the actual test section, instead of at the ground surface, in order
to measure the actual water injection pressure. It is also recommended to install
piezometers above and below the double packer to detect any pressure increases. This
could be interpreted as either a leakage around the packers or a rock mass with a good
connectivity between the fractures.
A shut-in device installed at the packer is also recommended to eliminate the borehole
storage effect.

4
The amount of sections needed for an analysis have not been investigated, from experience there
is a need for at least 20 to 40 WPT sections. This topic is planned to be investigated in future
research.

Pareto distribution
The Pareto distribution has turned out to be suitable for describing how hydraulic fracture
apertures and hydraulic transmissivities are distributed in hard crystalline rock (Gustafson 2012).
This distribution is well suited for data sets where a small quantity of fractures have a large
aperture while a large quantity have a small aperture.

With use of the Pareto distribution and combinatorics a fracture transmissivity distribution can
be calculated based on evaluated section transmissivities from a borehole together with
corresponding fracture data. The Pareto distribution is described by the function:

= Eq. 2

where C is a constant, T [m2/s] is the transmissivity and k [-] is the shape parameter. The value
of the k-parameter is vital as it describes the shape of the simulated aperture distribution as will
be shown later in the case studies. In the computational design tool a probability transmissivity
distribution is calculated and correlated to the Pareto function graphically and the k-parameter is
evaluated. This step is further described in Fransson (2002), Gustafson & Fransson (2005) and
Thrn et al (2015 and 2016).

If the k-parameter is smaller than 0.5 then the largest fracture is always in the same order of
magnitude as the section transmissivity. The larger k-values the smaller the spread are of fracture
transmissivities for individual fractures in a section. Large k-values are generally expected in
fracture zones or fractures with a high connectivity (Gustafson 2012). It has also been observed
that large k-values might be a result of data collected with not small enough lower detection
limit.

The transmissivity distribution can be translated to an aperture distribution with the cubic law
(Snow 1968):

12
= Eq. 3

where bhyd [m] is the hydraulic aperture and w [Pas] is the viscosity of water. The transmissivity
distribution is proportional to cube of the aperture distribution.

5
The following model assumptions need to be considered when using the SEG method:

The theory behind the cubic law considers individual fractures as smooth parallel plates.
The transmissivity of each individual fracture is assumed to be independent which means
that the sum of all individual fracture transmissivities corresponds to the total
transmissivity of the borehole (Fransson 2002).
The largest fracture in each section is assumed to be by far most conductive and is given
a flowrate (from the WPT) in the same order of magnitude as the total flow in the section
(Fransson 2002). The remainder of the transmissivities is shared among the remaining
smaller fractures. For a section where all fractures are hydraulically well connected
(dependent) it is suggested to set the number of fractures to one, since all fractures are
linked and acts as one.
If data is collected within an area where different hydraulic domains are present it is
recommended that the data is analyzed separately for each domain. Examples of
different domains can be different rock types, deformation zones etc.
Results from a WPT reflects the transmissivity of the rock mass in the close vicinity to
the borehole. But it is assumed that this data also represent a much larger rock volume.
Evaluation of transmissivity is assumed to be done from hydraulic tests that have
reached stationary steady state conditions, if not it is suggested to evaluate the test data
as a transient test which requires a more detailed and extensive data collection.

Estimation of water inflow

Gustafson (1986) gives an example of how to calculate a rough estimate of the water inflow into
an ungrouted tunnel which was further developed 2004 by Gustafson et al. for a grouted rock
tunnel. The groundwater level need to be assumed close to the ground surface.

The ungrouted, q [m3/s], and grouted tunnel inflow, qinj [m3/s], can be estimated with:

2 /
= Eq. 4
2
+

2 /
=
2 Eq. 5
+ 1 1+ +

6
Where Ttot is the sum of all fracture transmissivities and Tinj is the residual transmissivity after
fractures to a certain size have been grouted. L [m] is the total length of the tunnel section, H [m]
is the distance to the ground surface, rt [m] is the tunnel radius, [-] is the skin factor.

CASE STUDY

In the following case studies the computational design tool has been used with data from two
different projects that are to be constructed in the near future. The West Link project is presented
as a summary of the case study presented in Thrn et al. (2016) and the Stockholm Bypass is a
new case. These projects have been chosen since they have investigations with different
detection limits for the WPTs and it will be demonstrated how this influence the analysis and
conclusions.

The West Link Project

The West Link is a planned railroad tunnel under the city of Gothenburg that will improve the
capacity for commuting and regional trains in the area. Total length of the tunnel will be six km
of which four km is in rock and two in soil (Swedish Transport Administration 2016). The
geology can be described as a sparsely fractured gneiss that is frequently heavily foliated.

WPTs were performed during the course of two different field investigations. During the first
investigation a more rudimentary equipment was used with a lower detection limit for the flow,
set to 0.1 l/min (the actual lower detection limit was unknown). This was used early in the
project and corresponded to standard procedure at the time. The main reason for the second
investigation was to provide data for design with the SEG method, using a more refined
flowmeter with a lower detection limit of 0.005 l/min. There were also improvements made by
installing piezometers in the test section instead of at the ground surface as was the case for the
first investigation. The measurements for both investigations were performed in the same three
core drilled boreholes which makes the results comparable. Data from the first and second
investigation is hereafter denoted as data set 1 and data set 2.

The sum of the evaluated transmissivities, Ttot, for the boreholes is 410-6 and 510-6 m2/s for data
set 1 and 2 respectively. This indicates that both investigations manage to log the larger
hydraulic fractures in a similar way. But the difference between the two data sets is obvious
when performing an analysis and calculating the fracture aperture distribution, see Figure 1. For
data set 1 a majority of the fracture apertures (more than 90% of the fractures) are within the
range 15 to 40 m with a k-value of 1.1. The corresponding range is 2 to 20 m with a k-value
of 0.42 for data set 2.

7
Micro-fine cement is considered to properly seal fractures down to a range between 50 to 100
m (Fransson 2008, Gustafson 2012, Stille 2015). Micro-fine cement is according to the
European Standards (2000) characterized by a specific surface area larger than 800 m2/kg and a
d95 smaller 20 m. A tunnel inflow calculation, Eq. 4, results in a tunnel inflow, q, of
approximately 8 l/min100 m for both data sets. If grouting is assumed to be performed of all
fractures down to an aperture size of 50 m the residual inflow, Eq. 5, is estimated to 7.5 and 3.0
l/min100 m respectively (assumed that H= 30 m, L= 135 m, rt= 5 m and = 5.). Based on the
first investigation a standard grouting procedure with micro-fine cement would marginally
reduce the inflow, while the second investigation result in a significant inflow reduction if all
fractures down to 50 m are grouted.

Simulated aperture distribution

1
0.9
0.8
0.7
0.6
P(b<bn)

0.5
0.4
0.3 First data set
0.2 Second data set
0.1 50 micrometer
0
1 10 100 1000
b(m)

Figure 1: Show the aperture distribution for the first and second data sets. The dotted line
represents 50 m.

E4 the Stockholm bypass Project

The Stockholm bypass is located west of the city of Stockholm in order to relieve the major
existing connection between the southern and northern suburbs. The two main tunnels is
designed as a six lane 21 km highway of which 19 km are tunnels. The tunnel is divided into two
stretches, 16.5 km and 1.8 km respectively, and will be one of the worlds longest road tunnels
when completed in 2025. The geology can be described as sedimentary gneiss that is sparsely
fractured with occasional parts of highly fractured rock.

A total of seven cored boreholes have been investigated. WPTs were performed during different
phases of the pre-investigation program, hence two different flowmeters were used with a lower
detection limit of 0.05 l/min (two boreholes, data set 3) and 0.002 l/min (five boreholes, data set
4).

8
As opposed to the West Link project, data set 3 and 4 cant be directly compared since the
investigations were performed in different boreholes. Although all boreholes are considered to be
drilled in the same rock domain some differences are still expected. The sum of the
transmissivities are 410-5 m2/s and 210-5 m2/s and the k-values are 0.52 and 0.61 for data set 3
and 4 respectively. As seen in Figure 2 both data sets have a similar shape which is also reflected
by the k-values.

The inflow to a tunnel can be estimated to an inflow, q, of 55 and 19 l/min100 m for data set 3
and 4 respectively (Eq. 4). If all fractures are assumed to be grouted down to 50 m the residual
inflow, qinj, is estimated to 16 and 10 l/min100 m respectively (assumed that H= 30 m, L= 135
m, rt= 5 m and = 5, with Eq. 5). This can be compared to the actual inflow requirements for the
Stockholm Bypass that varies between 7-20 l/min100 m depending on tunnel segment. If
fractures down to 50 m can be grouted successfully it is likely that the inflow requirements will
be fulfilled based on data sets and given assumptions.

Data set 3 and 4 have k-values in the same range, 0.52 and 0.61 which is evaluated as both data
sets are in the same rock domain. Even though the lower detection limit were different, 0.05
versus 0.002 l/min, both data sets produce a similar aperture distribution which ranges over the
same order of magnitude. The difference in estimated inflow to a tunnel before and after grouting
can be explained with the investigations being performed in different boreholes. Which implies
that the detection limit of 0.05 l/min seems to be suitable for the inflow requirements that are
presented above.

Simulated aperture distribution

1
0.9
0.8
0.7
0.6
P(b<bn)

0.5
0.4
Data set 3
0.3
0.2 Data set 4
0.1 50 micrometer
0
1 10 100 1000
b(m)

Figure 2: Show the aperture distribution for the third and fourth data sets. The dotted line
represents 50 m.

9
CONCLUSION

An understanding of the fracture aperture distribution is a key element to get a better

understanding of the hydraulic properties when doing a grouting design in crystalline rock. As
seen in the case studies an aperture distribution can be estimated based on transmissivity and
fracture data from core drilled boreholes.

For the Swedish grouting industry it is generally assumed that grouting can be performed with
cementitious grouts. By using the analysis and theories of inflow estimates to tunnels the
designer can get valuable input to whether a standard grouting design with cementitious grout is
reasonable or if it need to be reconsidered.

When it comes to the data collection the West Link project is a good example of how different
level of input data will affect the analysis. If only the first data set would have been considered
during the grouting design it is possible that grouting with cementitious grout might have been
rejected considering the limited sealing effect since the majority of the hydraulic fractures
appeared to be smaller than 50 m. It is obvious that the lower detection limit wasnt specifically
considered during the preparation of the investigation program for the WPTs. In this case data set
1 wasnt planned to be analyzed according to the SEG method presented in this article.

The detection limits should be designed according to an expected inflow requirement. If the
requirement is yet to be determined the detection limit should be assumed on the conservative
side, i.e. a small limit. A small enough lower detection limit should be used in order to not lose
resolution. A loss in resolution will reflect on the transmissivity and aperture distribution, which
is seen in data set 1 for the West Link case. It is observed that the aperture distribution only
range over one order of magnitude, while data set 2 (with a lower detection limit) span over three
order of magnitudes and therefore show accordance with the general notion of a Pareto
distribution, i.e. the data collection shouldnt be physically censored/limited. The rock quality is
evaluated, from core logging, as sparsely fractured which in terms of k-values should result in a
value of 0.5 or lower (Gustafson 2012). With this knowledge data set 1 can be deemed
inappropriate with a k-value of 1.1 which hints of a highly fractured rock. This implies that a
data set that isnt physically limited (by detection limits) will result in a more reliable description
of the hydraulic properties of the rock mass. Early time estimates such as those presented here
should be followed up and reviewed during further investigations and construction.

The research conducted at technical universities is of great importance in order to push the
industry to a higher level of understanding of how grouting can be used in a more cost effective
manner. The computational design tool is an example of such an effort. With more refined
models and analyses there is also a need for high quality data that is collected in a uniform and
controlled way with the use of the latest technology to log and visualize the data in the field.

10
REFERENCES

European Standards (2000). Execution of Special Geotechnical Work Grouting. CSN EN

12715.
Fransson, . (2002). Nonparametric Method for Transmissivity Distributions Along Boreholes.
Ground Water. 2002 Mar-Apr; 40(2):201-4
Fransson, . (2008). Grouting design based on characterization of the fractured rock -
Presentation and demonstration of a methodology. SKB, Report R-08-127, Stockholm,
Sweden.
Fransson, ., Funehag, J., Thrn, J. (2016). Swedish grouting design: hydraulic testing and
selection of grout. Proceedings of the Institution of Civil Engineers Ground
Improvements. DOI: http://dx.doi.org/10.1680/jgrim.15.00020
Gustafson, G (1986). Geohydrologiska frunderskningar i berg. Bakgrund Metodik -
Anvndning. BEFO, Report 84:1/6, Stockholm, Sweden.
Gustafson, G., Stille, H. (1996). Prediction of groutability from grout properties and
hydrogeological data. Tunneling and Underground Space Technology. 1996, Vol. 11,
No. 3, 325-332.
Gustafson, G., Fransson, ., Funehag, J. and Axelsson, M. (2004). Ett nytt angreppsstt fr
bergbeskrivning och analysprocess fr injektering. Vg och Vattenbyggaren 4,
Stockholm, Sweden.
Gustafson, G., Fransson, . (2005). The use of Pareto distribution for fracture transmissivity
assessment, Hydrogeology Journal, 14, 15-20.
Gustafson, G. (2012). Hydrogeology for Rock Engineers. BEFO, Stockholm, Sweden.
Runsltt, E., Cretz, M., Hssler, H. (2013). Groutability of the rock mass using fracture
statistics. Nordic Grouting Symposium 2013, Gothenburg, Sweden.
Snow, D. T. Rock fracture spacings, openings and porosities. Journal of the Soil Mechanics and
Foundations Division. 1968, Vol 94, 73-92
Stille, H. (2015). Rock Grouting Theories and Application. BEFO, Stockholm, Sweden.
Swedish Transport Administration. (2016-09-07). Retrieved from
http://www.trafikverket.se/nara-dig/Vastra-gotaland/projekt-i-vastra-gotalands-
lan/Vastlanken---smidigare-pendling-och-effektivare-trafik/skiss-sa-bygger-vi-
vastlanken/
Thrn, J., Kvartsberg, S., Runsltt, E., Almfeldt, S., and Fransson, . (2015). Berkningsverktyg
fr bergkaraktrisering vid injekteringsdesign Teori och anvndarhandledning. BEFO,
Report 143, Stockholm, Sweden.
Thrn, J., Kvartsberg, S., Runsltt, E. and Fransson, . (2016). Calculation tool for hydraulic
characterization during grouting design. Nordic Grouting Symposium 2016, Oslo,
Norway.

11