Guideline for preliminary design of Spiles

Jeffrey Oke
Mine Design Engineering, Kingston, Ontario, Canada

ABSTRACT
Umbrella arch systems are defined as a pre-support method that is installed during the first pass of excavation from
within the tunnel, above and around the crown of the tunnel face, that reinforces through the interaction between the
support and the ground. There exist no universal design specifications for umbrella arch systems, however,
recommendations have been made regarding the design components of the eleven sub-categories. Three of such sub-
categories consist of spile elements. This paper will present a guideline for the preliminary design of two of these three
spile elements, when installed to mitigate structure or to manage convergence/strain. These guidelines were developed
from the collection of published empirical data, analytical solutions, empirical verification, laboratory results, as well as
observed limitations/restrictions of installation equipment.

1 INTRODUCTION
An umbrella arch is defined as a pre-support method that
is installed during the first pass of excavation from within Oke et al. (2014a) provided an empirical selection
the tunnel, above and around the crown of the tunnel chart (Figure 1) as well as a range of design properties
face, that reinforces through the interaction between the (Table 1) for each element (spile, forepole, and grout) and
support and the ground (Oke et al. 2014a). Steel the eleven sub-categories of the umbrella arch systems.
elements that are used for the umbrella arch are Oke et al. (2014b) provided some design
described as spiles when they are shorter in length, L, recommendations for umbrella arch systems that employ
than the height of the excavation, He, and described as forepole elements in squeezing ground conditions.
forepoles when they are longer. There currently exist no However, there are still no concrete design guidelines in
universal design specifications for umbrella arch support existence for spile elements. Guidelines for the
systems (Carrieri et al., 1991; Hoek, 1999; Volkmann, preliminary design of spiles will be presented within this
2003; Kim et al., 2005; Volkmann et al., 2006; Volkmann paper.
and Schubert, 2006, 2007, 2010; FHA, 2009; Hun, 2011;
Peila, 2013; Oke et al. 2016). These support systems can Table 1. Suggested range of design parameters for spile
easily be overdesigned, or worse under designed, elements of an umbrella arch system (Modified after Oke,
because of the lack of design standard. An insufficient 2014a). Note: SpCUA: Spile Confined Umbrella Arch
design is easy to identify as it leads to failure, but an SpGUA: Spile Grouted Umbrella Arch.
overdesign is much more different to identify. Such
overdesign of spile elements (L < He) was illustrated in Umbrella Arch with Spiles
Oke et al. (2014a) at the Niagara Tunnel, Canada. The SpCUA SpGUA
over design was minor, as it called for grout to confine the
Length, L 3L -H (3L )
elements in their locations (SpGUA), however, due to the u e u

high stress environment, the ground material squeezed Angle (°) 5-20 (15)
and confined the spiles in location which removed the Overlap (m) L/3-2L/3 (L/3)
requirement for grout. Once this confinement was
Initial center-to-center (cm) <50 (30) <30 (30)
observed, the design was modified the exclude grout
(SpCUA). Similarly, the forepole elements (L > He) at the Diameter (mm) 25-101.6 (50)
Driskos Tunnel, Greece, were slightly over designed, as Grout Pressure (MPa) NA 0.20-1 (1)
shown in Oke et al. 2014b. A study done by Langford et Notes: XX-XX: Range of data (XX): Typical value
al. (2016) echoes this overdesign at the Driskos Tunnel, Lu: Unsupported Span NA: Not applicable
as it was shown that the optimum support systems He: Height of excavation L: Length of support
(without considering the influence of umbrella arch in the
analysis) was typically a design category lower (less 2 BACKGROUND
support installed) than the design category employed.
Figure 1. A condensed version of the UA Selection Chart (UASC) for Convergence Management Ground Conditions and
Structural Control Only. Terminology: Spile Grouted UA (SpGUA); Spile Confined UA (SpCUA); Forepole Confined UA
(FpCUA); and Forepole Grouted UA (FpGUA). (Modified after Oke et al. 2014a).

The umbrella arch system originated over a century ago, 1857 (Szechy, 1966)) steel became the preferred material
when the primary support material was timber and the of support (i.e. introduction of rock bolts). However, the
support installation equipment was powered by humans. first instance of steel-based forepoling/spiling (in pipe
Upon encountering problematic ground conditions form) was not until the mid-1970s in Italy, as cited in
concerning face stability, mining excavation support Carrieri et al. (1991). Oke (2016) illustrates that there has
practices suggested the employment of forepoles been increasing demand for umbrella arch systems with
(wooden boards), as shown in Figure 2, in order to steel and grout elements since their introduction in the
advance through to the next section (Stauffer 1906, 1970’s. An illustration of the modern configuration of spile
Richardson and Mayo 1941, Proctor and White 1946, elements with the remaining typical support elements is
Lewis 1954, and Szechy 1966). Such a support method found within Figure 3.
protected the workers, who were excavating the project
by hand, from the possibility of falling ground material due 3 PRELIMINARY DESIGN OF SPILES
to the newly exposed crown of the excavation. With the The following preliminary design for spile elements was
advancement of technology (such as hydraulic drilling in based on conservative estimations, the response of which
should always be verified in the field or through advanced
calculations (e.g. three-dimensional numerical analysis). It
is important to note that water has not been taken into
consideration for the design. If water is encountered, the
spiles would need to be designed to mitigate the water
flow through ground improvement (a different section of
the UASC, missing from Figure 1).

3.1 Spile Capacity

The bending moment of forepole elements is the most
critical design criteria (Song et al. 2013), and should be
the governing factor for spile elements as well. Therefore
for preliminary design, the simple approach of a fixed
beam with a pin reaction loaded by a uniform distributed
load, UDL, by John and Mattle (2002) will be used to
assess the bending moment capacity (embedded in
Figure 4). The UDL is based on the equation [1] derived
Figure 2. Drift advancement with forepole boards. from Terzaghi silo theory.
Modified after Szechy (1966). Based on case studies, the loading condition
(equation [1]) has been found by Pelia and Pelizza (2003)
 (2003) suggest that the Terzaghi silo theory load should
be reduced by 50-75%. Additionally, Pelia and Pelizza
(2003) state that a simply supported beam (pin-roller
connection) can also be employed to approximate the
bending moment. However, Pelia and Pelizza (2003) do
not specify which analysis configuration to employ for a
given ground condition.

2
௠ ு

      1   
ିଶఒ ௧௔௡ ఝ ௧௔௡ ఝ
௠ 2  
ோ೘
[1]
ு
௦ ∙ 
ିఒ ௧௔௡ ఝ ௧௔௡ ఝ
ோ೘

Where:
3
γ: specific weight of ground [N/m ]
c: cohesion [Pa]
Rm: mean radius of the silo, which should be calculated
assuming the silo as an ellipse with the small diameter
equal to 1.5 times the unsupported span (1.5Lu), and the
large diameter being the width Dt (or excavation width) of
0.5
the tunnel: Rm = 0.5(Dt∙1.5Lu) as shown in Figure 3.
λ: horizontal pressure coefficient, which can be assumed
as λ = 1-sin ϕ
ϕ: friction angle [Deg]
H: overburden [m]
2
qs: Stress on ground surface [N/m ] (if applicable)

It has been found in an investigation by Oke et al.

(2014c) that the analytical solution of John and Mattle
Figure 3. Illustration of typical spiling elements of the (2002) can capture the maximum moment, however, it is
umbrella arch arrangement and other supports layout. mechanistically incorrect. Complex analytical solution,
The assumed silo shape is based on Equation [1] with such as that of Oke et al. (2016), can provide
H=Height of failure, Lu=Unsupported Span. mechanistically correct representation of the support

Figure 4. Spile capacity based on varying height of failure, cohesion, and spile spacing for a constant excavation
width, unsupported span, and friction angle.

to overestimate the loading condition. Pelia and Pelizza,

element. It is important to note that different simple and as they are easily available. Additionally, their connection
complex analytical solutions are available in literature, but capacities (the limiting factor for segmented elements)
it is outside the scope of the paper to discuss them have been researched extensively by Volkmann and his
further. However, it is the author’s opinion that complex collaborators.
analytical solutions should not be used for preliminary The design chart in Figure 4 can be used to quantify
design, due to the inherent high variability of ground the selection of the cross sectional properties required to
conditions. These complex analytical solutions should meet the moment caused by the straining/failing ground.
only be used for verification of response of the support Additionally, the design chart will also quantify the
system when coupled with instrumentation. selection of spacing between the centers of the spile
Figure 4 provides a sample solution which utilizes elements. However, this chart only considers the global
the analytical solution of John and Mattle (2002), with a arching effect (stresses being distributed around the
Terzaghi silo theory load reduction of 50% as suggested excavation) on the spacing, and does not consider the
by Peila and Pelizza (2003). Figure 4 assumes that the local arching effect (stresses being distributed between
excavation width is set at 10m, the unsupported span is each spile element) on the spile elements, as considered
1m, the friction angle is set to 25°, and that there is zero in the subsequent section.
influence of stresses acting on the ground surface. Figure
4 illustrates the influence of increasing the height of 3.2 Spacing and Angle of Installation
failure, which has insignificant influence past 15m The center-to-center spacing of the spile elements is
(because of the assumed zero stress at surface). Figure 4 typically defined by block size, not by geomechanics. A
also shows the influence of cohesion for 0, 19kPa, and geomechancis related solution was proposed in Oke et al
23kPa. Greater than 24kPa, based on assumed silo 2014b, involving a numerical trap door test to assess
shape, the ground will be self-supporting. Additionally. spacing. However, this analysis can prove too complex for
Figure 4 displays the capacity of the different type of preliminary design. An upper bound of 30cm initial center
elements based on DSI specifications (DSI, 2015) to allow to center spacing is empirically/typically chosen, based on
for selection of an appropriate section size. Other spile the possible damage larger blocks could cause.
types are available from different suppliers, but DSI Designers need to additionally understand that the radius
products were used for the bending moment capacities, of the crown, the angle of installation, and deviation of

Figure 5. Spacing and angle of installation assessment based on length between umbrella arches.
drilling (0.5% Mager and Mocivnik, 2000, to 2% John and length between umbrella arches without considering the
Mantle, 2002) will cause the spacing between the spiles required overlap, as explained in the following section.
to increase with length. As illustrated in Figure 5, the
center-to-center spacing can increase to 40cm from an 3.3 Overlap
initial 30cm after 8m of length for a 10m diameter crown The required overlap for an umbrella arch with spile
and a deviation of drilling of 1%. elements depends on the mitigation method they are
However, if the overlap of the umbrella arch spiles is employed for. Within the scope of the design
greater than 1/2 of the length it is possible to increase the recommendation presented in this paper, Spiles are
spacing, as long as the next array of spiles is offset installed to mitigate two different stabilities (Figure 1): i. to
resulting with a spile being installed between two existing increase the stability of the unsupported ground between
spiles and that the relative spacing is maintained. the support and the excavation face, and/or ii. to increase
Additionally, Figure 5 illustrates the required angle of the stability of the tunnel face. If spiles are installed solely
installation based on length and the required support to support the unsupported span, an embedment length,
clearance. Clearly, the smaller the length of the support based on empirical data are typically 1m. However,
element, the higher the required angle of installation. John additional analysis (e.g. wedge stability) needs to be
and Mattle (2002) state that the drilling equipment completed in order to quantify this length.
typically requires 40cm of clearance. If it is both the face and unsupported span that
It is important to note that the angle of installation require increased stability, the overlap can be based on
design (Figure 5) is based on a “niche” or “saw-tooth” the Rankine failure block (based on the material angle of
profile that requires enlargement of the excavation profile friction, Φ, only) as shown by the red dashed line within
as the ground breaks back to the spile elements. In some Figure 6. Figure 6 illustrates the required overlap based
cases, the ground is stable, and the requirement for on different heights of excavation. The grey zone reflects
enlarging the profile is not required. If no enlargement is the range of applicable overlap based on the empirical
needed, then the free end of the spile elements may need design recommendations within Table 1 while remaining
to be trimmed/cut/decoupled if they cannot be installed true to the definition of a spile as having a length of the
flush enough to the excavation crown. element less than the height of excavation.
However, it remains difficult to select the exact The overlap design chart (Figure 6) does not

Figure 6. Required overlap on umbrella arch based on friction angle and height of excavation.
consider any other additional pre-support, such as core- 12. Coverage angle based on the recommendation
reinforcement, nor does it consider the influence of water. within Figure 1.
Engineering judgement will be required when determining 13. Optimization of design parameters (during
the overlap of the umbrella arch composed of spiles when operation):
additional pre-support is employed. a. The center-to-center spacing can be increased if
stability of the ground between spiles is stable.
4 DESIGN METHODOLOGY This increase in spacing can be verified during
The design methodology for spiles are illustrated in Figure operation, or through numerical analysis, as
7 and are as follows: described in Oke et al. 2014b;
1. Define the unsupported span based on: b. The required overlap can be reduced if the face
a. Lattice Girder/Steel Rib spacing; and/or can be shown to be stable, but should not be
b. Rock bolt spacing. less than 1m; and
2. Based on the mode of failure, predict the height of c. If the length of the spile has increased to be
failure if spiles are not installed; greater than the height of the excavation, then
a. If failure depth is unknown, assume the greatest the element should be treated as a forepole. It
of the excavation width or height. can be more economical to employ the longer
3. Solve analytical solution of John and Mattle (2002) elements due to less set up time (Volkmann
(or other simple analytical solution) with a 50% 2016).
reduction of Terzaghi silo theory load as illustrated in
Figure 4: Unsupported Span, Lu Estimate Height of failure
a. For drill and blast operations, assume cohesion • Lattice Girder/Steel Rib spacing • If failure depth is unknown, assume the
• Rock bolt spacing greatest of the excavation width or height
of 0 due to blast damage;
b. For mechanical excavation (i.e. road header),
use the cohesion of the ground material; and Analytical solution of John and Mattle (2002) – Figure 4
c. If close to surface, estimate surface load. • 50% reduction of Terzaghi silo theory load
• Cohesion
4. Select an appropriate spile size that meets the • 0: for drill and blast operations
moment capacity. • Ground material: mechanical excavation (i.e. road header)
5. Estimate the length of the spile element based on • If close to surface, estimate surface load

equipment and standard lengths and assume an

angle of installation of 15°: Spile Cross Section Overlap of Spile, Lo
a. Length of the spile < Height of the excavation; or • Moment capacity • Face stability required:
• Assume 30cm spacing • Rankine Failure (Figure 6)
b. Length of the spile > 3 times the unsupported • Face stability not required:
span. Length of Spile, L • Lo > Lu/3
6. Select the required overlap of spiles: • L < He • Lo > 1m (wedge analysis)
• L > 3 Lu
a. If face stability improvement is required, ensure
overlap is greater than Rankine failure block for
the given height of excavation and friction angle Center-to-center spacing and angle of installation – Figure 5
• The spacing between spiles < 30cm
of the ground material as illustrated in Figure 6; • Support thickness + Jumbo Clearance
b. If face stability improvement is not required, • Assume jumbo clearance = 40cm if unknown
ensure that the overlap of the umbrella arch is
greater than 1/3 of the length of the spile Grout specification (if applicable) Coverage Angle
• Suppliers recommendations • Figure 1
elements. If this value is greater than 1m, the
overlap could be further reduced based on Figure 7. Illustration of the design methodology of spiles.
additional calculations.
7. Select an initial center-to-center spacing so that the 4 DISCUSSION
spacing between spiles (center-to-center spacing Components of the design methodology may not be
minus the diameter of spile) shall not exceed 30cm mechanistically correct, but empirically has worked in the
based on the length between umbrella arch past. In order to further optimize/streamline spiling design
installations (Horizontal length of the spile elements the following items require further research:
minus overlap) as illustrated in Figure 5. • Boundary Conditions
8. Modify length of spile elements, if necessary (check o Quantify when to select different boundary
step 6a). conditions (pin vs fixed reaction) for the
9. Calculate the angle of installation based on the analytical solutions.
length between umbrella arches: • Loading condition
a. Assume jumbo clearance is 0.4m, if unknown. o In which situation can a 75% reduction be
10. Check overlap and length between umbrella arches employed to the loading condition.
based on new angle of installation; go back to step o Harazaki et al. (1998) proposed two different
6a if applicable. loading conditions (non UDL) for alluvial and
11. Grout design specifications (if applicable) should be diluvial material zones. Harazaki et al.
site specific and based on suppliers (1998) state that further research is needed
recommendations. to quantify the linear distributed load.
• Center-to-Center Spacing
 o Spacing currently set based on empirical Design and Construction of Tunnels. Geotechnical
safety guidelines (30cm). Further research and Geological Engineering. (DOI 10.1007/s10706-
is required to quantify when it is applicable 013-9697-4)
to increase the spacing. Oke, J., Vlachopoulos, N. and Diederichs, M.S (2014b).
It is important to note that the reaction forces found from Numerical Analyses in the Design of Umbrella Arch
the analytical solution should be included in the analysis Systems. Journal of Rock Mechanics and
for the selection of proper lattice girder/steel set. Geotechnical Engineering. (DOI
5 CONCLUSION 10.1016/j.jrmge.2014.09.005)
The design methodology presented in this paper allows a Oke, J., Vlachopoulos, N., Diederichs, M.S., (2014c).
designer to easily quantify the design parameters of spile Semi-analytical model of an umbrella arch employed
elements during preliminary design. These design in hydrostatic tunnelling conditions. In: 48th US Rock
parameters should be verified during operations in order Mechanics/Geomechanics Symposium. American
to further optimize the design. Rock Mechanics Association, Minneapolis.
Oke, J., Vlachopoulos, N. and Diederichs, M.S. (2016).
6 ACKNOWLEDGEMENT Semi-Analytical Model for Umbrella Arch Systems in
The author would like to thank N. Boultbee and K. Squeezing Conditions Tunnelling and Underground
Falkenstein for their input and discussion on this paper. Space Technology. 56: 136-156 (doi:10.1016/j.tust.
The author would also like to thank his previous thesis 2016.03.006)
supervisor Dr. N. Vlachopoulos and Dr. M. Diederichs, Oke, J. (2016) Determination of nomenclature,
who provided the author guidance with the previous mechanistic behavior, and numerical modelling
research presented in this paper. optimization of umbrella arch systems. PhD Thesis.
Queen’s University.
Proctor, K., & White, R. V. (1946). Rock Tunneling with
5 REFERENCES Steel Supports with an Introduction to Tunnel
Geology. Youngstown, O.: Commercial Shearing and
Carrieri G, Grasso P, Mahtab A, Pelizza S. (1991) Ten Stamping Co.
years of experience in the use of umbrella-arch for Richardson, H. W., & Mayo, R. S. (1941). Practical Tunnel
tunneling. In: Proceedings of the international Driving. New York and London: McGraw-Hill Book
congress on soil and rock improvement in Company.
underground works; 99-111. Song KI, Cho GC, Chang SB, Lee IM. (2013) Beam-
Dywidag-Systems International (DSI). (2015) AT – Pipe spring structural analysis for the design of a tunnel
Umbrella System. Retrieved 05 02, 2016, from pre-reinforcement support system. International
https://www.dywidag.co.uk/uploads/media/DSI- Journal of Rock Mechanics and Mining
ALWAG-Systems-AT-Pipe-Umbrella-System-en.pdf Sciences;59:139-150.
Federal Highway Administration (FHA). (2009) Technical Stauffer, D. M. (1906). Principles of Tunnel Timbering and
manual for design and construction of road tunnels Driving. In Modern Tunnel Practice. New York:
civil- elements. Retrieved 04 22, 2014, from Engineering news Publishing Co.
http://www.fhwa.dot.gov/Bridge/tunnel/pubs/nhi09010/ Szechy, K. (1966). The Art of Tunnelling (Vol. II).
errata.cfm. Budapest: Akademiai Kiado.
Harazaki, I., Anon, H., Matsuda, A., Aoki, T., Hakoishi, Y., Volkmann, G. (2003). Rock mass - pipe roof support
1998. Field observation of large supported by umbrella interaction measured by chain inclinometers at the
method: case of Maiko Tunnel in Kobe, Japan. In: Birgltunnel. International Symposium on Geotechnical
Tunneling and Metropolises, Proceedings of the World Measurements and Modeling, Proc. Karlsruhe: A.A.
Tunneling Congress, pp. 1009–1014. Balkema. 105-109
Hoek, E. (1999). Putting numbers to geology – an Volkmann, G.M., Schubert W., Button E.A. (2006) A
engineer’s viewpoint. Quarterly Journal of Engineering contribution to the design of tunnels supported by a
Geology, 32(1), 1-19. pipe roof. In: 50 Years of Rock Mechanics Landmarks
Hun Y. (2011) Stability and collapse mechanism of and Future Challenges, Proceedings of the 41st US
unreinforced and forepole reinforced tunnel headings. Symposium on Rock Mechanics. Golden, Colorado;
PhD Thesis. Singapore: National University of Volkmann, G., & Schubert, W. (2006). Optimization of
Singapore; excavation and support in pipe roof supported tunnel
John, M., & Mattle, B. (2002). Design of tube umbrella. section. 32nd ITA-AITES World Tunneling Congress.
Magazine of the Czech Tunnelling Committee and Seoul.
Slovak Tunnelling Association, 3. Volkmann GM, Schubert W. (2007) Geotechnical model
Lewis, R. S. (1946). Element of mining. London: John for pipe roof supports in tunneling. In: Proceeding of
Willey & Sons, Inc. the 33rd ITA-AITES World tunneling congress,
Mager W, Mocivnik J. Modern casing technology sets a underground spaced the 4th dimension of
milestone in drilling and ground anchoring. Felsbau metropolises. London: Taylor & Francis Group; p. 755-
2000;18:43. 760.
Oke, J., Vlachopoulos, N. and Marinos, V.P. (2014a). Volkmann GM, Schubert W. (2010) A load and load
Umbrella Arch Nomenclature and Selection transfer model for pipe umbrella support. In: Zhao J,
Methodology for Temporary Support Systems for the Labiouse V, Dudt JP, Mathier JF, editors. Rock
 mechanics in civil and environmental engineering,
proceedings of the European rock mechanics
symposium (EUROCK) 2010. Rotterdam,
Niederlande: CRC Press/A.A. Balkema; p. 379-382.
Volkmann GM. (2016) The Hirschhagen Highway Tunnel
(BAB 44) in Germany: Pre-Support in Extremely
Difficult and Inhomogeneous Ground Conditions.
World Tunnelling Conference.