Excavation of Swelling Rock by Measuring The Displacement Under Roadbed

Download as pdf or txt
Download as pdf or txt
You are on page 1of 10

Tunnels and Underground Cities: Engineering and Innovation meet Archaeology,

Architecture and Art – Peila, Viggiani & Celestino (Eds)


© 2019 Taylor & Francis Group, London, ISBN 978-1-138-38865-9

Excavation of swelling rock by measuring the displacement


under roadbed

K. Miyazawa & M. Fukushi


East Nippon Expressway Company Limited, Tokyo, Japan

T. Akiyama & H. Kinashi


Obayashi Corporation, Tokyo, Japan

ABSTRACT: Recently, in some in-service tunnels, roadbed heaving caused by swelling and
squeezing has damaged the invert concrete. In many of these cases, it is presumed that road-
bed displacement was undergoing during construction and the displacement should have been
measured and managed. However, it is difficult to accurately grasp the roadbed displacement.
Therefore, in the tunnel under construction, the displacement convergence of the whole tunnel
is judged from the inner section displacement of the arch part and the like. Then, we newly
developed a system which can accurately measure the roadbed displacement during construc-
tion, and applied it to the Sankichiyama Tunnel. It was consequently discovered that only the
roadbed was significantly raised after inner section displacement of the arch part converged.
In this paper, we report the case of roadbed displacement measurement using this new system,
and also quantitatively evaluate the measures against heaving by analytical method using
measurement data.

1 INTRODUCTION

The Tohoku-Chuo Expressway is a 268-km national automobile highway running from


Soma City in Fukushima Prefecture through Fukushima City, Yonezawa City, and Yama-
gata City in Yamagata Prefecture, before arriving at its northern terminus in Yokote City,
Akita Prefecture.
The Sankichiyama Tunnel (hereafter “this tunnel”) is a 2980m long provisional two-lane
highway tunnel part of Tohoku-Chuo Expressway located 7–13 km to the south of Yamagata
Station on the JR Ou Main Line. It is the longest tunnel in the section from Nanyo Takahata
to-Yamagata Kaminoyama of the Expressway.
In this tunnel, localized pockets of springs and tuff breccia containing large quantities of
expansive clay minerals were found during excavation from the tunnel face, and inner dis-
placement of more than 300 mm occurred. Therefore, remedial construction work was con-
ducted to converge the arch inner displacement and breakthrough such sections.
However, in some cases with other tunnels already in use, long-term roadbed bulging
occurred as an effect of ground swelling, and the invert had to be reconstructed. To avoid
such repair works after commencement of service, a newly developed invert displacement
meter was additionally applied to this tunnel for the measurement of roadbed displacement
during construction. This measurement revealed that the roadbed was rapidly bulging even
when the inner displacement was moving toward convergence.
This paper reports the results of measurement of roadbed displacement during construction
and the application of these results in remedial construction to rectify the roadbed bulges.

992
Figure 1. Geological profile.

2 SUMMARY OF GEOGRAPHICAL AND GEOLOGICAL CONDITIONS

This tunnel is located in the west wing of Mount Zao, a member of the Ou mountain range,
which bisects the Tohoku region from east to west. “Mount Zao” is the general name for a
belt of volcanoes along the boundary between Yamagata and Miyagi prefecture. The tunnel
passes through the foothills of a mountain spur that splits off the west wing of the ridgeline of
this volcano belt and the alluvial fan formed in the adjoining valley with a shallow rock cover-
ing. The rock covering the tunnel is 156.3 m deep at the thickest point and 14.5 m deep at the
thinnest point, excluding the tunnel mouths.
The bedrock in the area is composed of the Yoshino Layer, dating from the Miocene Epoch
of the Neogene Period, and the Hayama Layer, dating from the Pliocene Epoch, covered by a
quarternary stratum on the surface. The Hayama Layer consists of strata of tuff breccia layers
(Htb) and andesite layers (Han). The tunnel cuts through these layers; near the middle, the
tunnel encounters extrusions of rhyolite (Yry) from the Yoshino Layer (Figure 1).
The andesite is charcoal grey in color and is slightly porous but very hard; it is used to develop
cooling joints such as columnar joints and block joints. The lighter grey tuff breccia is a non-
stratified block composed of andesite fragments, ranging in diameter from a few centimeters to a
few dozen centimeters, locked in a weakly compacted matrix, and while the rock fragments
themselves were soft, the bedrock was deemed favorable for supporting the tunnel face.

3 STATUS OF DISPLACED SECTIONS

3.1 Topography/geology
The displacement in this tunnel occurred along a 65.4-m extension (STA.205+16.80-STA.205
+82.75) approximately 2300 m from the tunnel mouth. The rock covering this section is 66 m
deep at its thickest point, and during the design phase, it was predicted that this section would
be composed of hard andesite, with an elastic wave velocity of 4.0–4.2 km/s. However, during
excavation, localized protrusions of tuff breccia (Htb), which were not predicted during the
design phase, appeared crossing the tunnel face (Figure 2). This tuff breccia was greenish grey
in color, and the matrix had been transformed into clay through hydrothermal alteration, ren-
dering its characteristics evidently different from those of the tuff breccia (Htb), which had
appeared previously. In addition, approximately 200 L of spring water/minute flowed out of
the juncture with the andesite (Han) layers, causing frequent cave-ins at the tunnel face and a
steep drop in the stability of the cutting surface.

993
Figure 2. Geological condition in the displaced section.

Figure 3. Deformations in the sprayed Figure 4. Side wall displacement measurement results.
concrete.

3.2 Tunnel support distortion


Although no significant distortion was detected in November 2016, soon after this section of
the tunnel was excavated, after a while, various kinds of deformation were detected in the
tunnel supports, including cracks in the sprayed concrete and distortion of the rock bolt plates
(Figure 3). Most of these deformations were concentrated in the side walls, and when the loca-
tion of the upper support feet of the steel arch was measured, a maximum deformation of 381
mm to the inner side was observed (Figure 4).

3.3 Consideration of displacement status


To investigate the causes of displacement in this section of the tunnel, the bedrock
and support deformation status at the time of tunnel excavation were categorized as
follows.
(1) The bedrock in this section of the tunnel was Neogene tuff breccia containing a large
amount of spring water.
(2) There was a correlation between the distribution of the tuff breccia and the range of
deformation in the steel arch supports.
(3) Displacement was concentrated in the side walls, and the inner space displacement along
the horizontal traverse line exceeded the roof subsidence.
Accordingly, it is believed that the displacement in this section of the tunnel was caused by
multiple factors, including stress release from excavation of the Neogene tuff breccia and
swelling and slaking owing to the spring water.

994
3.4 Displacement rectification work
Rock bolts are effective against ground swelling, and can restrain the ground if poured as
early as possible after excavation. However, it is considered that, in conjunction with the
increased looseness of the ground, the pattern bolts (3.0 m) poured during excavation were
not sufficiently effective, and longer rock bolts (6.0 m) were deemed necessary. As deform-
ation was greater along the horizontal traverse line, three longer rock bolts were poured
between each of the pattern bolts on the left and right side walls.
In addition, the sprayed concrete invert was installed at a thickness of 20 cm at the bot-tom
edge of the invert concrete. While tunnel supports are not closed by invert concrete, structural
resistance is not sufficient to retain the ground. Therefore, it was believed that the tunnel sup-
port needed to be closed as quickly as possible. Then, sprayed concrete in-vert, which is easily
workable and can quickly develop support function, was used to close the cross-sectional sup-
port structure. The remedial construction is shown in Figure 5.
Figure 6 shows the results of the STA.205+65 measurements. Immediately after excavation,
displacement from the inner displacement had already surpassed 50 mm (management level
III). The additional longer rock bolts reduced the speed of the displacement slightly, but
owing to an increase in the creep displacement regardless of the distance from the tunnel face,
the sprayed concrete invert was introduced to stop the displacement.
Ultimately, the inner displacement showed a convergence trend, and the displacement did
not increase even after tunnel excavation was resumed. Therefore, the ground was deemed sta-
bilized, and back-stitching was conducted to secure the inner needed space.

3.5 Swelling ground determination test


Tunnels built in swelling ground indicate many specific behaviors different from regular tun-
nels, and hence, it is desirable that a detailed ground survey be conducted to gather da-ta for
use during construction and after the highway is opened to the public.

Figure 5. Diagram of remedial construction during tunnel excavation.

Figure 6. STA.205+65 measurement results.

995
Figure 7. Octagonal diagram of laboratory rock test results.

Among methods for judging whether an invert should be installed in relatively hard CII
bedrock, an index to determine swelling potential has been proposed (Ohshima et al. 2001).
According to this proposal, the laboratory rock test was conducted for this tunnel to deter-
mine the swelling potential of the ground via the shape of an octagonal diagram showing the
level of water impact and amount of clay minerals contained. Tests were conducted on two
samples of tuff breccia (①, ②) and one sample of andesite.
The results of the tests are shown in Figure 7. The results of both X-ray diffraction and CE
tests show very high levels of expansive clay minerals (montmorillonite: Ca-type) throughout
the samples, and the slaking erosion degree also exceeds the standard values. In addition, the
moisture absorption rate of the tuff breccia (①, ②) shows a value approximately equal to the
standard, which is relatively higher than that of the andesite (③). Thus, it was judged that, for
this section of the tunnel, the possibility that the tuff breccia would constitute swelling ground
was high. While, it is believed that the andesite did not occur tunnel displacement owing to its
high strength level of the rock fragments.

4 INVERT DISPLACEMENT MEASUREMENT

4.1 Additional measurement process


In some previous cases, on both highways and railways, roadbeds in public tunnels have
bulged over long-term use owing to the effect of ground swelling, and remedial construction
or invert repair has become necessary. In highway tunnels, invert repair or new installation is
necessary, which has an adverse effect on road users owing to the need to restrict lanes or stop
traffic (Miyazawa et al. 2016).
To prevent this problem, invert bulging must be assessed and remedial construction under-
taken while the tunnel is still under construction. Conventionally, displacement convergence for
the entire tunnel was estimated based on the roof subsidence and horizontal inner displacement
measurements obtained as part of the usual A-measurements. Because it is difficult to measure
invert displacement using methods that involve an everyday reference point, as heavy machinery
and vehicles pass through the tunnel during construction. However, for tunnels in swelling
ground, such as the one detailed in this report, even when displacement convergence is confirmed
through A-measurements, displacement of the roadbed alone was observed to progress markedly.
In most cases where roadbed displacement actualizes after the tunnel is opened to the public, it
was estimated during construction that roadbed displacement might already be in progress.

996
Figure 8. Outline diagram of invert displacement meter.

Owing to these factors, a system was developed to measure roadbed displacement during
construction without posing a hindrance to the passage of heavy machinery and vehicles, and
the new system was practically applied to this tunnel.

4.2 Measuring system outline


An outline of the newly developed invert displacement meter is shown in Figure 8. This meter
is buried in the protection tubing around the water pipe connected to the water pressure gauge
located under the roadbed during construction and measures the amount of bulging based on
the head differential with the standard water tank installed above ground. This system allows
for continuous, automatic measurement of vertical displacement of the ground under the
invert. In addition, by measuring the height of standard water tank with a total station, it is
possible to measure the level of absolute displacement.

4.3 Meter installation process and measurement accuracy


The invert area was excavated, and protective tubing was installed for later re-burying in
earth or in sprayed concrete invert. After re-burying, the meter was inserted into the protective
tubing, and electric lines were connected to the meter.
In advance, meter accuracy was verified through laboratory testing. The invert dis-place-
ment meter was subject to an incrementally increasing level of displacement, and when the
measurement values were compared with and verified against the measurements obtained
from a high-precision CDP displacement meter, the margin of error of the invert displacement
meter was less than 0.5 mm (0.5%) per 100 mm of deformation.

5 RESULTS OF MEASUREMENTS IN THIS TUNNEL

In this tunnel, invert displacement meters were installed at two locations (STA.205+55,
STA.205+73) where large displacement occurred at the time of excavation (STA.205+55,
STA.205+73). The results of the measurements are shown in Figure 9. Bulging exceeding
60 mm was recorded over a 50-day period, and cracks and other damage were observed when
the status of the sprayed concrete invert was checked.
In addition, the system also included installation of a meter to measure the distribution of
displacement, which attempted to measure vertical displacement every 0.5 m. The results, as
shown in Figure 10, demonstrate that the greatest amount of displacement was observed in
the center portion of the invert span. Thus, it was observed through this measurement that it
would be appropriate to install a sensor in the center, where the displacement was greatest.

997
Figure 9. Invert displacement measurement results.

Figure 10. Invert area vertical displacement distribution.

6 REMEDIAL CONSTRUCTION FOR ROADBED BULGING

6.1 Remedial construction selection


Based on these measurement results, it was deemed that bulging was progressing in this
section of the tunnel, and it was predicated that, if the tunnel was left as it was, remed-
ial construction would be required after it was opened to the public. Therefore, remedial
construction was undertaken.
To suppress the roadbed bulging, it was necessary to create a structure that could op-pose
the load (swelling pressure) on the invert. First, past experience for correcting invert dis-
placement were referred, and as shown in Figure 11, a design was chosen whereby the shape
of the excavated cross-section would be made more circular, and after sealing off the area
with high-spec steel supports (HH-100) and sprayed concrete invert (thickness 150 mm),
the displacement would be suppressed by the weight of invert concrete with a maximum
thickness of approximately 1.0 m.

6.2 Remedial construction verification process flow


Subsequently, an analysis of the appropriateness of the design was conducted through the pro-
cess shown below.

Figure 11. Diagram of remedial construction for roadbed puckering.

998
First, the residual displacement within the primary seal-off construction was calculated
using the Voigt approximation formula. Second, the strength of future external force was pre-
dicted by frame structure analysis. Assuming that the external load would added uniformly on
the arch area and invert, the load value was calculated inversely which occurred the residual
displacement calculated above. Finally, the load was applied to the post-repair construction
design to check the initiation stress of the components.

6.3 Predicting the amount of residual displacement


Based on the past measurement results, it was believed that the amount of invert deformation
over time would converge to a certain value over an unlimited time period. These changes
over time can be expressed through a Voigt model (Yamaguchi & Nishimatsu 1977), in which
springs and dashpots are arranged in parallel. This model expresses time-dependent elastic
deformation for a given amount of stress, as expressed by the following formula:


σ ¼ Eε þ η ð1Þ
dt

where E represents the spring constant and η is the viscosity.


Time-dependent action under a given level of stress σ0 is shown in the following formula:
σ0

1  e  η t ¼ ε0 ð1  e αt Þ
E
ε¼ ð2Þ
E

where ε0 shows strain at t=∞. By using Equation 2, it is possible to predict the displacement of
an invert in swelling ground over time. The forecast below was generated by substituting
load–displacement in place of stress–strain.
The change in displacement speed over time based on the results of this round of measure-
ments (Figure 9) is shown in Figure 12. The curves shown are matched as closely as possible
to the actual data for the two locations.
Based on the measurements from cases where an invert was repaired after the tunnel was
opened to the public, displacement speed ranges from a few millimeters per year to a few
dozen millimeters per year, approximately 1/10 the speed of the displacement in this tunnel.
Thus, it is believed that the amount and speed of displacement during construction were high
and were therefore easier to measure.
When this displacement speed approximation formula is integrated to calculate a displace-
ment–time relationship formula corresponding to Equation 2, a curve can be generated as
shown by the dotted line in Figure 13. Based on these results, the highest value predicted for
future residual invert displacement is shown to be 46 mm (Figure 14).

6.4 Remedial construction verification results


In the frame analysis, when multiple uniform loads were set to act on supports and an analysis
was conducted, the relationship between acting load–displacement and acting load–strut

Figure 12. Change in displacement speed over time and best-fit curves.

999
Figure 13. Change in displacement speed over time and best-fit curves.

Figure 14. STA.205+55 residual displacement forecast results.

Figure 15. STA.205+55 strut initiation stress.

initiation stress was observed to be nearly linear (Figure 15). From these results, it was
observed that the strut initiation stress in the event of the forecast maximum residual displace-
ment of 46 mm was approximately 60% of the yield stress. In addition, the sprayed concrete
initiation stress was approximately 20% of the design standard strength, thus verifying the
appropriateness of this remedial construction.

7 MEASUREMENT RESULTS FOLLOWING ROADBED PUCKER REMEDIATION

Subsequently, remedial construction was undertaken according to the design in Figure 11.
The spring water that appeared during the remedial construction work at the time of tunnel
face excavation was drained, causing a drop in the underground water table, and hence, no
spring water was detected in this round of construction.
After remedial construction on the roadbed bulging was completed, invert displacement
was measured, along with invert strut stress. Figure 16 shows the measurement results for the
significant bulging trend shown in Figure 9 alongside the measurement results following
remedial construction on roadbed bulging. It can be observed that the remedial construction
suppressed invert displacement to a significant degree. In addition, as shown in Figure 17,

1000
Figure 16. STA.205+55 invert displacement measurement results following roadbed pucker remediation.

Figure 17. STA.205+55 strut initiation stress measurement results.

though a maximum bending tensile stress of approximately 100 N/mm2 is initiated on the
center periphery invert struts, it is less than 25% of the yield stress, and this value is only
approximately 38% of the predicted value shown in Figure 15, and hence, the remedial con-
struction can be deemed sufficiently effective.

8 CONCLUSION

Significant roadbed bulging on swelling ground was measured during construction using the
newly developed invert displacement meter. Although remedial steps were taken at this loca-
tion at the time of tunnel excavation and inner displacement was brought to convergence,
these measurements demonstrated that the roadbed would continue to bulge significantly over
long-term use. Originally, ground displacement convergence was judged through A-measure-
ments, but this case study fully demonstrates the importance of the in-vert displacement
meter. In addition, it is believed to be very important for displacement to converge in order to
prevent roadbed deformation after the public opening of the tunnel by measuring invert dis-
placement at the points where large displacement occurs during tunnel excavation and under-
taking remedial construction reflecting these results during construction.
Breakthrough on this tunnel was reached on April 26th, 2017. It is hoped that the know-
ledge gained from this study will help accelerate the opening of this tunnel to Tohoku Chuo
Expressway traffic and be of use in the construction of future similar tunnels.

REFERENCES

Ohshima, K., Shiroma, H., Itoh, T., Muraji, E. & Kubota, T. 2001. Suggestions for Invert Installation
Standards Based on a Cause Analysis of Deformation in Tunnels. Proceedings of the Japan Symposium
on Rock Mechanics 11(1):329-334.
Miyazawa, K., Yasuda, K., Suyama, K. & Watanabe, A. 2016. Installation of New Invert Using the
Whole-Surface Continuous Pressing Method under Day and Night Continuous Traffic Stoppage, Tun-
nels and Underground 47 (8):7-18.
Yamaguchi, U. & Nishimatsu, H. 1977. An Introduction to Rock Mechanics, Version 2, Tokyo: University
of Tokyo Press.

1001

You might also like