Tunnelling and Underground Space Technology 50 (2015) 68–78

Contents lists available at ScienceDirect

Tunnelling and Underground Space Technology

journal homepage: www.elsevier.com/locate/tust

Excavation cycle times recorded during sinking of a deep shaft in

crystalline rock – A case example at Ventilation Shaft of Mizunami URL,
Japan
H. Sanada a,⇑, T. Sato a, Y. Horiuchi a, S. Mikake a, M. Okihara b, R. Yahagi b, S. Kobayashi b
a
Japan Atomic Energy Agency, 1-64 Yamanouchi, Akiyo, Mizunami, Gifu 509-6132, Japan
b
Shimizu Corporation, 2-16-1, Kyobashi, Cyuoku, Tokyo 104-8371, Japan

a r t i c l e i n f o a b s t r a c t

Article history: Realistic and properly planned excavation cycle times developed prior to tunnel construction are an
Received 2 September 2014 important issue for dependable project management, including cost estimation and time saving.
Received in revised form 1 May 2015 Existing design studies, conducted to estimate the total cost for geological disposal, have been limited
Accepted 30 June 2015
to estimates of approximate construction total time schedules. There were no case studies which com-
pared the detailed excavation cycle time used in design studies with the actual results from an excavation
taking into account each excavation cycle operation and task. In this paper, analyses of actual cycle time
Keywords:
recorded during the sinking of the Ventilation Shaft of the Mizunami Underground Research Laboratory
Geological disposal
Mizunami URL project
and comparison with the planned cycle time developed in the design stage are intended for an evaluation
Crystalline rock of the baseline work plan and schedule and for optimization of the calculations for application to shaft
Shaft sinking construction in any future repository in Japan. Actual times to perform three operations, drill and blast,
Excavation cycle time muck removal and concrete lining emplacement were much larger than expected and represent the lar-
gest portion of the total excavation cycle time and had the largest impact on the efficiency of the exca-
vation cycles and schedules. Operating at increasingly greater depths and excavating harder, more
competent rock combined to significantly affect the shaft sinking speed. The underlying reasons for
increases in actual times for the excavation cycle operations compared to their design estimates fall into
three categories: constraints on construction; to ensure safety; and maintenance.
Ó 2015 Elsevier Ltd. All rights reserved.

1. Introduction
Realistic and properly planned excavation cycle times (here-
after, Ect) developed prior to tunnel construction are an important
issue for dependable project management, including cost estima-
tion and time saving. Especially, if the design phase Ect signifi-
cantly underestimates the actual Ect, there could potentially be a
huge, negative impact on the project, i.e., potentially, a big financial
loss in a large tunnel project such as one for the geological disposal
of HLW (High-level radioactive waste), with inherently huge
underground excavations in a rock mass. Therefore, minimizing
Ect differences between plan and actual are required in design
studies of tunnel excavation.
Existing design studies (Harmon and Lazur, 1984; Kim and Choi,
2006; Kukkola and Saanio, 2005; NIREX, 2005; Read, 2011),
⇑ Corresponding author.
E-mail addresses: sanada.hiroyuki@jaea.go.jp, sanada0924@hotmail.com
(H. Sanada).
conducted to estimate the total cost for geological disposal, have
been limited to estimates of approximate construction total time
schedules. There were no case studies which compared the
detailed Ect used in design studies with the actual results from
an excavation taking into account each excavation cycle operation
and task as discussed in this paper. Moreover, recent examples of
the excavation of deep shafts similar in size to those for HLW
repositories are decreasing in Japan since there have not been
any new coal mines developed in Japan for over 20 years
(Sakurai et al., 2009).
In this paper, analyses of actual Ect recorded during the sinking
of the Ventilation Shaft of the Mizunami Underground Research
Laboratory (MIU) and comparison with the planned Ect developed
in the design stage are intended for an evaluation of the baseline
work plan and schedule and for optimization of the Ect calculations
for application to shaft construction in any future repository in
Japan. The comparison includes not only a review of the Ect but
also a review of the method and underlying assumptions used to

http://dx.doi.org/10.1016/j.tust.2015.06.012
0886-7798/Ó 2015 Elsevier Ltd. All rights reserved.
 H. Sanada et al. / Tunnelling and Underground Space Technology 50 (2015) 68–78
calculate the cycle times in the design process and comparison
with the actual construction results.
The Mizunami Underground Research Laboratory (MIU), an
‘‘off-site’’ URL, constructed and operated by Japan Atomic Energy
Agency (JAEA) at Mizunami City in Gifu Prefecture, central Japan,
involves a comprehensive research program of investigations in a
deep underground crystalline rock environment for R & D into geo-
logical disposal of High Level radioactive Waste (HLW). The MIU is
being developed and operated in three overlapping phases:
Surface-based Investigation phase (Phase I), Construction phase
(Phase II), and Operation phase (Phase III) with a total duration
of around 20 years. The overall project goals of the MIU Project
in Phases I through III are: (1) to establish techniques for investiga-
tion, analysis and assessment of the deep geological environment
and (2) to develop a range of engineering technologies applicable
to deep underground environments in Japan (Saegusa and
Matsuoka, 2011).
The MIU consists of a circular Ventilation Shaft (4.5 m diameter)
and a Main Shaft (6.5 m), both with 0.4 m thick concrete lining, and
horizontal tunnels connecting the shafts every 100 m depth
(Matsui et al., 2011). Excavation of the Ventilation and Main
Shafts reached GL-500 m in FY2011. During Phase I, the design
work for surface and underground facilities in the MIU was com-
pleted with the aim of providing a foundation for engineering tech-
nologies to safely construct and operate a deep underground
facility built for the purpose of data acquisition and demonstrating
disposal technologies (Saegusa and Matsuoka, 2011). The design
work incorporated development of Ect estimates prior to MIU con-
struction as a basis for estimating overall costs and schedule
requirements for the proper operating management of the MIU.
Repositories for the geological disposal of HLW will include ver-
tical access shafts and/or ramps and disposal rooms, which are an
assemblage of disposal tunnels comprising the underground facil-
ities (NUMO, 2013; Svensk Kärnbränslehantering AB, 2009).
Vertical access shafts will be utilized to transport the HLW, buffer
and backfill materials and for mucking operations during construc-
tion and for access to the repository. Use of vertical shafts will be
essential from the viewpoint of ventilation and personnel access
and vertical shafts make it possible to effectively excavate at
increasing depths during the construction phase and to transport
the HLW to the disposal panels during the operation phase.
Discussion of the cycle time during shaft sinking is important for
the future geological disposal project, therefore the authors under-
took the challenge of discussing, recording and validating the Ects
in the MIU with the goal of providing relevant input to shaft con-
struction in a future HLW geological disposal project.
2. Construction of the MIU
2.1. Geological setting of the study site
The bedrock at the MIU site consists of the widely distributed
Toki granite (Upper Cretaceous to Lower Paleogene) and overlying
Miocene sedimentary rocks, the Toki, Akeyo and Hongo
Formations of the Mizunami Group. Pleistocene gravel of the Seto
Formation overlies the bedrock (Tsuruta et al., 2013a). Fig. 1 pro-
vides a vertical cross-section of the Ventilation and Main shafts
and horizontal Sub Stages connecting both shafts in the MIU below
the unconformity at GL-168 m between the Mizunami Group and
the Toki granite. The general geology of the Main shaft is also shown,
indicating the distribution of faults and fault zones. There are many
differences in the geological conditions in the Main and Ventilation
Shafts with respect to hydrothermal alteration. Intrusive rock (lam-
prophyre) exists along the fault plane only in the Main Shaft.
69
Extensive hydrothermal alteration of the intrusive rock and in the
fault zones in the Main Shaft is evident by the abundance of clay
minerals. Rock strength, i.e., the uniaxial compressive strength, is
usually less than 3 MPa on average, in the hydrothermally altered
zones in the Main Shaft. This is much less than in the rock from
the Ventilation Shaft where it is in the order of 173 MPa (Hirano
et al., 2009; Tsuruta et al., 2009). In addition there is no soft,
hydrothermally-altered rock as in the Main Shaft (Fig. 1). This paper
discusses Ect during sinking of the Ventilation Shaft where sound,
high strength granite, without hydrothermally-altered, low
strength rock, is distributed.
From the pattern of structural development and spatial distri-
bution of fractures, the Toki granite in the MIU is divided into
two domains, namely the Upper Highly Fractured Domain
(UHFD) and the Lower Sparsely Fractured Domain (LSFD)
(Tsuruta et al., 2013b). According to Tsuruta et al. (2013b), there
is a high frequency of low angle fractures in the UHFD due to pres-
ence of sheeting joints. This is reflected in the results of well log-
ging during Phase I in which investigations indicated there is
high porosity in the UHFD. Fracture frequency is lower below
GL-460 m, the boundary between the UHFD and the LSFD in the
Ventilation Shaft (Fig. 2).
2.2. Excavation method and procedures
The excavation cycle and component operations are illustrated
in Fig. 3 and are as follows:
(i) Drill and blast operations
A shaft jumbo with two hydraulic drifters is used for drilling
blast holes. After they are drilled, rock mass blasting is done using
non-electric detonators and water gel explosives. All equipment
(e.g. dewatering pumps) are removed from the shaft, and protec-
tive mats are placed prior to blasting to prevent muck scattering.
Moreover blasting is only carried out after the scaffold has been
raised from 20 to 30-m above the shaft face to prevent damage
from fly-rock and pressure shocks induced by blasting to the scaf-
fold. After blasting, a fan at surface in the soundproof housing over
the Ventilation Shaft and local fans in the Sub Stages were used to
ventilate the air in the shaft. Workers are allowed to enter five
minutes after blasting.
(ii) Muck removal operations
The shaft mucker attached to the bottom of the scaffold and the
kibble removes muck after each drill and blast cycle for disposal at
surface. A mini-backhoe is transported from the surface to the
shaft face for scaling to dislodge any loose rock from the shaft wall
and trim rock from the walls to the design shaft diameter. And then
the kibble, filled with muck, is hoisted from the shaft face to sur-
face. Muck is deposited in a temporary yard through the chute in
the derrick over the shaft before loading into dump trucks for
transport off-site.
(iii) Shaft wall mapping, drainage mat installation
Geologists map the shaft wall by sketching and photographing
fractures on the shaft wall and recording information on lithologies
and structures and thus obtain data necessary for rock mass classi-
fication (e.g. RMR (Bieniawski, 1989) etc.) and hydrogeological
characterization. Drainage mats installed along the shaft wall help
to remove any groundwater inflow into the shaft. Anchor bolts are
installer quickly with two to secure the required piping.
70 H. Sanada et al. / Tunnelling and Underground Space Technology 50 (2015) 68–78

Fig. 1. Cross-section view of the Ventilation and Main Shafts with generalized geological map of the latter.

(iv) Concrete lining rounds, and the other activities mentioned above. First, the con-
crete formwork still in place from the previous pour is removed
Concrete lining, 2.6-m high and 0.40-m thick, is poured in place and moved up and out of the way for cleaning. Then the cleaned
against the shaft wall using slip forms after the two 1.3-m blast slip forms are moved into place using an electric winch prior to
 H. Sanada et al. / Tunnelling and Underground Space Technology 50 (2015) 68–78
Fig. 2. Fracture map in the Ventilation Shaft spanning the boundary between the
UHFD and LSFD.
71
pouring the new concrete lining. Concrete, transported by truck
from the surface mixing plant is placed in the kibble for transport
to the hopper in the shaft scaffolding.
3. Excavation cycle times in design
Ect derived in planning during the design phase was based on
construction results for shafts in Japan (e. g. Ventilation Shafts in
express highways for Japan Highway Public Corporation) and
interviews with several contractors in Japan. Ect in the design
phase also took into consideration the rock mass classification
developed by the Central Research Institute of Electric Power and
referred to as the ‘‘Denken system’’ in English (Japanese
Geotechnical Society, 2004; Cai et al., 2004). According to the
‘‘Denken system’’, rock mass in the Ventilation Shaft in the MIU
is categorized as ‘‘CH class’’ between GL-200 m to GL-460 m depth
and ‘‘B class’’ between GL-460 m to GL-500 m depth. Thus ‘‘CH
class’’ and ‘‘B class’’ correspond to the UHFD and LSFD, respectively.
Fracturing in ‘‘B class’’ is less well developed than it is in ‘‘CH
class’’; in other words, the rock mass condition in ‘‘B class’’ is better
than ‘‘CH class’’.
Table 1 is a summary of the planned and actual Ect in the gran-
ite area of the Ventilation Shaft. Table 1 provides the excavation
cycle in terms of a work breakdown structure, that is, how the
work was broken down into work packages (operations) and tasks
as follows.
(a) Drill and blast: (1) preparation, (2) cleanup, (3) drilling blast
holes using shaft jumbo, (4) blasting (includes placing explo-
sives) and ventilation after blasting.
(b) Muck removal: (5) removal of fly-rock from lining, (6) prepa-
ration, (7) muck removal using shaft mucker and kibble, (8)
survey work, (9) cleanup, (10) shaping bottom wall.
(c) Shaft wall mapping: (11) clean shaft wall, (12) preparation,
(13) shaft wall mapping, (14) cleanup.
(d) Drainage mat: (15) preparation, (16) installing drainage mat,
(17) cleanup.
(e) Concrete lining: (18) clean and install formwork, (19) prepa-
ration, (20) pumped concrete, (21) cleanup.
(f) Other operations: (22) additional operations and/or tasks not
included in the above, and not done in every excavation
cycle (e.g. niche excavations) or auxiliary operations (e.g.
scheduled and unscheduled maintenance).
4. Recorded excavation cycle times
Comparison of the planned Ect and the actual excavation
results, on average, are provided in Table 1. Fig. 4 illustrates Ect
for the operations performed in each 2.6-m excavation cycle and
others at specific locations in the shaft and how the total times var-
ied with depth. The main difference in the planned cycle times
developed in the design phase compared with the actual is in
two of the main activities (#3) Drilling and (#7) Mucking opera-
tions in the UHFD and LSFD.
With respect to (#3) Drilling, since the number of blast holes
per m2 of excavated area required for effective rock breakage
was 3.0 in the UHFD and 3.2 in the LSFD per cycle, excavating in
the LSFD required more overall cycle time than in the UHFD.
With respect to mucking operations, since the volume of broken
rock to be removed (muck volume) was 109.4 m3 in the UHFD
compared to 112.7 m3 in the LSFD per cycle, mucking in the LSFD
added more to the overall cycle time than in the UHFD. The sum
of the average recorded actual times for all tasks in the UHFD
72 H. Sanada et al. / Tunnelling and Underground Space Technology 50 (2015) 68–78
Fig. 3. Shaft sinking procedures using the short-step method at the MIU.
and the LSFD is 3507 min and 3910 min, respectively, and the
actual work time was twice as long as the estimate in the design
plan of 1819 min (Table 1). In particular, the total time required
was long in special excavations such as for shaft insets, that is,
the locations where the shaft and horizontal sub-stages intersect,
and for pump niches, and for auxiliary work such as for pipework
extensions and maintenance work (Fig. 4).
The proportion of actual cycle times recorded, the sum of aver-
ages, is drill and blast (20%), muck removal (44%), shaft wall map-
ping (7%), drainage mat (2%), concrete lining (15%) and other
operations (12%) for each of the main operations as a percent of
the total Ect from GL-200 m to GL-500 m. The operations that took
the longest, occupying the largest area are drill and blast, muck
removal and installation of concrete lining. These three operations
have a major influence on total Ect and hence on schedule. Almost
all actual operations and tasks took longer than estimated in the
design plan (see percent increase of result to design in Table 1).
Fig. 5 shows the relationship between excavation depth and
date, and average excavation rate per month during sinking of the
Ventilation Shaft. The average excavation rate per month shown
in Fig. 5 does not include special operations such as construction
of shaft insets, pump niches, and grouting. The rates between
GL-200 m to -300 m, GL-300 m to -400 m, GL-400 m to -460 m,
and GL-460 m to -500 m were 18.3 m/month, 19.3 m/month,
17.0 m/month, and 13.6 m/month, respectively (Fig. 5). Therefore,
the excavation rates deeper than GL-300 m decreased as depth
increased, and thus extended the schedule. In particular, the exca-
vation rate below GL-460 m, was slowest of all. It was assumed that
both increased depth and more competent rock mass conditions
contributed to reduction in excavation rates.
Fig. 6a presents the actual drill and blast times with depth.
Average times in the UHFD and the LSFD are 676 min (137% of
plan) and 871 min (172% of plan), respectively (Table 1). In the
Ventilation Shaft, depth increased 2.6-m every excavation cycle,
with each cycle consisting of two consecutive 1.3-m drill and blast
rounds. Depending on the shaft sinking progress after each excava-
tion cycle, adjustments could be made to the number and length of
blast rounds and thus adjusting the excavation cycles to maintain
consistency in the shaft deepening progress. Concretely speaking, if
excavation progress in two consecutive 1.3-m blast rounds sub-
stantially exceeded 2.6 m, the number of blast rounds was reduced
from two to a single blast round in the next cycle with length of the
blast round increased in proportion to the extra length of the prior
excavation cycle, thus ensuring that excavation progressed in
2.6-m deep excavation steps. Times less than the design times
shown in Fig. 6a correspond to single-blast cycles (purple circles
in Fig. 6a). Average time was 749 min for the double blasting.
Special activities such as pump niche and shaft inset construction
required more time. The average of 871 min in the LSFD is longer
in comparison to the 676 min in the UHFD (Fig. 6a). The reasons
could be due to a combination of excavating in more competent,
intact rock due to decreased fracture frequency below GL-460 m
 H. Sanada et al. / Tunnelling and Underground Space Technology 50 (2015) 68–78 73

Table 1
Summary of design estimates and average actual work time in minutes for each task in an excavation cycle. Average time calculated does not include time for special operations
such as construction of shaft insets, pump niches and extension of pipe work.

Design Result
GL 200 m to 460 m GL 460 m to 500 m GL 200 m to GL460 m GL 460 m to 500 m
(CH class; UHFD) (B class; LSFD)
Design Design Result Increase rate of Result Increase rate of
(avg.) result to design (avg.) result to design
Drill and blast (1) Preparation 80 80 142 178% 150 188%
(2) Cleanup
(3) Drilling blast holes 222 236 311 140% 403 171%
(4) Blasting and ventilation 190 190 223 117% 318 167%
Subtotal 492 506 676 137% 871 172%
Muck removal (5) Removal of fly-rock from lining – – 156 – 190 –
(6) Preparation 60 60 72 120% 74 123%
(7) Muck removal by shaft mucker and 662 686 1215 184% 1272 185%
kibble
(8) Survey work 50 50 16 32% 60 120%
(9) Cleanup – – 57 – 72 –
(10) Shaping bottom wall 30 30 26 87% 21 70%
Subtotal 802 826 1542 192% 1689 204%
Shaft wall (11) Clean shaft wall – – 13 – 19 –
mapping (12) Preparation 180 180 20 23 127%
(13) Shaft wall mapping 188 127% 185
(14) Cleanup 20 20
Subtotal 180 180 241 134% 247 137%
Drainage mat (15) Preparation 7 7 15 214% 16 229%
(16) Installing drainage mat 28 28 52 186% 33 118%
(17) Cleanup – – 14 – 15 –
Subtotal 35 35 81 231% 64 183%
Concrete lining (18) Clean and install formwork 70 70 153 219% 122 174%
(19) Preparation 41 41 65 159% 51 124%
(20) Pumped concrete 109 109 227 208% 226 207%
(21) Cleanup 30 30 115 383% 126 420%
Subtotal 250 250 560 224% 525 210%
Other operations (22) Additional operations and/or tasks not 60 60 407 678% 514 857%
included in the above
Subtotal 60 60 407 678% 514 857%
Total 1819 1857 3507 193% 3910 211%

and the increased travel time with depth. The actual drill and blast
times per cycle, for (#3) Drilling and (#4) Blasting and ventilation,
shown in Table 1 are due to the differences in the fracture domains.
Fig. 6b shows actual muck removal times per cycle with depth.
Average muck removal times in the UHFD and in the LSFD were
1542 min and 1689 min, respectively, and approximately twice
as long as the design estimates (Table 1). It can be seen in
Table 1 that ‘‘(#7) Muck removal by shaft mucker and kibble
required the most time, 1215 min and 1272 min in the UHFD and
the LSFD, respectively, of all the tasks in all operations identified
in Table 1.
The variations in the actual durations of shaft wall mapping per
cycle, with depth, are shown on Fig. 6c. The average of the actual
times, 243 min exceeded the design estimate of 180 min between
GL-200 m to GL-500 m. The actual shaft wall mapping times per
cycle are the closest to the design estimates in the total excavation
cycle (subtotal: 134% in UHFD, 137% in LSFD).
Fig. 6d shows drainage mat installation times per excavation
cycle vs depth. The average actual time per cycle of 77 min is more
than twice the design estimate of 35 min for installations between
GL-200 m to GL-500 m. The average time for the drainage mat
installation task is the shortest of all the operations in the shaft
excavation cycles.
Fig. 6e shows the actual and design estimates for the concrete
lining operation expressed in terms of time vs depth. Similar to
the other excavation cycle operations, the actual times and their
average of 551 min exceeds the design estimate of 250 min. Note
that special activities such as niche and shaft inset construction
required additional form work and concrete than for the standard
2.6-m excavation cycle and thus increased the Ect and conse-
quently, extended the schedule. The concrete lining operation
proved to be the least accurately estimated operation in the plan-
ning stage, having actual times, on average, exceeding the planned
duration by more than 200%.
5. Discussion – differences between design and actual
excavation cycle times
5.1. Excavation
The average number of blast holes and actual explosive charge
volumes used in the UHFD per excavation cycle were 168 holes
and 2.11 kg/m3, respectively. In the LSFD, 212 blast holes and explo-
sive charge volumes of 2.90 kg/m3 were required per excavation
cycle, somewhat more than in the UHFD. Although the planned
blast hole length was 1.6 m, the actual blast hole lengths were mod-
ified from 1.6 m to 1.8 m during the actual construction due to blast
efficiency considerations. These factors caused an increase in work
with respect to drilling requirements and placing explosives; there-
fore, considerable additional time was required for drill and blast
activity in this cycle representing a significant increase in time in
74 H. Sanada et al. / Tunnelling and Underground Space Technology 50 (2015) 68–78
Fig. 4. Depth distribution of excavation cycle times in the Ventilation Shaft with
excavation cycle operation represented by colored squares. Times for special and
auxiliary operations (e.g. construction of pump niches and shaft insets every 100-m
depth, maintenance) that caused an increase in total cycle time are included in this
graph.
comparison with the planned drill and blast activity in the excava-
tion cycle. The reason why the number of blast holes, explosive
charge volumes and blast hole lengths increased in excess of the
design, as mentioned above, are summarized as follows:
(1) High drill angle for blast holes in the confined area of the
Ventilation Shaft
Since the range of movement and the workability of the shaft
jumbo were substantially restricted due to the presence of the con-
crete lining and the forms, resulting in the small diameter of the
Ventilation Shaft, a drill angle of 70° is the shallowest angle possible
for the center-cut (V-cut) around the middle of the shaft excavation
(Fig. 7). In contrast, a standard drill angle of 60° is preferred for sink-
ing shafts. For example, the shaft diameter of 6.5 m in the Main
Shaft at the MIU, being larger than the Ventilation Shaft diameter,
allowed a drill angle of 60° to be utilized for the Main Shaft. As
shown in Fig. 7, if a line is drawn orthogonally from the bottom of
the blast hole (V-cut), it would intersect the shaft wall at point
‘‘a’’ below the shaft face as shown in Fig 7. A short drill length of
1130 mm was generally, just achieved. In other words, drilling
and blasting in the constricted conditions in the Ventilation Shaft
reduced workability. Therefore, more blast holes, larger explosive
charge volumes and blast hole length were required.
(2) Loading muck bucket
The capacity of the muck bucket (the kibble) used in the
Ventilation Shaft is only 0.2 m3. It is therefore necessary that the
muck be finely comminuted to ensure efficient use of the small size
kibble. Therefore the number of blast holes and explosive charge
volumes were increased and thus the size of the muck was reduced
to enhance the loading efficiency. The larger number of blast holes,
increased explosive charge volumes and blast hole length due to
(1) and (2) above lengthened the timeline for the drill and blast
operation, thus increasing the shaft Ect.
5.2. Muck removal operation
The number trips made by the kibble and the rock volume con-
version rate (the bulking factor), L, will be described with depth. As
a rule, rock volumes will increase compared to in situ rock volumes
due to blasting and fragmentation of the rock and generation of
loose rock conditions in the muck. In this paper, rock volume con-
version rate of normalized value was used to estimate the muck
removal volume. Volume of mucked rock and rock volume conver-
sion rate were calculated as follows:
T ¼ M  Vk ð1Þ
where T is total volume of muck (m3), M is number of kibble trips,
and Vk is volume of muck in the kibble (m3).
T
L¼ ð2Þ
C þ Asi  l
where L is rock volume conversion rate, C is designed concrete
placement volume (m3), Asi is designed cross-sectional area of the
shaft (m2), and l is designed progress length of shaft (m).
In design, the assigned Vk for estimating was 2.0 m3 but the
actual Vk was reduced to 1.6 m3 (80% of design) for work safety
considerations, that is to minimize kibble overloading and the risk
of muck falling from the kibble while being hoisted. Therefore, the
number of actual trips for the kibble was expected to increase by
about 20%. However, the actual average number of trips to remove
muck using the kibble in the UHFD and in the LSFD was 54 and 53,
respectively. These are less than the design estimate of 53 in the
UHFD and 58 in the LSFD. This is not what was expected.
The explanation lies in the rock volume conversion rate. The
actual results, on average, were 1.33 and 1.38 in the respective
fracture domains, representing a decrease of about 20% compared
to design. The reason is that the large blast energy released due
to increase in the number and length of blast holes and explosive
charge volumes greater than the design, as described in
Section 5.1, resulted in improved fragmentation efficiency. Thus,
the drill and blasting operation is the factor resulting in an
improved rock volume conversion rate and suppression of the
number of kibble trips to a level lower level than design, despite
the decreased volume of muck in a kibble, Vk. From the reason
mentioned above, the factor causing a major difference between
actual working time and design estimate is not related to increas-
ing muck removal volume and limitation of loading volume in
kibble.
 H. Sanada et al. / Tunnelling and Underground Space Technology 50 (2015) 68–78 75

Fig. 5. Excavation progress through time and average excavation rate per month in the Ventilation Shaft.

Fig. 6a. Depth distribution of cycle time for each operation with estimated cycle
Fig. 6b. Depth distribution of cycle time for each operation with estimated cycle
times from design and actual total cycle time averages. Drill and blast time vs depth
times from design and actual total cycle time averages. Muck removal time vs.
for each excavation cycle.
depth for each excavation cycle.

Differences between actual and planned work in design are (2) Use of mini-backhoe with a hydraulic breaker for scaling the
described as follows: shaft wall

(1) Limitation of kibble volume
Although the maximum kibble volume is 2.0 m3, muck loads in
the kibble were limited to 1.6 m3 during the actual construction for
safety considerations as mentioned above.
To maintain stability of the shaft wall after blasting, removal of
loose rock from the shaft wall is required. This was conducted
using a mini-backhoe with a hydraulic breaker for safety reasons.
This is a change from the design plan, in which scaling would be
done manually using a pick hammer. Therefore, extra time was
76 H. Sanada et al. / Tunnelling and Underground Space Technology 50 (2015) 68–78

Fig. 6c. Depth distribution of cycle time for each operation with estimated cycle Fig. 6e. Depth distribution of cycle time for each operation with estimated cycle
times from design and actual total cycle time averages. Shaft wall mapping time vs. times from design and actual total cycle time averages. Concrete lining installation
depth for each excavation cycle. time vs. depth for each excavation cycle.

required for moving the mini-backhoe in the shaft, attachment of

the hydraulic breaker, in addition to scaling of loose rock from
the shaft wall.

(3) Limitation on maximum winch hoisting speed

Although design maximum winch hoisting speed was

300 m/min, the speed limit was reduced to 200 m/min for safety.

(4) Supplement for removing blast rock

Before muck removal could be done, fly-rock ejected onto the
concrete lining was removed. The work time for this was over
150 min on average. This work was necessary from a safety per-
spective to mitigate falling rock hazard.

(5) Muck removal from the shaft outer periphery under the con-
crete lining

The removal of muck from the shaft outer periphery under the
concrete lining proved too difficult using the shaft mucker.
Therefore, the mini-backhoe was used for removing the muck from
that region. This required extra time not included in the design
plan to move the backhoe to the shaft face to perform this work.
The five items mentioned above are reasons for significant dif-
ferences between the design and the actual work. With the excep-
tion of (1), these could be the main reason for differences between
Ect derived in design compared to the actual excavation.

5.3. Shaft wall mapping

Actual shaft wall mapping time in almost all excavation cycles

Fig. 6d. Depth distribution of cycle time for each operation with estimated cycle
was longer than allowed for in design. The reason is that although
times from design and actual total cycle time averages. Drainage water pipe shaft wall mapping time of 180 min was considered in design plan-
installation time vs. depth for each excavation cycle. ning, work for tasks ‘‘(#11) Clean shaft wall’’, ‘‘(#12) Preparation’’,
 H. Sanada et al. / Tunnelling and Underground Space Technology 50 (2015) 68–78
Fig. 7. Cross-section of the excavation face in the Ventilation Shaft showing blast
hole angle at center V-cut on the shaft face.
‘‘(#14) Cleanup’’ in Table 1 required more effort than expected,
that is, about 50 min extra in addition to shaft wall mapping. The
average time of about 187 min for actual shaft wall mapping in
each excavation cycle was almost the same as the design time of
180 min. Time allocation for shaft wall mapping should include
more than 50 min for any auxiliary tasks, such as mentioned above,
in addition to shaft wall mapping, in any future scheduling of Ect.
5.4. Drainage mat installation
Iron plates were installed on the shaft wall with anchor bolts to
fasten drainage mats. The effort to install the anchor bolts
exceeded the estimated effort due to the greater than expected
hardness of the rock.
5.5. Concrete lining operation
The design concrete volume for the concrete lining per excava-
tion cycle is 25.0 m3, which took into consideration the possibility
of over-break at the periphery after blasting. The average concrete
placement volume of 23.3 m3, with the exception of the shaft
insets and pump niches, is almost the same as the design specifica-
tion of 25.0 m3. Therefore, the basis for the design estimates,
including likely overbreak estimates, were fundamentally correct.
The reason why concrete lining work was more time consuming
than was allowed for in design is because of the difference between
the construction method proposed in design and the actual method
employed, since the actual concrete placement volume is lower
than the design estimate.
(1) Maximum hoisting speed
Although maximum hoisting speed was 300 m/min in design, as
noted for the mucking operations, the winch speed was limited to
200 m/min for safety reasons.
(2) Cleanup
Removal of slip forms for concrete lining actually required sig-
nificantly more time than estimated in the design. As shown in
Table 1, the actual time required is around 400% more than in
the design.
6. Recalculation of cycle time for drill and blast considering
actual blasting conditions
In Section 5, the difference between design and actual effort are
described. With respect to blasting efficiency, actual drill and blast-
ing has been modified. This chapter provides a description of the
recalculation using the actual blasting circumstances. Consider
the following with regard to (#3) Drilling blast holes in Table 1:
77
Total drilling time for blast holes Td can be calculated by follow-
ing equation.
Dl  Dn
Td ¼ ð3Þ
Dv
where Dl is drill length of blast hole (m), Dn is number of blast holes
per drifter and Dv is drilling rate (m/min).
Number of blast holes per drifter, Dn, can be defined by the fol-
lowing equation.
As  Dnr
Dn ¼ ð4Þ
J
where As is design shaft excavation area (m2), Dnr is number of blast
holes per m2 and J is number of drifters.
T m ¼ 1:5  Dn ð5Þ
where Tm is the time required for drifter moves between holes. 1.5
(min./hole number) means drifter move time between holes.
‘‘(#3) Drilling’’ in Table 1 is the sum of Td and Tm. Since Td and Tm
are time per excavation length of 1.3 m, the sum should be doubled
to be representative of a 2.6 m long excavation cycle. The drilling
component of Ect in design for the UHFD and the LSFD were
222 min and 236 min, respectively. The recalculated estimate is
294 min and 372 min in the UHFD and the LSFD. The actual drilling
took 311 min in the UHFD and 403 min in the LSFD, fairly close to
the recalculated estimates based on more realistic actual values for
the drilling.
Now, consider the following with regard to ‘‘(#4) Blasting and
ventilation’’ in Table 1
Dnr Dlr
T br ¼ T bd   ð6Þ
Dnd Dld
where Tbr is recalculated explosive charge and wire connection, Tbd
is explosive charge and wire connection in design, Dnr is number of
blast holes (actual), Dnd is number of blast holes (design), Dlr is
length of blast holes (actual) and Dld is length of blast holes (design).
The design estimate was 190 min (blasting and ventilation of
80 min, and others of 110 min) for this task in the drill and blast
operation. Recalculation of the estimate yields effort of 233 min
and 263 min in the UHFD and the LSFD. Actual work time was
223 min and 318 min in the UHFD and the LSFD. Thus, recalcula-
tion using actual values is much closer to the actual blasting and
ventilation times in the excavation cycle.
7. Concluding remarks
This paper describes the results of the analysis of Ect and a
breakdown of excavation cycle time by task during the sinking of
the Ventilation Shaft in the MIU and discusses differences between
design and actual times. The results obtained are summarized as
follows:
(1) Average cycle time for all excavation cycles in the UHFD
required about 3500 min, which is twice as long as the
design estimate. Actual times to perform three operations,
drill & blast, muck removal and concrete lining were much
larger than expected and represent the largest portion of
the total excavation cycle time and had the largest impact
on the efficiency of the excavation cycles and schedules.
(2) Excavation rate declined with increasing depth below
GL-300 m. In particular, large decreases in excavation rate
occurred at depths greater than GL-460 m, the boundary
between the UHFD and the LSFD where there is a substantial
78 H. Sanada et al. / Tunnelling and Underground Space Technology 50 (2015) 68–78
decrease in fracture frequency. It is apparent that the
increase in depth and in rock strength had a significant role
in the decline in excavation speed.
(3) For the drill and blast timeline, recalculation of the estimate
using the actual number of blast holes and their length
yields much more realistic estimates, that is, times that are
closer to the actual times expended in the excavation cycles.
It is suggested that realistic operation times can be esti-
mated by evaluation of accurate drill and blast work vol-
umes and methods prior to the actual excavation.
(4) The underlying reasons for increase in actual times for the
excavation cycle operations compared to their design esti-
mates can be ascribed to three categories: constraints on
construction; safety assurance; and maintenance.
(a) Constraints on construction;
– Decrease in blasting efficiency because blast hole
angles were steep due to restricted work area for
the drill jumbo. Therefore, the number of blast holes
and explosive charge volumes had to be increased
over and above the original design plan.
– Difficulty removing muck from just below the con-
crete liner. An additional task, the removal and rein-
stallation of the backhoe from the Shaft mucker was
required to improve mucking efficiency.
(b) Safety assurance;
– To ensure safety, the hoisting velocity of the winch
was self-regulated and lower than the design plan;
therefore time required for muck removal and con-
crete placement for the liner increased.
– Scaling operations entailed substitution of the back-
hoe with a hydraulic breaker for trimming the shaft
wall. Removal and reinstallation of the backhoe and
set up of the hydraulic breaker increased mucking
time.
– To ensure water inflow reduction, unplanned drilling
of pilot boreholes for pre-excavation grouting was
required.
(c) Maintenance;
– Routine equipment maintenance services (e.g. for
Shaft mucker, Pumps etc.) increased as construction
progressed.
Acknowledgements
The authors thank Glen McCrank, ex-JAEA International Fellow,
and anonymous reviewers for improving the manuscript.
References
Bieniawski, Z.T., 1989. Engineering Rock Mass Classifications. Wiley, New York.
Cai, M., Kaiser, P.K., Uno, H., Tasaka, Y., Minami, M., 2004. Estimation of rock mass
deformation modulus and strength of jointed hard rock masses using the GSI
system. Int. J. Rock Mech. Min. Sci. 41, 3–19.
Harmon, L.M., Lazur, E.G., 1984. Testing a predictive model for repository cost – the
waste isolation pilot plant experience. In: Proc. of WM1984, pp. 345–351.
Hirano, T., Nakama, S., Yamada, A., Seno, T., Sato, T., 2009. Mizunami underground
research laboratory project (rock mechanical investigations) – MIZ-1 borehole
investigations. Technical Report JAEA, JAEA-Research 2009-031, 58 p.
Japanese Geotechnical Society, 2004. Method for Engineering Classification of Rock
Mass (JGS: 3811-2004). Japanese Geotechnical Society, 36 p.
Kim, S., Choi, J., 2006. Assessment of the cost of underground facilities of a high-
level waste repository in Korea. Nucl. Eng. Technol. 38 (6), 561–574.
Kukkola, T., Saanio, T. 2005. Cost estimate of Olkiluoto disposal facility for spent
nuclear fuel. Posiva Working Report, Working Report 2005-10, 65 p.
Matsui, H., Noda, M., Nobuto, J., 2011. Status of the Mizunami URL construction and
study on engineering technology. In: Proceedings of ITA-AITES 2011 World
Tunnel Congress/37th General Assembly, pp. 1248–1255.
NIREX, 2005. Cost estimate for a reference repository concept for UK high-level
waste/spent nuclear fuel. NIREX Technical Report, 484281, 22 p.
NUMO, 2013. Safety of the geological disposal project 2010 – safe geological
disposal based on reliable technologies. Technical Report NUMO, NUMO-TR-13-
05, 125 p.
Read, J., 2011. APM conceptual design and cost estimate update. SNC Lavalin
Nuclear Report, 020606-6100-REPT-0001, 159 p.
Saegusa, H., Matsuoka, T., 2011. Final report on the surface-based investigation
(Phase I) at the Mizunami Underground Research Laboratory Project. Technical
Report JAEA, JAEA-Research 2010-067, 377 p.
Sakurai, S., Shimizu, N., Akutagawa, S., Yoshida, H., Matsui, H., Sato, T., Yamachi, H.,
2009. Investigation on collapse mechanism of super-deep vertical shafts in
Japan. In: Proceedings of the 7th International Symposium on Rockburst and
Seismicity in Mines, August 21–23, Dalian, pp. 1314–1324.
Svensk Kärnbränslehantering AB, 2009. Underground design Forsmark Layout D2.
SKB Technical Report, R-08-116, 128 p.
Tsuruta, T., Takeuchi, S., Takeuchi, R., Mizuno, T., Oyama, T., 2009. Results of the
pilot borehole investigations conducted from bottom of shafts at the Mizunami
Underground Research Laboratory. Technical Report JAEA, JAEA-Research 2008-
098, 116 p.
Tsuruta, T., Sasao, E., Kawamoto, K., Kuboshima, K., Ishibashi, M., 2013. Study on
geology on the Mizunami Underground Research Laboratory Project – geology
and geological structure from G.L.-300 m to G.L.-500 m. Technical Report JAEA,
JAEA-Research 2013-014, 78 p.
Tsuruta, T., Tagami, T., Amano, K., Matsuoka, T., Kurihara, A., Yamada, Y., Koike, K.,
2013b. Geological investigations for geological model of deep underground
geoenvironment at the Mizunami Underground Research Laboratory (MIU). J.
Geol. Soc. Jpn. 119, 59–74.