Formed by the Tunnel Engineering Committee of Japan (TECJ), the Special Committee has the mission of assessing the damage to underground structures in the wake of the 2011 earthquake. As T&TI goes to press, the committee is working on making a list of all tunnels within the disaster area, though determining the disaster area is not as simple as getting a set of compasses and a map, as committee chair and Kyoto University professor Asakura explains later.

Inspections
The tunnel inspections are also delayed, as Asakura puts it, by the ‘disturbed’ environment. Carried out by the respective tunnel operating companies, inspections involve a walkthrough observation of the tunnel, then further detailed inspections.

The only reports of notable damages when T&TI spoke with Asakura were those of two Shinkansen (Bullet Train) running tunnels, Fukushima and Shiga, after immediate inspection by the East Japan Railway company. In these tunnels, the track bed concrete and invert concrete was damaged by strong lateral ground pressure, causing serious but not extremely serious cracking. The damage in these tunnels has already been repaired by grouting the cracks and service has resumed, though as a result of this the rail level has been brought up by some tens of millimetres.

“The Special Committee will compile and analyse [all data generated through inspection] as the mechanism behind earthquake damage to tunnels requires further study,” Asakura says.

“There are many unknown conditions related to such damage. Although mountain tunnels are assumed to be earthquakeresistant structures, previous studies have shown that they may sometimes suffer damage such as cracking or spalling caused by flexural compression failure.”

Although no papers exist yet on the 2011 Tohoku earthquake, prior to this, in August 2010, a paper was published by Asakura and a number of his colleagues: Naritoshi Fukazawa, director of the Design and Technology Department for Japan Railway Construction of the Transport and Technology Agency; Yoshiyuki Kojima, laboratory head of the Tunnel Engineering Laboratory, Structure Technology Division and Kazuhide Yashiro, senior researcher of the same. Titled ‘The Mechanism behind Seismic Damage to Railway and Mountain Tunnels and Assessment of their Aseismic Performance’, it gives details of a study in which case studies and model experiments are used to investigate this mechanism and its effects.

Past damages
At the time of the paper’s publishing, 19 tremors had caused damage to Japanese mountain tunnels since the Kanto earthquake of 1923, with five causing serious damage (see figure 1). Some 25 tunnels collapsed during Kanto, needing reconstruction and earthquake countermeasure work to reinforce against future tremors.

The Izu-Oshima-Kinkai (Izu) earthquake of 1978 caused serious damage to tunnels on the Izukyu line, particularly the Inatori tunnel, the lining on which was seriously damaged by fault displacement, closing the tunnel for half a year during repair works.

The 1995 Hyogoken-Nanbu (Hyogo) earthquake was a near-field tremor in an urban area resulting in the serious damage of 12 tunnels at the foot of or within the Rokko mountain range near Kobe City in southwest Japan. The damage was particularly bad in the Higashiyama tunnel, part of the Kobe electric railway and also in the Rokko tunnel, which forms part of the 553.7km Sanyo Shinkansen Line between Osaka and Fukuoka.

In 2004 the Niigataken Chuetsu (2004 Niigata) earthquake seriously damaged 11 tunnels including the lining of the Myoken tunnel, the Uonuma tunnel and the Wanatsu tunnel. All are on the Joetsu Shinkansen Line running between Tokyo, on the Pacific side, and Niigata, on the Sea of Japan side of Honshu Island. The Wanatsu tunnel took two months to repair.

The 2007 Niigataken Chuetsu-Oki (2007 Niigata) earthquake seriously damaged four tunnels, three of which were on the Shinetsu Main Line. In particular, the First Yoneyama tunnel experienced serious flexural compression failure and shear failure of the lining.

Damage categorisation
Tunnel damage analysis in Japan is categorised in four main ways. The level of damage; modes of damage, which are divided into three types discussed below; distance and magnitude relationships and also whether the tunnel is subject to any of the ‘special conditions’ discussed by Yoshikawa in his treatise ‘Earthquake Disaster Survey on Railway Tunnels’ published in the Research Report of the Railway Technical Research Institute in September 1979.

Damage levels
Damage level assessments range from no damage to serious damage, meaning a tunnel in need of serious repair or reinforcement work. Of the five earthquakes shown in Figure 1, the percentage of tunnels displaying serious damage averaged around 12 per cent, with medium damage at 7 per cent, slight damage at 23 per cent and an average 58 per cent of all affected tunnels exhibiting no damage.

Modes of earthquake damage

Type I
Cracks in the arch shoulder, characteristic to shallow tunnels in soft ground, is believed to be caused by shear deformation in the ground inducing shear deformation in the tunnel, causing an increased bending moment at the lining arch shoulder that results in failure.

Type II
Flexural compression failure and spalling at the crown are typical in tunnels through poor geological conditions. A probable result of the poor, usually soft geology is the exacerbation of deformation caused by the ground motion during an earthquake, as well as the loads caused by loosening ground pressure and squeezing ground pressure that could be present before the tremor strikes.

Type III
Damage caused by fault displacement, including lining fracture resulting from local forced displacement accompanied by fault displacement.

Asakura also notes in his paper that type two damage in ground with poor geology, and type one in shallow tunnel situations, accounts for the majority of damage reported, if Slope disasters are excluded as a matter of secondary damage. Tunnels in these conditions are highly likely to suffer earthquake damage.

Distance and magnitude
An increasing magnitude and a decreasing distance both escalate the probability and severity of tunnel damage. Although a general guideline is given stating that for a magnitude eight earthquake, the possibility of medium or serious damage greatly increases when the distance from the quake epicentre is less than 30km, and 10km for a magnitude eight, Asakura points out that this is flawed, and at least one tunnel further than 100km from a tremor epicentre has been damaged. As there are so many other factors, and because the 2011 Tohuku earthquake was the largest yet detected in Japan, this guide is unreliable.

Yoshikawa’s special conditions Yoshikawa listed seven special conditions in his survey: landslide, deformed tunnel, poor geological conditions, shallow tunnel, under construction, fault slide and lining defects. Figure 2 shows the percentage of tunnels damaged by the five major earthquakes that fall under the special conditions. This shows that tunnels with special conditions make up the overwhelming majority of cases of serious damage and are at most risk.

“Before 2011 Tohoku Japan Earthquake, railway companies in Japan checked the tunnels that had special conditions, and reinforced where required,” says Asakura. “We now intend to check all the tunnels because we have to know what is the dominant condition. We need to know what the mechanism of damage is more exactly.

“Although the special conditions are not unique to Japan, the geology in this country is relatively more complicated and of lower quality than that found in others.”

Past experiments and studies
Experiments were undertaken jointly by Kyoto University, the Railway Technical Research Institute and the Japan Railway Construction Transport and Technology Agency, under the ‘Program for Promoting Fundamental Transport Technology Research’ of the Japan Railway Construction, Transport and Technology Agency (JRTT). The model experiments focus on earthquake damage to mountain tunnels and place an emphasis on tunnels with shallow cover and in ground with poor geology.

One of the experimental investigations into type one damage was conducted through the use of a two-dimensional shear box (see figure 3). A model tunnel was placed in the box and then shear displacement was applied to the tunnel through the boxed ground by three movable jacks. The ground was represented using dry silica sand. Compaction brought the sand to a relative ground density of 80 per cent. The model was made with mortar possessing a uniaxial strength of around 26MPa.

The model tunnel was equipped with a webcam and displacement transducers. During the experiment, the lowest jack was fixed with the upper two actuated to provide displacement with a triangular distribution pattern. Loading was applied under displacement control to statically provide alternate loading up to six per cent of the shear strain of the ground. The shear strain of the ground was increased at every loading cycle.

Displacement of the tunnels inner surface was measured and any cracking was observed through the webcam. This was done for models with sound and defective linings (with a void and lack of thickness).

Conclusions
“The experiments and studies further clarified the damage mechanism and seismic performance of the [case] tunnels in question,” says Asakura. “Sound tunnels and those with inverts are less susceptible to seismic damage. It was also confirmed that such damage tends to be greater when tunnels have voids above the lining and a lack of thickness, no invert or when local displacement acts on them.

“In the future work we intend to execute investigation of numerical analysis and design methods, and plan to reflect the results of these investigations—in addition to those of our present study—in guidelines and manuals for use in the aseismic design of lining in mountain tunnels.

“We are in the process of gathering more information and will be able to contribute more detailed and accurate information in the future,” he explains.

Despite all the experimentation and study, more is needed. As Osamu Kiyomiya—a colleague of Asakura—told T&TI, “Even though various countermeasures were arranged for tsunamis, the damage was past imagination, and beyond engineering judgment. Seismologists never imagined such a strong earthquake. The worst possible scenario had happened. We now have to consider the possible reoccurrence of such a strong tsunami, and countermeasure to protect underground structures in the cities.”

“It is important to maintain tunnels carefully at all times, not just in the event of a tremor,” adds Asakura.

“The conditions in which tunnels are damaged easily by earthquakes will cause them to be deformed easily even without regard to earthquakes.”


Figure 1, five of the earthquakes in Japan that caused serious tunnel damage since 1923 Figure 2, the percentage of tunnels damaged by the five major earthquakes that fall under special conditions Figure 3, an investigation into type one damage that models a tunnel with shear displacement