1 INTRODUCTION

1.1 The Tsunami on 3.11

The 2011 off the Pacific coast of Tohoku Earthquake

occurred on 11th March in 2011. Its magnitude was

9.0 and it caused a massive disaster of tsunami. From

the report of the ministry of land, infrastructure and

transport (Mlit report 2011), the tsunami height from

the normal tide level was more than 20m in many

areas along the Sanriku coast line, and in some areas,

it was more than 30m up to 40m. The death toll was

more than 15,000. It is said more than 90% of the toll

was death by drowning due to the tsunami.

From the report by the Weather News Company in

September 2011 (Wethernews 2011), they conducted a

hearing investigation soon after the earthquake from

those who survived in the tsunami disaster area from

May 18

to June 12

. From this report, the average

evacuation time which means the time when one

started evacuation from when the earthquake

occurred was 19 minutes for those who survived,

whereas 21 minutes for those who deceased. This

means only a few minutes earlier action can save

many lives. It is said that most of the people there

thought the tsunami wouldn’t come as usual because

they experienced big earthquake several times before

with tsunami waning every time but no tsunami

hadn’t come before. So it is no wonder that they

thought the warning was just a routine work of the

news program like before and evacuation starting was

possible to be late. In addition, from this report we

can find several reasons to prevent people from

evacuating to the supposed safe place. The main

reason was there were so many obstacles on the road

like traffic jams or landslides that they cannot move

smoothly (Fujiu et al. 2016). The tsunami speed was

faster than their walking speed, so many were not

able to reach the safe place. Another main reason was

some of them went back to a dangerous area because

they were going to help others like the aged people or

their family (Mikami 2014).

Considering these data, we think if there is a safe

shelter in the garden or the parking lot in the house or

a small company, we can overcome those bad factors

Development of a Small Tsunami Shelter and Its Sea

xperiment of Towing and Drifting

K. Watanabe

School of Marine Science and Technology, Tokai University, Shizuoka, Japan

S. Mizuno

Mizuno Marine Co.Ltd., Osaka, Japan

ABSTRACT: We developed a small Tsunami shelter. The design characteristics of this shelter are, its floater

keeps the shelter floating even if the cabin space is fully flooded, the shelter can be self-recovered from the

upside down situation when it rolls in the sea. When the shelter is washed away from the shore, it starts

drifting. In that case, passengers might have to stay in the small cabin for several days. The length of the shelter

is around 2m, which is much less than a typical life boat. So we carried out the first sea experiment using a real

shelter with riding 8 passengers. In this paper, we’ll show the experimental results of motion sensor, towing

force, as well as lessons learnt.

http://www.transnav.eu

the

International Journal

on Marine Navigation

and Safety of Sea Transportation

Volume 14

umber 1

March 2020

DOI:

10.12716/1001.14.01.08

mentioned above. They can just go into the shelter as

soon as a big earthquake occurs whether the tsunami

is coming or not. It will take less than 10 minutes. A

family can be in the same place. If they can be in the

shelter, they don’t need to use a car for evacuation so

that the traffic jam will be alleviated.

On the other hand, there have been many projects

that were supported by Japanese government or

municipal governments like building Tsunami

evacuation towers, building a long seawall along the

coastline where is supposed the next Tsunami coming

(Kihara et al.2014). But it is difficult to cover all area,

so individual effort to prepare for the disaster by

oneself is important.

Many concepts of Tsunami shelter have been

proposed by many house makers or venture

companies. Most of these shelters are designed as a

temporary evacuation capsule to prevent people from

drowning at encountering the first Tsunami, however

few of them are designed considering the situation of

surviving when the shelter starts floating on the sea

after being washed away by the first or second

Tsunami. So we developed a small lifeboat type.

2 THE TSUNAMI SHELTER

2.1 Background of design concept of the shelter

The shelter is designed reflecting some lessons learnt

from the 3.11 earthquake. The massive flow of

tsunami includes everything on the ground like

crashed houses, cars, debris from factories or shops,

and even big ships from near ports. If one can drift

with these debris, the external force exerted to a

drifter seemed endurable and the possibility of

survival can be increased. Actually, there were some

people who were on a floatable debris survived

drifting with the flow. On the other hand, at one

community center where was designated as an

evacuation place when a big earthquake occurs, many

elderly people gathered were drowned because

tsunami was not supposed to come at that time. In

this case, one of the authors heard soon after the

tragedy from a student from the area that the victims

were not only drowning. The aged people were

whirled up to the beams of the roof of the community

hall when the hall had filled with sea water. As the

water ebbed rapidly while they were holding the

beams that they had to hang down from the beam and

fallen down from several meters high.

Considering these lessons, we decided to design

the shelter can be floatable and should be strong

while its drifting enough to protect people from the

debris which can hurt human body like a sharp edged

debris of metal or glass and so on.

Another lesson was prevention from hyperthermia

and store provisions and necessities. Most of the

evacuee lost not only their own house but also any

community center by being washed away by the

tsunami. More than 118,822 houses were totally

broken. More than 184,615 houses were half broken.

More than 386,000 of people had to spend several

nights without any shelter under the condition that all

route had shut down.

So our shelter is designed as a kind of small life

raft and also to be habitable as a temporary house.

2.2 The shelter

The shelter is designed as follows.

− The buoyancy must be always positive even when

the space inside is fully flooded, which means it

never sinks.

− The shape is round not to stack a structure like a

building or a bridge column while it is being

flowed by the tsunami flow.

− It is watertight even if it submerges under 20m.

− If the shelter turns upside down, it recovers by

itself. This is important because there is a very

high chance of rolling on the ground while being

washed away with the strong current.

− It has enough space for 8 passengers including

provisions like water and food.

Considering above requirements, the principal

dimensions is set as Table 1 and the final design are as

shown from Figure 1 to Figure 4.

Table 1. Principal dimensions of the shelter.

_______________________________________________

Length 2.55 m

Breadth 2.14 m

Height 2.01 m

Number of passengers 8

Weight in air 4900 N

_______________________________________________

Figure 1. Front view of the shelter

Figure 2. Side view of the shelter

Figure 3. Passenger arrangement (side view)

Figure 4. Passenger arrangement (top view)

The cabin is made of FRP. The lower overhanging

part is made of urethane which yields enough

buoyancy even when the cabin is fully flooded. There

are two water tight hatches where the evacuees get

onboard at front side and back side. A hatch can open

and close both from outside or inside. For air

ventilation, each hatch has a ventilating opening

which can be closed from inside. It is possible that the

evacuees get off the shelter on the sea to get on the

rescue ship of the coast guard or defense force. So the

hatch position is determined to avoid flooding while

it is opened.

As shown in Figure 3, maximum eight passengers

can get onboard. There are two benches inside and

eight seatbelts are installed. The bench has a lid and it

is used as a storage. The survival kit that includes

water, emergency provisions, fishing tools etc. is

settled in the bench. There are also a set of paddle

inside the bench and the hull has two holes for

paddling that can be closed from inside, so we can

thrust the shelter manually using these paddles.

The hull was fabricated in a Chinese company

which makes SOLAS adapted lifeboats. The wall

thickness is almost twice as thick as a normal lifeboat.

We designed the wall stronger than a life boat

because the shelter is possible to contact with many

obstacles like a floating car, destroyed houses or a

column of a bridge while it is being washed away by

the tsunami current.

The strength of the shelter body must be tested

especially from the viewpoint of impact force exerted

during tsunami. This is our future work.

3 THE SEA EXPERIMENT

3.1 The purpose of experiments

The shelter is designed to be able to survive in the

tsunami current. So we believe the evacuee in the

shelter can be much safer compared to the situation of

3.11 tsunami. The best story supposed is the shelter is

transported to the higher place by tsunami current

and it runs aground somewhere. The evacuee can live

there for the time being.

However, if the shelter is flowed off the coast with

the strong back wash, the evacuee in the shelter must

drift on the sea until a rescue ship comes to tow them

to the shore. As the size of the shelter is around 2 m,

sea sickness might be a big issue. In addition, as the

cabin is very narrow so psychological situation of the

people in the cabin might be an issue. Furthermore, if

many shelters are on the sea, a rescue ship has to tow

several shelters and how many shelters can be towed

at one time might be estimated to make a plan of

collecting shelters in an appropriate period.

So the purposes of the sea experiment in this

research are set as,

1 Evaluating roll, pitch motion in the sea.

2 Estimation of towing force for one shelter.

3 How do the evacuees feel in the shelter while

drifting.

4 Temperature difference measurement between

inside and outside.

The experiment was carried out on 16

January in

2019 in the Suruga Bay. The weather was fine but

windy and around 1～1.5m swell was observed. It is

towed by Tokai University’s research ship

“HOKUTO”. We made a motion and position

recording unit. The GARMINE GPS 18x 5Hz receiver

is used as a GPS receiver. The roll, pitch and yaw

acceleration, velocity, and angle are recorded with the

GPS position. The GPS antenna is attached on the top

of the shelter as shown in Figure 5. The radar reflector

is also attached as shown in Figure 5.

Figure 5. GPS antenna and a radar reflector

Figure 6. Trajectory record in the experiment

The trajectory recorded in the experiment is shown

in Figure 6. We confirmed the shelter can be found

from more than 4 miles by radar.

Before sea trial, we measured how many seconds it

takes for eight passengers to get into the shelter.

When we get onboard using only one hatch, the

average was about 55 seconds. If we use both hatches,

the average was 35 seconds. As we tried several times,

the time becomes shorter.

3.2 The temperature comparison

The temperature and air ventilation are very

important for the shelter habitability. As mentioned

earlier, more than 386,000 people spent a week

outside in the cold air. To provide a shelter which

protect people inside from cold wind or rain is very

important. To know heat retaining property of the

shelter, we measured both outside and inside

temperature. Figure 7 shows the position of the

thermo sensor inside of the shelter. Outside thermo

sensor is attached adjacent to the GPS antenna.

Figure 7. Thermo sensor arrangement inside

Figure 8. Temperature record of first 100 minutes

Figure 9. Temperature record of next 100 minutes

Figure 8 and Figure 9 are a series of temperature

data. The orange line is the inside temperature. The

blue line is the outside temperature. Eight passengers

were onboard.

As shown is the Figure 8, we closed both hatches

at 9:44 and started towing. After 20 minutes, around

10:04, everyone felt stifling. As the hatches have an air

ventilation hole, we regarded ventilation would be

enough and we would not feel stifling before this

experiment started. However, we all felt bad air

condition in around 20 minutes. The hole is small to

be able to be closed easily to make it watertight. At

10:52, we all felt the inside air was endurable so we

agreed to open the rear hatch. When hatch opened we

all felt the air rapidly refreshed. From this

experiment, we found the ventilation hole should be

larger. As for heat retaining property, we found it

satisfactory. As the Figures show, the inside

temperature was kept around 20℃ regardless outside

temperature was less than 15℃ in addition to the sea

temperature was low in winter. When both hatches

were closed, the inside temperature increased around

25℃. This is due to it was sunny on the day of

experiment and eight men were inside.

As shown in Figure 9 if both hatches were opened

the temperature drops around the outside

temperature in about 8 minutes. This also means the

air inside can be ventilated within 10 minutes if we

open both hatches.

From this experiment we confirmed that this

shelter can be useful not only protecting evacuees

from drowning or crashing in debris but also

providing them with a warm and rain-sealed shelter.

As the shelter stores at least minimum emergency

food and water, it will keep evacuees from

hyperthermia as well as dehydration.

3.3 Motion characteristics while drifting on the sea

The size of the shelter is determined as it can be

placed in one’s garden or a parking lot with keeping

eight passengers. As a result, the size became as

shown in Table 1.

This size around 2 m is so small that its motion on

the sea is supposed to be relatively large. This will

cause evacuees seasickness. The priority of this shelter

is to survive from tsunami. So ride quality must be

sacrificed. However, while waiting on the sea until

the rescue ship from the coast guard is coming, the

evacuees must drift freely on the sea.

So we conducted the free drift experiment and

recorded motion data. For this experiment, we made a

motion recording unit as shown in Figure 10. The

motion sensor is AMU light 9 axis MEMS of Silicon

sensing Japan Inc. All data were recorded in 5 Hz.

The module was set around the middle of the shelter

as shown in Figure 11. We can monitor whether the

sensor and GPS are working from outside PC.

Figures 12-14 are roll, pitch and yaw angles while

drifting. As Figure 12 and Figure 13 show, both roll

and pitch amplitude angles were around 10 degrees.

Those motion periods were from around 2 seconds to

4 seconds from FFT analysis. Figure 14 shows yaw

angle. We can see it kept rotated. The combined

motion made everyone feel sick. To improve its ride

quality, we need to evaluate RAOs and compare it

with our experimental results. This is our important

future work.

Figure 10. Sensor module architecture

Figure 11. Motion recording unit setting

Figure 12. Roll angle while drifting

Figure 13. Pitch angle while drifting

Figure 14. Yaw angle while drifting

Figure 15. Students feeling sick while drifting

3.4 Towing force measurement

The shelter must be recovered as soon as possible

after a disaster ends. If many shelters are on the sea,

how to collect them is an important agenda for a

government. One possible story is gathering several

shelters together and tows them by a ship from the

coast guard or defense force. Towing force estimation

is important for this story. So we measured the

towing force while our shelter was towed as shown in

Figure 16. The towing force was measured by a force

sensor as shown in Figure 17. We read the sensor

value every 600 seconds with the speed of the boat.

The measured result is shown in Figure 18. The

towing speed was kept constant around 1.5 m/s to 2.0

m/s. However, we measured fluctuation of the towing

force, especially when the towing ship turned or the

shelter front plunged into the water. To estimate the

typical drag force coefficient of the shelter, we

assumed the representative towing force can be

estimated by the next simple drag force equation.

ACF

towing

≡

(1)

where, ρ kg/m

is water density, u m/s is towing

speed, A m

is a representative area, breadth*draft.

Figure 16. Towing experiment

Figure 17. Relation of towing speed and force

Figure 18. Towing force and speed measured

By using the approximate polynominal in Figure

18, we evaluated the typical towing force was 314 N

at 1.5 m/s and 461 N at 2.0 m/s.

The weight of shelter itself is 4900 N and total

passenger weight was around 6860 N. So we

calculated the draft as 0.3m, then the submerged area

was around 0.6 m

. From equation (1) and the

estimated typical towing force values above, we

estimated the representative Cd value around 0.4 or

0.5. From this experiment, it turned out that the

estimated Cd value of the shelter was much lower

than we supposed.

When the towing speed was faster than 2.0 m/s,

we in the shelter observed the front of the shelter

sometimes plunged into the water. This seemed to

increase the exerted area A in equation (1) so that the

drag force becomes more than twice of the typical

value. As the estimation here is very simplified one,

so further investigation including 6 DOF simulation

and experimental data accumulation should be our

future work. As a primary study, we found the

towing force was not excessive and a rescue ship can

tow many shelters to the shore simultaneously with

gathering shelters drifting on the ocean.

3.5 Upside down test

We can easily suppose that the first tsunami water

might roll over the shelter and the shelter rolls several

times on the ground before it is afloat. Personal safety

inside is very important. As shown in Figure 19, eight

seatbelts are implemented in the shelter. We planned

the roll-over experiment in the shallow rapid river,

but permission has not authorized yet.

Figure 19. Passenger seatbelts in the shelter

Figure 20. Upside down test in the pier

Instead, we carried out several upside-down tests

in the pier. We suspended the shelter from a crane

and rolled it over. While this test, we sat inside the

shelter with seatbelts fastened. As the shelter was

designed to be able to recover by itself on the sea, we

were not able to turn it down 180 degrees because it

somehow rolled back. However, we confirmed we

managed to bear the upside down rolling situation

inside without falling down from the head.

We realize more accelerated rolling situation

should be investigated further. It is a little bit

dangerous for us to ride on the rolling shelter, so we

need to use dummy dolls to collect data exerted on a

human body. This is also our important future work.

4 CONCLUSION AND FUTURE WORK

We conducted a sea experiment using the real

tsunami shelter. It turned out that the shelter drag

force is small enough, it can keep inside temperature

higher than outside air in winter. The rolling and

pitching motion were relatively large but we

managed to bear. We also confirmed that its self-

recovery from upside-down rolling situation. The

hatch height seemed high for the elderly, so we need

to modify it. Through the experiments in this study,

we confirmed basic function of the shelter and it

seems it will be helpful for protecting people from

supposed tsunami.

This is our primary study and we found several

important future work agendas as follows. Impact

force test should be carried out to confirm its

structural strength. RAO evaluation and motion

simulation in various sea conditions should be

studied. Fluid dynamical parameters like a drag

coefficient should be investigated further to estimate

its 6DoF motion and towing force. Inertial force

applied to a human body while the shelter is rolling

or smashed should be estimated by experiments.

ACKNOWLEDGEMENT

We would like to thank those laboratory members,

K.Utshnomiya, K.Harada, Y.Iwata, S.Nakajima, T.Furuya, S.

Tsuru and R.Katoh for their dedication to this research. We

sincerely appreciate Capt. I.Hashigaya and H.Sakata of the

HOKUTO crew for their extensive cooperation. This

research is supported by Toyonaka small and medium-sized

enterprises challenge business subsidies.

REFERENCES

Fujiu et al.2016. Analysis of countermeasures for car

evacuation against tsunami using traffic-micro

simulation. Journal of JAEE, vol.16 No.8:135-143.

Kihara et al.2014. Large-scale experiment of tsunami

hydrodynamic load on vertical tide wall. Journal of JSCE

B2, vol.70 No.2:826-830

Mikami 2014. The survey analysis about the victims by the

tsunami under the great east Japan earthquake. Journal

of JSCE A1, vol.70 No.4:908-915.

Mlit report 2011. http://www.mlit.go.jp/common/

000208803.pdf.

Weather news 2011. http://weathernews.com/ja/nc/

press/2011/pdf/20110908_1.pdf.