TOPPLED VEHICLES DURING KUMAMOTO EARTHQUAKE:

RESPONSE ANALYSIS & PGA ESTIMATION IN NEAR FAULT ZONE

Rishi Ram PARAJULI1
Junji KIYONO2
Yoshihiro Okumura1
1

Dept. of Urban Management, Graduate School of Eng., Kyoto University

2
Graduate School of Global Environmental Studies, Kyoto University

1.0

INTRODUCTION

April 16, 2016, M7.3 earthquake struck in Kumamoto of Kyushu region in Japan at depth of about 12 km following the
two major foreshocks of M6.4 and M6.5 in 15th and 14th of April. At least 49 people lost their life with tens of thousands
moved to the evacuation centers where thousands of structures got collapsed. Seismic intensity due to the main shock
was recorded up to 6.5 in Mashiki area[1], about 7 km far from the epicenter. Tateno area in Kumamoto, about 26 km far
from the epicenter was hit by the large landslide that triggered due to earthquake. Kurokawa, adjacent to Tateno was
severely affected where rupture line exposed to the surface. During the damage survey in that area we spotted some
vehicles toppled on the side due to the earthquake ground motion where shinkansen and JR train were also derailed at
nearby Kumamoto station and Akamizu station respectively. We used the google earth to find the more cases where we
found five cars and a pickup van in Kurokawa area along with a truck and a car in Tateno area got toppled. Red arrows
in the figure 1 show the toppling direction. All of the vehicles toppled toward rupture line (fault normal direction) where
pickup van seems toppled in opposite direction that might be due
to the pushing of collapsed building on the side. Locations of
such phenomenon are found with in the distance of about 500
meters from the line of rupture that exposed to the surface.
We previously proposed Circular Path and Linear Momentum
(CPLM) method for seismic response analysis of vehicles that
can be used for the cases of stationary and running condition of
vehicle [2] where the model was validated with response of the
stationary vehicle during Gorkha earthquake 2015 having long
period ground motion in Kathmandu [3]. Vehicle response and
risk associated with location during earthquake, running on the
elevated structures were also studied [4]. In this paper we present
Figure 1 Location of toppled vehicles, rupture line and
the analysis of toppling vehicle with a car model. We use the
direction of vehicles toppled (in red arrow)
concept of angular momentum for the lateral rotation of vehicle
that supposed to be parked with the locked brake.
2.0

VEHICLE MODEL

Suzuki MR wagon car is taken as the sample car model for analysis having
3.395m, 1.475m, 1.625m total length, breadth and height respectively. Total
weight of the vehicle is 790 kg where the wheel base and trades are of 2.425m
and 1.29m. A simple model for rolling analysis is shown in figure 2 where b is
the trade, h is the height of CG, g is the angle from lower corner to CG, is roll
angle. We define the vehicle model and rest of the parameters of car as specified
in the previous study [2]. In this study we supposed the car is parking with
locked brake hence we do not consider the rolling of tyre effect.
3.0

ANALYSIS

Figure 2 Simple model of car for rolling

Longitudinal and lateral analysis of the model are followed as specified in CPLM method[2]. Here we introduce the
angular momentum method in calculation of rolling angle when the vehicle tilted enough to take load in one side only.
Considering the up-down component of the ground motion, threshold value of the lateral acceleration for that condition
of vehicle is given by /2( ) where g is the gravity. We calculated the total angle including the rolling angle for
location of CG in each step of calculation. Changes in angular momentum due to the vertical and lateral acceleration are
calculated by using the following equation in each step.
= ( sin( ) ( ) cos( )
Where is change in angular momentum, M is total mass of the vehicle, R is the distance from bottom corner to the
35th Annual Meeting of Japan Society for Natural Disaster Science, 20-21 September, 2016, Shizuoka

locus of CG (AO in figure 2) is the total angle ( + ) representing the position of CG. SIGN define the direction of
lateral acceleration as it is positive for anticlockwise when we consider the left bottom corner as the support. Summation
of angular momentum in every step of calculation is now used to find out the linear velocity of body in lateral direction
centered on lower corner = /. Now change in rolling angle can be calculated as = ( )/. When the
rolling angle reached 90 degrees, we stop the calculation as the vehicle toppled on the side.
4.0

RESULT AND DISCUSSION

Vehicle response can be taken as a reference to estimate the

PGA values hence we used sine waves to analyze the general
relationship of PGA and the frequency of input wave for the
vehicle toppling mechanism. Frequency of the wave started
from the 0.4 Hz in difference of 0.4 up to 16 Hz (total 40
samples), same amplitude of wave used for each 5 seconds and
then increasing amplitude at scale of 0.5 m/sec2 for next 5
seconds until the vehicle toppled. Toppling mechanism of the
vehicle mainly dominated by the lateral acceleration where
similar effect of vertical component is also exists. We used the
similar sinusoidal waves in vertical direction of same frequency
in all cases of lateral input. There is no vertical motion in the
first case and the other 14 cases have increasing values of
amplitudes upto 14 m/sec2 were analyzed for all 40 cases of
lateral acceleration. Figure 3 shows the relationship of frequency
to lateral PGA for the toppling of vehicle under different cases
Figure 3 PGA - Frequency relationship of lateral and
of vertical ground motion. Legend in the figure denotes the
vertical acceleration to topple the vehicle
amplitude of vertical acceleration for corresponding cases.
NIED station KMMH16 located supposed to be near the fault in the Mashiki town recorded the strongest ground motion
with PGA of 11.57, 6.53 and 8.73 m/sec 2 as EW, NS and UD components [1]. This station is about 7.6 and 18.7 km far
from the epicenter and the site respectively, having the similar geological condition with Kurokawa. Slip in the station
area is about 1 meter where the site Kurokawa had maximum range of about 5 meters[1]. We used the ground motion
recorded in the surface of KMMH16, transferred it to the axis of fault parallel and normal supposing the strike angle of
fault is 40, PGA values now changed to 12.2, 5.34 and 8.73 m/sec2. In the vehicle analysis we used the fault normal
ground motion in lateral and fault parallel in longitudinal direction of vehicle and used the multiplication factors for
normal component until the vehicle got toppled where fault parallel and UD components kept same. When the factor is
3.38 the vehicle topples on 10.56 seconds with PGA of 18.05 m/sec2.
5.0

CONCLUSION

Vehicle analysis using the angular momentum along with CPLM method defines the toppling condition of vehicle. PGA
required to topple the vehicle is depend on the dominating frequency along with the amplitude of vertical acceleration.
When the ratio of lateral to the UD component decreases, the lateral PGA value need to topple the vehicle getting down
but for increasing dominant frequency condition PGA value also increases. Relationship of PGA and frequency for
toppling of the vehicle using the sine waves are shown in Figure 3 which can be used cautiously in estimation of the
PGA in the near fault area in case of earthquakes like Kumamoto where vehicles were toppled. Preliminary analysis of
toppled vehicle during Kumamoto earthquake in Kurokawa area near the Minami Aaso bridge suggested the PGA value
of fault normal is about 18.05 m/sec2, 3.38 times in reference with the Mashiki area. When we select the calculated PGA
and ratio of NS-UD as 2 then the dominant frequency in the Kurokawa area is nearly 10 Hz. Evidences of large slip and
surface rupture along with the damages to the buildings in that area could agree with the higher frequency dominance.
6.0

REFERENCES

[1]

NIED, Strong-motion Seismograph Networks (K-NET, KiK-net), 2016. [Online]. Available:

http://www.kyoshin.bosai.go.jp/.

[2]

R. R. Parajuli and J. Kiyono, Circular Path and Linear Momentum Method for Seismic Response Analysis of
Vehicles, Front. Built Environ., vol. 2, no. 16, 2016.

[3]

R. R. Parajuli and J. Kiyono, Ground motion characteristics of the 2015 gorkha earthquake, survey of damage to stone
masonry structures and structural field tests, Front. Built Environ., vol. 1, no. 23, 2015.

[4]

R. R. Parajuli, J. Kiyono, and H. Yatsumoto, Safety analysis of vehicles running on bridge during an earthquake, in
16th World Conference on Earthquake, 2017. [Submitted]

35th Annual Meeting of Japan Society for Natural Disaster Science, 20-21 September, 2016, Shizuoka