SSSSSSSSS 08

UA RTERLY

EPURT
V01. No.1

RAILWAY TECHNICAL RESEARCH INSTITUTE

 A Method for Dynamic Characteristics Estimation of Subsurface using
Microtremor on the Ground Surface
Yutaka NAl<AMUFiA*, Dr. Eng.

As methods for dynamic characteristics estimation of surface layers, investigation with

boring and a method which employs microtremor are well known so far. Boring investiga-
tion, one of the most accurate methods is costlyaaqd time consuminggqnd is not available
all the time. The method that employs miéilr/r!:’iti'remor is handy britihas not produced
satisfactory result to this day.
This paper describes a new processing method that employs microtremor observation
yet producing accurate estimation of characteristics of the ground motion. The method
uses vertical component and horizontal components.
As a result, rrra;‘§pe¢rrum ratioof the horizontal components andfvertical component of
microtremor bears irefseirnblance to transfer function for horizontal motion of surface
iilaryers. "ii tr;/rt
1. Introduction completely in order to eliminate the effect of a tremor
A survey on the earthquake damage shows that, whose source can be identified. But this procedure will
considering seismic intensity equal, the effects on substantially detract the benefit of the micro tremor
structures vary considerably: some are free from (i.e., readiness for measurement).
damage almost completely while others suffer heavy This paper proposes a new method to estimate the
damage. This phenomenon occurs due to difference in dynamic characteristics of surface layers by measuring
the seismic response characteristics of the structures. solelythe microtremor of the surface. According to this
The surface layers carrying these structures also differ method, stable estimation of the predominant frequency
substantially in the effect of earthquake due .to and amplification factor can be made even in the
difference in their seismic response characteristics. The presence of a certain degree of artificial tremor and there
effect range is specific to the structure as far as the is no need any more for time restriction on microtremor
structural characteristics are concerned. ln the case of measurement.
surface layers, however. the effect covers a certain Q

spread of the area. Consequently, understanding of 2. rEIffectof-Surface Layers on Seismic Motion V

characteristics of surface layers may be said to be an
important review item in analysis of the earthquake
disasters.
ln this context, the characteristics of surface layers
(particularly, the seismic response characteristics of
surface layers) have been investigated positively by the
administrative agencies and other authorities.
Boring exploration will surely offer highly accurate
data concerning the dynamic characteristics of-surface
layers. To understand the surface layer characteristics
over a wide area, however, lots of borings must be made
in high density. Boring exploration is therefore not a
ready means because it demands considerable manpower
and substantial time as well as tremendous cost.~ From
this view’ point. a study has been extensively made on
the method to estimate the dynamic characteristics of
surface layers using the microtremor (readily measura-
ble). If large structures are not to be considered, the
microtremor of a frequency range of 0.5- 20 Hz will
be measured. This frequency range tends to include
tremors induced artificially, but it is necessary for
investigation of characteristics of surface layers that the
effect of a specific source not be large. Accordingly,
measurements have to_ be carried out during midnight
(around 3:00 a.m.) when the social activities stop almost
* Geotechnical Engineering & Disaster Prevention
Laboratory
Earthquake involves release of the strain energy
accumulated in the focal region and propagation of a
part of the energy thus released to the surrounding. The
dynamic characteristics 0 (fl, f means frequency, observ-
ed at a certain point include all of the wave motion
radiation characteristic F(f) at the focal region, dynamic
characteristic T(f) of the wave motion propagation route
up to the observation point, and dynamic characteristics
S(f) of the surface layers at the observation point.
I-1-iii-ishows
how the difference in earthquake and
observation point is reflected in the tremor waveform.
The acceleration waveform for 5 seconds around the
maximum value is indicated vertically and horizontally,
withsimilar earthquakes arranged vertically and same
observation points horizontally. The magnitude of
earthquakes rises toward the left of the figure. Evident-
ly, ma seismic acceleration waveforms at observation
pointsare generally quite similar without much variation
between different tearthquakeé. though there exists a
.-strend that a high-frequency tremor prei/ails, when the
magnitude of earthquake’ is small and the lgwiiffiequency
wrrltrilltremor prevails when the magnitude is large. This means
thifrt the observed seismic waveform is similar for the
same" observation point (i.e., the same surface layers
characteristics SIf))ieven' when the radiation character-
istic F(f) or propagation characteristic T(f) is different.
In other words, it may be said that the effect of surface

OH of RTFTI, Vol.30, N0. I, '89. Feb- 25

 layers is most critical among the three factors influenc- the importance of understanding the vibration character-
ing dynamic characteristics. This fact in turn leads to istic of the surface layers.

" 78094 it, 7T2Tti o2TTY irrccs CD C3 C3 LO f isU4kms CD TTTT7rc02 T2acfrTs TiT2Te0oTsTjT T" 14002 macs r

E
lI—|
Hl7.I]l'lnD kmilf 2 .on zorur rr-6 sh 10km
I V W I *__.
I i “Q 010'! “LO:rD OD . Ac: '_.C:r :rC'J
.3. 3
3 Oi :;rC.J , . CDCO
fiflss, IhI~lUk2Lr1IS_§hnZ[]hf"!'tiHe§, She. 2oik|“I1l£‘r’lrLr 7K7n:§i, §|'l|lQlm‘1nl,:6f!fp iljllwlw
I '.‘I u-\.l \.I'tlTJ‘ '. I
‘ T TT r1'\lx-ITTDTCJI1 ‘ 36. I 3? or-Egn- ea
i'I‘\IlI5TITTcli$T‘ minim: TIT: II 7:TCjls 9° 3,8’ l rl-ran=TT§TTE;iTsT‘ 4,3 ‘

rHiratsul<a l ‘ I r "_...
l I I 1 I ,
I
LT’ TT )r~i-- mg miii imam FniiTT'45T(T;;r;i Fm: :7 at ma:ri€TC.|r;TTTT JT‘ T TTTTT,iTn\ax- §;|;l T I, T TT pea
1l
I
\\Kamonomiya I
I I

I ‘‘I, mu- Tc.TT

ls rnurr so
sis; mu- §'lrTTc5iii -is . J. z E? )Tr¥\iiiTIT.7'€fi;|iTT _ ‘_
T _ Tc A _
lTrnu:-T41 iiiiitl‘__ 7* 7*. I
A

I
TT

‘I Numazu

T
TTTTTT

TTT
)
I
I
‘

tiiists1 is
Tl Tr'r\l:i§7TGal:T Tinlnngb

T L i‘i‘\lTI?:lT'.TTTGeTl$ atI - eitrtsrsleT-stI

r
I

-
T

Tf‘T'\IITI
"III

2 vie ls
TT
T To‘Gals
I
"181

Ii-TIT
5
5 T
Tl man the is-s@T

Timon; at3|: H
‘

T
I

T
F maTiT6TTTGars i\T\iTiTIT§TiTTBnls Tr|'~|i§fiTC§IsTTTTT”TTTTT
“

1)
L inn ~
i
TTCZJQTIT moi G
I‘

TTCg|T; I GU12” Deli T

It
TTiT-lT§TiTTCaTtsT)
L T Th
‘

TTTTTTTTTI TTTT
“T is ‘mm
‘
T
Jl
,-

as es i-
l lwabuchi wL

TT TTTTT min?!‘ Calif I ll l"l\ll is mnTT mu

‘tom
,lShimizu _ I
is.iaie
2;
TT TT T

, Yaizu
TT TI Fiilinrb Cal;

~
s s s TTT T T} n'uI=7!TTGihI 1?: 1

l
I T;
___ _L____ ___
manifi Gaul)
,
__:_

l
5
l
__

T
E
3 Q. 53
\

IT
Shin-Kikugawa
r Irsarr-T1t""c;r.“

if is ‘
i s r
ems»-r:'"c.r,‘": mr;.ra ca. *a.T.irrs;r.l

ii
so I

S
ca. It l T: r r----rs Gm p

T cm»--H Gals’

T
T T 1 "‘Tma=if§'t‘;5r;§” first\ rt»---no-car. y ~51 .r.t mu -- —- r —~ .~..m...t

E s
y , T mu
“ l, I I

I Shin ' lwata V& il;*'tvAV5oi\ I Al I l I 1‘

TT TT T T "‘ T "LTTT *:T‘* TTT” T‘T ma-H’;-cilit T“ ,‘ T TT T r ‘ mi»-1 Gm 1.
l , l -
‘Shin-Takatsukai i I I, ‘ * . l MW

1' Q ti»-elirrt are t_ttr*l9rj_tTlr r tater -j t .-:-_

TT ‘=— r:r'f 31-L...,; 3*‘-*1 III -.1.-tJTt 3E|_t.rr=.Er i,T§T g EL- =:_t..t5Ii1 l_;t_r..1.§I
lIL‘_U"l

Fig. 1. Differencein seismic waveform due to difference in earthquake and observation point

3. %riLong-term .Ccir1firruo'us‘, ,~O,bserv(a’tilojrfr(li 3.1 %QQb,seNat,ioni('l.at~l<amonomiya *

The surface layers are nonnally exposed to tremor by The N~ value of surface layers at this point is 15 or
natural forces (storm, sea waves) and artificial forces more and exceeds 50 at a depth of more than 35 m. The
(Plant, automobile, train and etc.). rsea-tvavr-srnaaca ah N value increases gradually while repeating an increase
tremor era relatively long period I2i- 31sec. ormore), and decrease down to this depth. As a whole, the
the socalled microseisms, while the sstonrr and artificial surface layers are silty.
forces do a tremor of short) period, the so+calledmicro- Measurement of microtremor was made from the
tremor. night of February 8 (Saturday) to the morning of 10
(Upon wave motion comprising the microtremor, (Monday). 1986. Though this point is surrounded by
there-are two theories: §'th(e.'¢isolidTwave-it-th'eorys' and A-The plants, their principal equipment was considered to be
surfacewave theory. Both theories have tried to prove f
shut down while the measurement“ was made. The
the respective validities on the.basis of measurements of measuring time span is considered to be cover a period
microtremor,‘i"but- are not-"yet-I recognized as established. with fewer artificial tremors.
The microtremonthough" not known fully, has often
Kamonomiya NS component: February 9, 1986
been chosen" for engineering purposes such as estimation 3 r r r~1—T+T1 ~rT—T+*T+#-'-YT] 1TTrTr tT*1-TT-rTT+*Tt=1 1 I -
of the prevailing frequency of surface layers with micro-
tremor, etc. 3 }%

For the purpose of utilizing the microtremor in the E ,l

tude -s
nun- 3. ._.
engineering’ fi@ld- ar1...9t>.ss.o@_t.i_9.n,--.ei-.-r1rist91£s£r19_t-..t~.e§ -|r_.

mD
I—4 .5 ‘F

fl1§.§I.§_-§.€2_i}.'ELlU.9.!!§lY_IQK,-EQJTQ;-.lI1§I]..§Q...I19.E..[$..I.9._lg9Pt.iTY II-O

"st Jr

_I.I1§W§,l1§re,9,t§[istjps.. The observation toTok plaTceTaTtTtThTe tcay (mk

0ne:
$2.9

°=I ,0- P\
Kamonlorniya substation of the Tokaido Shinkansen and T‘ 1
‘ h
‘<1. P. I.‘--= t>.____ ‘WT I
‘=1--I
Veo v rt .fl._ 6/ ‘Q
the Tabata substation of..the Tohoku Shinkansen. A \ F?*.
V yr. 8 A
velocity meter of specific period of 1 sec. equivalent was
7/HT?‘
used for the observation. spectrum discuss- /Fm

0 4 1-4 4. u,.1__i 1 L_-4 ,4, T_|__1 r- r r__l_r-r_.r L4 I .

ed below was obtained by applying the hanningwindow -5 0 5 12 18
II1Fee.>times.-to waveformvoata (2048 "points; 20sec.) .... . Time Ihour)
after frequency analysis with FFT. Fig. 2 Time change of the horizontal amplitude
of the microtremor
Z6 QR of FiTFil , Vol.30, No.1. '89. Feb.
 itr 1 it r"
{%Fig. 2 shows the time change of the horizontal 3:00 a.m. and large around noon. The extent of velocity
velocity amplitude. Needless to say the tremor due to amplitude change is around 0.5 — 2 mkine (10'T5 m/s),
the Shinkansen train is far larger than the one shown with the ratio between maximum and minimum ampli-
here, but it is not include in this figure. The figure tudes being about 4 times. This corresponds to the quiet t/‘I
shows that the microtremor is smallest from 2100 to time at weekends. and the amplitude might be mifch
larger during weekdays when tremors of plants are
added. V
3---shows spectra of hourly microtremors which
Kamonomiya : 24 hours
NS component are overlapped for 24 hours. Spectra indicate fluctua-
tion according to the time and it is difficult even to read

gilt T
the prevailing frequency of the surface layers. iFig. 4
shows the amplitude characteristic of incident waves to
8'5
surface layers estimated from b0l'iI"lg investigation and
kn
P8 velocltv logging The ml<1'0i£.?.m9.L..!tl1L*3Q-i..$..99D§.iQ§t.:
.m!ar.m.s.-aq--J. l ....._.__..-_._._l...........
fe uenc _(about 15 Hz of ‘/
surface la ers did not b,ec_om_e,_remar_l<a_b,le duringgthge
\@w4wwww:___. wee--- »___s.s__,_,_ -___ _______ _ ____ q_.,,,,,___:,,¢ _______.,_,_,_,____.;___
P U1
SHS
ude
m b——ll- time span when_?__t”he_ampli”t,u\d_e_" becomes smallest, but itm,_
L1
.._ __‘
_.__l_ m was ebsewed during 8 . re'.at.i!s!x.,.9.Pi.s£ .£i.£ns;..ses.0-,tr.9m..--
20100 P -m- t2.i.§§.i'-iI95?!Y...i.9.§.i.QQ.8.-
Four
erampt 3.2 ggasawarran at Tabata I
The surface layers. at this point are 18 m in thickness
0.05 Jr--~—~ - ~ ~ ' and sandy as 8 whole IIJ.?...I§I..\!.i=.i.I..l.!§..i..Q_§...§i§.QII1-Q.1.£t.mJ§_
0.1 0.2 0.5
- Frequency (Hz)
1.0 2.0 5.0 10.0 20.0 ..sLev.e.d....3..s.r1§...restsassas.ta9sal.Lv.-.as...ths-.dsr2tb_J.nc.rsasss
s.e£.£.§.e£b.§.§.a.!;<.>.es!;l.§9..ét,,!hs-d?.eIh_.0f..1.§-m~ Below this
level, the soil has an N value of 50 or more, with the
Fig. 3 Time change the horizontal spectrum of the shearing wave velocity at around 500 m/s.
microtremor Measurement of microtremor was made from the
night of June 24 (Thursday) to the morning of 26
(Friday), 1987. Tabata is located in a district of intense
10 _
I
T T T TTT T
i social activities and_f_l,_a_g@_Tr;1_§w%}th~ microtremors. Besides,
Karnonornry the measuring time PspanT’T‘ir\r%%T\one considered to be
5 i—III _ heavily laden with artificial tremors in the week. s
XIII I V! =5-"2 2 ! |lIr 1
ilI_III

-“_-n |lI||I| I
tamp
ude
10
I

it
iii
EA!-.=._ s s .: s IIi"l l _IlIl l -@552!!IIITI
i ihjppg
--In-5
‘ I_EIl l is-.-s.I|IuIu| :| nnrrir
i
i
V ' I
- Tabata: June 25, 1937 °* Horizontal
er?“ °°"'p°"°"t
component
-1r
Ii
IIHIIM
" 5 Illi- -s
U

T iTfT
-o
o.
-T- -i=0 T'_‘:~o ,0. +444-.4 +
(m10
kirn/s)
ne
3
___“ .-°T
T 1 ud 1 _-=o o=--2'1":
QT”

1- - . ‘ ‘ 7?. —-—-——“— ,—.———.—,*.r* . =7 c:;""' oi-_-anQ--. __

T E .,/.
‘I l I 1 i
ill,‘ , ‘I
l ‘

I l
0;.“-Q
--QT ""0-._ 1:0
¢".-oT

~ \ ~ o <T5Tt><>°‘.
E -

ll yampt
e
._. '5TT "x , ITT

IIIIIIr--r
sar- i: illlllT‘
Tl‘
_

-
IZ c" ashi =.4....: tr _ _ T‘ --o L44l
1-:-
cul- TTTT _ a 8 be ed;t
5I -I IE;InngT;To I

ii.
-n-n

Ampf
for
magn'
cat
onnc'dent'cfation I _III ZfllTilT-'l‘Bll A I::::=
Ileiiii
- .-
TATTTT
TI-.-g I‘Tun Ii -Ins“.:1 '"::-:.-':!:2:' —I‘- -~ I=iIlI,g
r:g::::i
-IIIIIkq.||| I
T inc!‘
1. -_ ‘U-u
Veoct I

5
i

0
I

s
I

12
I

re
I

24 30
.1

Time (hour) £1‘ //’/A-I, 5'-I

1 __
0.1 0.2
Fig. 4 Amplitude amplification characteristics for
incident seismic motion on surface layers at
each point, which is estimated from results
- of boring and PS prospecting
Fig. 5 Time change of microtremor amplitude I
0.5 1.0 2,0 5U to 5: @
I--I“pi OI
Frequency (Hz)
15"shows the time change of velocity amplitude.
Tabata is a place subject to substantial effect of the
train-induced tremor because the Tohoku Shinkansen,
Tohoku Line, Keihin-tohoku Line. Yamanote Line, and
freight line run nearly. The velocity amplitude was read
avoiding the train-induced tremor as much as possible.

OR of FiTFil , Vol. 30, No. 1, '89. Feb. 27

 As in the case of Kamonomiya, the velocity amplitude indicates that 4‘ to 8 Hz frequencies are the specific
of microtremor was smallest from 2:00 to 3:00 a.m. and ?’[_i'6I1_10l‘"-Of surface layers at the Tabata point. It is also
largest around noon. The horizontal velocity amplitude, known that these frequencies did not become prevailing
on the other hand, fluctuated between 5 and 30 mkine, during the time span when the microtremor amplitude
which is more than ten times the case of Kamonomiya was small. Frequencies of 2 to 3 Hz prevailed during the
(refer to :Fig.-12).» T-he l'8iI_Q_,Q§+'D{\r88[l:[T1aX*lE1LllTlW8l1d quiet time span, but did not prevail for the estimated
-m.inims'D_§.F‘lPaf(6..I. TT amplification characteristic of surface layer,
>r=ig. shows spectra of hourly microtremors overlapped
for 24 hours. It is known that substantial fluctuation is
observed within the frequency range of 3 Hz or more.
The mean prevailing frequency was read every hour on
the basis of the above spectra and plotted against time as
II -r

frequencies» and 4' to 8) Hz’ frequencies “prevail-Tin.r'imost of

Tabata EW component Tram
the cases. The amplification characteristics of incident F

5° |||| _ .rrr|T T
wave to the surface layers estimated from boring investi-
gation and PS velocity logging are shown in*Fig.,4.T+ -This
i- ,, I.%lI;gll|IIIl_l!!i1:;;:":::1::.:;
k
mne-s
lligrr iiililffl.Hie-=" —I.,.-_bait - . ._."‘=_ _*r:-=_.- 5

Tabata EW component! 24 hours

' iii I lrllnrfl
I‘'T.
-r.— I -_ I-
in:._'f-—- - er-I: T'-*"—5";-."—'_=-._\—._ ."_'?":_‘-"1—- _-_- 5* i.- .- ". id
-El
100} ; ; T
i-— ‘ fi-- ‘_

F3
- n-'ch.-=;'

T II
¢I'Ib

5-OTIII -- E_
III
3 00
ai-s-
-- I

- i T ' Iritri.
__ _ _ it
$52)":
5 _ qg‘i;

Four
S53
teramp
ude
I‘l

'r,ifT'T=,;ry ‘ll 0.1 I IIIIC

- .-4 4-

__ Q
*1
"I-‘I-Q
- ,Q—_ ,
:"
TTT
El1
....
-.
1-0

0 5...
,__
’ J;-§l\\!crri§r
’J?.'?-Ii)’-\'\VJ!‘... I,
4' .I/Ji\ -. it‘{'TIh: fs\$L'0Tfavlt III K
1* S=\1[l_!1I_}1-
I *n.\.\rwa rl. Q Ie-.z-s»*-._|I"_'-ra- '-"~-.
':3-TIT.’
""l!'lrl
r “Ii
3% int
t-4:r!-1.;."L:'-...9:*.."‘“?a=i‘gi ~;-=-:.':*____,'=-.—_:=I
0.05
an9-itg |T~ | '":.lll==
: : FiI=flIi:!
s':Ti-|:::Fi I- =13".
5"
-'1-T ‘.4 T="-lg
Frequency (Hz)
TT.5 TIT° lllll ‘"ITT|||..
IIIII_
fliflgg
Is»T‘TIIIII_.
Q-ERE : 552":
:2il%_.P 0.:’-=-7 E8.I‘-T- 1::
I9E-3

" r I '-- mrrelrfirr

4 I -- I+-- ‘ . .-t=-':_—:=*-ril]
'- -|__-
__;-\?-._1-IaJn-"-4-I ._

"" A 4nrr;.;rrrr_. - “'-1.‘

;-"41.-_".,:q-._¢-g.Ii_:;',:._

\--III,
Fig. 8 Comparison of tremor spectra between train
IMII - _ .-'1 ;-3"-=‘,_ -,.8-g.:7-._
.Iu‘ ‘

Four
S53
teramp
umkne-
de s
TT"'
'}-T1i-1-.'—1"|‘Tt“i-'Tp¢i4;:,T induced tremor and quiet time span
5-T'Z‘.L'\- ;-4-r~
'A-E.ZI Q -if-:'=_r-n|
;_I
"'lI'f". I".£;0;f;-
Ti%;_'i, _'T‘,:"--TT_: .@';:-r
(intermission of train operation and 3.00 a.m.)
0.1)--
' re-
‘‘I.i‘ll1i7,] I . E?-""‘
TTT='‘T-' .Li.*".';_.:"-.?'-_: -., -‘
TTT i IT IQ
I IOI, +1!_ ‘ Io 01 i'_-_-c.i t.h ¢li_»b
. -Qi1$—I l - 2
ft-.If_4!‘ IE" '€I‘t§'\"="U,'.-:.;'=¢I:-;.‘-*I"I-Iti;—q:t—-A-.—"T:D"vT“*'@
Ti-'iT-. i__:
=‘Gp-.*'"
E1"1: '-1."—'‘I“'9"' 1-;?_i$—‘rO
._Q
!:.''_?»:‘q;;_.-.:‘.~tl=_a¢1.-?—-I-In-Q'.--I"-‘-5.-:':-!'nQ.
-..fi‘|§I"‘'"l.-:|_,"-.'5-'~"=at.-“"'=_-=-‘.--"
—\=_r1;:=-."“.=-_ B . ca
TTT‘-‘TT'-T- -P-TTTT

Frequency (Hz) ows comparison of frequency analysis results

between the train-induced tremor component and other
Fig. 6 Time change of Horizontal spectra of components with the observation results of microtremor
microtremor at Tabata es.'s.@9.\t<9..fre91I.*r..i§..s9rI!esr,is9!rr,.s-.t,rs.usr.ei
4 Hz or more was induced by train to cause a considera-
12 L
A; O
Tabata: JUI18 25, 1987

0 Vertical component T.)

Ty!-.<.=.tr-nseiliiii re -streetthis
inTd0ceTdTtr.emor was remarkable in the vertical compo-
11 0 T 0 Horizontal component nent. Such fluctuation of the prevailing frequency of
10 ,_ . 8 O0 1 microtremor is considered due to a substantial effect of
r 0 o o O --5_f the artificial tremor in the vicinity.
0 0 co 0 0 00 0 O
duo
0
T)-----’—'0------- 00000 e-- o-o.----O-- I
4. i-'_E§tirna'tion Method of Dynamic Characteristic of the
6- 0 0
\-- . e o 0 o . L.It..aJ
A
at-..-.-, f Laye rs “-
Q
.

O C. -'*
-4-

lt is said thatthe microtremor must be measured

fre(Hz)
.._ng
quency U1
CO
C0
-Pb
.

H» -~
14-

during the quiet time span. Indeed, the observation

1
‘T1
iF1"“T0 .
Q
-4-
result at Kamonomiya shows that the tremor like a
3 *—- . . 0 0-o‘Qe 50-I
Preva 2L
-—-----oooe0o----
8'0 0
----0 ~ — - - --0--0-cc-0*
O6 2 ._.\
specific vibration of the surface layers prevailed during
‘-

'I the quiet time span. At Tabata, however, what prevails

1"" "1
-
0..4_r_r_1.,,.,,|,,,,,|,,,,,|,,,,,|,,,,,|,g,,,‘
4 during the quiet time span appears to be the tremor re-
-6 0 6 I2 18 24 30 lated to the geological structure far deeper than the soil
Time (hour) which is considered to be the bearing soil in terms of
engineering. A tremor with a frequency whose amplifi-
Fig. 7 Fluctuation of prevailing frequency of cation magnification increases as the so-called estimated
microtremor transfer function of the surface layers was not observed

28 on of arm _ Vol. 30, No. 1, '89. Feb.

 frequently. In this way, it often occurs that the specific
frequency of the surface layers concerned cannot be
found from the frequency analysis only of the micro-
tremors during the quiet time span. It is also generally
said that the prevailing frequency of microtremor
fluctuates less, but the prevailing frequency of micro
tremor at Tabata and Kamonomiya fluctuated sub- ’e/Z;_@I"specif.lc direction prei/_ails_, on the hard soil, with the
stantially depending on the measuring time span.___Such
.£,I..sis.t.ria..t.Lo.0,._stay. ...l2§,._..€ik!.§¥.-19....T£.I1§...I§.£§.’3).9.E--§.Q.&1.EEE.€-
..e...e_i.glJ-b.orl19.9r;i.-.e.L-ths..m In other Words-
_ the tremor from such source may be-considered a noise
-*’ from a stand;p_oi_nt- of estimating)-(the tdyrramic» character-
isti cs of rface ayefs sin'g= the rm icrotremor.
The artificial tremor source -has mostly the prevailing
- .~.TLi-1-Til vertical motion and tends to induce the Rayleigh waves.
fad
T This may be the case of Tabata where the effect of train-
... fr; induced tremor was remarkable in the vertical compo-
.,
‘ I‘.
4
“I nent. _l_r_3__111i__s__cg,r1te,x_t,.,,the Rayleigh wave is assumed as
‘..L‘
"=‘.'-:\1“h
“~
?f'Fig.r=9{;shows the ratio (AH/A V) of maximum values‘
between horizontal and vertical motions of earthquake
for each observation point or area. Evidently, the AH/
A Vvalue of earthquake is related to the soil conditions
ofithie observation point and AH/AV is close to "1" for
thejirm soil; From this viewpoint, no amplitude in a
tremors 'ev.en-in all directions. ="T""-'1' -it
Dividing the tremor into horizontal and vertical
directions, it is considered that horizontal and vertical
tremors are similar to each other. Such tremors are
amplified by the soft surface layers accumulated on the
hard substrate. The horizontal tremor may be consider-
ed, to a certain accuracy, to be amplified through
multi-reflection of the S wave while the vertical tremor
is through multi-reflection of the P wave. The propaga-
tion velocity of the P wave is generally higher than 1000
m/s, and the tremor of around 10 Hz or less may not be

T‘ noise of microtremor and a method to eliminate the amplified so much through multi-reflection within
effect of Rayleigh wave is studied. surface layers of several ten meters at most in thickness.
‘Rayleigh wave, on the other hand, H ,,
' -» 4.1 Basic Concept rleipbears-remarkably in the vertical tremor. Accordingly,
if There are various research attempts, but there is no the degree of its effect. may be known by determining
established theory concerning what I<lI'ldS ' of wave )the’#1ia'tio").of-vertical tremor between the surface and
motions the microtremor is made up from. Namely,%-the is-suibstrarte. Namely, the effect of Rayleigh wave is nearly
microtremor can be considered to consist of various Zero when the ratio is approximately "1". With an
wave motions. (increasing ratio, ‘the effect of Rayleigh wave may
become more critical. Elirnination of the effect of
after r r rs rTrr r:Tr r ts TTI fiayleigh wave. isstudited by using this ratio.
)-

4)....-
I’
O

_l
1(-

I
s 4.2 New Transfer Function Estimation Method for
Surface Layers
Sf;-yof
surface layers is generally
I O

1
2fl_#__6 I O
I
C
OOIOO I

I
O
2r" O
0
r _-u ‘I73
"TM:
'1“

8 SH3 and SHE are respehtively the horizontal tremor

for
AH/
max.earthquAakVe 0030)
Q0
i
I -L
spectrum on the surface and the horizontal tremor
spectrum incident from the substrate to surface layers.
—-
3FBIO 9.24-1! 1.1.1 I 1 J ,,___-J. r . »-sq But .211, this
-,!!?.‘!.§-
' IIIIQ

3?
nI"I\ .—-|‘u3IunII—'I|-it-nu‘ 4|-In P-lh
". Since th_eiTa.oi.f.ir;iai-._ooj§§.I.-.i.§...II19§i..LY_..--9.[9E§_9§i€€i
3 rock rock soft
,.Eta.rrl.e,i9.I1...waxe. -§'H.Si.,-of. .,m.i§=.rQ1tem0r, -..poss,i.bIv
be
/vert
c (rock .:-. softi Iii

§.fi§9I9.d.,.i2.!._,.B§Y.l§l.qh.Y!.B.\@..,. The effect Of Fiav leigh wave

i
1|‘-pr 'iF

arock
It III
mi
xtsoftra j

should be included in the vertical tremor spectrum SV,5'

i ‘II:-II‘
USU
M'yako Ii
on the surface, but not included in the vertical tremor
Horzonta
It

spectrum SV3 in the base ground. -‘Assuming that the

Pen‘
zu
Se
kan
tunne Eaanrthquake
-_p

C'excoety
var-mart tremor is not amplifiedby the surface layers, the
Morokasubstaton
below should represent the effect of
Mex
C'ty
arof
coea
1-;
__-
I1
the vertical tremor:
I
M
arofea KotTokyo
arextsoft
eoaof
ra
en-
X ig, L-‘*1 -1
\T~..
e . .4 M" H,

Mex'ty
ntof
ercmoC
edatearea
-1.
HCLISTFD8 =4 1 LE5‘ will take a
nu-no vsruéi-’|argerrrrhan..‘51','*-.w,rrh increasing effect of Rayleigh
Od
riwave.
Mex
smcc
the
arof
¢ea enter Assuming that the effect of Rayleigh wave IS equal
<3 for vertical and horizontal components‘, ST/ES ITIBV be
Fig. 9 Ratio of maximum values between horizontal “I.

and vertical components of earthquake

OR of RTRl,\/ol_ 30, No. I. '89. Feb 29

 considered to offer a more reliable transfer function S77 horizontal maximum value by surface layers can be
after elimination of the effect of Rayleigh wave. estimated from the ratio of horizontal and vertical
maximum values on the surfaoe.
Namely: S77‘ =ST/ES. lt is also considered that the tremor in the substrate
=Rs/Ra is equal in all directions because various wave motions
where {RS ’f =1‘-SH5‘/S VS repeat local refraction after principal tremor of
RB =6'HB/SVB I-
earthquake. "t-ii_,,.rff,.;gjs.t.gi5F.tnererqre"prossible to estimate the
0‘*ieharaicterisrtcsrlk‘oi surface layers using the
R5‘ and RB were obtained by dividing the horizontal surface according to the above
tremor spectrum by vertical spectrum, corresponding
'@$Pe°il"e'Y i°.§.&',.C.f.==3£¥?..§.fid S"bsttai§,2§.E!h.9.!.!a.ke..1te..m0rs.
As shown in ire 5. §*e.rif.icatioiri'ifbfithe‘Proposed Method l
relatiTvélys*iwideffrééiuéfistwfzslsrangeii firm 71':F':iQSi,T'f1_';'1"'8l'id 1t2"show R5 (obtained from continuous
seestrate, propagation is-gviien insii-direatans. is observation results of microtremor for every hour) over-
lapped for 25 hours, respectively for Kamonomiya and
Namely RB 3'—:1 Tabata. When compared with plottings ofFigs. 4 and 7,
Therefore S77 l-=R_5' RS varies F less with the time, and the prevailing peak
J3 _- . 1-,’. “L -__’
position and magnitude are stable. riiiig.-r-.;_ =
This means that the transfer function of surface Karnonomiya NS component: 24 hours
layers may be estimated from the tremor on the surface
only. In other words, the vertical tremor on the surface s ~\lU.U _ I
W is
retains the characteristics of horizontal tremor of the
substrate, thereby substituting the latter. i%*?lfhel>xrRay.leigh at‘c(_S\\_.5.0 4
$5.‘-.,g§=| in-=i="l ='-" IIIIIIZSIIII2efiii i“-'=-alnln -i":'- -II L‘:-ai‘l
""-"o7-J.-__

- wave acts to hulliéfyisuch substitution. But R51 becomes 4? l ' ‘I _

‘
more or less than '1“ in the frequency range where the
*’a’>‘-1%
Rayleigh wave prevail 0 s "fd thus it isnot remarkable in q . iln,
l ea!
. -ease4.IIIl.l l
. __ 0'-";ii’/l"i“I~‘*.1'-itil mi il=
-.-, .4 ",i‘:.:§‘
~"{,i i~.;‘i!-l 1',i',i;
J‘ L

the estimated trar"is1?er5§{fi?€ction. ,lt-may1thusibe.pre'sum-. ilierl -

_-
C3SPBCT
f FUTH
‘ :1‘-D

‘t-*-\3*~=‘*i~ §.;§~;\’-lg}-I-IIIIII .*~.':,I~'-n-;~S..-'t*-l1;,.u_\'--T:_2‘l-~"''1!_.*—:|-.-':i"~:‘>":r‘"

I’.
Q‘ w ‘I:

li".ii.'-*-'-..1-.
_ ‘,1 I

M 1-0 .. i I I1"
. .'n r"
ed that RS includes the effect to eliminate the effect of L‘; ti"

Rayleigh wave. t
f. . -'
VEIT \a_g»;f-:2 Hi ll
W
ti.II no i ll l];l;'tl._li
h -I
' ~
_ _§I __I
— . il'.'l

-fill‘.
I r |ll 1 ll
I
Iii.
1':-=
5.. “"- -an
l i rnlmi g
-2 I
5-'i 7 -:..~..:
' a:;. n. , ;\-IF:

SQ,
‘
5.0 W at ' **+‘*‘*‘"'*“
--- Kamonomiya l24-hOUf average)
*1 l frat“-* ._
!IIl:€ |
_. iii
. _ _ _ i
.
____ _
f _
it

l
Kamonimrya
I’
' 4.; i __l
rm 1 IIIIII Surhor
I zontfacea Q:= si ii
"';is
I
§ :@-.-Ie6-= 4-i es s:-=- -
‘In_ner.
>-
____--
i!-‘is.
'‘ 44_‘-‘I

5 ' HUcs‘ Li
—A ; _._ P1-B._ -M F4 4: \i5::
&-L
‘:.,_ 9“IIl
U‘ —QII
- __ ‘r0:
4;“AT___ ll
!‘°
4.,as . _ 4 Qln
y s.0 10.0 20.0
""Q ‘Ii’ I \\ iU r 1
pi .'1ll_I1—tI.l
3
_
\ I. ' - Frequency (Hz)
II?
' s'a'5'i-2'.¢IiI-'o'~l'.I'Jl
""

'
_
W H
__
llll
UH
Ill__
.,_~
'll'i!l'
Prll
_.__.
a0== "-I-\fli =. ; Fig. 11 ‘ Time change of horizontal/vertical spectrum

0-II
5- 52>
P U‘
’ *' ._.\l—* r__1___ _ :, "
1;’
5-s0=-‘i i: :i:§!;:: 1:5?-.
Fungi -'25’
i_hg_c:5;§:
I!§i;i‘-=._z': =-: ratio on the surface
‘l

I * *_ : r
l
___;i L _.__ _ . _ _ _ '1-;s1w,-=Q5
0.1 0.2 0.5 1.0 2.0 50 =5
"'- Q3: 8
=- @

"' _ FFBQUBHCY (Hz) Tabata NS component: 24 hours

Horzontcamt'c@
/vert a U1 Q

"- . _,. ./
L Fig, 10. Spectrum. ratio of horizontal and vertical
. components in the substrate RE
P
(Microtremor: Kamonomiya, Tabata)

Note, however, that theiestimation accuracy drops

when there exists a noise itremorfagreeing with the L“ as
ratio\pRe_gctrum
H A Q gpirevgiling frequency in the estimated transfer function. ’-‘ 0-5 l i l‘ 7 ‘i

i l l ll i ll l‘
' RB should be f_reeg”_f_r%”n1W{?;éffect of Rayleigh wave and .1 ;_, L.’ —.~#. H V .--i"+‘:=- ,_~,_ Te ~ s . 4‘
“T
‘. l -w-»—
l

have the characteristics unique to the point (and also the ll l , ii

‘,l l ,
I
4
characteristics unique to the earthquake when the /vert
zonta
c
"'-
, l l
l‘
‘ I ‘i i
i i i ll
l 3‘
‘
i
.i
‘l ‘
1 l‘
l
,

l i i ‘ it it “ .i
l y . 1 ii. y l\ p, . y 1

tremorto be handled is an earthquake tremor).

Consequently, a more reliable ST)" can be estimated by °"
1 .. -.__ -,~._.l.tu . i,
multiplying with 1/R3 as a compensation term when Surhor
face 0.054,»-~: eel ‘‘
data oif§tl5i.i'€=§*§‘ground are provided. l 01 02 05 10 2.0 5.0 10.0 20.0
. - t Frequency (Hz)
So_far described is a discussion in terms of the
frequency range. Thesame applies to the maximum Eig._ 12 Time change of horizontal/vertical spectrum
value. Namely, the amplification magnification of the ratio on the surface

30 QR of FiTRl , Vol. 30, No. 1. '89. Feb-

 J

l° I1--i *_p3;ii.i;i:i:1:n n .5;

' ——-—-— Rs —---- s.,.,
'°~° --1 it
1-I

d
. §§@!!
': “_*“_ """ STT
130
_I'
“:0.
tr-_:;..!i3_°Ii
—I- ms.- .= : : -Rf
_

1 ‘“III'- IIIcan.-§z!e.%E§==:l- Elia-aéii l l

!:. :._." li ji1.—_.-

ilI-ll
1'“
sm“1*-£3ii:§fl_i—imilIlnimi=:'._:-srn un o I!iii!!-
!"lIl!.!_l|:r=w—-.9Q i ssi.=-55252
I_..J"'-z.-‘i
‘

*55;; "-1:
.-.-5 __4_ T Amp
'f'onmagn'
ccaton
at f F---
‘;-
.
I-' H°"Pg Emgz=-
aagezggn-' 1 0 2.0 5.0 10. Q i-*5 C3
‘f
f‘mp
cmagn
onaton
at L
§l AT I. -ii
.5. Frequency (Hz)
A §§
QEQIEIEEEE -ii III. I-I-It —-I !-i
N —II -_ EQEE1 -ISL: lllll /vi-‘c l" i"f'rr--»-E.‘ 1"
“~51 ||||1
‘U

.—
II tflII Fig. 15 Comparison between the horizontal/,,
0.1 0.2 0.5 1.0_ 2.0 5.0 10.0 20.0 K
vertical spectrum ratio (average) RS of
Frequency (Hzjfk AJ0‘l:°\i3‘-’-H‘!-. .i-i .-\fi,_'.'i.r.?»); train-induced tremor and intermission
l ii-{I o'1.»<€'E'L_$?/71-¥.?"*-*" /‘I/"""~"
at
tremor and the estimated transfer function
F lg. 13 Comparison of estimated transfer function 527' (_"__I_:gb____‘a_g)
(Ka_monomiy_a): RS (24-hour average) and
/Y zwiiira-\fir"‘é:'*-‘P-’i"'
S
TTR. p_tL~»<~Ln»_ic
s -

___ as
Maud

at
s
_ll%ll@_ST
_
'5CJ .:_ W7

ant i i ‘
W W W lil
- ._, . II.-‘I
==:=
1- :-I T . 5 .llil|
_(. Ililll. t
C ‘ii '**"“'”’
;
1-In
"if~"v mu l
Am
f
P
f'ccatmagn
at'onn
.
0_5
If.
‘:1‘,T T"IiIIISI
)-4!‘-=J5 : :L. § E§ i
II‘.
‘
$53:
“'

5i1.E:-".,I. . .
ii”; Amp
fcaton
magnfcation
ind
nu-

—IIIIfiI§VIl.I1_
IIIPJFIII-I
IlillllI
: r.; -' —:H.|!.|!.l: |'?.—-".5
-Ii-Ill!‘-{I
:g: :===:nIuI__'Ill'I_K
i1=:
.__F:-"—._T _
05 inrnnlllllnullnnn
0.1 0.2 0.5
Frequency (Hz)
1.0 2.0 5.0 10.0 20.0 0.1 0.2 0.5 1.0 2.0 5 b:_':::I
I;:g5z=:-.n=: n"3'nI u:|=
. =E§ § "“ €-=i'5 %:.-i|r.e-r="i_-
§g"'n-E‘£%Z$'§=: l"'_b
- ink-§!_5£§iH Q
Frequency (Hz)

Fig. 14 Comparison of estimated transfer function Fig. 16 Comparison of the estimated transfer
RS (24»hour average) and S17‘ function flshjbashgi) F, -,
"F')41.;§Yu,r_,__V_£W gv ' ?';);;_\mJg':{rf__ulh0y '"""' K '; '-E hi I oi rt‘/1 3.; i.

. lshibashi substation of the Tohoku Shinkansen. RS was

‘$5.-=!.t ! enemas....as__.se.0:.eatsa. calculated from observation results of microtremor at
-V: >$I?1i..esti.t0.ate.d....f.r.Qm...Ih.e...se.i.sr.0.i.c..0.b.s.e.r:-'..a1i.0t1.resuJ1.. The this substation while ST was estimated from simulta-
figure indicates that-R5 is similar "toSTT not onlyin neous microtremor observation results at the surface and
terms of the"peal<
5 .
position and zmagnitude, but also in at a d9[_?__1_Z_)_1___Of)_)2_‘_f1Ti_:) R5 and ST are quite similar to each
the whole shape. "other. Thiewaimplification characteristic of the wave
(Fig. 14 shows the 24-hour average of _R_g at Tabata as incident to surface layers as estimated from boring
compared with the transfer function SIT determined investigation and P5 velocity logging is shown in#:Fig. 4.
from another measurement data. This estimated transfer shows comparison of RS and ST at four
function was obtained from simultaneous observation of points which (re istered ,very strong earthquake, by
microtremor on the surface and at a point’ 25 m_;t3_e_[oyv_ using seismic of Mexico City.
the surface. In sp_ite*Tof.variation, thesemafeisifsimilar ST of Mexico City was estimated by assuming the
roughly. This fact proves validity of ‘the proposed tremor on theggrrxoclc within the campus of UNAM
estimation method, ‘ (Universidad Nacional Autonoma de Mexico) to the
; Fig. 15‘sh0ws the average of three RS of train- . a__s%.Ia su.b.stra.te_,tre.mor. Lo_a.ll_cases.£,g'
south of the city
induced tremor and two RS, of microtremor during £'?...'?.9....,.§2’.......=@.'.!.'.!€-*.....§.i.e.i...i.§.. ...QIl1e.r..,..ind.i.<;a:.i.00....tl3a.t,)
. ,, I ’
it stir‘-iii‘-intermission (measurement at Tabata) as compared with
flI$ bi
Tiitheiistransferfunction of the surface layers can be
iii} if
the estimated transfer function STT. STT is similar to estimated with Rgalso by using the seismic waveform.
the means R5, which indicates that the approximate ;?iFig(;' 18”shows oomparison between the maximum
transfer function can be estimated using the train- acceleration ratio of horizontal and vertical components
induced tremor. I 0 of the tremor on the surface and the amplification
Fig. 16 shows comparison of R5 and ST at the magnification of maximum acceleration estimated from

on of RTRI , Vol. 30, No. 1, '89. Feb. 31

L-
 O

C (")itv0)(r'€-titer’
8 i F V "T t ' Irror fir“ rt
---- R5 ----— S, EW component
- .1. -- ~+-i—ei-i-i-i-t---
CDAO point
0 ,,,
-L

0' i
o ‘ ;
4 .-l-
I1 H
-_-
A
'T
conat o .° 1
l
III 1
0 ..
I-in in
cu-
c ‘
\ .

n
magn'f I \

l l
.... 2 0- ' o -i
Qi l . O ie

--— l O O
l 0 -I

T _
CDAF point Ampf
cat
o » 0 oo siilvsxicsciifi‘
5 _ ‘ 0/ Along the Tokaido ‘e
/ L ShinkaT_nsen:_:T ~
i i
T
1101") ION 511- '

g r.
I
1 0 0 4 4 1 4 -- 1 I i
1
,__

OH
i
___
l
‘
‘ I
Q
I
I .
T if 41 0; ifs
gn'f'ca l ‘ I
Ho rizontal/vertical maximum value ratio AHS/A VS
f
mp
ca
seismic tremor on the surface
Fig. 18 Relationship between the horizontal/vertical
0.5 .
maximum value AHS/A VS and amplification
10 “Fm M magnification AT of tremor on the surface
SCT pOlflt
Inmj
_ ___ llllll stood at the seismic observation point if the seismic
"""'-2-"IlI (III tremor observation waveform of 3 components, 2
;; .:f'ca E _._
nil E I
horizontal components and vertical component, is
available.
‘I lt has generally been said that the microtremor iit;%E%E‘l$i 3

observation should be made at;;_3__1fl,i_<.:_l_t1i,ig.l_1t. when the social

caLn
FHTT F ' activities are stopped nearly completely, in order to C , er;
/(pt = ’i'i 1=i1'£'
SXVI point m "f;.".f v‘) "
avoid effect of artificial noise. But the proposed method
ll:::: \Ԥ_<.-
enables observation of microtremor at all times without @355‘I-4.\‘>- . ____\L_ :1‘-"I
-31-.‘

F Ii
-.
-—i-

-a-.1
-(restriction of the time span, thereby allowing measure-
Q1 ‘_-
_--—-§!
"_:'=I-n*iE".P-'~_-.—g._|-—i_1$_——--quasi-‘ --- (‘ ments within a short period at multiple points. This
nuu _-
§,1ipplie’men_‘(the J boring investigation results
Ill
Bi
I Ian.I I-..Il‘IQ-A=mi?T-n’-n—un:lIl|_._
1

P gn'f
"ii-fo‘r»‘j’prop§§z*i|nd /minute. ifiimation of the characteristics

/2
Amf
Ampf
A
fAmp
'¢a~;cat'Qc'oatn'on B OH H0'
"''5
llllll
ll
”‘ow-—
IlIlIEl-.1"-Il
I
"li
-|||‘“fl||||||l'"Il"i|~ | | l l
Eu.
III.
I | !
I -!—_ill!
§.
‘Q:
1_
u-§
_ Ii:-gI.5:Ii|-FJuIIIIl
= i'I5;_Iill!
| il-_"l|||
“iI=$l.Ba-1-IEl' !I‘
‘
I
IIIII I III_'
-I :iII
-
lHI f_-
III
ll
‘
K.a l< :|I $
-...-
I|'i5lE'Si"*E"?-F3»-?"5iigigg
i ! ~
“-”"i’|“-"i'0|I”l-"§“|‘§'”"~“§=§i;-E'5§~i!5'i| '|i' g_§5s:_asz:;=%|=:§:i i |iill
|iIfI| .|- | II||%
I'e=z-s=a-5*"
'_--_—!-
nuc- ulna?-a_|$1|IIII—=i
s~.@-_".-* - Ei"Si:nffEim§l§¥n:-Z:“~E:T.—"-" 5Il a=s-e_r."-,—.;:%.=_:5-. §_I .:z=.:
—n: : l
-erseie
ivalaeiozn
ayersiand itwis expected to demonstrate
collection of fundamental data for estimation
s==
U!
IitsIli gliIxarlmn-IIIIIJ!
'=q@‘lr=IIIIIin‘-=_n—2gIsIll.r"“:-;Qa1""..Ij|—‘$u-I".e=n--i|§.u5;:lI'\i"-|1oI.-=:5,=-3--_
.In9-.n_l"i-.3u—"nI1l-=.;‘nuim——_In'.ai-u—-a:_u5nt?-—|u1-_I n?2u-:_'~.4:-.|—l -Iliu‘m. ' !§ £1 .... ._.,
/. @ E H 5 I-\3 CD ‘U1 III C3 Q Q 5*" Ii 9'Q1.: 3::--- s==- 20.0 etaf i ledieairth quake damages.
' Q Frequency (Hz)
0 On the basis of above results, the dynamic seismic
Fig. 17 Comparison of the estimated transfer function. characteristics of the soil along an about 1500 km sec-
-. ,s"-1___:.- 2.-\' _.‘
using seismic waveform (Mexico City): tion of the Japan Railways (JR; formar Japanese
R5 and ST National Railways) line in the suburb of the metro-
“'‘\

J/$0 es-L-.3.-=;-?=r‘.*= ‘-’°‘ F-swam .. . politan district have been measured by using micro-
simulfgneous ‘observations in the ._g_rgund _ and on . thew»)
.7 tremor for four years from 1987. This isa part of the
surfaoe. Evidently, the maximumQelocjtjbamplification research project sponsored by the Ministry of Transport
magnification of surface layers can be estimated roughly, and involves measurements of microtremor at an interval
by using the maximum racéeilefatiojiratio of horizontalpwq of about 100 m along the line. In this project, measure-
and vertical Ktféflmvoifflong the surface although there exists ments were made at top and bottom of a structure in
a considerable variation. gm ,~~, order to understand the dynamic characteristics of the
X l Z) fie. {Q structure. The results will be reported as soon as
e.-cenelasiaas". F summarized.
Characteristics of s§__i§_nj_i__Q____’t[§_‘__!]1Q_[ and _rr]_ig;_r__ojtr_ern_gr
were reviewed and a new estimation method for Acknowledgment
dynamic characteristics of surface layers-was proposed. The seismic observation waveforms in Mexico City
This method is based on the assu motion that the ratio of were obtained by UNAM. The seismic observation data
horizontal and vertical spectra of surface tremor as an of the lzu Peninsula and Koto area of Tokyo Metro-
approximate transfer function. The validity of this politan were obtained by the Earthquake Research
method was proved by using microtremor observation lnstitute, University of Tokyo. The author would like
results at several points. lt was also shown that dynamic to express deep gratitude. He would also like to thank
characteristics of surface layers can be roughly under- Mr. Y- Watanabe of Fukuyama Consultants C0,, Ltd. for
this cooperation in measurement and analysis.
32 QR of RTRI , Vol. 30, No. 1, '89. Feb.
 Reference
(1) Nakamura, Y. and Saito, A.: Estimation of the
acceleration amplification characteristics and
maximum acceleration of surface lay.ers on the basis
of strong earthquake motion records (in Japanese),
Proceedings of 17th Meeting for Earthquake
Engineering Research, July 1983.
(2) Nakamura, Y. and Ueno, M.: Attempt to estimate
the dynamic characteristics of surface layers using
vertical and horizontal components of the tremor
on the surface (in Japanese), Proceedings of 7th
Japan
December 1986.
' Nakamura, Y. and Watanabe, Y.: Characteristics of
surface layers and seismic fault as observed in
seismic observation waveform (in Japanese),
Proceedings of 19th Meeting for Earthquake
Engineering Research, July 1987.
Nakamura, Y.: A Method for Dynamic Character-
istics Estimation of Surface Layers using Micro-
tremor on the Surface (in Japanese),.Railway
Technical Research Institute Report, Vo. 2, No. 4,
April 1980 _
Earthquake Engineering Symposium,

o u u Q o o Q Q u o u Q a u u Q o - I o I o I o c q Q Q o n Q I Q o n u u Q u n Q Q Q u u Q u o n o o 0 I o o v c a Q u Q a n Q o o I u u u u u n Q Q u o no o o o I n I o o o Q . . . , . . . ' - ¢ q ¢ ¢ Q - Q Q Q Q Q I u Q Q u o o c o u I A a a I Io
unnu- - ¢ Q a n Q p o u a n a Q u u o o n Q u u Q Q I n o Q ¢ Q Q Q - ' q 4 . ' ' - Q - Q n Q Q Q a | q - Q q Q - - - q ¢ Q q ' q Q q ¢ q . ; . q. ' ‘ ' ' ' ' ' ' ‘ *
. I . Q | q Q Q Q o Q Q n I I I 4 o I o I I I I I O O G 1 I I I I I 0 Q o o n Q n I o Q I I I 0 I 0 0 I 0 I I 0 0 I n 0 Q Q o Q Q c Q I Q I o o Q Q no o u n - n - - ¢ - - Q - e ¢ - ¢ - a ¢ | 4 Q n - c e - a ¢ ¢ u u v n Q a o u o a o u Q u a Q u | ¢ o u a Q Q »| Q - a - a n ¢ e . u | o u a ¢ o I I n 0 o o I 0|
on
p o I o o Q 0 Q o Q c Q - I a o Q Q a o u o u 1 o Q o n 0 0 I r u Q Q o Q a - q q Q u u u Q u 1 q Q u v 0 q o Q Q Q ¢ Q Q Q u n Q Q Q u Q Q u q Q Q In o o o o n o u q u Q Q - Q a c Q u u 0 o n o Q e Q Q o n n u q Q ¢ u o q Q u o o u o Q o on
o o o a o Q I a Q o a u | u o Q I n Q o w o o n I I I I 0 I 0 I J 0 o u o u u u Q u I u o Q o Q Q Q 0 I Q Q n Q u u Q Q u u a a Q n I o Q u Q o 0 Q no ' . ' ' ' ' ' ' a o o 0 I o o \ I o o n o n o o o u u a I 0 U o Q I I o o Q u O o J Q IQ
o o Q o o I Q Q Q Q n a u u u o o J I 0 I 0 I ¢ Q I I Q 9 I i I I I I ¢ I Q 0 I o Q o n Q Q - u q u I u Q Q Q Q Q Q Q Q q Q Q | q Q Q Q Q Q q Q Q ; an ' ' ‘ ' ‘ ' ' ' I Q 4 I Q o I o 0 Q Q o I Q Q o I o o o 0 o Q O o Q o G Q o o o I I l an

Portable Intelligent Collector

PIC 87
_,‘
-..- in-
=
E
‘i'- >t,'_-1'; . ;
--2'31 I-:"."-I'.- '
’ ' it "l.j -4'
-r -

.-_:-;_ , I This is a portable vibration measuring instrument for

-..-1'-.i-719':
'=‘ _. . --.
. . microtremor, train vibration and etc..
fit,5 _
:I:t?., 1

_ ._ . . ,,
Dimensions and Functions
'-A '.
1' >_- .- _;
~
Dimension 560mm(W) x 330mm(D) x
\ _ - 5
l95mm(H)
4-,-
U

it _. Weight 14kg (main body)

' 5'
Amp. Channels 6 ch.
Gain 0 — 120 dB (5 dB step)
Freq. Char. 0.18 -- 20 Hz
Main body of PIC87 L.P.F. 5 Hz/10 Hz/20 Hz
I
I
A/D Converter 12 bit, 24 us
Multiplexer
16 ch.
16 bit CPU pPD702l6 (8 MHz)
Display 640 x 400 dot Liquid Crystal
Floppy Disk Drive3.5 in. single drive (built-in)
Interface RS232C
Centronics
Power Supply 12V (2.6 AH) battery and 12V
(7 AH) battery (both built-in)
Power Consumption 10W maximum (FDD access)
Senser 3 component x 2 set
f

Extention cable 50m x 2 set

Sensers and Extention cables for PIC87
Since 1987, RTRI has been measuring the micro-
tremor along JR line in Tokyo metropolitan area
(approximately 1500 km length) at every 100 m using
PIC87.

View of the microtremor measurement

along JR line using PIC87

OR of RTRI, \/0|. 30, No- 1, '89. Feb. 33