A Method for Dynamic Characteristics Estimation of Subsurface using Microtremor on the Ground Surface

Yutaka NAKAMURA", Dr. Eng. As methods for dynamic characteristics estimation of su rface layers, investigation with boring and a method which employs microtremor are well known so far. Boring investigation, one of the most accurate methods is costly and time consuming and is not available all the time. The method that employs microtremor is handy but has not produced satisfactory resu It to th is 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 resu It, the spectrum ratio of the horizonta I components and vertical component of microtremor bears resemblance to transfer function for horizontal motion of surface layers.
1. Introduction

A survey on the earthquake damage shows that, considering seismic intensity equal, the effects on structures vary considerably: some are free from damage almost completely while others suffer heavy damage. Th is phenomenon occurs due to difference in the seisrn ic response characteristics of the structures. The surface layers carrying these structures also differ substantially in the effect of earthquake due to difference in their seismic response characteristics. The effect range is specific to the structure as far as the structu ral characteristics are concerned. In the case of surface layers, however, the effect covers a certain spread of the area. Co nseq uentlv , understanding of characteristics of surface layers may be said to be an important review item in analysis of the earthquake disasters. In this context, the characteristics of surface layers (particularly, the seismic resp onse 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 measurable). 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 socia I activities stop almost " Geotechnical Engineeri ng & Disaster Prevention Laboratory QR of RTRI, Vol. 30, No.1, '89. Feb.

completely in order to eliminate the effect of a tremor whose source can be identified. But this procedure will substantially detract the benefit of the micro tremor (i.e., readiness for measurement). This paper proposes a new method to estimate the dynamic characteristics of surface layers by measuring solely the microtremor of the su rface. According to this method, stable estimation of the predominant frequency and amplification factor can be made even in the presence of a certain degree of artificial tremor and there is no need any more for time restriction on m icrotremor measurement.
2. Effect of Surface Layers on Seismic Motion

Earthquake involves release of the strain energy accu mu lated in the focal region and propagation of a part of the energy thus released to the surrounding. The dynamic characteristics 0 (f). f means frequency, observed 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. Fig. 1 shows how the difference in earthquake and observation point is reflected in the tremor waveform. The acceleration waveform for 5 seconds around the maximum va lue is indicated vertically and horizontally, with similar earthquakes arranged vertically and same observation points horizontallv. The magnitude of earthquakes rises toward the left of the figu reo EvidentIy, the seismic acceleration waveforms at observation points are generally quite similar without much variation between different earth quakes, though there exists a trend that a high-frequency tremor prevails when the magnitude of earthquake is small and the low-frequency tremor prevails when the magnitude is large. This means that the observed seismic waveform is similar for the same observation point (i.e., the same surface layers characteristics S(f)) even when the radiation characteristic F(f) or propagation characteristic T(f) is different. In other words, it may be said that the effect of surface

25

.:0 80009 69003 krn H=ii.. or more).sh=ar.layers is most critical among the three factors influencing dynamic characteristics.?OKM 72002 M=6.. The measuring time span is considered to be cover a period with fewer artificia I tremors.lh=aOkm 80018 l'I:g. '13 .A. Upon wave long period (2 . .. 2Gl.. -0.)km 1'1=6..}C" ern Hiratsuka Atami Numazu Iwabuchi Shimizu Yaizu max » _ [j~15 ma". -0'. ·'\y. The N value increases gradually while repeating an increase and decrease down to this depth. .SI-] . 1 Difference in seismic waveform due to difference in earthquake and observation point 3. Sea waves induce a tremor of a reiatively forces do a tremor tremor.1 Observation at Ka monomiya The N value of su rface layers at th is point is 15 or more and exceeds 50 at a depth of more than 35 rn.. . .m M.). IS > V . Though this point is surrounded by plants. Both theories have tried to prove the respective validities on the basis of measurements of rnicrotrernor." t\ p. The observation took place at the Kamonomiya substation of the Tokaido Shinkansen and the Tabata substation of the Tohoku Shinkansen. As a whole. has often been chosen for engineering purposes such as esti mati on of the prevailinq frequency of surface layers with rnlcrotremor. Lonlrterm Continuous Observation of Microtremor 3. 'b. .) after frequency analysis with FFT. 'tI OJ <. r u km 11"'6. 6 .= • Gals max :l:l Gals m(t~..-j':l Gals Shin-Kikugawa • .7. the so-called microcomprising the rnlcrotrernor . while the storm and artificial of short period. This fact in turn leads to the importance of understanding the vibration character- istic of the su rface layers. 1986 I'.61'1. .' I ..1 • . "J. .-.. an observation of microtremor was made continuously for more than 30 hours to identify the characteristics.~ OJ ~ r S 2 . The micro rrernor. though not known fully. ".. 11=).''0'-"" 0 Qj 9.d i .-- p tl'A '0-' o 6 (hour) 12 Fig.1. 20 sec..I'. 26 . Kamonomiya NS component: til February I II' 9. Sh=2Gkm 7BOll9 7EllOB H~:..3 sec. their principal equipment was considered to be shut down while the measurement was made. '89. train and etc... sea waves) and artificial forces (plant. \I'~rv~~r" ~~ 'Ii Shin-Takatsuka Fig.:.... motion there are two theories: the solid wave theory and the surface wave theory. The Fourier spectrum discussed below was obtained bv applying the hann ing window three times to waveform data (2048 points. the so-called microseisms.61'1=:) 74iJ02 111. equivelent was used for the observation. The surface layers are normally exposed to tremor by natural forces (storm.. A velocity meter of specific period of 1 sec. No. Vol.. . 5h~30km )2iJ(Jfl H=5 . 30. but are not yet recognized as established... If '\ Time 4. Measurement of micro tremor was made from the n iqht of February 8 (Satu rday) to the morning of 10 (Mondav}.nh. For the purpose of utilizing the microtremor in the engineering field. 1986..:km "1_4. I . Feb.. 71002 Ih"'~Okrl'1 M=S.0.. the surface layers are silty. etc." I' . 2 Time change of the horizontal amplitude of the microtremor QR of RTRI.01l=0 78004 km "". . 31 ' .!I.I ! ::3 '" ~ E '" <= :. Ul a.t! • . automobile.9h: 74008 l.:S. "Q_ . .

.0 5.0 10. • Horizontal component '0 .00 p. Frequency (][z) Fig. '" o 5 « ~ 1 0. e V . '89.2 Observation at Tabata 0.. and freight line ru n nearly..m.e 0 '+0 . The ve locitv amp litude was read avoiding the train-induced tremor as much as possible.2 0. 30. c OJ V 1\ 1/ . 5 shows the time change of velocity amplitude. did not become remarkable during the time span when the amplitude becomes smallest. the soil has an N value of 50 or more. Fig. Besides. No. "0 '" - E 10 5 ~ .1 0. and the amplitude might be much larger during weekdays when tremors of plants are added. 3. 2 shows the time change of the horizontal velocity amplitude.. but it is not include in this figure.5 = Tabata ____. E .1 0. which is estimated from results of boring and PS prospecting QR of RTR1.~ ::: Q. of Saturday to 6:00 a. Needless to say the tremor due to the Shinkansen train is far larger than the one shown here.. 1987.0 10. Fig. Feb. 3 shows spectra of hourly microtremors which are overlapped for 24 hours.:./ I Ij I' 5.Fig.0 The surface layers at th is point are 18 m in th ickness and sandy as a whole. = Ishibashi ~ . .5 LO V \ > -6 o 6 18 24 30 Time (hour) 2. E S 10 .5 Hz) of surface layers. Tabata is located in a district of intense social activities and laden with microtremors. 5 t== Kamonomiya / -- "0 ::l .0 '" . 4 shows the amp litude characteristic of incident waves to surface layers estimated from boring investigation and PS velocity logging..2 Frequency (Hz) 20. Fig.2 mkine (10-5 rn/s). This corresponds to the quiet time at weekends. and large around noon.:: '" u 'c en E . 27 . June 25. The figure shows that the microtremor is smallest from 2:00 to 3:00 a. Vol. Measurement of microtremor was made from the night of June 24 (Thursday) to the morning of 26 (Friday). The extent of velocity amplitude change is around 0. with the shearing wave velocity at around 500 m/s. the measuring time span was one considered to be heavily laden with artificial tremors in the week.rn.5 La 2. of Sunday. Tohoku Line. but it was observed during a relatively quiet time span from 20.. 4 Amplitude amplification characteristics for incident seismic motion on surface layers at each point. Tabata: .05 -I---+---1--+-t-++-f-H-----1--l---l-+-++-I+if------1 0.5 . Keihin-tohoku Line. 3 Time change the horizontal spectrum of the m icrotremor 10 r-.rn. Yamanote Line. 5 Time change of microtremor amplitude Fig..0 0. which is considered to be the prevailing frequency (about 1.0 200 Fig. Tabata is a place subject to substantial effect of the train-induced tremor because the Tohoku Shinkansen.. with the ratio between maximum and minimum amplitudes bei ng about 4 times.!:: Q. Spectra indicate fluctuation acoording to the time and it is difficult even to read the prevailing frequency of the surface layers. 1987 o Vertical component . The N value to a depth of 4 m is around 3 and increases gradually as the depth increases and reaches around 50 at the depth of 18 m.0 5.1.. 10 5 "0 ::l . Below this level.!:: Q. The microtremor.!<: '" c . -.

0 Fig. As known from this comparison. 8 Comparison of tremor spectra between train induced tremor and quiet time span (intermission of train operation and 3. E . 7 Fluctuation of prevailing frequency of microtremor 28 . The ratio between maximum and minimum amplitudes is more than 6. •• •• • •• 0-- . The amplification characteristics of incident wave to the su rface layers estimated fro m boring investigation and PS velocity logging are shown in Fig.0 20. No. '89..•• -. 4. Surface Layers It is said th at the microtremor must be measured during the quiet time span.rn.'. 6..1 3.I • • __ .1 0.. 30. Vol. This 5.~.0 F requency (Hz) Fig. It is known that substantial fluctuation is observed within the frequency range of 3 Hz or more..1 0. '! 005 0.0 10.. Indeed.0 0. 7. . ~ll 1'1: !~ IIA ~ ~ .0 20. c e c •• ·. 4. ~~ --. c [\~ /.0 Frequency (Hz) 5. Estimation ~1ethod of Dynamic Characteristic of the .~. 0 LW~~LLUU~~~~UU~~LLLW~~LO~U • • • .[ 0. the velocity amplitude indicates tremor that 4 to 8 Hz frequencies are the specific It is also of microtremor was smallest from 2 :00 to 3:00 a.1. o8 ~ _ o [ -6 o 6 12 18 24 30 Time (hour) Fig.5 ~~. I .5 1. 1. "1J ::I f1tf_ \ ¥ A J i Interml!lsion of ~\train operatio n r ' . Such fluctuation of the prevailing frequency of microtremor is considered due to a substantial effect of the artific ial tremor in the vicin itv.m.: a. what prevails during the quiet time span appears to be the tremor related to the geological structure far deeper than the soil which is considered to be the bearing soil in terms of engineering.--. and largest around noon.0··. 6 shows spectra of hourly microtremors overlapped of surface layers at the Tabata point. The mean prevailing frequency was read every hour on the basis of the above spectra and plotted against time as shown in Fig. 2)._ • • 4 f3 2 ------._ i\ IW \ I ~1IIl(.• r •• _ •• •• - o ••• Q. E Tabata EW component: 24 hours :.0 Tabata EW component )~ Trai nlM I '" . but did not prevail for the estimated amplification characteristic of surface layer..00 a.m.!.0 - "8 0 •• 0 0 •• - ::c N 9 8_ __ 76 r5 f- ~_.1987 - 12 11 10 I o I I I ° • • i 0 o Vertical component • Horizontal component 0° 0 ...Il~. Fig. known that these frequencies did not become prevailing during the time span when the microtremor amplitude was small.6 Time change of Horizontal spectra of microtremor Tabata: June 25.00a. the observation result at Kamonomiya shows that the tremor like a specific vibration of the su rface layers prevailed during the quiet time span. This figure shows that 2 to 3 Hz frequencies and 4 to 8 Hz frequencies prevail in most of the cases.As in the case of Kamonomiya. a tremor of 4 Hz or more was induced by train to cause a considerable change in the microtremor.0 5. which is more than ten times the case of Kamonomiya (refer to Fig. . The horizontal velocity amplitude. .0 2. on the other hand.2 0. Fig. fluctuated between 5 and 30 mkine.. however. Frequencies of 2 to 3 Hz prevailed during the quiet time span.0 2.: 0 U::I '" ~ 0.0 [0.2 05 1.. 8 shows comparison of frequency analysis results between the train-induced tremor component and other components with the observation resu Its of microtrernor at Tabata.) 0. The effect of this traininduced tremor was remarkable in the vertical component. Feb. for 24 hours. At Tabata. --. A tremor with a frequency whose amplification magnification transfer function increases as the so-called estimated of the surface layers was not observed QR of RTRI.

the Rayleigh wave is assumed as noise of microtremor and a method to eliminate the effect of Ray leigh wave is studied. the AHI A V va lue of earthquake is related to the soi I conditions of the observation point and AHIA V is close to "1" for the firm soil. but the prevailing frequency of microtremor at Tabata and Kamonomiya fluctuated substantially depending on the measuring time span... <..( m '" ..~ n. 0 J: ~ III ::0- '(jj !J) '" "" .i "" o ~ III . The horizontal tremor may be considered.• I r • • • • The transfer function defined as follows: ST of surface layers is generally 8 SHS and SHB are respectively the horizontal tremor spectrum on the surface and the horizontal tremor spectrum incident from the substrate to surface layers. ~ . 30. . but not included in the vertical tremor soectrum Sun in the base oround Assuminn that the .J 0 " ... to be amplified through multi-reflection of the S wave while the vertical tremor is through multi-reflection of the P wave. 0 III 0 amount ES defined below should represent the effect of Rayleigh wave on the vertical tremor: Es=EvsIEvB If there is no Rayleigh wave. frequency found from In th is way.. o ... C N ... t x ....... The effect of Rayleigh wave should be included in the vertical tremor spectrum SVS on the surface.. 4. No. .. It is also generally said that the prevailing frequency of microtremor fluctuates less.. 29 . Namely.i. From this viewpoint. .. This may be the case of Tabata where the effect of traininduced tremor was remarkable in the vertical component. i0 O Q~ •• --- •• - '.. to a certain accuracy..2 New Transfer Surface Layers Function Estimation Method for tremors during the quiet time span.:M! C- >- ~ :J . The effect of Rayleigh wave..x U ... Since the artificial noise is mostly propagated as Rayleigh wave. the effect of Rayleigh wave is nearly zero when the ratio is approximately "1". 4... ----y r •• --... on the other hand. Such fluctuation may be due to the tremor source in the neighborhood of the measuring point.1.. E . the microtremor can be considered to consist of various wave motions.. With an increasing ratio. C to OJ .. The propagation velocity of the P wave is generally higher than 1000 mis. 9 shows the ratio (AHIA V) of maximum values' between horizonta I and vertical motions of earthquake for each observation point or area. the tremor from such source may be considered a noise from a standpoint of estimating the dynamic characteristics of the surface layers using the microtremor.. . Dividing the tremor into horizontal and vertical directions. it often occurs that the specific of the surface layers concerned cannot be the freq uency analysis on lv of the micro- Fig. The artificial tremor source has mostly the prevailing vertical motion and tends to induce the Rayleigh waves. appears remarkably in the vertical tremor. In other words.. Assuming for vertical that the effect and horizontal of Rayleigh wave is equal components. '" ~ >. 9 § Ratio of maximum values between horizontal and vertical components of earthquake STIES may be QR of RTR1. Va\...2 Jl Fig.LJ --_ ... 0 . 0 ~ ~ I!! . '" ~ c . Feb. --= . Evidently. Eswill take a value larger than "1" with increasing effect of Rayleigh wave. -c > 4 o • 01:.. 0 4:' . r: $ -5 '0 . But SHS is readily affected by the surface wave. ... OJ () 8 8 8 o 0 0 0 c ..... '89.s ~ " i5 -0 E 0. ~ -.. :2 0 >U .>t. . --r. ...1 Basic Concept There are various research attempts. 0 i 0 I "" g ... 0 g 0 .J _g 4:' e 0 . Such tremors are amp lified by the soft surface layers accumulated on the hard substrate. .: . OJ vertical tremor is not amplified by the surface layers.....( . and the tremor of around 10Hz or less may not be amplified so much through multi-reflection within surface layers of several ten meters at most in thickness. ES"" 1. 0 '" x .s c 0 . Elimination of the effect of Rayleigh wave is stu died by using th is ratio.. Accordingly. '" :2: 0 III '" " ~ ~ :2: . ... i .. 2~ ~ '3 C OJ 0 ~ '. In this context.... 6r ..~ .. it is considered that horizontal and vertical tremors are si milar to each other.. no amplitude in a specific direction prevails on the hard soil. the degree of its effect may be known by determining the ratio of vertical tremor between the surface and substrate. Namely.. the I!! OJ ~ . o 0 .. . . the effect of Rayleigh wave may become more critical.....2 X . . ..frequently..5:1 X _g '§ " . . .. e C ::I c '" ~ ~ t i.:M! <. .0 '2 2 <. SHS of microtremor may possibly be affected by Ray leigh wave.:M! J: --« . with the tremors even in all directions. but there is no established theory concerning what kinds of wave motions the microtremor is made up from..Q "" II> :i!: :5 C III '" ~ '0 ...... m t.

0 2. RS and RB were obtained by divid ing the horizontal tremor spectrum by vertical spectrum.- ._ Kamonimiya ----. Tabata) Note. Kamonomiya NS component: 24 hours E ti a.-"T'TTTrr~_r-n. But RS becomes more or less than "1" in the frequency range where the Rayleigh wave prevai Is. "' RB should be free from effect of Rayleigh wave and have the characteristics unique to the point (and also the characteristics unique to the earthquake when tremor to be handled is an earthquake tremor). . 0. Kamonomiya (24·hour average) -.----~-.0 10. Namely.1 1. E B :::l . -- ~ . 5.0 5. a. So far described is a discussion 0. As shown in Fig..0 200 It 0.. respectively for Kamonomiya and Tabata. thereby substituting the latter. ~ 'E !!! . on the firm substrate. .2 0.0 LJ!~~I I. The same applies to the maximum value.0 Frequency (liz) 10. t I 0. a more reliable SIT can be estimated by multiplying with llRB as a compensation term when data of the ground are provided.0 2. ~ I..9 ~ n. Namely. '89. that the estimation accuracy drops when there exists a noise tremor agreeing with the prevai ling frequency in the estimated transfer function. RB becomes nearly 1. The Rayleigh wave acts to nullify such substitution. .r.0 2.0 frequency range. Namely Therefore This means that the transfer function of surface layers may be estimated from the tremor on the surface only. however. Namely: where STT horizontal estimated maximum from the value by surface ratio of horizontal layers can be and vertical SIT =ST/ES =RS/RB RS =SHS/SVS RB =SHB/SVB maximum values on the surface. 10.0 Frequency (Hz) 10..0 Fig.0 -. corresponding respectively to surface and su bstrate earthquake tremors. 2 . \ . v .2 0. 5. It is therefore possible to estimate dynamic characterist ics of su rfaee layers usinq tremor method... RS varies less with the time.. When compared with plottings of Figs.considered to offer a more reliable transfer function after elimination of the effect of Ray leigh wave..1.5 in terms of the 1. \ j '. and thus it is not remarkable in the estimated transfer function.--_r-. No.Tabata ~ . 12 Time change of horizontal/vertical ratio on the surface spectrum 30 OR of RTRI. and the prevailing peak position and magnitude are stable. propagation is even in all directions. the Consequently. 30. It is also considered that the tremor in the substrate is equal in all directions because various wave motions repeat local refraction after principal tremor of the the earthquake. Vol. 11 Time change of horizontal/vertical ratio on the surface spectrum § N 20.0 1--_ I-- observed on the surface according to the above 5.5 0. :E Fig.0 .0 Tabata NS component: 5. -1..0 for a relatively wide frequency range. In other words. the amplification magnification of the Fig. .5 1.. 4 and 7. It may thus be presurned that RS includes the effect to eli rninate the effect of Rayleigh wave.2 05 -~ . Verification of the Proposed Method Figs.. Feb.: CJ . the vertical tremor on the surface retains the characteristics of horizonta I tremor of the substrate.0 Frequency (Hz) 5. 11 and 12 show RS (obtained from continuous observation resu Its of microtremor for every hour) overlapped for 25 hours. . 10 Spectrum ratio of horizontal and vertical components in the substrate (Microtremor: Kamonomiya.

: .1---1-1+1 1---1 ----l'I'-1--1 0. indicating that the transfer function of the surface layers can be estimated with RS also by using the seismic waveform. 17 shows comparison of RS and ST at four points wh ich registered a very strong earthquake. 13 shows the 24-hour with average the transfer of RS at estimated from the seismic observation result. No.. ~ « ::-: 0. vJfo [t' 1. 13 Comparison of estimated transfer function (Kamonomiya): RS (24-hour average) and SIT (Tabata) 50 SIT ----SrT ~- Rs . n» ~ ----- SrT 11. Vol.0 0. Fig. '" E _5 <C ~ . i . .1....' l~ 1 20.::~~t1m'''~li~:I~ l----_-+-.0 Frequency '.: '[.0 10.5 0..5 .. .~rl'. I. 14 Comparison of estimated transfer function (Tabata): RS (24-hour average) and SIT Fig. III !' i . In all cases. Feb.0 E 0.2 ~ ~ §..0 5. . 15 Comparison between the horizontall vertical spectrum ratio (average) RS of train-induced tremor and intermission tremor and the estimated transfer function Frequency Fig. ¥ '[..0 Frequency (Hz) --11--+-1+1-1'1'4-1"j1'I----'--1' I j 5..0 2..'_'.5 1 0. SIT is similar to the means RS. 30.0 Frequencv Fig. I¥~ :a 1.1 I'-. . . of Mexico City was estimated by assuming the tremor on the rock within the campus of UNAM (Universidad Nacional Autonoma de Mexico) to the south of the city as a substrate tremor.5 1. Fig.:.1 0. "" . . which indicates that the approximate transfer function can be estimated using the traininduced tremor. 18 shows comparison between the maximum acceleration ratio of horizontal and vertical components of the tremor on the surfaoe and the amplification magnification of maximum acceleration estimated from the surface.Z.0 Fig.:.1 '-. " '--. RS and STare quite similar to each other. lYt. ~ 'M 1. ~ .r--. 2. « E <.2 I .0 10. '' I IJvt~ ~ 5.:! a.0 20.I 10 Rs ----- .I • . The figure indicates that RS is similar toSITnot only in terms of the peak position and magnitude..~ . Fig.'~ .Q 5. '~ 0. RS was calculated from observation resu Its of m icrotremor at this substation while ST was estimated from simultaneous microtremor observation resu Its at the surface and at a depth of 12 m. '" '" E ~ 0 .0 ~ . -- -----== . 10.0 0. . . '" SrT " ~j . ~jfI 'I!! N lA i Ci E .0 - Rs .0 2. !j :~ I 11:.0 (Hz) 0. roughly. Th is fact proves val idity ST estimation method. RS and ST are quite similar to each other. 1 l' 1'1 1 1 "" .s. i' i'V :' I~ ~ I !~ '!lII""J nHr - 1-- ~ A 1' 2. '89..::. The amplification characteristic of the wave incident to surface layers as esti mated from boring investigation and Pg velocity logging is shown in Fig.' /. . . 16 shows comparison of RS and ST at the QR of RTRI.a . but also in the whole shape. 1 1 lI'q 10. Fig. In spite of variation..':! ~l: ~ I! : '! 1:1:111 Ir'l" i'" 10. 4.0 1 (Hz) . 31 .'~ r~H''t%'H'fI'-t--l'tJ~-!t.~ s c <. 16 Comparison of the estimated transfer function (Ishibashi) Fiq. :"':'L.0 50 .! .2 0.0 (Hz) .5 1.' \\ .0 20. This estimated transfer function was obtained from simultaneous observation of microtremor on the surface and at a point 25 m below these are similar of the proposed srr Kamonomiya as compared function Ishibashi substation of the Tohoku Shinkansen. :\\ \.1 f. \\ .2 0. e -10 5 I Rs I IIIIII ----.2 0. 15 shows the average of th ree RS of train- induced tremor and two RS of micro tremor during intermission (measurement at Tabata) as compared with the estimated transfer function SIT./J . ~ §-mi':'-m'I.I '" '" E B a .5+-1----11-+1--+-1--11'--1'4.111 f~ . by using seismic observation data of Mexico City. Fig.L~--:-\+-++++-J>tN-'+-<:'f.5 1. 14 shows the 24·hour average of RS at Tabata as compared with the transfer functionSrr determined from another measurement data. ~ Ili~~I .0 0.0 20.~ :!: a. 0.5 0.+'1 1 mti~'i. ~ " ..

IV 0 00 v IVI ~~ c. the maximu m velocity amplification magn ification of surface layers can be estimated roughly ratio of horizontal by usi ng the max imum acceleration National Railways) line in the suburb of the metropolitan district have been measured by using microtremor for fou r years from 1987.::~ / i / rv ~. Y.--- 10 5 'I I'~~. The results will be reported as soon as su mmarized.. in order to avoid effect of artificial noise.0 2. 6.. 17 Comparison of the estimated transfer function using seismic waveform IMexico City): RS and ST simultaneous observations in the ground and on the surface.2 10. The seismic observation data of the Izu Peninsula and Koto area of Tokyo Metropolitan were obtained by the Earthquake Research Institute.. In this project. E '" • • -= - -- J I"lft-I . ~. is 1 1)/ JII. !'i I 6 4 2 Horizontal/vertical maximum value ratioAHs/A seismic tremor on the surface 8 VS Fig. '1/ ~ I.1 0. for this cooperation in measurement and analysis.~ '~ .0 Frequency Fig. C~ _.5 ~. '89.0 20. University of Tokyo.. .. This method is based on the assurnption that the ratio of horizontal approximate method and vertical transfer spectra of surface tremor function. Conclusions Characteristics were reviewed of seismic and a new tremor and microtremor method for estimation dynamic characteristics of surface layers was proposed.. This is a part of the research project sponsored by the Ministry of Transport and involves measurements of microtremor at an interval of about 100 m along the line.2 • 11 /I I" 1 r-' 0.. formar Japanese . and vertical tremors on the surface although there exists a considerable variation. :CDAF pail II . the dynamic seismic characteristics tion of the of the soi I along an about 1500 krn secJapan Railways (JR.'.0 (Hz) It has generally been said that the microtremor observation should be made at midnight when the social activities are stopped nearly completely. The validity as an of this was proved by using microtremor observation results at several points..IIII OS e c:: .. On the basis of above results.~ .'. But the proposed method enables observation of microtremor at all times without restriction of the time span.. !i ~E E .1. -- 0.ti 0. Evidently. The author would like to express deep gratitude. Watanabe of Fukuyama Consultants Co. Ltd.. 'g. \. No.j:: .5 0 5 C::c:: '. • o '" ~2 ii -c E s . components and vertical component. 2 horizontal available. thereby allowing measurements within a short period at multiple points.II 5..-~--~ 10 5 Rs ----- Sr EW component CDAOpoint ~ h c:: 6 o 4 o o .5 1. QR of RTR I. He would also like to thank Mr. . measurements were made at top and bottom of a structure in order to understand the dynamic characteristics of the structure.0 . Acknowledgme nt The seismic observation waveforms in Mexico City were obtained by UNAM. It was also shown that dynamic characteristics of surface layers can be roughly under- 32 .". 18 Relationship between the horizontal/vertical maximum value AHS/A VS and amplification magnification AT of tremor on the surface .... This method can supplement the boring investigation results fo r proper and min ute estimati on of the characteristics of surface layers and it is expected to demonstrate val id ltv in co Ilection of fundamental data for estimation of detailed earthquake damages. :.:i'~ 00 BB =a. Feb. Vol. 30..1 0.CT pail ::: I I r'~! stood at the seismic observation point if the seismic tremor observation waveform of 3 components. E . ~E 1 05 o <> VI __/ / \\1' - :.

Proceedings of 7th Japan Earthquake Engineering Symposium. No. vibration measuring instrument for train vibration and etc . July 1983. 2411S 16 ch. L. and Watanabe. Y. and Ueno. and Saito.: Characteristics of surface layers and seismic fault as observed in seismic observation waveform (in Japanese). Val. Feb. Y. single drive (built-in) RS232C Centronics 12V (2. CharacterRailway Layers using MicroReport. Channels Gain Freq. Y. December 1986. (4) Nakamura..6 AH) battery and 12V (7 AH) battery (both built-in) lOW maximum (FDD access) 3 component x 2 set 50m x 2 set has been measuring the micro- Main body of PICS7 JR line in Tokyo metropolitan area 1500 km length) at every 100 musing View of the microtremor along JR line using PIC87 measurement QR of RTRI. 3D. Research Institute Portable Intelligent Collector PIC 87 This is a portable microtremor.P. 33 .: Attempt the dynamic characteristics of surface layers using vertical and horizontal components of the tremor on the surface (in Japanese). Proceedings of 19th Meeting for Earthquake Engineering Research. 0. tremor along (approximately PIC87.5 in. Y. Dimensions and Functions (1) Dimension 560mm(W) Weight (2) Amp. 2. I1PD70216 (8 MHz) 640 x 400 dot Liquid Crystal 3. Char.Reference (1) Nakamura. Y.120 dB (5 dB step) 0. cable RTRI x 330mm(D) x 195mm(H) 14kg (main body) 6 ch.20 Hz 5 Hz/l0 Hz/20 Hz 12 bit. Vo. No.18 . July 1987. '89.F. for Earthquake to estimate Proceedings of 17th Meeting Engineeri ng Research.4.: A Method for Dynamic istics tremor Technical April 1988 Estimation on the of Su rface (in Surface Japanese). (3) Nakamura. A: Estimation of the acceleration amplification characteristics and maximum acceleration of surface layers on the basis of strong earthquake motion records (in Japanese). (2) Nakamura.1. (3) A/D Converter Multiplexer (4) 16 bit CPU Display Floppy Disk Drive Interface Power Supply Power Consumption (5) Senser Sensers and Extention cables for PI C87 Extention Since 1987. M.

Sign up to vote on this title
UsefulNot useful