Journal of JSCE, Vol. 1, 307-321, 2013 Special Topic - 2011 Great East Japan Earthquake (Invited Paper)
STUDIES ON SOIL LIQUEFACTION AND SETTLEMENT IN THE URAYASU DISTRICT USING EFFECTIVE STRESS ANALYSES FOR THE 2011 OFF THE PACIFIC COAST OF TOHOKU EARTHQUAKE
Kiyoshi FUKUTAKE1 and Jiho JANG2
1Member of JSCE, Institute of Technology, SHIMIZU Corporation (Etchujima3-4-17, Koutou-ku, Tokyo, 135-8530, Japan) E-mail: [email protected] 2Material & Energy Research Team, GS E&C Research Institute (Cheoin-gu, Youngin-si, Gyeonggi-do, 449-831, Korea) E-mail:[email protected]
The 2011 off the Pacific coast of Tohoku Earthquake caused soil liquefaction over a wide area. In par- ticular, severe soil liquefaction was reported in the northern parts of the reclaimed lands around Tokyo Bay, even though the seismic intensity in this area was only about 5 on the JMA scale with low acceleration. The authors surveyed the residual settlement in the Urayasu district and then conducted effective stress analyses of sites affected and not affected by liquefaction. The analyses results were compared with the acceleration waves monitored with K-NET Urayasu or ground settlements surveyed. They were based on the acceler- ation observed on the seismic bedrocks in earthquake engineering in some other districts adjacent to Urayasu. Much of the settlement was due to the long duration of the earthquake, with further settlement resulting from the aftershocks. The study shows that the effects of aftershocks need to be monitored. The
simplified liquefaction prediction methods using the factor of safety, FL, also need improvement.
Key Words : The 2011 off the Pacific coast of Tohoku Earthquake, liquefaction, settlement, effective stress analysis, aftershock
1. INTRODUCTION Aerial laser surveying was carried out over Urayasu City before and after the earthquake, and the change The 2011 off the Pacific coast of Tohoku Earth- in the ground surface was indicated by the difference quake (hereafter, Earthquake 3.11) caused liquefac- in the elevation values.2) The errors, however, were tion over a wide area. Focusing on the reclaimed land large and the values did not necessarily correspond to in the northern part of Tokyo Bay, the seismic in- the actual amount of settlement. Konagai et al.21) tensity was about 5 (ground acceleration of about 100 conducted a LiDAR survey and drew subsidence to 150 Gal), thus the accelerations were not very maps of the Tokyo Bay shore areas, and the measured high, but the duration of the seismic motions was results contained the effects of the structures. long, and this caused liquefaction and settlement seen The authors measured the amount of settlement in in many areas. On the other hand, even though the Urayasu, Shin Kiba, and Tatsumi a few days after the ground conditions were similar, some places did not occurrence of the earthquake, and prepared drawings undergo large settlements. As a result, the Japanese of the distribution of the settlement. They then in- Geotechnical Society and the Ministry of Land, In- vestigated the causes of the variation in the distribu- frastructure, Transport and Tourism Kanto Devel- tion of the settlement, the amount of the settlement, opment Bureau jointly summarized the liquefaction and the difference in settlement. and non-liquefaction areas on a map as shown in In addition, focusing on the Urayasu area, typical Fig.1.1) However, indices directly related to damage, ground models were prepared in each area from the such as settlement, etc., were not shown on the maps. N-values and borehole data of ground that liquefied
307 and ground that did not liquefy. Response analysis by The results were compared with the K-NET effective stress analysis was carried out based on the Urayasu acceleration records and the amount of set- acceleration records measured at the engineering tlement measured at the locations of liquefaction. In bedrock near these areas, including the main shock addition, an improvement to the FL method of de- (M = 9.0) and the largest aftershock (M = 7.7), which termining liquefaction in cases where the duration is occurred 29 minutes later. long as in multi-segment earthquakes was proposed.
2. CHARACTERISTICS OF SETTLEMENT IN RECLAIMED GROUND IN THE NORTHERN PORTION OF TOKYO BAY
Toky (1) Characteristics of settlement distribution Figures 2 and 3 show the amount of horizontal Tokyo Bay ground surface settlement in the Tatsumi - Shin Kiba Kanagawa Pref. Chiba Pref. - Urayasu area located in the northern part of Tokyo Liquefaction Area Bay shown in Fig.1 within a few days after the Non-Liquefaction Area earthquake. These values were measured using bearing pile structure as a benchmark. The numerical values of the colors are: blue, 0 to 10 cm; orange, 10 to 30 cm; and red, 30 cm or more. Measurement of the amount of settlement was carried out using points Shimizu Institute of Technology that had not moved, such as pile-supported buildings Tatsum where no settlement occurred. Settlement of the whole area did not occur, but places where the set- tlement was large and places where there was almost Shin Kiba Urayasu no settlement were mixed. Figure 4 shows the geological section along the ―Liquefaction ― survey line and the improved areas by preloading and Tokyo Bay Non-Liquefaction 17) ■Liquefaction Area sand drain . ■Non-Liquefaction Area Fig.1 Developmental distribution of liquefaction in Kanto region and north Tokyo Bay in Earthquake 3.111).
JR Keiyo Line
Metropolitan Expressway
Shin Kiba station
Rinkai Line
400m
Fig.2 Settlement in the Tatsumi - Shin Kiba area (cm).
308 Figure 5 shows the distribution of settlement along However, according to local taxi drivers, the noise the survey line indicated in Fig.3. Plots and arrows levels are greater when they drive now compared mean measured values and their ranges. Hatching with before the earthquake; likewise, the vibrations zone covers all measured values. With the Tokyo seen to be greater, thus it is considered that a certain Metro Tozai Line as the origin, there is almost no amount of settlement had occurred. From about evidence of settlement up to 2000 m.
K-net URAYASU JR Keiyo Line Reclaimed land Phase I
Reclaimed land Phase II Shin Urayasu station
Metropolitan Expressway Old Coastline
Ito- Hinode Yokado
Takasu Maihama station Survey Line
Tokyo Disneyland A Minato
B
Fig.3 Settlement in the Urayasu area (cm).
Alluvial lowland Reclaimed land (Phase I ) Reclaimed land (Phase II)
Fill preloading (Hb=1.7-4.0m)
F
Sand drain
Fig.4 Geological section along the survey line in Fig.3 and the improved areas by preloading and sand drain17).
309 Tozai Line of Tkyo Metropolitan JR Keiyo Ito‐Yokado Co., Ltd. Metro Expressway Line Seawall 0
10
20
Old 30 Coastline Settlement (cm) 40
50 0 1000 2000 3000 4000 5000 Distance (m) Reclamation Phase I Reclamation Phase II (1971) (1978) Fig.5 Settlement along the survey line in the Urayasu area (cm).
τ 0.30.3
2400 m, the settlement becomes larger, and is largest σ’mo at around 4000 m, reaching a maximum of about 50 Silty Sand cm. From that point until the sea wall, the settlement 0.20.2 A becomes smaller, thus overall, the distribution of Loose Sand Liquefiable settlement is shaped like a spoon. The settlement B 0.1 becomes smaller towards the sea wall because the Stress Raio 0.1 σ’mo=100kPa Xl =0.1 ground has been compacted as a mitigation measure γDA=5% Not-liquefiable against liquefaction from about 4500 m to near the 0.00.0 sea wall (near point A in Fig.3). In addition, the 1 10 10 100 1000 1000 construction methods taken to promote consolidation Number of cycles to liquefaction (sand drains + preloading) are considered to have a Fig.6 Liquefaction strength curves and liquefaction certain effect in reducing liquefaction near the sea strength lower limit value Xl. wall.17) The effect of construction methods taken to promote consolidation on reducing liquefaction has also been reported at Port Island and Rokko Island thick, thus the amount of settlement is comparatively after the 1995 Kobe earthquake.3), 4) If this type of small. In addition, small differences in the shear mitigation measure had not been taken, then the stress that acted many times are considered to have amount of settlement in the area reclaimed between had a major effect on the resulting phenomenon. In 1975 and 1978 would have appeared to be a bit larger other words, once liquefaction has occurred, the soil compared with the area reclaimed between 1968 to thereafter has been subjected to many repetitions 1971. However, the correlation between the amount resulting in severe liquefaction. The next section of settlement and the year of reclamation is not very considers the cause of this phenomenon. clear. As can be seen from the settlement distribution in (2) Liquefaction characteristics during cyclic Figs.2, 3 and 5, there is large variation in the amount small amplitude shearing at the element level of settlement. The amount of settlement varies con- The characteristics of liquefaction in this case siderably with 100 m separation, and there is large were caused by comparatively small accelerations variation even within narrow areas. This variation in that had continued for a long period of time, and as a the amount of settlement is due to whether the ground result, small shear stresses were repeated a large has been improved or not as described earlier, and number of times. differences in the surface layer soils and thickness of the reclaimed ground. If the surface layer is a thick clay layer, the settlements tend to be small; for ex- ample, near point B in Fig.3, the surface clay layer is
310 62 cycles 68 cycles 15 10 Loose Sand 5 0 -5 -10
( kPa ) -15 -0.30 -0.25 -0.20 -0.15 -0.10 -0.05 0.00 0.05 0.10 0.15 0.20 0.25 0.30 132 cycles 15 145 cycles 10 Silty Sand Shear Stress 5 0 -5 -10 -15 -0.30 -0.25 -0.20 -0.15 -0.10 -0.05 0.00 0.05 0.10 0.15 0.20 0.25 0.30 Shear Strain
Fig.7 Stress-strain relationship near the liquefaction strength lower limit value Xl (stress ratio = 0.11).
Therefore, the reason why the small differences in approached with fewer cycles compared with silty the acting shear stress could result in an extreme sand, thus this is still more reason to take this ap- difference in the phenomenon was investigated from proach. Incidentally, the following relationship is the behavior of soil elements. shown in Fig.6. A calculation was carried out for cyclic shearing with small shear stress amplitudes that were just Xl ≒ R100 ≒ 0.8×R20 (1) sufficient to cause liquefaction. Figure 6 shows the calculation results for the liquefaction strength curve Figure 7 shows a stress-strain relationship close to (relationship between the shear stress ratio when the Xl (for a case where liquefaction just barely occurs). double amplitude shear strain DA reaches 5% and the In the element calculation, the shear stress ratio Xl number of cycles) of loose sand and silty sand that was set to 0.11, which is slightly larger than Xl. In the contains much fines. The constitutive equation used case of sand, it required 62 cycles for the strain to in the calculation was the Ramberg-Osgood model start to suddenly increase, but once liquefaction oc- expanded to three-dimensions with a multiple shear curred the strain accumulated with few cycles. This mechanism as the stress-strain relationship5), and the corresponds to severe liquefaction even though the bowl model used as the strain-dilatancy relation- acting shear stress is small, as in the Earthquake 3.11. ship.6), 7) The difference between sand and silty sand In the silty sand, the number of cycles until the in- is expressed by swelling index Cs in the calculation crease in strain is reached is greater, and the rate of (sand: Cs/ (1+e0)=0.006, silty sand: Cs/ (1+e0)= increase in the strain after liquefaction is slower 0.010). compared with sand. Therefore, if the acting shear Xl is the lower limit of shear stress ratio (lower stress ratio is slightly larger than Xl, there is a possi- bound value of the liquefaction strength)6) that will bility of liquefaction occurring accompanied with not cause liquefaction even with multiple repetitions. large strain, but if the shear strain ratio is smaller than The figure shows the case where Xl = 0.1. If the act- Xl liquefaction does not occur. In other words, the ing shear stress ratio is Xl or less, liquefaction does subsequent behavior greatly differs depending on not occur, but if it is greater than Xl, liquefaction will whether the acting shear stress ratio is greater than or eventually occur due to multiple repetitions. In less than Xl. This is considered one reason why the conventional design, the liquefaction strength R15 or locations where settlement occurred and the locations R20 under 15 or 20 cycles, respectively, is used (the where settlement did not occur were mixed. In other value near A in Fig.6). In the Earthquake 3.11, a words, even in a small area, the acting shear stress liquefaction strength (for example, R100, in which ratio fluctuated greatly depending on small variations liquefaction is reached after 100 cycles, etc.) should in the ground stiffness and the circumstances of the be used, such as near B. In the case of sand, Xl is nearby topography and structures. As a result, at
311 certain locations where the acting amplitude was Table 1 Stratigraphic composition and soil constants at the smaller than Xl, the settlement was small, and at other two assumed locations. locations where the acting amplitude was larger than (a) Location-1: non-liquefaction-prone ground based on the boring data of K-NET Urayasu Xl, the settlement was large. Therefore, in addition to the differences in the surface soil stratum, the mag- (1st natural period=0.82sec.) Thickness Fine nitude of the shear stress was considered one cause of GL γt V s N of Strata Soil Profile Content m 3 value the extremes in the phenomenon of liquefaction and m Fc (%) kN/m m/s the variation in the settlement. 1.0 1.0 Fill (F) 17.5 140 3 This discussion has been based on a theoretical 2.3 1.3 Silt (Asc) 60 17.5 140 4 model having a lower limit value of liquefaction 4.6 2.3 Sand (As) 10 17.5 140 19 7.9 3.3 Silt (Asc) 40 17.5 140 7 strength at Xl. Therefore, although it cannot be said 20.0 12.1 Clay (Ac) 16.7 125 3 that the model as it is represents the behavior of the 27.0 7.0 Silt (Asc) 17.5 230 3 soil in Urayasu, it can be said to be one explanation 33.0 6.0 Silt (Asc) 18.5 370 30 for variations in settlement. In the future it is neces- 35.0 2.0 Sand (Ds) 19.0 400 >50 sary to verify this experimentally using on-site test specimens. (b) Location-2: liquefaction-prone ground based on the bor- ing data of the southeast of the reclaimed area of Urayasu (1st natural period=2.18sec.) Thickness Fine GL γt V s N 3. EVALUATION AND ANALYSIS OF of Strata Soil Profile Content m 3 value LIQUEFACTION BY EFFECTIVE m Fc (%) kN/m m/s STRESS ANALYSIS 0.4 0.4 Fill (F) 16.0 130 3 Silty Sand 1.7 1.3 35 16.0 100 3 (F) (1) Analysis conditions Silty Sand 3.8 2.1 35 18.0 110 3 Data were collected from several boreholes in the (F) 10.9 7.1 Sand (As) 10 18.0 120 10 Urayasu area, and two types of FEM ground models 15.6 4.7 Silt (Asc) 60 18.0 140 2 were produced: location-1 where there was no liq- 36.7 21.1 Clay (Ac1) 16.0 140 0 uefaction, and location-2 where there was liquefac- 42.8 6.1 Clay (Ac2) 17.0 170 2 tion. Location-1 was the ground at K-NET Urayasu 44.8 2.0 Sand (Ds) 19.0 400 >50 near the Urayasu City Office. Location-2 was set based on the data from several boreholes in the southeast of the reclaimed area (the Takasu area and The constitutive equations used in the analysis the Hinode area). The settlement at both locations were the R-O model as described in Section 2.(2) in was 30 to 50 cm. Effective stress analysis was carried combination with a bowl model. Figure 8 shows the out with the ground models for locations-1 and 2, and results calculated from the constitutive equation for the acceleration response and the extent of liquefac- the liquefaction strength curve of the strata that have tion (amount of settlement, etc.) were investigated. the potential for liquefaction as determined from the Table 1 shows the stratigraphic composition at the N-values and the fine fraction content Fc. The two assumed locations. The blue portion indicates N-values and Fc values for the borehole data for the strata that could liquefy. For location-1 ground showed variations with depth, thus there was (non-liquefaction-prone ground), the K-NET also variation in the liquefaction strength. Therefore, Urayasu borehole data were used to the depth of for the sand strata at location-2, RL was calculated as GL-20m, and deeper strata were set by reference to it is with variation based on the Japanese design 19) nearby borehole in-company data. At location-2 specifications for highway bridges , and the mini- (liquefaction-prone ground), an average ground mum value (about 0.2) was taken to be R20. Based on model was determined based on data from several this, the lower limit value of liquefaction strength, Xl, boreholes: Chiba prefecture data20) and in-company was set to 0.80.2 = 0.16 in Equation (1), to provide data. The differences between the ground at the two the curved line of the liquefaction strength in the locations were as follows: at location-2 the N-values figure. The other strata were also set by the same of the sand strata were small, the liquefaction-prone method. The difference in liquefaction strength stratum in the silty sand stratum F was thick, and the curves in Fig.8(a) and (b) are based on Fc and strata were deep. N-values. Also, the dynamic deformation properties (the
so-called G/G0 - , h - relationship) were set to standard curves for sand, silt and clay based on Ref- erence 8) approximately.
312 continuing (after 29 minutes from the main shock), 0.4 the largest aftershock occurred (M = 7.7). Therefore in the analysis, it was assumed that during the period from the main shock to the largest aftershock, there 0.3 was no pore water pressure dissipation, and the main τ shock and the aftershock were input as continuous. σ ’mo The groundwater level was the same in the main 0.2 shock and the aftershock. In this analysis, the vibrations in two directions were applied simultaneously in a three-dimensional Silt (GL-1.0～-2.3m) Stress Raio 0.1 model. Therefore in addition to the shear strain for Sand (GL-2.3～-4.6m) Silt (GL-4.6～-7.9m) each component, the resultant shear strain and the cumulative shear strain G* given by the following 0.0 equations were used6), 7). 1 10 100 1000
Number of cycles to DA=5% 2 2 2( ) 2 ( ) 2 ( ) 2 (2) (a) Location-1: non-liquefaction-prone ground zx zy xy x y y z z x
0.4 Silty Sand (GL-0.4～1.7m) G* G * Silty Sand (GL-1.7～3.8m) Sand (GL-3.8～10.9m) 2 2 2 2 2 2 0.3 ( ) ( ) ( ) (3) Silt (GL-10.9～15.6m) zx zy xy x y y z z x τ
σ’mo
0.2 (2) Analysis results and discussion a) Non-liquefied ground (location-1) Figures 10 and 11 show a comparison of the ef- Stress Raio 0.1 fective stress analysis results for the acceleration wave form and the acceleration response spectrum for the main shock at K-NET Urayasu measured at 0.0 location-1. Liquefaction was not confirmed in the 1 10 100 1000 Number of cycles to DA=5% soils at this location. In the analysis, the excess pore (b) Location-2: liquefaction-prone ground water pressure in the sand strata increased by about Fig.8 Calculation results for liquefaction strength curves for 50%, but did not liquefy. Both wave forms are sim-
liquefaction-prone strata ('m0 = 98kPa, DA = 5%). ilar, and although in the acceleration response spec- trum the analysis values become small at a period of about 2 seconds, there is general agreement. From The input seismic motion was the acceleration this it can be seen that the engineering bedrock wave record for the main shock and aftershock measured at from Shimizu Institute of Technology can be applied engineering bedrock (GL-41m, Vs = 430 m/s) at the to the engineering bedrock wave in the Urayasu area. Shimizu Institute of Technology. The distance be- b) Liquefaction-prone ground (location-2) tween the measurement point and Urayasu City of- Figures 12 and 13 show the maximum excess pore fice is 9.4 km. Figure 9 shows the acceleration rec- water pressure distribution, the maximum resultant ord and the response spectrum measured at engi- shear strain distribution, the acceleration wave form, neering bedrock. The NS component and the EW and the excess pore water pressure ratio wave form at component of the engineering bedrock wave at the location-2. The liquefied strata were the sand stratum bottom surface of the analysis were input simulta- (As) and the silty sand (reclaimed stratum F). In neously as E+F waves. The primary natural period at stratum F, a of 9% or greater was produced, and in the point of measurement was 1.18 seconds, which is the As stratum it was 4 to 6%. close to the natural period at location-2. (Gal=cm/s2) While liquefaction caused by the main shock was
313 60 GL-41m Observed（EW） Max.=48.3Gal 30 EW 本震EW Comp.-EW 本震NS Comp.-NS 0 100 -30 -60 60 GL-41m Observed（NS） Max.=55.1Gal spectrum (Gal) ponse 30 NS
Acceleration (Gal) 0 -30
-60 Acceleration res 10 0 50 100 150 200 250 300 0.05 0.1 0.5 1 5 10 Time (s) Period (s) (a) Main shock: MW=9.0 (14:46) 60 100 GL-41m Observed（EW） Max.=22.9Gal 30 EW EW余震 Comp.-EW NS余震 Comp.-NS 0 -30 -60 60 （ ） 30 GL-41m Observed NS Max.=23.8GalNS
Acceleration (Gal) 0 -30
-60 Accelerationresponse spectrum (Gal) 10 3500400 50450 100500 150550 200 0.05 0.1 0.5 1 5 10 Time (s) Period (s)
(b) Largest aftershock: MW=7.7 (15:15, after 29 minutes from the main shock)
Fig.9 Measured acceleration wave form and response spectrum at bedrock (Vs = 430 m/s) at Shimizu Institute of Technology (Gal=cm/s2).
1000 1000 200 500 Observed観測値 500 Observed観測値 Observed K-net EW Max.=157Gal 解析値 解析値 100 Calculated Calculated 0 -100 100 100
-200 50 50 200 Observed K-net NS Max.=125Gal NS Comp. 100 EW Comp. Acceleration (Gal) Acceleration
0 (Gal) spectrum response Acceleration 10 10 0.05 0.1 0.5 1 5 10 0.05 0.1 0.5 1 5 10 -100 Period (s) Period (s) -200 0 50 100 150 200 Time (s) Fig.11 Comparison of acceleration response spectra for the main (a) Observed wave of K-NET Urayasu shock at K-NET Urayasu measured at location-1. 200 Calculated EW Max.=162Gal 100 0 -100 Looking at the excess pore water pressure ratio -200 wave form, it can be seen that the speed of increase in 200 Calculated NS Max.=134Gal the water pressure is slow, and liquefaction occurs 100
Acceleration (Gal) Acceleration after the main seismic motion has passed (in other 0 words, liquefaction did not occur for about 80 se- -100 conds after the seismic motions). Thereafter the vi- -200 0 50 100 150 200 bration continued, thus a large shear strain was pro- Time (s) duced. This is consistent with the element calculation (b) Calculated wave by effective stress analysis results shown in Fig.7. This process towards lique- Fig.10 Comparison of acceleration wave form of the main shock faction is different from past records and simulation at K-NET Urayasu measured at location-1. results.
314 shock."
From this, the largest aftershock was considered to 19.9cm GL-0.4m 1.0 9.1 have contributed to the liquefaction and settlement. Silty Sand (F) GL-1.7m Silty Sand (F) 0.8 7.3 GL-3.8m To verify the effect of the aftershock, the analysis
0.6 5.5 results for the main shock and the largest aftershock Sand (As) 0.4 3.7 were input continuously at location-2. This was be- GL-10.9m cause 29 minutes after the main shock the excess pore Silt (Asc) 0.2 1.8 water pressure ratio was still close to 100%, thus the GL-15.6m 0.0 0.0 liquefaction state continued. Pore water Γmax pressure ratio Generally, the magnitude of the shear strain that is （％） produced during an earthquake corresponds to the extent of liquefaction or the magnitude of the ground Clay (Ac1) deformation. To check the effect of earthquake du- ration and the aftershock on the ground deformation, such as settlement, etc., Fig.14 shows the time his-
GL-36.7m tory of the resultant shear strain . Focusing first on the main shock, the maximum value of , max, is Clay (Ac2) produced after the maximum value of the input ac- GL-42.75m Sand （Bedrock）(Ds) celeration. As shown in Fig.13, this is due to the EW NS excess pore water pressure ratio reaching 1.0 after Ground model Max. water pressure ratio Max. resultant shear maximum acceleration. strain Γ max The amplitude of including the aftershock was Fig.12 Distribution of maximum values at location-2 (liquefaction-prone ground). greater in the aftershock than in the main shock for elements 5, 12, and 16. This is because a small shear force due to the aftershock acts on a stratum where the shear stiffness has become very small in the main For example, in the 1995 Kobe Earthquake, which shock, and induces a large shear strain; thus in this was an active-fault-type earthquake, the "killer type of stratum, the liquefaction and settlement due pulse" of the main motion caused sudden liquefac- to the aftershock are promoted. This is consistent tion, producing the maximum shear strain, and with the results of the interviews as described above. thereafter the vibration time was short, and there was In calculating the ground settlement after the not much accumulation of strain after the main mo- earthquake, a method frequently used is to obtain the tion (for example, see the analysis results of Refer- volumetric strain after the earthquake from the ence 6)). maximum shear strain experienced during the For unidirectional components (NS or EW com- earthquake, which is then integrated with depth. This ponent only) liquefaction was limited to a part of the method has been proposed for sand foundations9), 10) sand stratum, and it was not possible to explain the and clay foundations,11) but in each of the proposed severe liquefaction that occurred. In the seismic mo- equations, after undrained cyclic shearing test as tion conditions here, the acceleration amplitude is not element tests, a cock is opened to drain the water, and very large, and the phenomenon is close to the lower the volumetric strain after cyclic shearing is obtained limit value for liquefaction, thus the effect of the based on test results. bi-directional input on the occurrence of liquefaction In the Earthquake 3.11 a large quantity of sand was is considered to be particularly large. ejected, and in some places was deposited 30 to 40 cm deep. If the above empirical equations are applied (3) Effect of aftershock on the liquefaction and to evaluate the settlement at locations like these, the settlement settlement will be underestimated. This is because in In interviews with the local public regarding the the element tests, only water is ejected, and the effect largest aftershock on this occasion (15:15 hours of ejection of soil particles is not included. March 11th, M = 7.7), the following evidence was obtained. "At Shin Kiba, mainly water was ejected during the main shock, but sand was also ejected in the aftershock." "In eastern Kanto (the Tone River catchment area), sand was ejected in the main shock, and the ground flowed during the aftershock." "In the backfilled ground of the old iron sand quarry, sand was ejected in both the main shock and the after-
315 120 60 GL-0.4m EW Max.=98.4Gal Silty Sand (F) 0 No.5 GL-1.7m -60 Silty Sand (F) -120 No.7 GL-3.8m 120 60 NS Max.=84.2Gal Ground Surface Surface Ground
Acceleration (Gal) Acceleration 0 No.13 Sand (As) -60 -120 0 50 100 150 200 No.18 1 GL-10.9m Element No.5 0.5 Silt (Asc) 0 No.22 1 GL-15.6m Element No.7 0.5
0 Pore water 1 Element No.13 pressure 0.5 ratio 0 1.0 1 Element No.18 Clay (Ac1) 0.5 0.8 0 Excess poreExcess waterpressure ratio 1 Element No.22 0.6 0.5
0 0 50 100 150 200 0.4 60 30 EW Max.=48.3Gal GL-36.7m 0 0.2 -30 -60 Clay (Ac2) 60 0.0 30 NS Max.=55.1Gal GL-42.75m 0 -30 （ ） Input(Gal) Acceleration -60 Sand Bedrock (Ds) 0 50 100 150 200 EW NS Time (s)
Fig.13 Acceleration wave forms (main shock) for bedrock and the ground surface at location (2), excess pore water pressure ratio wave forms for sand, silty sand, and silt.
Therefore, in calculating the settlement, the ob- Case-1: Calculation of the settlement using max jective here was to determine the extent of the main produced by the main shock and the aftershock: shock and the aftershock, rather than to quantitatively "Stotal". (Using over the whole time; 0 to 580 se- evaluate the settlement. The settlement was calcu- conds) lated by obtaining the volumetric strain from the Case-2: Calculation of the settlement using max maximum shear strain from the equation of Ishihara produced by the main shock only: "S ". (Using 10) main and Yoshimine for sandy soils, and the equation of over 0 to 350 seconds) Shamoto et al.11) for clay soils. Case-3: Calculation of the settlement using max Figures 15 and 16 show the distribution with produced by the aftershock only: "Safter". (Using depth of the amount of settlement S calculated from over 360 to 570 seconds) the main shock + aftershock continuous analysis Case-4: Calculation of the sum of the settlement in results, and a comparison of the actual measured the main shock (Case-2) and the settlement in the ground surface settlement and the analysis values. aftershock (Case-3) separately: Ssum = Smain + Safter The maximum shear strain during the seismic motion The settlement in the main shock from Case-2 is used to calculate S was obtained using the maximum 16.6 cm, the settlement in the aftershock from Case-3 value of the resultant shear strain max, and the set- is 11.1 cm, thus the settlement in the main shock + tlement S of the ground surface was obtained for the aftershock from Case-4 is 27.7 cm. following four cases.
316
6 5 要素No.5 4 Element No.5 3 2 1 0 10 8 Element要素 No.6 6 4 2 0 10 8 Element要素 No.7No.7 GL-0.4m 6 Silty Sand (F) 4 No.5 GL-1.7m Silty Sand (F) 2 No.6 GL-3.8m No.7 0 6 No.12 (%) 5 No.13 Element要素 No.12No.12 No.14 Sand (As) 4 No.15 G 3 No.16 2 GL-10.9m 1 0 9.1 Silt (Asc) 6 5 Element No.13 8.2 要素No.13 GL-15.6m 4 7.3 3 2 6.4 1
ResultantShear Strain 0 5.5 6 4.6 5 Element要素 No.14No.14 4 3.7 3 2.8 Clay (Ac1) 2 1 1.8 0 6 0.9 5 Element要素 No.15No.15 0.0 4 3 G (%) 2 1 0 GL-36.7m 6 5 Element要素 No.16No.16 4 Clay (Ac2) 3 2 GL-42.75m 1 0 Sand （Bedrock）(Ds) 0 100 200 300 400 500 600 Time (s) EW NS 60 40 EW 20 0 -20 -40
ation (Gal) ation -60 Main shock Aftershock 60 40 NS 20 0
Input Acceler Input -20 -40 -60 0 100 200 300 400 500 600 Time (s)
Fig.14 Resultant shear strain time histories for the main shock and the aftershock input continuously.
317 Tozai Line of Tkyo Metropolitan JR Keiyo Ito‐Yokado Co., Ltd. Metro Expressway Line Seawall 0
Main shock 10 60%
20 40% Aftershock
Old 27.7cm 30 Coastline Settlement (cm) 40
Measured values (plots) and calculated values (solid line) 50 0 1000 2000 3000 4000 5000 Distance (m) Reclamation Phase I Reclamation Phase II (1971) (1978) Fig.16 Comparison of actual measured values (red dots) of ground settlement along the Urayasu survey line and the analysis results (solid line) for Case-4. values that are smaller compared with the actual Distribution of Settlement S （ｃｍ） settlement. The settlement at the location of the 0 5 10 15 20 25 30 boreholes based on Table 1(b) was 30 to 50 cm, and 0 Silty Sand the settlement from Case-4 was somewhat smaller （F) 5 than these values. As stated previously, this could be Sand （As） because when evaluating the settlement, the quantity 10 of sand and soil ejected was not taken into consider- Silt （Asc） ation. 15 The amount of settlement obtained in Case-1 was
20 17.0 cm, which was considerably underestimated. The reason for this is that amplitudes close to the 25 Clay (Ac1) maximum value of were produced several times during the main shock and the aftershock, as can be Depth Depth （ｍ） 30 seen from the time history of . There was only an
35 interval of 29 minutes between the main shock and the aftershock, but they were separate earthquakes, Case-1:Stotal 40 Clay （Ac2） thus rather than calculating the settlement consider- Case-2:Smain (Main shock only) ing the main shock and the aftershock to be a single (Bedrock) 45 Case-3:Safter (Aftershock only) earthquake as in Case-1, the settlement should be Case-4:Smain＋Safter calculated separately as in Case-4. This would be the 50 same even if the interval between the main shock and Fig.15 Distribution of vertical displacement with depth the aftershock were, for example, 1 minute. Taking according to the analysis at location (2) (liquefac- this idea to its logical conclusion (assuming there is tion-prone ground). no time interval between the main shock and the aftershock), they can be considered to be a single
earthquake. Therefore, in earthquakes for which the If this final settlement is taken to be 100%, then duration time is long, as in multi-segment earth- 60% of the settlement was produced by the main quakes, the method of calculating the settlement shock, and 40% of the settlement was produced by should be as follows: divide the total duration of the the aftershock. The fact that a considerable amount of earthquake into several time periods, calculate the the settlement was produced during the aftershock is consistent with the interview results described earli- settlement in each of the time periods using max, and er. Comparing the final settlement with the actual add the calculated settlements. This concept is simi- measured settlement, the calculated results provide lar to the cumulative damage theory.
318 termination results showed weak liquefaction for FL 0.0 0.5 1.0 1.5 2. 2.00 silty sand (FL = 0.82 to 1.0) and borderline liquefac- 0 tion for sand (F = 0.96 to 1.06), thus even though the 液状化 L Water Pressure Raio Liquefy Silty Sand F シ acceleration was set on the high side, the extent of (Fig.11) 2 liquefaction was underestimated compared with the GL-0.4m 砂 (F) GL-1.7m SiltyR=R20 Sand F シ analysis results. To obtain consistency between the 砂(F) GL-3.8m R=0.85×R20 two, it is necessary to review the liquefaction R 4 =0.85×R ) 100 20 R20 strength R and the acting shear stress L in the FL
GL-10.9m method, and a proposal for this is shown below. 深度 (m) ト(Asc) 6 Depth (m) Sand As 砂 (1) Review of liquefaction strength R
8 In earthquakes with long duration, the value of the R20 R100 liquefaction strength R should be reduced to take into consideration the effect of the number of cycles. 20 100 10 Instead of the liquefaction strength R20 at 20 cycles, it Normal Earthquake Multi-segment Silt Asc シ is possible to consider, for example, the liquefaction Earthquake 12 strength R100 at 100 cycles (= 0.9 to 0.8R20), or to Fig.17 Results for application of the liquefaction strength use the lower limit value Xl of the liquefaction strength (see the following equation). method (FL method) based on the Japanese design specifications for highway bridges19) to location (2) (for acceleration 120 Gal at the ground sur- R ( 0.9～0.8) R 20 (5) face). X l
Therefore, in earthquakes with long duration, The value of varies with the density and the fine there is a possibility that calculating the settlement by fraction content Fc. The higher the density and the obtaining a single max will underestimate the set- higher the fine fraction content Fc, the smaller the tlement. It is not possible to say that the shear strains value of . Figure 18 shows the liquefaction strength 12) that are smaller than max all contribute to the set- curve for Sengenyama sand (Fc=2.4%) for various tlement, but if the concept of "accumulation" of strain relative densities Dr, and the coefficient in the is not introduced, it is possible to underestimate the above equation. The results for Toyoura sand ac- 13) settlement. The concepts of accumulation include, cording to Toki and others are: =0.89 (Dr=80%), for example, the cumulative shear strain G* as shown 0.94 (Dr=50%). Neglecting the influence of Fc, R100 in Equation (3). and Xl are given approximately as follows:
D r (6) 4. DISCUSSION OF THE APPLICABILITY R100 X l 1.0 0.3 R20 100 OF THE METHOD OF DETERMINING
LIQUEFACTION Based on these results, the liquefaction determi- nation results using 0.85R as R are shown in The results of applying the method of determining 20 Fig.17, corresponding to the analysis results for the liquefaction (F method) based on the Specifications L excess pore water pressure ratio. Therefore, when for Highway Bridges Part V Seismic Design by the there is a large number of cycles at small acceleration Japan Road Association for location-2 in Table 1 are (in the case of ground distant from a multi-segment shown as the blue line in Fig.17. Here the value of FL type earthquake), it is necessary to correct R as in the is given by the following equation, where R is the equation above. liquefaction resistance value (liquefaction strength), and L is the acting shear stress ratio. (2) Review of the acting shear stress L
When the magnitude M is large, it seems to be R (4) FL necessary to take into consideration the effect of the L earthquake duration (number of cycles) in another way. The maximum acceleration amplitude on the ground surface was set to 120 Gal, which is slightly larger than the result in Fig.13 (100 Gal). The de-
319 earthquakes or multi-segment earthquakes. The fail- ure mechanism at the epicenter, the transmission τ path, and the ground structure, etc., should be taken σ Sengenyama sand (Fc=2.4%) ’m into consideration. σ’mo=138kPa
5. SUMMARY
Stress Raio Extrapolation of measured Values12) In this paper, liquefaction of the ground mainly in the Urayasu area in the Earthquake 3.11 was studied. Effective stress analysis was also carried out for a Number of cycles to liquefaction location where there was liquefaction and a location where there was no liquefaction, inputting the main (a) Liquefaction strength curve shock and the aftershock continuously. From the 1.00 results, the following can be summarized. Measured Values (●)
1) The amount of settlement in the Urayasu, Shin
α 0.90 Kiba, and Tatsumi areas was measured a few days after the occurrence of the earthquake, and a map of 0.80 Equation (6) the distribution of the settlement was produced. The results showed that the amount of settlement varied Coefficient 0.70 greatly from spot to spot, even within the same area. The causes were due to: (1) differences in soil and 0.60 differences in the surface thickness of the reclaimed 0 20406080100stratum, and (2) whether the ground had been im- Relative density Dr (%) proved or not. In addition, according to trial calcula- (b) Coefficient and relative density D r tions using the constitutive equation (bowl model) Fig.18 Liquefaction strength curve12) and coefficient used here, another cause was the small differences in for Sengenyama sand at various relative densities the acting shear stress due to the large number of Dr. cycles causing extreme phenomena (after liquefying once, the subsequent large number of cycles resulted In the Recommendations for Design of Building in severe liquefaction). Foundations18), the effective number of cycles of the 2) The accelerations were not very high, but the envisaged seismic wave form is corrected with a number of cycles was large, thus the rise in excess coefficient rn to take into consideration the ground pore water pressure was slow, and slowly increased density. However, doubts remain whether the rela- after the main seismic motion had passed, resulting in tionship14) between the magnitude M, the number of liquefaction. Thereafter, when the oscillations con- cycles, and the correction coefficient rn can be used tinued, severe liquefaction occurred. up to M9.0. 3) When the total settlement due to the main shock In a method proposed by the National Center for and the aftershock was analyzed, based on the as- Earthquake Engineering Research (NCEER)15), the sumption that the settlement was the sum of the set- FL value is obtained with a M7.5 earthquake as cri- tlements due to the main shock and the aftershock terion, thus the FL value is corrected using a correc- separately, it was found that the settlement was 27.7 tion factor MSF. The relationship between the cor- cm, which was smaller than the values (30 to 50 cm) rection factor MSF for the liquefaction safety factor found in the site survey. This was because when and the magnitude M shown in that method goes up evaluating the settlement, the ejection of sand was to M8.5, but is not proposed up to M9.0. not taken into consideration. If instead of the amplitude of the acceleration, the 4) The aftershock also caused a considerable Arias intensity is used, which is defined as an energy amount of the settlement, 40% of the total settlement based on the two horizontal components of the being due to the aftershock. This was consistent with seismic motions and has the units of velocity, both the testimony obtained from site. the amplitude of the acceleration and the duration are 5) In earthquakes with extremely long durations evaluated over each frequency range, thus in the and large number of cycles, by reducing the lique- opinion of some this is more logical.16) In any case, in faction resistance value to about 0.85, it is possible to the current method of determining liquefaction, the obtain a value of FL that is consistent between the duration and the number of cycles are determined actual liquefaction phenomena and the analysis re- from M only, and this is not appropriate for giant sults.
320 In the Tokyo Bay area including Urayasu, detailed neering, pp. 509-512, 1986. soil investigations and level survey for long-term 9) Shamoto, Y. and Zhang, J.-M : Evaluation of Seismic Set- 17) tlement Potential of Saturated Sandy Ground Based on settlement of clay layer are in progress . In the fu- Concept of Relative Compression, Soils and Foundations, ture we intend to confirm the validity of the analysis Special Issue on Geotechnical Aspects of the January 17 parameters, etc., using this type of soil test data, and 1995 Hyogoken-Nambu Earthquake, Vol. 2, pp. 57-68, carry out estimates of damage in multi-segment 1998. 10) Ishihara, K. and Yoshimine, M.: Evaluation of settlements earthquakes that are postulated to occur in the future. in sand deposits following liquefaction during earthquakes, Soils and Foundations, Vol. 32, No. 1, pp. 173-188, 1992. ACKNOWLEDGEMENTS: In this investigation, 11) Shamoto, Y., Sato, M. and Zhang, J. M.: Simplified esti- we received the cooperation of Hideyuki MANO, mation of earthquake-induced settlements in saturated sand Hiroyuki HOTTA, Yoichi TAJI, Akira ISHIKAWA, deposits, Soils and Foundations, Vol. 36, No. 1, pp. 39-50, 1996. and Tadashi SAKAMOTO of the Shimizu Institute 12) Tatsuoka, F.: Agendas of soil test and evaluation of test of Technology in the site investigations/survey. In results, Workshop on Recent Soil Mechanics and Founda- addition, we were allowed to use the earthquake tion Engineering, Japanese Geotechnical Society, pp. observation records of K-NET of the National Insti- 55-103, 1986. (in Japanese) tute for Earth Science and Disaster Prevention. The 13) Toki, S., Tatsuoka, F., Miura, S., Yoshimi, Y., Yamada, S., Yasuda, S. and Makihara, Y.: Cyclic Undrained Triaxial authors express their gratitude for these valuable Strength of Sand by a Cooperative Test Program, Soils and forms of assistance. Foundations, Vol. 26, No. 3, pp. 117-128. 1986. 14) Tokimatsu, K.: Seismic design N-Value, Foundation En- REFERENCES gineering & Equipment, Vol. 25, No. 12, pp. 61-66, 1997. 1) Ministry of Land, Infrastructure, Transport and Tourism of (in Japanese) Kanto Region Development Bureau and Japanese Ge- 15) Youd, T. L., Idriss, L. M., Arango, R. D., Castro, G., otechnical Society: Investigation of liquefaction phenom- Christian, J. T., Dobry, R., Finn, W. D. L., Harder, L. F., ena of Kanto region in Great East Japan Earthquake, 2011. Hynes, M. E., Ishihara, K., Koester, J. P., Liao, S. S. C., (in Japanese), http://www.ktr.mlit.go.jp/bousai/bousai Marcuson, W. F., Martin, G. R., Mitchell, J. K., Moriwaki, 00000061.html Y., Power, M. S., Robertson, P. K., Seed, R. B. and Stokoe, 2) Urayasu City: Strategy on how to cope with Great East K. H.: Summary report of the 1996 NCEER Workshop on Japan Earthquake, Document 1-4, 2011.07.22 (in Japanese), Evaluation of Liquefaction Resistance, Salt Lake City, http://www.city.urayasu.chiba.jp/menu11324.html Utah, pp. 1-40, 1997. 3) Yasuda, S., Ishihara, K., Harada, K. and Shinkawa, N.: 16) Kayen, R. E. and Mitchell, J. K.: Assessment of liquefaction Effect of soil improvement on ground subsidence due to potential during earthquakes by Arias intensity, Geotech- liquefaction, Soils and Foundations, Vol.36, Special Issue, nical and Geoenvironmental Engineering, ASCE, Vol. 123, pp. 99-107, 1996. No. 12, pp. 1162-1174, 1997. 4) Tsuboi, H., Takahashi, Y., Harada, K. and Nitao, H.: 17) Nigorikawa, N. and Asaka, Y.: Long-term settlement of Comparing Remedial Measure against Soil Liquefaction on holocene, clay ground after The 2011 Great East Japan Reclaimed Lands, Tsuchi to Kiso, Japanese Geotechnical Earthquake, 10CUEE, Conference Proceedings, pp. Society, Vol. 44, No. 2, pp. 67-69, 1996. (in Japanese) 371-378. 2013. 5) Nishimura, S. and Towhata, I.: A three-dimensional 18) Recommendations for Design of Building Foundations, stress-strain model of sand undergoing cyclic rotation of Architectural Institute of Japan, 2001. principal stress axes, Soils and Foundations, Vol. 44, No. 2, 19) Specifications for highway bridges, Part V Seismic design, pp. 103-116, 2004. Japan Road Association, 2012. 6) Fukutake, K.: Studies on three-dimensional liquefaction 20) Chiba prefecture. Information bank of geological envi- analyses of soil-structure system considering mul- ronment. (online), available from
321