-
39JR EAST Technical Review-No.26
Special edition paper
From a perspective of effective use of over-track space, we
proposed a frame form that secures installation space for seismic
isolation devices such as laminate rubber and dampers while
reducing the seismic isolation layer height as much as possible.1)
We carried out structural tests using scaled-down models of the
structure around that seismic isolation layer. And based on the
test results, we further proposed a design method and verified the
validity of that method.
Overview of Frame Form2Fig. 1 shows a frame form where a seismic
isolation construction method is employed for an aboveground
two-story building (first level: track level, second level:
concourse level, hereafter “assumed building”) having an upper
structure of pure steel rigid frame structure. This was designed as
a mid-story isolated building where a seismic isolation layer was
set up directly above the track level so as to prevent displacement
of the seismic isolation layer from interfering with train
operation on the track level and passenger flow on the platforms.
In order to secure overhead clearance of the track level so as to
not obstruct clearance gauge, the seismic isolation layer was of a
form where its lower beam, which is ordinarily straight, was bended
upward. The upper beams of the seismic isolation layer were
replaced with two large beams (hereafter “double girder”). By
placing the bent lower beam between the beams of the double girder,
the height of the seismic isolation layer was reduced while
securing space for the damper.
In such details, the double girder is not fixed to the column
directly above the seismic isolation layer; instead, it fixed to
the column with the large beam at a right angle to the double
girder. In this structure, the double girder cannot transfer the
applied bending moment as the torsional resistance of the
perpendicular large beam to the column because the double girder
cannot directly transfer that bending moment to the column.
Therefore, in order to confirm the structural performance of the
part around
Introduction1
the seismic isolation layer, we carried out static loading tests
using a scaled-down specimen that reproduced the part shown in Fig.
1.
Outline of Tests33.1 Overview of the SpecimenFig. 2 shows the
shape and dimensions of the specimen modeling the frame form of the
upper seismic isolation layer. The 1/1.6-scale specimen modeled the
seismic isolation layer and its upper column up to its point of
contraflexure and its beams at the both ends also up to their
points of contraflexure. The column was 300 × 300 × 16 mm square
(STKR400), the beam directly fixed to the column was a 400 × 200 ×
8 × 13 mm H-beam (SS400), and the double girder fixed to the column
with the large beam was 500 × 200 × 10 × 16 mm H-beams (SS400). The
laminate rubber was seismic isolation and vibration control
laminate rubber (RV30 - 520 - 27.2 × 3), of which lower surface was
supported in a secure manner.
Research on Structural Performance of Frame Forms in Seismic
Isolation and Vibration Control of Over-track Buildings
•Keywords: Over-track building, Seismic isolation, Laminate
rubber, Static loading test
Past research suggests the possibility that employing seismic
isolation technology could reduce response of a building in an
earthquake and accordingly allow for smaller lower structures. JR
East has also proposed jointly with Takenaka Corporation a frame
form called “double girder form.” The aim of that form is to secure
clearance gauge at the track level in over-track spaces, which have
much design and construction constraints, and at the same time to
minimize the space required for the seismic isolation layer. In
this research, we conducted static loading tests using a
scaled-down specimen consisting of laminate rubber and a double
girder so as to clarify the structural performance of the double
girder. By simulating the test results, we established a design
method for double girders.
*Frontier Service Development Laboratory, Research and
Development Center of JR East Group
Madoka Yamataka* Tsutomu Hoshikawa*
Scope modeled
倉庫
北口歩道レベル
基準RL
基準HL
最高高さ
下り線
上り線
ホーム
(a) Cross-sectional view
(b) Detailed planar view (c) Perspective drawing of frame
form
Seismic isolation layerupper beams (double girder)
Seismic isolationlayer lower beam
Seismic isolation layer
Laminate rubberUpper beam
Lower beam
Upper beam
Fig. 1 Overview of Seismic Isolation of an Over-track
Building
-
40 JR EAST Technical Review-No.26
Special edition paper
We controlled the load by shear deformation angle of the
laminate rubber, and we applied positive and negative progressive
loads twice at each of shear deformation angles of +/-50%, +/-100%,
+/-150%, +/-200%, +/-250%, and +/-300%. Here, shear deformation
angle 300% is about 1.2 times larger than the maximum shear
deformation angle of the laminate rubber that can be obtained from
the seismic response analysis in the case where Level 2 seismic
vibration specified in the railway aseismic standard2) was applied
to the assumed building.
3.3 Measurement MethodWe measured the horizontal displacement of
individual parts of the seismic isolation layer of the specimen
using a retractable displacement gauge, and we measured the angle
of gradient of the upper surface of the laminate rubber using a
clinometer. We also measured the deformation angle between the
layers of the upper frame and the torsion angle of the
perpendicular beam.
We further measured the strain near the joint of the ends of the
double girder and the beam and the strain caused by the torsion of
the perpendicular beam to which the double girder was fixed. The
bending strain at the ends of the beam directly fixed to the column
and the bending strain and axial strain of the column were measured
too.
Test Results4This article covers the test results of Patterns 1
and 3 from Patterns 1 to 4. Fig. 5 shows the relation of horizontal
force and deformation angle between layers for the whole frame as
deformation in the range from the lower end of the laminate rubber
to the loading point. Fig. 6 shows the relation of horizontal force
and laminate rubber shear strain, and Fig. 7 shows the relation of
horizontal force and deformation angle between layers of the upper
frame. To the +/-300% amplitude of the shear strain of the laminate
rubber, both of the laminate
Table 1 lists the tested patterns. The four parameters were the
position of the column (center column or side column) and the span
of the double girder (short or long). In the case of a center
column, the beam fixed directly to the column was assumed to be
supported by a pin and roller, and in the case of a side column,
free ends were to be supported by the pin and roller. The span
length of the double girder was determined by the position of the
pin-roller support on the double girder.
Table 2 shows the mechanical properties of the steel material
used for the specimen.
3.2 Loading MethodFig. 3 shows an overview of the force
applicator. In design of the assumed building, we applied to the
column of the specimen using a hydraulic jack axial force
equivalent to the average surface pressure of the laminate rubber
(269 kN, excluding deadweight), and we also applied horizontal load
at the point of contraflexure of the column. As the point of
contraflexure of the beam was a pin-roller support, it gave
resistance to vertical force only. By measuring that vertical force
using a pin-type load cell, we calculated the bending moment that
acted on the beam.
Fig. 4 shows the appearance of the whole specimen and force
applicator and the state of the laminate rubber after applied with
the axial force (269 kN).
Linear slider
Pin-type load cell
Seismicisolator ø520
Pin-type load cell
Linear sliderLinear slider
Vertical jackHorizontal jack
Upper frame form
Pin-roller support point
Column: 300 × 300 × 16 mm square
Pin-rollersupport point
Double girder: 500 × 200 × 10 × 16 mm H-beams
Elevation view
Planar view
Laminate rubber ø520
Double girder: 500 × 200 × 10 × 16 mm H-beams
σ ε μ σ εPart Steel type
Degree ofyield stress Yield strain Tensile strength Yield ratio
Elongation
Column
Panel
16 mm
19 mm
22 mm
19 mm
9 mm
12 mm
32 mm
Double girder500 × 200 × 10 × 16 H-beams
Beam400 × 200 × 8 × 3 H-beam
Upper diaphragm
Lower diaphragm
Large beam web
Reinforcement rib
Seismic isolation layer joint
Diameter orplate
thickness
Table 2 Mechanical Properties of the Steel Materials
Fig. 4 Overall View of Specimen (left),Laminate Rubber after
Applying Axial Force (right)
Fig. 3 Overview of the Force Applicator
Fig. 2 Overview of the Specimen
Test pattern Double girder span Column
positionLongShortLongShort
1234
Center column
Side column
Table 1 Test Patterns
-
41JR EAST Technical Review-No.26
Special edition paper
rubber and the upper frame form largely expressed elastic
behavior.
Fig. 8 shows the relation of horizontal force and laminate
rubber angle of gradient. This shows largely elastic behavior,
while the rigidity varied depending on whether it as positive or
negative. The reason for the difference in rigidity would be
because the horizontal deformation of the laminate rubber increased
the vertical downward displacement of the upper laminate rubber and
additional stress was generated there. Fig. 9 is a photo of the
deformation in Pattern 1 at a laminate rubber shear deformation
angle of +300%.
Examination of the Design Method5When designing a double girder
frame by replacing members with straight materials as usually done
in designing a rigid frame structure, we have to take appropriate
account of torsional deformation at joints generated due to the
double girder beam ends and the column not being directly fixed to
each other. In this chapter, we will propose a modeling method for
replacing the double girder frame with straight members and verify
the validity of that method by simulating the test results.
5.1 Proposing a Simulation ModelOur policy in modeling is (1) to
model two beams of the double girder as one beam, and (2) to model
the effects of the torsional deformation at joints as rotational
springs (Fig. 10).
In order to model two beams of the double girder as one beam, we
double the specifications of each cross-section of the beam.
In terms of torsional deformation at joints, we assume that
bending torsion—so-called Wagner torsion—would stand out because
the cross-sectional shape at joints is similar to that of a H-beam
(Fig. 11).
Laminate rubber angle of gradient (degree) Laminate rubber angle
of gradient (degree)
Pattern 1 Pattern 3
(Center column, span 4 m) (Side column, span 4 m)
Boundary conditions di�erent between both ends of the torsion
sectionCross-section of
torsional resistance:�ickness di�erent between upper and lower
�anges Torsion span
Fig. 8 Relation of Horizontal Force andLaminate Rubber Angle of
Gradient
Deformation angle between layers R (1 / 1000 rad.) Deformation
angle between layers R (1 / 1000 rad.)
Pattern 1 Pattern 3
Laminate rubber shear strain (%) Laminate rubber shear strain
(%)
Pattern 1 Pattern 3
(Center column, span 4 m) (Side column, span 4 m)
(Center column, span 4 m) (Side column, span 4 m)
Deformation angle between layers R (1 / 1000 rad.) Deformation
angle between layers R (1 / 1000 rad.)
Pattern 1 Pattern 3
(Center column, span 4 m) (Side column, span 4 m)
Fig. 5 Relation of Horizontal Force andDeformation Angle Between
Layers for Whole Frame
Fig. 7 Relation of Horizontal Force andDeformation Angle Between
Layers for Upper Frame
Fig. 6 Rubber of Horizontal Force and Laminate Shear Strain
Relation
Column
Large beamModeling of two large beams replaced with one straight
member
Rotational springstaking into account torsional deformation at
joints
Replacing withstraight materials
Fig. 10 Concept Image of Replacing Double Girderwith Straight
Members
Fig. 11 Details of the Joint
Fig. 9 Photos of Deformation(Pattern 1, Shear deformation angle
+300%)
-
42 JR EAST Technical Review-No.26
Special edition paper
From the bending torsion constant EIw, rotational spring
rigidity KT is derived as follows. tf is the average of the upper
and lower flange thickness, and the boundary condition at the ends
is pinned support.
KT = 48EIw /L3
EIw = t fb3h2 / 24 E: Young’s modulus of steel materialh: Beam
depth b: Beam width L : Torsion spantf: Flange thickness (average
of the upper and lower flanges)
5.2 Verification of the Validity of the Simulation ModelAs
torsional deformation at joints is a problem related only to the
upper frame form, we will examine only the upper frame excluding
the laminate rubber. We will consider the joints as a rigid
body.
To be consistent with the specimen dimensions, we set members to
be 300 × 300 × 16 mm square for the column, 400 × 200 × 8 × 13 mm
H-beam for the left large beam, and two 500 × 200 × 10 × 16 mm
H-beams for the right large beam. The analysis program used is
Nastran.
Fig. 13 shows a comparison of the test and analysis results of
the horizontal force-deformation angle between layers relation.
With Pattern 1, the test results agree well with the analysis
results, while with Pattern 3, those agree well with each other on
the side with positive loading but largely vary on the side with
negative loading. Assuming the column as a side column, the
position of the column-beam joint is opposite the point of pinned
support from the viewpoint of the laminate rubber at negative
loading. We believe that reduces rotational force of constraint,
and consequently deformation progresses.
Fig. 14 shows the relation of the deformation angle between
layers and the moment generated at the beam ends. As the test
results for Pattern 1, we adopt data where the additional moment is
corrected considering vertical rigidity of the laminate rubber
lowering at an earthquake. The figure shows that the simulation
results using the proposed model largely agree with the test
result.
Fig. 15 shows the relation of deformation angle between layers
and bending moment before adjustment. As shown in Fig. 16 (a), the
effective load support part of the laminate rubber decreases, and
that reduces its vertical rigidity at horizontal deformation. In
such a case, the vertical deformation at the laminate rubber
increases, but vertical deformation at each support point does not.
Therefore, the stress shown in Fig. 16 (b) occurs. Adjustment of
that bending moment leads to the relation of deformation angle
between layers and bending moment as shown in Fig. 15.
Conclusion6In order to identify the structural performance of
the double girder frame form around the seismic isolation layer
devised for seismic isolation of over-track buildings, we carried
out static loading tests using a scaled-down specimen. And we
proposed a simulation model for designing double girders. The
simulation results for the tests largely agree with the results of
actual tests, so we could confirm the validity of the proposed
model.
Reference:1) Madoka Yamataka, Kazuaki Iwasaki, Tsutomu
Hoshikawa, Mitsuru
Shimizu, “Development of Low-rise Over-track Buildings Using
Thick Laminate Rubber Seismic Isolation Materials”, JR East
Technical Review, No. 21 (2011): 15 - 21
2) Railway Technical Research Institute, “Design Standards for
Railway Structures and Commentary (Seismic Design)”, Railway
Technical Research Institute
Test
Analysis
Test
AnalysisHor
izon
tal f
orce
Q[k
N]
Hor
izon
tal f
orce
Q[k
N]
Deformation angle between layers R [1/1000 rad.] Deformation
angle between layers R [1/1000 rad.]
Pattern 1 Pattern 3
Test
Analysis
Test
Analysis
Bend
ing
mom
ent [
kN ·
m]
Bend
ing
mom
ent [
kN ·
m]
Deformation angle between layers R [1/1000 rad.] Deformation
angle between layers R [1/1000 rad.]
Pattern 1 a�er Adjustment Pattern 3
Fig. 13 Relation of Horizontal Force andDeformation Angle
Between Layers for the Upper Frame
E�ective load support part(a) Load support part of
laminate rubber
Simpli�edstructure diagram
Momentdiagram
(b) Pattern diagram of added moment
Test
AnalysisBend
ing
mom
ent [
kN ·
m]
Deformation angle between layers R [1/1000 rad.]
Pattern 1 before correction
Fig. 15 Relation of Deformation Angle Between Layers and
Bending Moment at Beam Ends
Rigid range Rotational spring
Fig. 12 Model Image of Replacement with Straight Members
Fig. 16 Concept Image of Laminate Rubber
Fig. 14 Relation of Deformation Angle Between Layers andMoment
at Beam Ends (total of the values at the double girder
beams on the right of the model image)
