Effects of thermal aging of laminated rubber bearing on seismic performance of bridges

2.1. Specification of specimen

The characteristic and durability test and evaluation of specimens were carried out according to ISO22762 (2010) [8, 9]. The above standard consists of 3 parts including the test method and test and evaluation standards of seismic isolator in form of laminated rubber bearing used for protecting a bridge or building structure from an earthquake. Natural rubber has excellent elastic force and restoring force so that it has been widely used for the seismic isolator. The natural rubber bearing (NRB) having the elastic restoring force which was the inherent property of natural rubber and the lead rubber bearing (LRB) were used the specimens in this study. The diameter of specimen included the thickness of external rubber (10 mm), and the thickness of one internal laminated rubber layer was 3 mm each. 29 lamination layers for the rubber were used and the thickness of reinforcing steel plate was 3 mm. $S_1$is primary shape factor as the ratio of load area to the free surface area including one rubber layer hole, and $S_2$ is the secondary shape factor indicating the ratio of effective width to the total width of internal rubber. The detailed specifications of specimen are as shown in Table 1, and Fig. 1 shows the shape and sectional specifications of specimen. The specimens used for the thermal aging test were designed as same as the laminated rubber bearing which was actually applied to the bridges approximately 16 years ago, and the natural rubber with shear strain of 100 % and shear modulus of 0.4 MPa was used. As a result of examining the target bearing according to the current bridge design standard, the shape factor was too small so that it is unsuitable for the current design standard. It's because no design concept for the laminated rubber bearing was established in the past, so it does not satisfy the current design standard.

2.2. Test conditions and method

The tester used for the test of laminated rubber bearing specimen is the compression-shear tester which has the maximum vertical load of 2,000 kN and the maximum horizontal load of 500 kN. And, the maximum horizontal speed is 250 mm/sec and the maximum vertical displacement is ±200 mm. The shear characteristics of laminated rubber bearing before and after thermal aging were compared to determine the compression and shear characteristics according to thermal aging. For the aging test, the shear characteristics and extreme shear characteristics of the specimens were tested before aging, and then after aging the specimens in an aging oven for 168 hours at a temperature of 70 °C, the shear test and the extreme shear failure test were carried out. 3 specimens before and after thermal aging each were used for the ultimate shear failure test, and these specimens were not used again after the test. The specimens used in this study were 16 specimens including 8 natural rubber bearings and 8 lead rubber bearings. The specimens were divided into 4 specimens before and after thermal aging, and the extreme failure test was carried out after the compression and shear test, and the cyclic loading test was carried out using additional 1 specimen. Also, in order to reduce a manufacturing error of specimen and an error which may occur during the test, the average value was taken from the result of each test. The compression-shear test which was the basic test was carried out using the same specimens before and after thermal aging, and the extreme shear test was also carried out using the specimens before and after thermal aging separately.

Fig. 1Cross section of laminated rubber bearing

2.3. Test result and analysis

2.3.1. Compression test

After thermal aging the laminated rubber bearing specimens (70 °C, 168 hr) according to the test conditions and carrying out the compression test, the result as shown in Table 2 was obtained. At this time, the design compressive force of lead rubber bearings was 565 kN and the design compressive force of natural rubber bearings was 490 kN, and the design compressive force ±30 % three times each was loaded on each test. As a result, in case of lead rubber bearing specimen, its compression stiffness before aging was 478 kN/mm but its compression stiffness after heat aging was 462 kN/mm, showing an approximately 3.3 % decrease. And, in case of natural rubber bearing specimen, its compression stiffness before aging was 195 kN/mm and its compression stiffness after aging was 191 kN/mm, showing an approximately 2 % decrease. The load-displacement history curve after thermal aging is as shown in Fig. 2, and the deformation curve for each type of specimen shows that the compressive force after aging increased slightly in comparison to the compressive force before aging. Also, after heat aging, the vertical displacement decreased and the slope increased. This is because the elongation percentage decreased as the stiffness of rubber increased due to thermal aging just as the rubber material specimen test result.

Fig. 2Hysteresis curve of compression test

a) NRB

b) LRB

Table 2Result of compression test

Item	Compression stiffness (kN/mm)
	NRB	LRB
Virgin	195	478
Aged	191	462
Change ratio (%)	–2.05	–3.34

Table 3Result of compression-shear test

Item	Shear stiffness (kN/mm)		Equivalent damping ratio (%)
	NRB	LRB	NRB	LRB
Virgin	0.196	0.73	7.56	35.28
Aged	0.195	0.70	7.35	35.16
Change ratio (%)	–0.51	–4.10	–0.27	–0.34

2.3.2. Shear test

As a result of compression-shear test before and after thermal aging, the shear stiffness of lead rubber bearing was 0.73 kN/mm before aging and 0.70 kN/mm after aging, showing an approximately 4 % decrease as shown in Table 3, and the shear stiffness of natural rubber bearing was 0.196 kN/mm before aging and 0.195kN/mm after aging, showing an approximately 0.5 % decrease. And, the equivalent damping ratio of lead rubber bearing was 35.28 % before aging and 35.16 % after aging, showing an approximately 0.34 % decrease. Also, the equivalent damping ratio of natural rubber bearing was 7.56 % before aging and 7.35 % after aging, showing an approximately 0.27 % decrease. Fig. 3 shows the comparison of history curves of lead rubber bearing (LRB) and natural rubber bearing (NRB) specimens before and after thermal aging.

Fig. 3Hysteresis curve of compression-shear test

a) NRB

b) LRB

2.3.3. Cyclic loading test

In order to determine the cyclic loading characteristics of lead rubber bearing and natural rubber bearing before and after thermal aging, the compression-shear test was carried out 50 times repeatedly. Each specimen was installed and a 0.5 Hz sine wave of the design shear displacements loaded horizontally. The rate of shear stiffness and equivalent damping ratio change according to increase in the number of cyclic loading on each specimen was measured after applying shear loading for 3, 5, 10, 30, 50 times each. Changes in the shear stiffness and equivalent damping ratio of each specimen for each cycle according to cyclic loading were shown in Fig. 4. According to the previous studies, changes in the shear stiffness and equivalent damping ratio at the 1st to 3rd cycle in case of cyclic loading test for 50 times were relatively high, but the range of change decreased gradually after that [10]. For changes in the shear stiffness due to thermal aging, the shear stiffness of lead rubber bearing changed from 0.731 kN/mm before aging to 0.715 kN/mm after aging and the shear stiffness of natural rubber bearing changed from 0.194 kN/mm before aging to 0.187 kN/mm after aging, showing a 2.7 % and 3.6 % decrease for each specimen. The equivalent damping ratio of lead rubber bearing changed from 35.25 % before aging to 34.58 % after aging, showing an approximately 1.9 % decrease, and the equivalent damping ratio of natural rubber bearing changed from 7.5 % before aging to 6.68 % after aging, showing an approximately 10.9 % decrease. As a result of cyclic loading test before and after thermal aging, NRB received greater influences than LRB and the equivalent damping ratio decreased by approximately 10 % due to cyclic loading. Here, the average value of 50 cyclic loading test results of each specimen before and after thermal aging was compared and shown for the rate of shear stiffness and equivalent damping ratio changes.

Fig. 4Physical properties of cycle test

a) Shear stiffness

b) Equivalent damping ratio

2.3.4. Extreme shear failure test

The extreme shear failure test was carried out in this study in order to determine the shear deformation performance of laminated rubber bearing before and after thermal aging. The test was carried out by installing one laminated rubber bearing specimen. If the test was carried out in a high speed, the shear deformation performance of the specimen may be overestimated, so the horizontal displacement was loaded up to the maximum strain with the constant wave at a low speed of 0.52 mm/sec. After installing the specimen, the maximum design compression force was loaded, and the shear displacement was applied in one direction until the laminated rubber bearing specimen failed. At this time, the test was carried out continuously until the load decreased rapidly after its maximum point, and the test was ended when it was considered that the proper failure was achieved. After the shear failure test, the shape of failed specimen is as shown in Fig. 7, and the shear failure occurred at the bottom of the specimen. The test result of each specimen is as shown in Table 4, Fig. 5 and Fig. 6. In case of lead rubber bearing, no clear strain hardening of rubber occurred, but it began to show at the shear displacement of 60 mm and the failure began to occur after 130 mm. In case of the natural rubber bearing, the strain hardening of rubber began to show at the shear displacement of approximately 100 mm and the shear failure began to occur at the shear displacement of approximately 150 mm.

Fig. 5Hysteresis curve of ultimate failure test (NRB)

a) Virgin

b) Thermal aged

Fig. 6Hysteresis curve of ultimate failure test (NRB)

a) Virgin

b) Thermal aged

Table 4Result of ultimate failure test

Items		NRB		LRB
		Shear strain (mm)	Shear strain ratio (%)	Shear strain (mm)	Shear strain ratio (%)
Virgin	#1	157.96	170	132.91	152
	#2	159.03	171	125.58	145
	#3	155.51	167	136.40	157
Aged	#4	148.10	159	117.91	136
	#5	143.85	155	117.36	135
	#6	150.51	162	118.73	137

Also, the result of extreme shear failure test before and after thermal aging showed that the shear failure displacement before aging decreased after aging. In case of the lead rubber bearing, the shear strain decreased by approximately 11 % on average, and in case of the natural rubber bearing, the shear strain decreased by approximately 6.8 %. However, in comparison between the test result and the result of previous studies, the shear strain is relatively low. According to the study conducted by Jung et al., the ultimate shear failure of lead rubber bearing and natural rubber bearing occurred at the shear strain of 500 % and 467 % respectively, which was approximately 3 times higher shear strain in comparison to this test result [11]. Such difference in the shear strain was caused by the effect of shape factor. The basic property of rubber used in the study mentioned above and this test is same, and the primary shape factor ($S_1$) for the lead rubber bearing is and the rubber bearing is 28, and the secondary shape factor ($S_2$) is 7 for both cases. However, the secondary shape factor of specimen used in this test is 2.98 which is relatively low. In the studies conducted by Yasaka et al. and Oh, the critical nature test result of laminated rubber was carried out with shape factors as variables showed that as the secondary shape factor was smaller, the shear force and the shear strain significantly decreased in the event of large scale deformation, and especially in case the secondary shape factor was less than 4, the elastic restoring force decreased and the Buckling Failure occurred at the same time in the event of large scale deformation [12, 13]. This result is similar to the test result in this study.

Fig. 7Shape the ultimate failure

a) NRB

b) LRB

3.1. Calculation of elastic seismic forces

The target bridge is the 2@2-span continuous slab bridge located in the seismic zone I and the specifications and pier section characteristics of the target bridge are as shown in Fig. 8, Table 5 and Table 6. The seismic analysis condition is seismic rating 1, and the acceleration coefficient is 0.154 and the site coefficient is 1.2.

The modeling was carried out with the plate element for the upper structure of target bridge for analysis and the frame element for the pier. The effects of site applied to the incident seismic load and the modeling was carried out with the fixed end for the bottom of the pier. Modeling was carried out for the bearing using NL Link to match the actual behavior. The dead load of structure was calculated automatically in the program and all other dead loads such as the pavement were considered. The structure analysis model is as shown in Fig. 8 and the axial force of pier according to the dead load analysis is as shown in Table 7.

In order to calculate the sectional strength, Horizontal Displacement and effective stiffness of pier, the moment-curvature ($M-\mathrm{\varnothing }$) analysis was carried out by considering the axial force effects. The section characteristics of pier are as shown in Table 6. The Kent&Park model was used for concrete and the Menegotto-Pinto model was used for the reinforcing bar. $M-\mathrm{\varnothing }$ analysis result is as shown in Fig. 9, The effective stiffness $\left(E_C\mathrm{*}{I}_{eff}\right)=(M_y/{\mathrm{\varnothing }}_y)M-\mathrm{\varnothing }\mathrm{}$was calculated using the yield moment $M_u$ and the yield curvature ${\varnothing }_u$ obtained by $M-\mathrm{\varnothing }$ analysis, and the secondary moment on the effective section applied when calculating the elastic seismic force for the collapse prevent level $I_{eff}=(M_y/{\mathrm{\varnothing }}_y)/\left(E_C\mathrm{*}{I}_g\right)$ is 0.022761 m⁴. Here, $E_c=\mathrm{8,500}\sqrt[3]{f_{ck}+8}$ (MPa).

Fig. 8Bridge dimensions and analysis model

The effective stiffness calculated from the section analysis of pier applied to the structural analysis model and the seismic analysis was carried out with the design response spectrum in Fig. 10 as the load. The mode number was selected to have the cumulative mass rate of vibration mode over 90 % through the eigenvalue analysis. The elastic seismic force was calculated by applying the 30 % regulation for each seismic load direction and the combined Seismic force of fixed-end pier and period vibration are as shown in Table 8.

Fig. 9Curve of M-∅

Fig. 10Design response spectrum

3.2. Seismic performance evaluation of pier

The section strength means the maximum resisting capacity of framework member to the horizontal load, and the load applying point is the centroid with respect to the direction perpendicular to bridge. The Table 9 shows the Stiffness and ductility of Pier. Table 10 and Table 11 show the Pier supply capacity and the pier demand supply respectively. In case of axis direction, the displacement ductility factor is (required = 2.498 > supply = 1.488) showing that the seismic performance is insufficient, and in case of direction perpendicular to bridge, the displacement ductility factor is (required = 0.686 < supply = 1.478) showing the behavior in the elastic region. Therefore, the comparative analysis was carried out with focus on the improvement of seismic performance in the axis direction.

4.1. Structural analysis modeling

In order to carry out the nonlinear time history analysis, 4 artificial earthquakes were constructed using the artificial seismic wave generation program SIMQKE under the seismic design condition by the standard design response spectrum of the highway bridge design code (2012). The seismic wave time history and seismic spectrum used are as shown in Fig. 11. The analysis model is same as shown in Fig. 8, except that all 24 bridge bearings were replaced with the lead rubber bearings and natural rubber bearings. Modeling was carried out for NL Link using the Isolator1 element which could analyze the characteristics of laminated rubber bearing. The specifications of bearing used for the Isolator1 are as shown in Table 1, and the main characteristics before and after thermal aging due to the test above are as shown in Table 12.

Table 12Properties before and after thermal aging

Items	NRB		LRB
	Virgin	Aged	Virgin	Aged
Shear stiffness $K_v$ (kN/m)	195,233.9	190,983.3	478,193.8	462,269.6
Effective stiffness $K_{eff}$ (kN/m)	194.13	186.7	730.6	701.0
Characteristic Force $Q_d$ (kN)			20.42	20.42
Elastic Stiffness $K_u$ (kN/m)			3235.2	3223.6
Post-elastic Stiffness $K_d$ (kN/m)			226.5	214.9

4.2. Analysis of results

The primary modes shape of natural rubber bearing before and after thermal aging for the target bridge are as shown in Fig. 11. As a result of analysis, all natural periods are longer than the natural period (T) (1.277 sec) of non-isolated bridge, showing the characteristics of bridge using the seismic isolation systems properly. The natural period of natural rubber bearing before and after thermal aging was 2.822 sec and 2.772 sec respectively, showing a 1.79 % decrease after thermal aging. This is because as time passed, rubber becomes hardened so that the elasticity decreases and the hardness increases due to the characteristics of rubber aging. On the other hand, the natural period of lead rubber bearing before and after thermal aging was 1.559 and 1.588 respectively, showing a 1.65 % increase after thermal aging. This is because the shear stiffness of lead rubber bearing rather than the hardening effect due to thermal aging of rubber decreased so that the natural period became longer as the displacement of bearing itself increased.

The shear force at the bottom of pier means the shear force at the part where the column of pier and spread footing contact, and it includes the inertial force due to the dead load of upper structure delivered through the bearing and the inertial force of the column. Also, the moment at the bottom of pier is the combination of inertial force due to the dead load of upper structure delivered through the bearing and the product of column length to the inertial force of the column. The analysis result before and after thermal aging is as shown in Table 13. In case of natural rubber bearing and lead rubber bearing, the shear force, moment and the displacement at the top of pier decreased after thermal aging. As shown from the response of bearing before and after thermal aging although the level of change is low, increased displacement response means the increased energy borne by the bridge bearing, resulting in decrease in the shear force of pier.

The displacement of natural rubber bearing increased by up to 1.59 % in comparison to the displacement before thermal aging, and the displacement of lead rubber bearing increased by 0.45 %. The shear force of bearing is closely related to the economic efficiency, edge distance and height. The shear force of bearing is an element which has an influence on the size of bearing just as the vertical capacity, and if the shear force related to the displacement capacity becomes higher, the rate of material cost increase due to increase in unit size becomes higher rapidly [17]. However, in case of seismic isolation, the safety ratio for shear force over 2 is secured in general, so increasing displacement due to the aging of bearing can be negligible.

Fig. 11Primary modes shapes and natural frequencies

The analysis result after applying the thermal aged laminated rubber bearing to an actual bridge is as shown in Table 13 and Table 14. The natural period of laminated rubber bearing according to thermal aging decreased after thermal aging in comparison to the natural period before thermal aging, and the shear force, moment of pier and the displacement at the top of pier also decreased. On the other hand, the displacement of laminated rubber bearing increased after thermal aging in comparison to the displacement of laminated rubber bearing before thermal aging. This result is similar to the result of study conducted by Kitahara [18]. In the study conducted by Kitahara, the seismic analysis was carried out for RC bridges by considering stiffness change in case the aging of rubber bearings has been carried out for 50 years and 100 years in the area where an earthquake actually occurred. As a result, the moment and shear force of a bridge where aging has occurred for 100 years decreased by 0.8 % and 1.2 % respectively. As a result of analyzing two results above, thermal aging of the laminated rubber bearing had no significant effect on the structure system of the whole bridge.

Fig. 12Artificial seismic wave used in analysis