1st International Conference on Natural Hazards & Infrastructure 28-30 June, 2016, Chania, Greece

Numerical research on the seismic response of novel integral abutment bridge designs and comparison to the current design practice Andrea Caristo1 Mott MacDonald Group, United Kingdom

Anastasia Palaiochornou2 Siemens, TEMPTON Personaldienstleistungen GmbH

Stergios Mitoulis3 Department of Civil and Environmental Engineering, FEPS, University of Surrey, UK,

ABSTRACT Integral Abutment Bridges (IABs) are joint-less bridges, that have low-maintenance cost. Soil-structure interaction due to thermal expansion/contraction cycles and dynamic loads limit their use to short-span bridges. SSI effects developed between the abutment and the backfill soil lead to settlements and ratcheting effects, hence replacement of the backfill soil might be required. This paper studies the performance of a new isolation scheme for IABs by using Tyre Derived Aggregates (TDAs) which form a Compressible Inclusion (CI) that decouples the abutment from the backfill soil. Two long span-integral abutment bridge models have been studied; first a conventional IAB is subjected to nonlinear time-history analyses, then a novel isolated IAB is modelled following the results obtained from laboratory tests on compressible inclusions (CI) obtained from re-used Tyre Derived Aggregates (TDAs). A comparison between the dynamic responses for selected seismic motions of the two systems is carried out. The permanent vertical settlements of the backfill soil, the residual pressures behind the abutment and the actions (bending moments and shear forces) on the abutment were numerically evaluated. The responses of the conventional and the isolated IABs showed that the use of isolator results in considerable reductions of the vertical displacements of the backfill soil and of the pressures acting on the abutment. The proposed research can be of use for extending the length limits of integral frame bridges subjected to earthquake excitations. keywords: bridge; integral frame abutment; long-span; compressible inclusion

1

Corresponding author: Mott MacDonald Group, [email protected] Siemens, TEMPTON Personaldienstleistungen GmbH, [email protected] 3 University of Surrey, [email protected] Bridge Engineering Research Group: www.mitoulis.com 2

1 INTRODUCTION Bridges are key elements of the transportation infrastructure and pose a cost for their construction and maintenance. This is particular evident in developed nations and major investments are necessary to maintain the network operative [1]. Integral Abutment Bridges (IABs) are frame structures that have no expansion joints or bearings, therefore they require less or no maintenance [2]; this characteristic has recently attracted attention [3] [4], as the industry seeks more sustainable solutions for asset management. The interaction of the abutment and the backfill soil can lead to long-term effects such as deterioration of the backfill soil at bridge approach embankments, overloading of the abutment and permanent displacements of the abutment. The aforementioned SSI results in recurring costs for agencies and poses a limit to the use of IAB solutions for long spans. In recent years, many researchers [5] [6] [7] have faced the challenge in an effort to reduce maintenance requirements for IABs [8]. The aforementioned limitations to IABs are mainly due to the thermal movements of the deck and they become more pronounced when the total continuous span length increases. Thermal contraction and expansion cycles lead to short-term changes in the position of the bridge abutment, leading to movements, densification and settlements within the backfill soil [28]. Additionally, creep and shrinkage permanent movements of the deck impose a permanent shortening of the deck in the long-term, allowing for soil to flow towards the footing of the abutment. When the bridge undergoes a cycle of expansion the soil resists its movement, inducing increased pressures on the abutment. The process of movements associated with soil-flows and increasing pressure is known as ratcheting effect. Directly connected to it are the deflections of the backfill soil within the bridge approach, which lead to a loss of comfort for the end-user (bump-at-the-end-of-the-bridge) and recurrent maintenance costs to repair the approach pavement. The increased loads experienced by the structure may lead to excessive loads and ultimate failure [5]. Such effects can be amplified by dynamic loads, such as seismic events, which strongly deteriorate the performance and integrity of the backfill soil, the abutment and hence the bridge [9] [10]. Different approaches have been studied in literature to address these intrinsic problems [11] [12] [13] [14] [15]. Among these, Humprey et al. [16] [17] introduced the use of tyre derived aggregates as a compressible inclusion behind culverts to reduce the soil pressure on walls. Mitoulis et al. [18] introduced this concept to IABs. This paper provides results obtained through 2D finite element analyses in PLAXIS 2D (ver. 8.2). Novel compressible inclusions (isolators), whose properties were identified by laboratory tests, were used to isolate the abutment from the backfill soil. Comparisons between a conventional and an isolated abutment were conducted. The study showed that the TDA isolators reduce effectively the soil pressures on the abutment and the permanent deflections of the backfill enabling the design of longer IABs.

2. NUMERICAL MODELLING 2.1. Compressibility of the TDA isolator on laboratory tests The properties of the compressible inclusions/isolators TDA were defined by laboratory tests conducted at the University of Surrey. The properties of the compressible inclusions were validated by triaxial tests conducted at Aristotle University [19] [20]; the results obtained through the lab tests conducted have shown good agreement. Based on previous results [18] the Poisson ratio for the isolator was taken as 0.49 and a thickness equal to 300mm was considered for the isolated IAB analyzed herein. The modulus of elasticity of the inclusion was found to be 56.9 kPa, which corresponds to a measured oedometric modulus of Eoed=974.2 kPa. The unit weight of the tyre derived aggregates was taken equal to 6.1 kN/m3. 2.2. Backfill and foundation soil Compacted sand was considered for the backfill soil having an angle of friction of 420 and dilatancy angle of 10.90. Soil deposit of 30 m depth corresponding to Eurocode 8-1 [21] ground type B was considered for the

foundation soil. Both the backfill and foundation soil were assumed to have an elasto-plastic behavior and the Mohr–Coulomb criterion was considered. The soil nonlinearity for the low to medium strain range, the parameters of the soil modulus and damping were estimated based on 1D equivalent linear analyses. The stiffness and the damping were based on calibration performed on the strain level as per [22]. For higher strain levels the effect of nonlinearity was taken into consideration through the Mohr-Coulomb yield criterion that was used for the soil behaviour in the 2D numerical model. The soil properties are summarised in Table 1. The geogrids used for stabilising the backfill were considered to have elastic behaviour with axial stiffness EˑA=1.0E+05 kN. Table 1. Soil properties of the backfill and the foundation soil.

backfill soil

foundation soil

layer(s)

z (m)

γ (kN/m3)

Poisson v

c (kPa)

E (kPa)

Vs (m/s)

1-14 1 2 3 4 5 6 7 8 9 10

0.5 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0

18.5 19.5 19.5 19.5 19.5 19.5 19.5 19.5 19.5 19.5 19.5

0.43 0.35 0.35 0.35 0.35 0.35 0.35 0.35 0.35 0.35 0.35

0.01 300 329.3 343.9 358.5 373.1 387.8 149.8 164.0 178.3 192.5

3.89E+05 6.96E+05 9.92E+05 1.19E+06 1.37E+06 1.57E+06 1.74E+06 1.93E+06 2.09E+06 2.26E+06 2.48E+06

270 360 430 470 505 540 570 600 625 650 680

2.3. Finite Element Model in PLAXIS Analysis on the complex soil-structure interaction for Integral Abutment Bridges (IABs) under seismic loads was carried out using the plane strain finite element code PLAXIS 2D (ver 8.2) [23]. Two models of the bridge were developed; the first one considered the conventional abutment in contact with the backfill soil, which comprised the abutments, the piers, the foundation soil and the backfill soil; the second model considered the isolated abutment, in which the compressible inclusion layer and mechanically stabilized earth are considered behind the abutments. Dynamic absorbent boundaries were used to simulate the far-field behaviour of the medium and avoid boundary effects within the model. The model has a total width of 282m and a total height of 36m. The meshing is done using 15-nodes triangular elements, with a total of 47317 nodes. Preliminary sensitivity analyses showed that the width adopted is sufficient to minimise the boundary effects, without increasing significantly the computational costs. The prestressed concrete box deck has been modelled with the use of a plate element, to which axial and bending stiffness are directly applied. The total axial stiffness of the bridge was equal to EA=2.222E+7 kN, with a flexural stiffness equal to EI=4.56E6 kNm2. In the conventional IAB 14 layers of backfill soil were considered behind the abutments. Interface elements between the backfill soil and the concrete abutments were used, to better model the interaction between the two different materials. The isolated IAB has been modelled through the use of 14 layers of backfill soil, the layers of compressible inclusions behind the abutments and geogrid elements; the latter were applied both horizontally and vertically, in order to stabilise the soil.

For all models the initial geostatic stresses, along with the stresses induced at construction stages for the abutments, the compressible inclusions and the backfill soil, were taken into account, in order to consider the effect that the sequence of loads has on the long-term behaviour of the system [24]. A cluster to model air underneath the bridge deck was defined.

Figure 1. The 2D PLAXIS model of the proposed integral bridge with the compressible inclusion.

2.4. Selection of earthquake excitations The dynamic loads chosen are seven recorded ground motions, all scaled to 0.3g, shown in Table 2. All records were compliant with Eurocode 8-1 elastic spectrum for ground type B. A decoupling approach was followed to define the response of the abutment-backfill system due to both the seismic motion that is induced to the foundation of the abutment and the seismic response of the bridge deck. Table 2. Summary of the seismic signals [18] Earthquake name

Station ID – Station name

Magnitude Mw

PGA (g)

Parnitha, Greece, 9/7/1999

2472-Athens 4 (Kipseli District)

6.0

0.12

Kozani, Greece, 5/13/1995

ST1320 - Prefecture building

6.5

0.14

Aigio, Greece, 6/15/1995

Telecommunication building

6.5

0.54

Friuli, Italy, 5/6/1976

ST20-Tolmezzo-Diga Ambiesta

6.4

0.32

ST64

6.9

0.18

Santa Barbara, Courthouse

7.3

0.20

090 CDMG station 1498

5.5

0.19

Montenegro, former Yugoslavia, 4/15/1979 Kern County, (Taft,) USA, 7/21/1952 Trinidad, USA, 8/24/1983

The damping properties of the system were based on results obtained in other studies [25]. The seismic excitations were used as input time-histories at the base of the model, in order to obtain the longitudinal dynamic response of the system. The choice of seven records is justified by the possibility of considering average distribution of results, in accordance with Eurocode 8-2 [26].

3. RESULTS 3.1. Earth pressures on the integral abutment High residual pressures on the abutment walls following seismic motions, even if structurally acceptable in the immediate after-shock, may cause structural damages once coupled with thermal expansion cycles of the bridge. Therefore, it is very important to control and reduce the permanent pressures on the abutments, in order to avoid the worsening of long-term pressures build-up and the formation of ratcheting effects. In the present research effects from thermal movements, creep and shrinkage prior to the earthquake excitations were not taken into account. For the conventional system, the maximum earth pressure experienced on the abutment’s wall after the seismic event was found to be 269.1 kN/m2 for the Montenegro earthquake. For the same motion the use of TDA isolator aggregates yielded a maximum pressure of 14.9 kN/m2. Apart from the significant pressure reduction, it is important to notice that the pressure distribution changes from a non-linear distribution with the maximum value registered at about the wall mid-height to an almost linear distribution for the isolated system, with maximum value at the abutment’s footing. The change in shape of the distribution is very important, as it leads to a shift downwards of the pressure centroid and consequently to an overall reduction of the overturning moment applied to the bridge abutments, which is reflected in reduced materials usage during construction. Interestingly, the pressure distributions and the peak values obtained for the novel system are almost equal for all the earthquake excitations examined here. This fact shows that the introduction of the CI layer leads to an almost complete decoupling of the system from the backfill soil, as already evidenced from the response period identification (results are not presented here). A comparison of the results shows a highly non-symmetrical behaviour for the conventional system, with peak pressures different in position and entity. Such non-symmetrical behaviour is sensibly rectified when the CI layer is installed behind the abutments.

6

6

4

-1σ -1σ

3

+1σ

2 mean

-300

mean

-200 -100 Soil pressure (kN/m2)

0

Isolated IAB

abutment height (m)

5

Isolated IAB

+1σ

7

abutment height (m)

Conventional IAB

7

Conventional IAB

5 4

-1σ

3 +1σ

-1σ

2

+1σ mean

1

1

0

0

mean

0

50

100

150

200

250

Soil pressure (kN/m2)

Figure 2. Soil pressure on the left abutment (a) and on the right abutment (b).

The non-symmetrical behaviour shown by the conventional IAB is directly connected to the non-symmetricity of the seismic excitation. The isolated IAB does not reflect this non-symmetricity as its response is decoupled from the one of the backfill.

3.2. Permanent vertical deflection of the backfill soil One of the design challenges of IABs is the control of the vertical deflections of the backfill soil; such displacements are responsible for the reduction of user’s comfort when approaching the bridge, as well as an important factor to recurring maintenance costs [27]. Controlling such behaviour and finding means to mitigate it is important to effectively reduce the maintenance costs associated with IABs. From the seismic analyses it has been noted that all the ground motions lead to the swelling of the backfill soil behind the abutments for the conventional system, with uplifts as high as 9mm. The use of CI to isolate bridge leads to swelling and settlements, which are significantly reduced in comparison to the displacements experienced by the conventional system. Results obtained for each signal are presented in Table 3. Table 3. Maximum vertical displacements of the backfill soil behind the abutments

Earthquake Parnitha Italy Montenegro Kozani Aigio Trinidad Kern County

Conventional IAB (mm)

Isolated IAB (mm)

+17 +23 +78 +35 +13 +15 +90

-4 -4 -9 -6 -4 +4 +8

Based on the seismic analyses carried out the settlements and the swelling of the backfill soil has shown a significant dispersion. Results can be seen in Figure 3, where the mean distribution and the mean ±1σ standard deviation of the vertical permanent displacement for the backfill soil are shown. In it, the solid line is used for the conventional system, while the dashed line is used for the isolated IAB. From these distributions it is possible to see that the maximum swelling assumes value in the range between 10mm and 70mm. No settlement has been observed to a distance equal to the abutment height for the backfill soil, but swelling only. With regard to the isolated system, both swelling and settlements are observed; the dispersion in this case is highly reduced and the settlement values are negligible. It is also important to notice that the area affected by the swelling is larger for the conventional IAB as significant deflections were observed at distances equal to the abutment height. The significant reduction of vertical displacements after seismic events when the abutment is isolated renders a resilient damage free bridge in the after-shock phases. Although swelling is observed in some cases with the compressible inclusion layer, it is interesting to notice the differences in the swelling formation mechanisms between the two systems. In the conventional backfill system the swelling can be attributed to the inherent non-linear behavior of the backfill soil, which does not recover the imposed strain once the abutment moves towards and then away from it; therefore, the settlement is due to translation effects. The response of the isolated system is inherently different; indeed the swelling in this case is due to the rotation of the abutment about its footing.

0.05

swelling(m)

+1σ Conventional IAB

swelling(m)

0.07

0.03

mean

0.07

0.03

mean

-1σ

-1σ

0.01

-1σ

-0.03 Isolated IAB

-8

-6

-4

-2

-0.01 +1σ -1σ

-0.03 Isolated IAB

-0.05

-0.05 -10

settlement (m)

-0.01

settlement (m)

0.01

+1σ

Conventional IAB

+1σ

0.05

0

0

distance from abutment (m)

2

4

6

8

10

distance from abutment (m)

Figure 3. Vertical displacements of the backfill soil for the left (a) and right (b) abutments.

3.3. Actions on the abutments Beneficial effects given by the use of the isolating layer of CI were introduced in the previous sections, in terms of reduction of pressure on the abutment walls and in terms of reduction in the permanent vertical displacements of the backfill soil. Such beneficial effects have a direct impact in the design of the abutments in terms of dimension, strength and reinforcement needed, as well as in the maintenance of the approaching slab and the backfill soil. From the results obtained through the distributions of pressures presented in this paper the following bending moment diagrams have been obtained for the system. 7

7 Conventional IAB

Conventional IAB

5 4 3 Isolated IAB

2 +1σ

6

Abutment height (m)

Abutment height (m)

6

5 4 3 +1σ

2

1

1

0

0

Isolated IAB

-1σ

-1σ

-400

-300

-200

-100

0

100

Bending moment (kNm/m)

200

300

-300

-200

-100

0

100

200

300

400

Bending moment (kNm/m)

Figure 4. Bending moments on the left (a) and right (b) abutments. After-shock phase.

For the conventional abutment, the maximum sagging bending moment was found to be 173 kN*m/m, whilst the maximum hogging bending moment was found to be 258 kN*m/m. When the compressible inclusion is installed behind the abutment, the sensible reduction of permanent pressures developed after the seismic motion is reflected in the bending moments induced in the abutments, with a maximum sagging equal to 7.04 kN*m/m and maximum hogging equal to 12.94 kN*m/m. Hence, the peak values obtained for the isolated system are 5% of

the ones obtained for the conventional system. The standard deviation for the conventional system is equal to ±45 kN*m/m, whilst it is negligible for the isolated system.

4. CONCLUSIONS A new compressible inclusion (CI) of tyre derived aggregates (TDAs) was utilised to isolate long-span Integral Abutment Bridge (IAB) from the backfill soil. The CI has proven to be a valuable mean to alleviate frequent serviceability movements of the backfill soil, as well as a valuable mean to reduce significantly the pressure build-up over cyclic loads in IABs. To assess the efficiency of the CI, a numerical study conducted with PLAXIS 2D (ver. 8.2) was undertaken, considering a model that is representative of typical integral bridges. The structure was subjected to seven Eurocode-compliant recorded seismic motions; all signals were scaled to a PGA value representative of typical European seismic prone areas. Based on the results of this study, the following conclusions were drawn: The permanent soil pressures on the abutments after the seismic events were significantly reduced for the isolated IAB, with permanent pressures lower than the at-rest state of the backfill soil. The peak mean pressure value for the conventional system was found to be equal to 197.5 kPa at mid-height of the abutment wall, whilst for the isolated IAB the peak value was of 14.9kPa at the abutment base, whilst a linear distribution of pressures was observed in the latter case, an indication that the backfill soil is not disturbed by the moving abutment. The soil pressure reductions also lead to significantly lower bending moments and the shear forces acting on the abutment. The BM of the conventional abutment was found to have maximum value of 258 kN·m/m whilst the BM of the isolated IAB was found to be 12.94 kN·m/m for the isolated system in hogging and from 173 kN·m/m for the conventional system to 7.04 kN·m/m for the isolated system in sagging. Such reductions are significant in economic terms, as less steel reinforcement is required. Reductions in shear are in line with the reductions in pressure and bending moment, with a maximum mean shear force in the conventional system equal to 167 kN/m and equal to 10 kN/m for the isolated system. Significant reductions in the settlements of the backfill soil were also registered in the isolated system, with a maximum settlement equal to 0.009m. Reduced swelling was observed when the compressible inclusion is installed behind the abutment, against swelling up to 0.09m registered for the conventional system. Hence, the permanent vertical deflections of the backfill soil were significantly reduced through the use of the novel system based on the use of tyre-derived-aggregates and the geogrids reinforcement. It is believed by the authors that the implementation of the novel isolator is a promising path to successfully expand the use of IABs for longer spans, allowing the engineering community to deliver more efficient, sustainable and highly resilient infrastructure links to future generations, in the effort of optimizing the whole life cost of transport networks and improve their reliability. Future research is deemed necessary to study in depth the hysteretic behavior of the CI and the impact that their use will have on the design approach, process and costs.

REFERENCES [1] ASCE. Report card for America’s infrastructure, 2013. [2] Mistry VC, Al Mangus. Get in, stay in, get out, stay out. Federal Highway Administration, Department of Transportation, United States 2006; 70(3). FHWA-HRT-2006-001. [3] INTAB. Economic and durable design of composite bridges with integral abutments. Research Program RFSC RFS-P208065, European Commission, Brussels, 2010 [4] IABSE. Special Issue on Integral Bridges. Structural Engineering International 2011; 2. [5] Dicleli M, Eng P, Albhaisi SM. Maximum length of integral bridges supported on steel H-piles driven in sand. Engineering Structures 2003; 25:1491–1504. [6] Mitoulis SA, Tegos I.A. Two new earthquake resistant integral abutments for medium to long-span bridges. Structural Engineering International: Journal of the International Association for Bridge and Structural Engineering (IABSE) 2011; 21(2):157-161. [7] Baptiste KT, Kim W, Laman JA. Parametric study and length limitations for prestressed concrete girder integral abutment bridges. Structural Engineering International 2011; 21(2):151-156. [8] Mitoulis SA, Tegos IA, Stylianidis K-C. Cost-effectiveness related to the earthquake resisting system of multi-span bridges. Engineering Structures 2010; 32(9):2658-2671. [9] Kappos AJ, Sextos AG. Seismic assessment of bridges accounting for nonlinear material and soil response and varying boundary conditions. NATO Science for Peace and Security Series C: Environmental Security 2009; 3:195-208. [10] Mitoulis SA. Seismic design of bridges with the participation of seat-type abutments. Engineering Structures 2012; 44:222-233. [11] England GL. A thermal displacement compensation unit for integral bridges. Patents, Bridge structures 2005. P.N.: CA 2605437 A1. [12] Horvath J. Lateral pressure reduction on earth-retaining structures using geofoams: correcting some misunderstandings. ASCE Earth Retention Conference 3 2010: 862-869. DOI: 10.1061/41128(384)86. [13] Hoppe EJ. Field study of integral backwall with elastic inclusion. Final Report, Virginia Transportation Research Council, Charlottesville, 2005. [14] Pötzl M, Naumann, F. Fugenlose betonbrücken mit flexiblen widerlagern. Beton- und Stahlbetonbau 2005; 100(8). [15] Arsoy S, Duncan J, Barker R. Behavior of a semiintegral bridge abutment under static and temperature-induced cyclic loading. Journal of Bridge Eng. 2004; 9(2):193–199. [16] Humphrey D, Blumenthal M. The use of tire-derived aggregate in road construction applications. Proc Green Streets and Highways, 1st T&DI Green Streets & Highways Conference, ASCE, 2011. [17] Humphrey D. Civil engineering applications using tire derived aggregate (TDA). DRRR-2011‐038, California Integrated Waste Management Board, California, 2010. [18] Mitoulis S., Palaiochorious A., Georgiadis I., Argyroudis S. Extending the application of integral frame abutment bridges in earthquake prone areas by using novel isolators of recycled materials, Earthquake Engineering and Structural Dynamics (accepted), 2016 [19] Senetakis K, Anastasiadis A, Pitilakis K, Souli A. Dynamic behaviour of sand/rubber mixtures. Part II: Effect of rubber content on G/Go-γ-DT curves and volumetric threshold strain. Journal of American Society of Testing and Material International 2012; 9(2). Paper ID JAI103711. [20] Senetakis K, Anastasiadis A, Pitilakis K. Dynamic properties of dry sand/rubber (SRM) and gravel/rubber (GRM) mixtures in a wide range of shearing strain amplitudes. Soil Dynamics and Earthquake Engineering 2012; 33:38-53. [21] Argyroudis S, Kaynia AM, Pitilakis K. Development of fragility functions for geotechnical constructions: application to cantilever retaining walls. Soil Dynamics and Earthquake Engineering 2013; 50:106–116. [22] CEN. EN 1998-1 Design of structures for earthquake resistance - Part 1: General rules, seismic actions and rules for buildings, European Committee for Standardization, Brussels, 2004. [23] PLAXIS 2D (2008). Reference Manual, Version 8. [24] Argyroudis SA, Mitoulis SA, Pitilakis KD. Seismic response of bridge abutments on surface foundation subjected to collision forces. Proc COMPDYN 2013, 4th ECCOMAS Thematic Conf Comp Methods Struct Dyn Earthq Eng, 2013; Kos Island, Greece, 12-14 June. [25] Mitoulis S, Argyroudis S, Kowalsky M. Evaluation of the stiffness and damping of abutments to extend direct displacement based design to the design of integral bridges. Proc COMPDYN 2015, 5th ECCOMAS Thematic Conf Comp Methods Struct Dyn Earthq Eng, 2015; Crete Island, Greece, 25-27 May. [26] CEN. EN 1998-2 Eurocode 8 - Design of structures for earthquake resistance - Bridges, European Committee for Standardization, Brussels, 2005. [27] Maruri, R. and Petro, S. (2005) - Integral abutments and jointless bridges (IAJB) 2004 survey summary. [28] LaFave J. M., Riddle J. K., Jarrett M. W., Wright B. A., Savatora J. S., An H., Fahnestock P.E. - Numerical Simulations of Steel Integral Abutment Bridges under Thermal Loading, ASCE Journal of Bridge Engineering, 2016