SMART 2008: Overview, synthesis and lessons learned from the International Benchmark

Highlights

• A International Benchmark to compare seismic assessment methods was launched in 2008.

• A synthesis of the results is presented considering several indicators such as the floor response spectra.

• The main lessons and conclusions from the SMART 2008 benchmark are exposed.

Abstract

This paper reports the main results and conclusions of an International Benchmark jointly organized within the framework of a wide research program launched by the French Atomic Energy and Sustainable Energies Commission (CEA), Electricité De France (EDF) entitled “Seismic design and best-estimate Methods Assessment for Reinforced concrete buildings subjected to Torsion and nonlinear effect (SMART)”. This research program included first an experimental campaign on reduced scaled specimen and shaking table tests, and second, a Benchmark exercise based on the contest of numerical methods and methodologies referring to the experimental results from the tests. The objectives of the Benchmark exercise were to compare the results obtained by conventional seismic assessment methods with those obtained by best-estimate methods, to compare the methodologies of taking into account uncertainties in the numerical analyses when a probabilistic assessment should be performed and to create an event allowing the international community in earthquake engineering to discuss the aforementioned topics of common interest. According to the analysis of the results, it turned out that (a) conventional methods are clearly able to provide relevant seismic assessment at the design level; (b) when dealing with overdesign seismic loadings, advanced best-estimate methods seem to be able to capture several key indicators such as the frequency shift of the peak of the response spectra; and (c) the blind nonlinear computations demonstrate a high robustness level of the specimen based on both structural and local damage indicators (frequency shift and inter-story drift, respectively).

Introduction

Earthquakes are a major concern when dealing either with the design of new buildings or the periodic reassessment of existing ones. The seismic risk is carefully taken into account by structural engineers, especially when considering buildings requiring a high-safety level, such as those in the petro-chemical industry or electro-nuclear energy production, in a context of strengthening of seismic risk requirements. Recent international regulatory standards recommend designing reinforced concrete (RC) buildings devoted to nuclear activities as the assembly of shear walls and frames. Indeed, the assembly of such structural components offers significant advantages over structures purely based on frames or shear walls. The story-drifts can be controlled in lower levels of the structure due to the stiff nature of RC walls. On the contrary, the frames increase the dissipative capability of the whole building, which leads to an increase in the displacement response of the structure. When this type of structure is regular or even slightly irregular, a consensus on the confidence level related to the assessment methodologies is nowadays accepted in the international community. However, the case of highly irregular frame-wall structures needs to be investigated, especially in the nonlinear behavior range. Indeed, geometric irregularities may lead to three-dimensional effects, such as torsion coupled with bending, especially when the structure exhibits a non-negligible eccentricity between the torsion center and the mass center.

To assess the safety and the robustness of such complex RC structures and related equipment regarding the seismic risk, it is necessary (i) to quantify the seismic margins, (ii) to quantify with an acceptable confidence level the floor response spectra (FRS) and (iii) to take into account uncertainties related to the input ground motions and the input material parameters that feed structural models. As usual, engineering practices can be gathered in two different families: (i) the conventional analyses and (ii) the best-estimate methods. To quantify in a meaningful way the seismic margins, conventional analyses that are usually simple and conservative are not sufficient. Best-estimate methods with advanced nonlinear constitutive laws are required to carry out in an efficient way the seismic assessment of such structures. Furthermore, the relevancy of advanced nonlinear models is not only related to the material parameters to be considered, but also to the variability of the input ground motion used to realize the structural assessment. Indeed, either material parameters or input ground motions are subjected to uncertainties that should be taken into account. From the aforementioned discussion, it is clear that improvements in the fields of nonlinear and uncertainties modeling are of primary importance for the earthquake engineering community.

The past decades were marked by major events that gathered the earthquake engineering community along the same path of the improvement of knowledge in the field of structural dynamics of low span shear walls and related assessment methodologies. The former Nuclear Power Engineering Corporation of Japan (NUPEC) organized a similar event, under the auspices of the Nuclear Energy Agency (NEA) Organization of Economic Cooperation and Development (OECD), twenty years ago. The RC structure was regular and U-shaped with low span shear walls. The main conclusions were that advanced nonlinear dynamic methods still have to be improved, in particular, when dealing with overdesign seismic ground motion leading to the structure working close to its ultimate limit state [1], [2], [3], [4], [5]. Some years later, within the framework of the CAMUS research program carried out at the French Atomic Energy and Sustainable Energies Commission (CEA) between 1996 and 2002 on a symmetric, in-plane, five-story RC wall 1/3th scaled mock-up [6], [7], shaking table tests were carried out to better assess the dynamic behavior of RC structures [8]. In addition, the predictive capabilities of existing assessment methodologies were evaluated due to two international benchmarks held in 1998 and 2003. It appears that the seismic margins were frequency dependent; this conclusion was confirmed by the related numerical experiences. In 2006, a blind prediction contest on the seismic response of a seven-story full-scale RC building with cantilever structural walls acting as the lateral force resisting system was launched by the Network for Earthquake Engineering Simulation (NEES), Portland Cement Association and University of California at San Diego. The objective of that research program was to verify the seismic response of RC wall systems designed for lateral forces obtained from a displacement-based design methodology, with particular emphasis placed on the interaction between the walls, slabs and gravity system, and on the issues related to construction optimization [9], [10], [11], [12], [13], [14].

To move forward and to bring new knowledge when dealing with the seismic assessment methodologies of RC structures being highly irregular, the Seismic design and best-estimate Methods Assessment for Reinforced concrete building subjected to Torsion and nonlinear effect (SMART 2008) research project joining the CEA, Electricité de France (EDF) and partially supported by the International Atomic Energy Agency (IAEA) was launched in 2006. Within this framework, the International Benchmark, SMART 2008, was organized to reach the aforementioned objectives. A representative of a typical, simplified, half-part of an electrical nuclear building 1/4th scaled mock-up was designed, built and tested between June and October 2008 on an AZALEE shaking table, at the Seismic Mechanics Studies (EMSI) laboratory [15], [16]. It was composed of three walls with openings forming a U shape and three stories. The specimen was designed according to the French current nuclear engineering practice [16]. Design spectra correspond to the seismic loading of a low to medium seismic area, anchored at PGA equal to 0.2 g. Seismic inputs of increasing intensity (up to a maximum PGA of 0.9 g) were applied to the mock-up; these synthetic accelerograms were generated from the design spectrum.

The SMART 2008 International Benchmark was organized between 2008 and 2010; thirty-four participating teams were registered. A list of participants is provided in Table 1 including in which parts of the benchmark they were involved.

The SMART 2008 International Benchmark consisted of two major stages, namely stage 1 and stage 2, and two scientific workshops. The participants’ results remained anonymous. The objectives of the benchmark were (i) to share with the international community the current engineering practices to perform seismic assessments (in particular, to evaluate conventional design methods for structural dynamic responses and FRS calculations, and to compare best-estimate methods for structural dynamic responses and FRS evaluations) and (ii) to promote the use of probabilistic methodologies addressing random and epistemic uncertainties, quantifying vulnerability associated to variabilities. The intent was to assess design practices and improve engineering knowledge concerning the seismic response of RC structures subjected to low and far-beyond design seismic motions, and to share experience and improve probabilistic approaches in order to provide the engineering community with more efficient tools and guidelines. To reach these objectives, an extensive experimental campaign on an RC mock-up that was designed to exhibit torsional and nonlinear effects was carried out, and the measurements fed the benchmark [15], [16]. It is important to notice the lack of experimental materials in the published literature for such complex structures; the ambition of the SMART 2008 research joint project was to bring about new knowledge, not only on the experimental aspect, but also on the applicability of existing seismic assessment methods to improve engineering practices.

Stage 1 was split into two sub-stages, namely stage 1.A and stage 1.B. Stage 1.A was devoted to the comparison of conventional approaches and best-estimate methods under blind conditions, since no experimental results (only design data) were provided to the participants by the local organizing committee. Stage 1.B gave participants the opportunity to adjust, or, in some cases, to modify, their numerical models based on the measurements realized during the low-intensity seismic test (linear range). In particular, the full set of experimental data, including inputs and outputs related to the design input ground motion (PGA equal to 0.2 g), was provided to the participants to allow them to improve their model calibration. In addition, the participants were asked to use their updated numerical model to assess the dynamic behavior of the RC mock-up when subjected to an overdesign input ground motion (PGA equal to 0.4 g). For both parts of stage 1, both kinematic (displacements and accelerations) and static quantities (internal stress resultants and moments) were required. Stage 2 was also split into two sub-stages, namely stage 2.A and stage 2.B. The most sensitive input parameters, including the seismic loading, were identified, and the probabilistic assessment of the vulnerability of the RC mock-up was conducted. Then, in stage 2.A, the sensitivity analysis was carried out. In stage 2.B, the participants computed the so-called fragility curves by using the methodology of their choice to estimate under several failure criteria the median capacity and the log-standard deviation. In this paper, stage 2 is not presented. The use of probabilistic methodologies addressing random and epistemic uncertainties will be discussed in future work.

This paper is devoted to presenting an overview of the work carried out within the framework of the SMART 2008 International Benchmark. In addition, a synthetic presentation of the main results is made before lessons and remarks are drawn for further works. To reach the aforementioned objectives, this paper is outlined as follows. In the first part, an overview of the participants’ work is given, and synthetic descriptions of the output results are presented. In a second part, the data are analyzed in two directions. On one hand, the existence of seismic margins is discussed in the case of an RC structure subjected to torsion and nonlinear effects designed with current engineering practices and applicable standards in the nuclear energy field. On the other hand, the capability of nonlinear constitutive models used in best-estimate methods to assess the seismic behavior of such a complex RC structure is also discussed.