Improved ground motion description for structural analysis of industrial facilities based on scenario
F. Bonilla (IFSTTAR), P. Bazzurro (IUSS Pavia)
I. Zentner, G. Senfaute (EDF)
France: EDF R&D Lab Paris-Saclay, IFSTTAR (3 years, 2018-2021, 50% of time) Marne la Vallée Cedex 2, Paris
Italy: IUSS Pavia
Approximately 3,000 euros gross/month
Gloria Senfaute : email@example.com
Irmela Zentner : firstname.lastname@example.org
The aim of this Ph.D thesis is to develop and implement methodologies that would improve the description of ground motion to be used for seismic reliability analysis of nuclear power plants (NPPs) in France. In particular, it is intended to contribute to the development of:
- “Nocivity” (i.e., degree of harmfulness) filtering methodology based on Cumulative Absolute Velocity – CAV -or other ground motion Intensity Measure – IM. This filtering is useful to concentrate the analyses on earthquakes that are significant for risk assessment, done both deterministically and probabilistically;
- Conditional spectra approach to obtain three-component ground motions suitable for best- estimate response distribution assessment and for performance-based analysis
Figure 1: Illustration of Probabilistic Risk Assessment PRA framework. Legend: EPS stands for Etudes
Probabilistes de Sûreté, the French acronym for PRA.
The work could be conducted in two steps:
1) In order to focus the Probabilistic Seismic Hazard Assessment (PSHA) on significant events and to
exclude non damaging earthquakes for probabilistic risk assessment (PRA), it is proposed to investigate a nocivity filtering methodology that could be used as an alternative to defining Mmin in the hazard integral (see e.g. Bommer et al 2017). For this purpose the following issues might be addressed:
- Define pertinent nocivity parameter (CAV, or average pseudo-spectral acceleration, or other IM) and thresholds for filterin Threshold might be structure- or component-specific
- Theoretical implementation in the PSHA framework of the of the hazard calculations in terms of the selected nocivity parameter including the development of the variance-covariance matrix with other IMs of interest. It is the intent of this project to implement this methodology in the Openquake hazard engine of GEM.
2) The Uniform Hazard Spectrum (UHS) is a useful tool for communicating and comparing engineering design safety levels (e.g., for a 10,000-year return period). However, a UHS representation of the hazard is not suitable for best-estimate response or performance-based seismic risk analysis. Nevertheless, in current practice, a set of ground motions is often chosen so as to comply in its mean or median response spectrum with the target UHS. This compliance can be achieved by selecting and modifying natural time histories via spectrum matching or by generating entirely synthetic time histories with equivalent features. This approach can be conservative, since the UHS constitutes an envelope of the ground motion that can be expected by many different earthquake scenarios controlling the hazard at that probability level. However, it can also be non conservative when UHS on soil are developed from the corresponding ones on rock using soil amplification factors based on amplification factors obtained from 1D analyses that use spectrum-compatible records.
Since UHS-consistent ground motions cannot be generated by any single earthquake scenario, researchers have proposed methodologies for using ground motions compatible with multiple hazard- consistent target spectra that still comply with the design requirements provided by the UHS. In recent applications of safety analysis of NPPs, for example, the compliance with the UHS was rather to consider only one conditioning frequency. In this framework, Conditional Mean Spectra (CMS) and Conditional Spectra (CS) were considered for compliance with a set of UHS at different return periods (e.g., Renault et al. 2015).
While more technically sound, this CMS- or CS-based risk assessment approach requires a large number of time histories to fully cover the hazard, and, therefore, a large number of time-consuming response analyses. Besides, a great number of the analyses with lower intensity ground motions may not lead to considerable structural damage and, therefore, are of less interest for the structural reliability analysis. On the other hand, the ground motion intensity levels useful for the design and verification of the seismic performance of industrial plants are very rare, a condition which makes it difficult to define physically sound time histories.
The final goal of this study is to develop a comprehensive methodology that is both technically robust and feasible in an industrial context (e.g., more than a few hundreds analyses are not feasible). Besides, some particular issues, such as the large number of eigenmodes of relevance to the response of the different components present in industrial plants is of concern and should be addressed.
EPRI (2006), Use of CAV in Determining Effects of Small Magnitude Earthquakes on Seismic Hazard Analysis, Product ID 1014099,
Bommer J., H. Crowley (2017), The Purpose and Definition of the Minimum Magnitude Limit in PSHA Calculations. SRL 88(4).
Campbell, K.W. and Bozorgnia, Y. 2011, Prediction equations for the standardized version of cumulative absolute velocity as adapted for use in the shutdown of U.S. nuclear power plants, Nuclear Engineering and Design 241( 7), 2558-2569.
Lin T., Harmsen S.C., Baker J.W., Luco N. (2013). Conditional spectrum computation incorporating multiple causal earthquakes and ground-motion prediction models. Bull. Seism Soc Am 103(2A), 1103-1116.
Renault Ph., Proske D., Kurmann D., Asfura A. (2015), EVALUATION OF THE SEISMIC RISK OF A NPP BUILDING USING THE CONDITIONAL SPECTRA APPROACH. Proceedings of SMIRT-23, Manchester, UK.
Kohrangi M., Bazzurro P., Vamvatsikos D. and Spillatura A. (2017) Conditional spectrum-based ground motion record selection using average spectral acceleration. Earthquake Engineering & Structural Dynamics 46 (10), 1667-1685.
EPRI (2016), Development and Application of a New Seismic PRA Methodology for Determining Seismic Fragilities Using Scenario Earthquakes. EPRI Final report 3002008098, Palo Alto.
Zentner, I. (2014), A procedure for simulating synthetic accelerograms compatible with correlated and conditional probabilistic response spectra. Soil Dyn Earth Eng. 63(1), 226-233.
I. Zentner (2017), Simulation of hazard consistent accelerograms compatible with target UHS and application to the computation of in-structure floor spectra. WCEE.