Action 6.2 : Selection of time histories

State of the art / Background knowledge

Methodology for the selection of time histories matching a specific response spectrum vary according to the country, the seismological background, etc.

On one hand, reference guides (e.g. ASCE 43-05, ASCE 4-98…) issued from countries susceptible to strong earthquake generally limit the use of synthetic time histories for linear analyses, with a complementary need to demonstrate that the produced responses are “consistent with those developed from the use of modified, real recorded motions for the responses of interest”. For non-linear analyses, the use of natural recorded time histories, modified or not, is mandatory due to possible lacks of coherency between synthetic time histories and physic behind the seismic event in terms of ground motions.

On the other hand, several reference guides take into account the fact that for some regions, the availability of recorded time histories of strong motion earthquakes is not guaranteed. References such as IAEA Safety Guides NS-G-1.6 and SSG-9, French RFS2001.01 and ASN Guide-2-0, European Utility Requirement EUR Rev.D (…) allow to use synthetic time histories completing specific conditions (frequency content, spectral energy, duration and Ground Motion Parameters) for both linear and nonlinear studies

Limitations of current practice or knowledge (i.e. technological barrier)

Various practices tend to favor one or other, with a greater inclination towards natural time histories (modified or not). However, there is a lack of objective comparisons between both methods with regards to the damages to structures and facilities.

The aim is:

  1. To demonstrate that if the same effort is put on the selection of the time histories either modified natural or synthetic ones, then the use of synthetic time histories allows reaching at least the same level of quality of results in terms of damages to structures and facilities even for non-linear studies
  1. To make available to Engineering the selection process of sets of time histories.

Fig6 2 1

Figure 1: [Viallet et al. 2017] Process for comparing
                natural and artificial time histories

Work plan includes 2 main topics:

  1. Development of simple non-linear oscillators in Code-Aster from Humber et al. (2014) and Viallet et al. (2017),
  2. Evaluation of the criteria relevance for the selection or generation of time histories by comparison with ones from Strong Motion Database.

The deliverable is a tool for the assessment of the quality of sets of time histories.


  • Type: Post-doc
  • Collaboration: Pavia University (IUSS), Italy
  • Status: Not started yet