r4.05.05 Seismic signal generation#
Summary:
In « best-estimate » seismic studies, transient dynamic analyses are often required. The seismic signal is then modelled by a stochastic process. This process expresses acceleration on the ground as a function of time.
Temporal signals (accelerograms), which are achievements of the stochastic process, are called trajectories. These signals can come from databases (ground accelerations measured during earthquakes) or be obtained by numerical simulation (cf. also [bib17] )) *. The simulation of artificial seismic signals, as carried out by the operator GENE_ACCE_SEISME, is described in this document. The operator proposes two models:
model of DSP by Kanai-Tajimi (§2)
DSP model compatible with a target response spectrum (§3)
In both cases, it is possible to introduce a change in frequency content and amplitude (modulation) with the duration of the earthquake.
- 1. Modeling seismic excitation by a stochastic process
- 2. Evolutionary Kanai-Tajimi DSP model
- 3. DSP template to generate « compatible » signals with a SRO target
- 3.1. Identifying a DSP compatible with a target spectrum
- 3.2. Non-stationary parametric model for generating spectrum-compatible signals
- 3.3. Introduction of natural signal variability (record-to-record variability)
- 3.4. Simulation of seismic signals compatible with a SRO: functionalities implemented in GENE_ACCE_SEISME
- 3.5. Model of DSP by Kanai-Tajimi in accordance with a SRO target
- 4. Vector process cases
- 5. Simulation of process trajectories
- 6. References
- 7. Description of document versions
- 8. Appendix