Generation of non-synchronous earthquake signals

Abstract

In this chapter, we describe two procedures to generate earthquake asynchronous signals at different space points for the same seismic event. The foundations of long structures, such as bridges, are placed at distant space points. The earthquake signals at these points have different characteristics and their correct evaluation is important to define design actions. However, design codes around the world do not consider this complex type of action in a consistent manner. The point-to-point signal variation is due both to time lag, since the seismic waves move through the soils with a finite velocity among distant points, and to a change of the signal frequency contents. This depends on physical complex soil-wave interaction phenomena during wave propagation (reflection, refraction, filtering, amplification, etc.). In this chapter, two different generation procedures (PR1 and PR2) to determine the non-synchronous actions at different surface points are shown. Both procedures have been implemented in MATLAB. PR1 generates asynchronous signals at the soil surface. It starts from recorded signals at a few surface points for the same seismic event. PR2 produces asynchronous surface signals by amplifying the bedrock signals obtained by a bedrock propagation process. The inputs for the bedrock propagation are obtained via deconvolution of the recorded surface signals. These latter are also the inputs of the PR1 procedure. Detailed knowledge of soil characteristics is required (soil layers, shear wave velocity profiles, soil density, nonlinear materials shear moduli and damping curves), which relies on in situ tests. Deconvolution and amplification processes are performed by Equivalent-Linear Earthquake Site Response (1D soil model, SHAKE91 (Schnabel et al. 1972) and EERA (Bardet et al. 2000)). PR1 and PR2 are then applied to an example case. Asynchronous surface signals are generated at eight foundation points of a bridge placed in the Aterno Valley near the city of L’Aquila in Italy, where recordings are available at different recording stations (AQA and AQV) for the same earthquake. The EW component of the strong main shock of 4-6-2009 in L’Aquila is selected as input for the two procedures. Finally, the comparison between the signals resulting by PR1 and PR2 and the input signals recorded at the same points is discussed in term of effects on the structures (acceleration response spectrum) and characteristics of the generated signals (Fourier amplitude spectra, coherences for each frequency) to evaluate the differences between the two procedures and between the procedures and the actually recorded signals.