Statistical shape refinement and genetic algorithm calibration of design response spectra based on strong-motion...

Area of Science:

• Earthquake Engineering
• Seismology
• Geophysics

Background:

• Seismic design codes worldwide rely on ground-motion parameters and response spectra derived from historical data.
• Existing code-specified spectral shapes exhibit significant deviations from recorded earthquake spectra, limiting their accuracy for site-specific conditions.

Purpose of the Study:

• To develop an improved analytical expression for seismic design spectra.
• To enhance the accuracy of seismic design by better representing site-specific ground-motion characteristics.

Main Methods:

• Compilation and classification of 1227 horizontal acceleration records based on magnitude, source distance, and site class.
• Computation of mean dynamic amplification spectra and derivation of an improved spectral shape.
• Application of a genetic-algorithm-based calibration procedure to optimize model parameters.

Main Results:

• The refined spectral form demonstrates superior accuracy in capturing frequency-dependent characteristics of earthquake response spectra compared to conventional methods.
• Calibration using the proposed framework shows improved spectral parameter estimation.

Conclusions:

• The developed calibration framework provides a valuable reference for advancing seismic design response spectrum research.
• The findings suggest potential for updating seismic design codes to incorporate more accurate spectral representations.