Fig 1.
(Note: The systematic workflow for processing raw strong motion records, including filtering, baseline correction, and Konno-Ohmachi smoothing, to generate high-quality frequency-domain inputs for the inversion model.).
Fig 2.
Theoretical Decomposition Model.
(Note: Schematic representation of the frequency-domain generalized inversion basis, illustrating how the observed spectrum is mathematically decomposed into source, path, and site terms in the logarithmic domain.).
Fig 3.
(Note: The strategic framework designed to resolve parameter coupling, where Step 1 isolates the path attenuation term using linear inversion, and Step 2 resolves source and site parameters using the nonlinear GCPSO algorithm.).
Fig 4.
Site Effect Consistency (SEC) Mechanism.
(Note: The core objective function design where the CV of site terms is minimized. This constraint ensures the physical rationality of the decoupled site effects by enforcing statistical stability across multiple seismic events.).
Fig 5.
(Note: The iterative process flow of the GCPSO algorithm, detailing the particle position update mechanism enhanced by chaotic perturbation to escape local optima.).
Fig 6.
(Note: The forward simulation workflow that utilizes the inverted source, path, and site parameters to synthesize ground motion time histories, forming a closed-loop validation of the model’s engineering reliability.).
Fig 7.
Comparison of convergence performance of optimization algorithms for each model.
Fig 8.
Comparison of site effect consistency and inversion parameter robustness of various models.
Table 1.
Quantitative comparison of computational efficiency and runtime metrics.
Fig 9.
Seismic motion simulation and goodness-of-fit analysis of each model.
Table 2.
Cross-domain validation results on the KiK-net Dataset.
Table 3.
Ablation experiment setup.
Fig 10.
Ablation experimental results of GCPSO-GIT model components.
Fig 11.
Regional site classification and canonical spectrum results of the GCPSO-GIT model.
Table 4.
Correction results of seismic design response spectrum of GCPSO-GIT model for different site categories.
Table 5.
Probabilistic risk assessment results of seismic ground motion parameters for major engineering sites using the GCPSO-GIT model.
Table 6.
Epistemic uncertainty bounds () propagated to design spectra.