Maximum credible ground motion evaluation based on broadband stochastic finite-fault method: A case study for Baihetan dam in China

Abstract

For the 300-m-high dam, reasonable estimation of the maximum credible earthquake ground motion at the site is crucial to prevent dam failure under near-field strong earthquakes. The primary aim of this study is to develop a broadband ground motion simulation method based on the stochastic finite fault method. This method can capture the low-frequency components of ground motion while maintaining high computational efficiency, thereby rendering it suitable for engineering applications. Firstly, we incorporate a random slip, which follows a k-squared distribution, into the deterministic asperity model to better simulate the fault rupture process. Subsequently, we adopt a dynamic corner frequency model, incorporating a non-uniform stress drop across sub-sources, to refine the theoretical source spectrum. Lastly, a low-frequency filtering function is incorporated into the sub-source amplitude spectrum to mitigate low-frequency simulation inaccuracies arising from uncertainties, such as simplified path effects. The proposed methodology was validated through simulations of the Wenchuan earthquake records, demonstrating its effectiveness in accurately reconstructing the amplitude and spectrum of ground motion at the bedrock site. However, the reasonable estimation of low-frequency components in ground motion hinges on the choice of the prior target spectrum. The proposed simulation methodology was employed to comprehensively assess the maximum credible ground motions at the Baihetan dam site, yielding ground motion scenarios for various risk levels. This research offers a practical and efficient method for estimating ground motion, facilitating seismic safety assessments of high dams subjected to near-field earthquake excitations.