Analysis of leakage seismic attribute characteristics considering the discrete random properties of earth-rock dam media

Abstract

Leakage is a prevalent latent hazard in earth-rock dams. Accurate detection of the location and severity of leakage areas can provide the basis for treating leakage and reducing the damage to the dam. The key to detecting leakage using seismic wave methods is mastering the response characteristics of leakage reflection wavefields and accurately identifying them from the acquired seismic data. In this study, the dam models with different water content leakage areas are established using discrete random modeling method, based on the soil-rock mixture characteristics of the actual dam media. The effects of water content changes on the leakage reflected wavefields are elaborated, and the seismic attributes that can effectively identify leakage wavefields are summarized. Scattering caused by the heterogeneity of the earth-rock dam reduces the energy of leakage reflected waves and complicates the identification of the leakage wavefields. The characteristics of leakage converted waves, primarily low-frequency signals, are more advantageous for identifying leakage wavefields. With the decrease of water content in the leakage area, the energy of converted waves gradually weakens, and the continuity of the phase axis deteriorates. Wavelet transform can enhance the continuity of the leakage reflected phase axis and effectively suppress random noise interference. The combined use of instantaneous frequency and wavelet transform frequency slice allows for more accurate identification of leakage-reflected wavefields.