Nonlinear amplitude–frequency characteristics of attenuation in rock under pressure

Abstract

Laboratory experiments have been carried out to investigate the influence of change in strain amplitude on the frequency dependence of attenuation in samples of sandstone, smoky quartz and duralumin. The measurements were performed using the reflection method on pulse frequency of 1 MHz in the strain range ~(0.3–2.0) × 10−6 under a confining pressure of 20 MPa. The attenuation in rocks is nonlinearly dependent on frequency and strain amplitude. In sandstone for P-waves and in smoky quartz for P- and S-waves, the dependences Q−1p(f) and Q−1s(f) have the attenuation peak. With increasing amplitude, the peak frequency can shift towards both the lower and the higher frequencies. It depends on the location of the frequency of an incident (input) pulse with respect to the peak frequency on the frequency axis. For sandstone the peak frequency of P-waves shifts towards the higher frequencies. For smoky quartz the shift of peak frequency is absent in P-waves, and S-waves shift towards the lower frequencies. The attenuation at the incident frequency always monotonically decreases with amplitude, and the other frequency components have complex or monotonic characters depending on the location of the incident frequency in the relaxation spectrum. Q−1p(f) in duralumin has monotonic character, i.e. a relaxation peak in the measurement frequency band is absent. Attenuation strongly decreases with increasing frequency and weakly depends on strain amplitude. The curve Q−1s(f) has an attenuation peak, and its character essentially depends on strain amplitude. With increasing amplitude, the peak frequency shifts towards the lower frequencies. The unusual increase of peak frequency of the P-wave spectrum in the bottom reflection in comparison with peak frequency in spectrum of the initial reflection is detected. The unusual behaviour of attenuation is explained by features of the joint action of viscoelastic and microplastic mechanisms. These results can be used for improving methods of geological interpretation of seismic data.