A quality factor inversion method of Markov chain Monte Carlo integrating well-log and seismic data

ABSTRACT The Q-value quantitatively characterized the extent of seismic wave absorption in rocks, aided in the compensation of seismic data, and supported the detection of oil and gas properties. Conventional Q-value extraction methods are typically suitable for vertical seismic profiling (VSP). However, when applied to reflection seismic data, the accuracy of interval-Q extraction was relatively low. Concurrently, the extraction accuracy was severely compromised by interference from reflected seismic waves, making it difficult to accurately separate tuning waves. Therefore, a Q-value inversion method integrating well logs and seismic data was proposed. This method utilized logging data to synthesize unattenuated seismic records and formulated an objective function between the actual attenuated and synthetic unattenuated data using nonstationary equations. Owing to the strong nonlinearity of the objective function, which conventional inversion methods failed to address effectively, the Markov chain Monte Carlo (MCMC) algorithm was employed to solve it. After obtaining the Q-value at the well location, an interpolation objective function incorporating dip angle constraints was developed, ultimately yielding a complete Q-value volume through inversion. Additionally, the effectiveness of the method was validated using two complex models, and it was successfully applied to field seismic data.