Stochastic inversion of elastic impedance based on simulated annealing genetic algorithm

Chanjuan Liu,Xingyao Yin,Baoli Wang
DOI: https://doi.org/10.1190/segam2016-13861277.1
2016-01-01
SEG Technical Program Expanded Abstracts
Abstract:The aim of seismic inversion is to estimate the subsurface elastic properties. As one of the many inversion methods, stochastic inversion is becoming more and more attractive because of its high vertical resolution. It can make full use of the well logging data which contains the high frequency, and be constrained by the seismic data at the meanwhile. Besides, the choice of inversion algorithm is another primary element in inversion problems. Here, we integrate the simulated annealing (SA) algorithm with the genetic algorithm (GA) to solve the problem. It can avoid the conventional problems that GA and SA suffer and get the optimum as soon as possible. In this paper, we first introduce the basis of elastic impedance (EI), and use it as our inversion strategy. Subsequently, the Fast Fourier Transform_Moving Average (FFT_MA) algorithm has been explained to obtain the simulated information. Finally, the initial simulated results are optimized using the simulated annealing genetic algorithm until we obtain the solutions that agree with the geological structure underground. The inversion result accords with the model well, which demonstrates the effectiveness of our method. Presentation Date: Wednesday, October 19, 2016 Start Time: 8:25:00 AM Location: 156 Presentation Type: ORAL
What problem does this paper attempt to address?