基于改进粒子群算法的VTI介质裂隙预测

Prediction of fractures in VTI media based on the improved particle swarm optimization algorithm

裂隙产生的各向异性在地层介质中广泛存在,基于各向异性的裂隙参数反演和预测在一定程度上能提高裂隙反演的精度和预测的可靠性。本文通过建立基于VTI介质的反射系数方程,引入模拟退火算法中的Metropolis准则设置跳出概率对粒子群标准算法进行改进,实现对VTI介质中纵、横波速度及各向异性参数等属性的反演;并通过各向异性相关的属性、泊松比和泊松速度的联合,实现对裂隙充填物的预测。通过二层模型和Marmousi2模型对算法的稳定性和抗噪性进行测试,论证了方法的可行性;并进一步将方法应用于煤矿实际资料,对裂隙含水性进行预测,论证了方法的有效性。

The anisotropy caused by fractures is ubiquitous in formation media.The inversion and prediction of fracture parameters based on anisotropy can somewhat improve the inversion accuracy and prediction reliability of fractures.This study established a reflection coefficient equation based on vertical transverse isotropy(VTI)media.Then,it improved the standard particle swarm optimization algorithm by setting the exit probability based on the Metropolis criterion of the simulated annealing algorithm.Consequently,it obtained the inversion results of compressional-and shear-wave velocities and anisotropy parameters in VTI media.By combining anisotropy-related attributes,Poisson's ratio,and Poisson's velocity,this study predicted the fillers in fractures.The improved algorithm was tested for stability and noise resistance using a two-layer model and a Marmousi2 model,demonstrating its feasibility.Furthermore,the improved algorithm was applied to predict the water-bearing property of fractures using real coal mine data,validating its effectiveness.