基于随机分布共网格模型的重磁电震联合反演技术及应用

Joint inversion of gravity,magnetotelluric and seismic data based on common gridded model with random distributions

针对重力、大地电磁与地震资料联合反演存在的精度和稳定性问题,应用以下技术:1基于统一物性随机分布共网格模型的建模技术;2通过改进地震走时的二维射线追踪计算方法以适用于复杂地质模型;3引入Tikhonov正则化思想至反演中,并加入先验信息进行模型约束,实现重、磁、电、震同步联合反演;4通过模型试验的对比分析对关键参数进行优化;5采用完全非线性的全局寻优模拟退火算法,以提高解决复杂问题的能力。实际应用效果表明,加入有效模型约束的正则化联合反演,可以提高反演的稳定性和计算精度,获得更接近实际的结果。

In view of the inaccuracy and instability in conventional joint inversion of gravity,magnetotelluric,and seismic data,we propose in this paper a new joint inversion approach based on common gridded model with random distributions.This approach consists of the following techniques:A.Model building based on common gridded modelwith random distributions;B.2dimensional travel-time calculation with improved ray-tracing for complex geological models;C.Introducing the regularization conception into inversion and using priori information to constraint the model;D.Key parameter optimization with model tests;E.Nonlinear global optimization simulated annealing algorithm.Application examples show that the proposed joint inversion improves not only the inversion stability but also calculation accuracy to help us to obtain better results.