半航空瞬变电磁自适应正则化-阻尼最小二乘算法研究

反演问题一直是地球物理的重点研究问题之一。本文以长导线源半航空瞬变电磁为例,对时间域半航空数据进行一维反演研究。将自适应正则化(ARIA)反演方法和阻尼最小二乘(LM)反演方法相结合,提出一种改进的反演方法:自适应正则化-阻尼最小二乘(ARIA-LM)反演算法。本文选用三层理论模型对比阻尼最小二乘算法、自适应正则化反演算法与本文方法,并使用四层理论地层模型来验证算法的有效性和可靠性。结果表明:该算法在一定程度上对高阻分辨率增强,能有效提高反演的精度及稳定性。

Study on semi-aerospace transient electromagnetic adaptive regularization-damped least squares algorithm

Inversion has always been one of focused research issues in geophysics. This paper studies one-dimensional inversion of time domain semi-aerospace data by taking the long-wire source semi-aerospace transient electromagnetic as an example. Combining the adaptive regularization inversion method and the damping least squares inversion method, an improved inversion method is proposed: adaptive regularization-damping least squares inversion algorithm. This work chooses a three-layer theoretical model to compare the damping least squares algorithm, adaptive regularization inversion algorithm and the method proposed in this paper, and employs a four-layer model to test the effectiveness and robustness of our algorithm. The results show that the adaptive regularization-damping least-squares algorithm can enhance the resolution of high resistance, which can effectively improve the accuracy and stability of the inversion to some extent.