Rapid forward of gravity and tensor gravity data

TL;DR

This paper introduces a novel fast forward method for large-scale 3D gravity and tensor gravity data inversion, significantly reducing memory usage and computation time while maintaining high accuracy.

Contribution

It proposes an innovative equivalent geometric architecture and a translation equivalent technique based on Toeplitz matrices for efficient gravity data modeling.

Findings

Requires less memory compared to traditional methods

Achieves high computational efficiency

Maintains high precision in large-scale models

Abstract

The large-scale three-dimensional inversion of surface gravity / tensor gravity data is a very challenging numerical and practical problem, which is a highly physical memory usage, time-consuming computation and high precision for large-scale gravity models. Equivalent geometric architecture was introduced to avoid to calculate and to save sensitivity matrix. A new equivalent geometric architecture was not rely on symmetrical characteristic of gravity field by added three virtual grids into geophysical model. A new fast forward method was proposed based translation equivalent technique based toeplitz matrix, which a toeplitz matrix carried out fast matrix-vector multiplication by using the fast Fourier transform. Numerical experiments show that the method proposed in this paper is require little memory, high efficiency and high precision.