Multiple-linear regression to best-estimate of gravity parameters related to simple geometrical shaped structures

Abstract

A new interpretative approach is proposed to best-estimate of gravity parameters related to simple geometrical shaped structures such as a semi-infinite vertical cylinder, an infinite horizontal cylinder, and a sphere like structures. The proposed technique is based on the multiple-linear regression oriented towards estimating the model parameters, e.g., the depth from the surface to the center of the buried structure (sphere or infinite horizontal cylinder) or the depth from the surface to the top of the buried object (semi-infinite vertical cylinder), the amplitude coefficient, and the horizontal location from residual gravity anomaly profile. The validity of the proposed approach is firstly demonstrated through testing different synthetic data set corrupted and contaminated by a white Gaussian random noise level. The theoretical synthetic obtained results obviously show that the estimated parameters values, derived by the proposed technique are close to the assumed true parameters values. This approach is applied on five real field residual gravity anomalies taken from Cuba, Sweden, Iran, USA, and Germany, where the efficacy of this new approach is consequently proven. A comparable and acceptable agreement is noticed between the results derived by this proposed approach and those obtained from the real field data information.

Gravity profiles interpretation applying a metaheuristic particle optimization algorithm of mineralized bodies resembled by finite elements

“…Gravity anomalies are frequently characterized by variations in rock density, and the inversion of this data is crucial for determining the mass properties and depths of the subsurface. Over the years, numerous geophysical methods for solving inverse problems have been developed, including robust simultaneous joint inversion [ 1 ], fair-function minimization [ 2 ], the s-curves method [ 3 ], non-linear least-squares [ 4 , 5 ], derivative-based approaches [ 6 , 7 ], moving average methods [ 8 , 9 ], multiple-linear regression [ 10 ], R-parameter imaging [ 11 ], and the Lanczos bidiagonalization method [ 12 ]. Metaheuristic optimization algorithms, such as genetic algorithms [ 13 ], particle swarm optimization [ [14] , [15] , [16] , [17] ], simulated annealing [ 18 ], cuckoo optimization algorithm [ 19 ], and ant colony algorithm [ 20 ], have also been used recently for gravity data inversion.…”