Akademik Veri Yönetim Sistemi
ADVANCES IN MODELING AND INTERPRETATION IN NEAR SURFACE GEOPHYSICS, ss.155-188, 2020 (SCI-Expanded)
In this study, global optimizations of the data sets obtained from potential field method applications including self-potential (SP), magnetic and gravity have been presented through particle swarm optimization (PSO), differential evolution algorithm (DEA) and differential search algorithm (DSA), respectively. Both synthetically produced and real field anomalies due to various kinds of geological sources (e.g. sheet, horizontal cylinder, sphere and fault) have been used to show the capability of those population-based metaheuristic algorithms. Firstly, error energy maps have been produced for each model parameter pairs in every synthetic example to reveal the mathematical nature of the inverse problems under consideration. These maps have clearly helped us assess the resolvability of the model parameters for the given problem. Metropolis-Hastings (M-H) sampling algorithm has been used to perform uncertainty analyses. Produced histograms of synthetic data cases have provide insight to the reliability of the estimated parameters. Additionally, the reliabilities of the solutions have been also tested via probability density function (PDF) applications. After performing successful synthetic studies on model parameter estimations, three real data examples including an SP anomaly from a sulphide mineralization zone (India), a total field magnetic anomaly from a fold belt (Australia) and a gravity anomaly from a chromite deposit (Cuba) have been inverted through PSO, DEA and DSA, respectively. Satisfactory solutions in well agreement with the results of previous studies have been obtained.