DEVELOPMENT AND ANALYSIS OF A 2D DETERMINISTIC INVERSION PROGRAM OF LEAST-SQUARE METHOD WITH THE ADDITION OF WEIGHTING FUNCTION ON MAGNETIC ANOMALIES (CASE STUDY: PANDAN MOUNTAIN AREA, EAST JAVA)

• Main Author: Widangma Naige, Gerardus
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/80357
• Institution: Institut Teknologi Bandung

The value of susceptibility as one of the physical properties of rocks expresses the degree or degree of magnetization of rocks due to the influence of magnetic fields. The value of susceptibility can be measured using geomagnetic methods, especially at the inversion stage. The challenge of the geomagnetic inversion process is the limitation of the amount of data and complex subsurface models. To overcome these challenges, we developed a forward modeling and inverse modeling calculation program based on the 2D deterministic inversion of the leastsquares method. The application of the weighting function aims to obtain more optimal inversion results. The program that has been built is tested on synthetic models in the form of mineral chunks, intrusions, and rock layers represented by a collection of rectangular prism blocks. Furthermore, the forward modeling calculations made on Matlab R2023a were validated by comparing them to calculations by GravMagInv2D software, which showed conformity with an average correlation of 98.89% and an average calculation difference of 0,72 nT for each measurement point. At the inversion stage, we compare the Least-square inversion method with the weighting and inversion functions of the Under-Determined SVD (Singular Value Decomposition) method to see the impact of weighting on the success rate of inversion results. The comparison process is also carried out at inclination angles of 90°, 60°, and 30° in order to see the effect of the inclination value on the response of the model. Then we modify the weighting function, which includes depth, compactness, Kernel, and noise weighting, as well as the compactness value and damping value, to get better results. The calculation results show that the Least-square inversion program with a weighting function can inverse model synthetic data quite well, especially at an inclination value of 90°, which shows an anomalous response in the form of a monopole. From the research that has been done, we also concluded that the weighting functions that are ideal enough to be applied to various cases are depth weighting, compactness weighting, and noise weighting. To strengthen confidence in the results of previous studies, we also apply a program that has been prepared to regional magnetic anomaly data in the Gunung Pandan area, East Java. In general, the study showed quite optimal results in estimating the 2D structure below the surface in the form of intrusion with an average susceptibility of 0,2626 SI at a depth of ± 100 meters and a layer of sediment on the surface with an average susceptibility of 0,0034 SI.