INVERSI NON-LINIER MAGNETOTELURIK DUA-DIMENSI MENGGUNAKAN ALGORITMA RANTAI MARKOV

• Main Author: Dwi Hananto, Nugroho
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/5131
• Institution: Institut Teknologi Bandung

<p>Abstract:<p align="justify"> <br /> A formulation of non-Tinier inverse problem using Markov chain is given for the interpretation of magnetotelluric two-dimensional data. Solution to the inverse problem explained quantitatively in the Bayesian manner as the posterior probability density function (PDF) of given a priori model. This problem is than solved using transitional probability of non-homogenous Markov chain, which convergent to the marginal a posteriori PDF. This non-linear inversion algorithm is believed to have an ability to overcome the local minima entrapment, which occasionally occurs in the linearized inversion.<p align="justify"> <br /> <br /> Synthetic and actual field data of two-dimensional magnetotelluric are used both to demonstrate the inverse methods and to illustrate certain properties of the inversion methods. Resolution of the inversion method to synthetic joint TE and TM data gives satisfactory results, although inversion on TE modes or TM modes separately gives less focused image. Inversion of actual field data from 19 sites AMT (Audio-frequency Magxetotelluric) measurements on the south Bandung (West Java, Indonesia) volcanic area results subsurface resistivity images that agree well with previous investigations.