#TITLE_ALTERNATIVE#

• Main Author: ASTUTI (NIM : 12307055), WENY
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/20653
• Institution: Institut Teknologi Bandung

Spectral inversion is a new inversion method that can resolve a thin layer. In this study, the application of spectral inversion will be performed on synthetic data and field data. Some experiments with certain condition are performed in synthetic data test: noise-free data condition, data with added noise, and application of spectral inversion which analytic point is not done in centre of layer. While application of the spectral inversion on field data is performed without control of wells data. The purpose of spectral inversion application in both types of data is to separate the thin layer and estimate the thickness of a thin layer (at the time, seconds). Synthetic data test are performed by Matlab simulation, while field data test is simulated with software ChronoSeis by RockDok. <br /> <br /> <br /> <br /> <br /> <br /> <br /> Spectral inversion is a series of processes performed in the frequency domain where the amplitude of the reflectivity spectrum in the frequency domain is used in the estimation of the thickness T and constant k in cost function equation. The calculation of cost function for noise-free data condition involves an analysis of the Nyquist frequency while for condition of added noise data in a certain level involves frequency range that includes the dominant frequency of seismic trace. Based on the results of data processing, it can be concluded that the estimated thickness for noise-free data can be done without limit to the resolution while the estimated thickness T and k indicates the maximum results when bandwidth is narrowed to the condition of the data that has noise. The more noise in the data the narrower bandwidth used. Thus, spectral inversion gives more accurate results for the data that has high S / N ratio.