Masterprosjekt vår 2019 ved GEO - Henk Keers

Project description
The quality and quantity of seismic exploration data has increased significantly over the past 10 years. In order to process these data as accurately as possible it is important to use high quality processing algorithms. Full waveform inversion is considered to a crucial part of these seismic processing algorithms. However, efficient 3D full waveform inversion algorithms for complicated elastic media do not yet exist. The focus of this master project is to start work on such an algorithm.

Underlying all waveform inversion algorithms are modelling methods and the speed and accuracy of the modelling methods to a large extent the success of the waveform inversion algorithm. In this thesis an efficient modelling method, elastic ray-Born modelling, will be
compared to a slower, but often more accurate modelling method, finite differences. The ray-Born modelling method will then be used in full waveform inversion.

In this project the student will write, modify and improve computer programs that model elastic wave propagation. This will be done using the ray-Born approximation as well as finite differences. Initially the comparison will be done in 2D. If time allows also some 3D modeling and comparison between the methods will be done. The elastic ray-Born modeling code will then be used to invert a synthetic data sets. Various types of inversion will be investigated, including stochastic and non-linear inversion. If time allows a small 3D inversion will also be done. The inversions will be done in both the frequency and time domain and special attention will be given to find an optimal combination of these inversion methods.

Prerequisites
Bachelor in geophysics (mathematics direction); this project requires a strong
background and interest in mathematics and programming.

Proposed course plan during the master degree 
GEOV219, GEOV277, GEOV355, MAT160, MAT261, MAT234