4. Data Processing/Signal Processing
Prestack Depth Imaging for Seismic Reflection Data by MPI_PSPI
Seonghyung Jang, Sangyong Suh and Hyuntae Kim
Korea Institute of Geoscience and Mineral Resources, S. Korea.
Abstract
Prestack depth migration for seismic reflection data is currently one of the powerful tools for imaging complex geological structures such as salt dom. Subsurface depth imaging to huge seismic reflection data requires high performance computing power and techniques. Since PSPI (Phase Shift Plus Interpolation) is intrinsically data parallel, its simplicity, less computing efforts and efficiency, PSPI method for subsurface imaging is widely used in reflection exploration. Here we implemented parallelized PSPI for the prestack depth migration using MPI (Message Passing Interface) LAM. MPI is a library for high performance parallel programs that pass messages between processes to make up an overall parallel job. We perform MPI_PSPI to a synthetic geological modeled data and gas hydrate field data set of East Sea. Before imaging the field data set, MPI_PSPI was tested through applications to the synthetic model with 21 nodes of Linux clustering machine. The imaging results from a simple disk velocity model shows good subsurface depth image with less computing time. For the imaging gas hydrate data set, we made initial velocity model from stack image which was processed by in-house-processing tool. The computing time for a single image gather is 553 s when the velocity model size is 28 km x 4 km and grid size is 6.25 m. The result with 1001 shot gathers give us acceptable and well-imaged subsurface. Since the depth image is sensitive to the velocity model, if we get more accurate velocity model we could get more accurate depth image.
Last modified: Thu May 11 17:33:01 2006