#TITLE_ALTERNATIVE#

• Main Author: (NIM : 223 08 001), RAHADIAN
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/23780
• Institution: Institut Teknologi Bandung

An important task in seismic data processing is the identification and subsequent suppression of multiple reflections. One of type multiples that should be suppressed on the seismic data processing is surface multiples. Surface multiples, or more accurately, surface-related multiples, are events that have at least one downward reflection initiated at the water surface. In this research, we have been done the seismic data processing to suppress surface-related multiples at 2D marine seismic data in line BGR06-141, a survey result from SO-186 in north-west of Simelue Island, North Sumatera. <br /> <br /> <br /> Many techniques exist to suppress multiple reflections. One of these methods is Surface Related Multiple Elimination (SRME). The procedure is based on the concept that all surface multiple events are composed of convolved primary events (Verschuur et al., 1992). It is implemented in two steps: prediction of multiples via extended auto convolution and adaptive subtraction of predicted multiples from the data (Verschuur et al., 1992). To get a good result from multiple predictions, therefore SRME needs some requirements, as data should be noise free and should have regularized offset of shots and receivers. <br /> <br /> <br /> The implemented of SRME shows that this method adequately effective to suppress the existing of surface-related multiples. SRME’s method is more effective to remove surface-related multiple in the near offset until middle offset. In the far offset, there are still residual multiples which probably caused by geometry offset on this part is not as regularized as geometry offset on the near and middle offset. The feather angle caused distinctive of these. By applying Radon filter after SRME’s method, therefore nearly all surface-related multiples which appear on the stack data can be removed.