Managing Near-Simultaneous Uncompressed Crosstalk Interference on Land Seismic Data for Optimum Time-Frequency De-noising
Pecholcs, Peter I.*1; Huo, Shoudong 1; Wang, Xin 2
(1) Exploration Operations, Saudi Aramco, Dhahran, Saudi Arabia. (2) Earth Science & Engineering, Kaust, Thuwal, Saudi Arabia.
With modern vibroseis source controllers and high-channel count recording systems, it is now possible to acquire very high source density (1000 VPs/km2 or greater) 3D surveys in an equivalent amount of time as more conventional (400 VPs/km2) surveys. Multiple vibroseis fleets are distributed over a receiver super-spread (~20 km x ~6 km) and operate in a near-simultaneous sweep mode, with or without time and distance rules. When using near-simultaneous source recording, the success of random noise attenuation algorithms depends on the distribution of the inherent crosstalk interference. This distribution is now managed at the acquisition design stage to optimize crosstalk noise attenuation methods.
In 2010, Saudi Aramco acquired their first 3D Independent Simultaneous Sweep (ISS™) seismic test using 18 point vibrator fleets, distributed equally over 6 km x 3 km. In this test, the vibrators operated without time and distance rules. The only rules used were for the vibroseis starting positions and unique sweep definitions per fleet. Although the crosstalk interference was not managed directly, the distribution of different sweep lengths (over the same frequency band) naturally introduced time-varying source cycle time differences between neighboring fleets. This variation guaranteed a sparse distribution of crosstalk noise in different sorted domains. An additional level of sparseness was achieved when the blended up-going and down-going uncompressed crosstalk were decomposed into frequency banded time sections. When the sparse crosstalk interference condition is met, simple time-frequency de-noising filters can be applied within a time-offset sliding window, or a time-frequency banded section, to attenuate the crosstalk.
In 2011, a second test was performed where two cross-spreads were acquired using flip-flop and near-simultaneous recording modes utilizing five fleets of vibrators (two vibrators per fleet). This source recording method used new time, distance, and sweep length rules based on the lessons learned from the 2010 field test. We present a comparison between different time-frequency de-noising filters for the same sweep, and then for different sweeps in different sorted domains. All comparisons are referenced to an interference-free flip-flop cross-spread acquired during the test. We conclude with recommendations for future near-simultaneous vibroseis survey design and de-noising filters.
AAPG Search and Discovery Article #90141©2012, GEO-2012, 10th Middle East Geosciences Conference and Exhibition, 4-7 March 2012, Manama, Bahrain