An analysis of the distribution of time delays on simultaneous source separation

Abstract

Simultaneous source acquisition has been recognized as an important way of improving the efficiency and quality of seismic data acquisition. Recently, several methods have been developed for separating simultaneous sources and to provide data that can be utilized in conventional processing streams. The key is to introduce randomness in time delays among simultaneously fired shots to make interferences appear incoherent in common receiver, common offset and common midpoint gathers. In this paper, we study the separability of simultaneous source data based on different distributions of fire time delays. We conduct Monte Carlo tests to analyze the relationship between different firing schemes and the quality of the separation via the iterative rank reduction (IRR) deblending method. We also adopt the fast simulated annealing (FSA) method to estimate an optimal empirical firing scheme by assuming a priori knowledge of the unblended data. Insights can be gained from these tests towards optimal acquisition design for simultaneous source acquisition.