Combining 1C & 3C Sensors: An Evolution in Passive LFS Low Frequency Seismic Exploration

Abstract

Passive low frequency seismic exploration is a technology for specifying the geological attributes of the earth's interior. Known examples of its application for tasks, for example, determining of the contour of hydrocarbon deposits (direct hydrocarbon indicator DHI) [2, 4, 8, 14, 18, 21, 23, 26, 35, 36, 37]. However, most of the results are qualitative. In this paper, we propose a physic-mathematical method for interpreting this phenomenon, based on the recognized approaches of extracting the Green function from passive observations [27, 45, 46] and methods for solving the partial problem of full-wave inversion [16]. It is known that the correlation function of observations of a random field carries information about the Green function, which is a response to the structure of the geological setting [15]. We additional filter out Rayleigh surface waves and inclined waves using linear prediction algorithms [10]. It is assumed that the properties of the medium are known; it is necessary to determine the location of a small-thickness formation with anomalous absorption. Hydrocarbon deposits with high permeability are recognized with high absorption. Abnormal absorption at low frequencies occurs as a result of friction of the liquid phase of hydrocarbons on the skeleton of the rock during the passage of longitudinal waves [7, 10]. Contrast of anomalous absorptions can cause additional reflections (resonance). The task is reduced to a one-dimensional partial task of the full-wave inversion, when the base model is generally known, and it is necessary to find individual deviations (oil-saturated reservoirs). A numerical simulation is performed with different possible placements of a search object. Field record spectra, after applying filtering procedures, are matched with synthetic ones. We solve this problem in the Born approximation of single scattering with regularization in the form of restrictions on the coefficients, this problem is solved by quadratic programming. The proposed method has been applied to solve problems of hydrocarbon prospecting [41]. Combining 1C (single component) sensors and 3C (three components) ones is considered as an evolution to solve problems concerning scarcity of passive LFS data due to our long-time reliance on 3C sensors only to collect such valuable data. Quantitative approaches will merge soon!