Reservoir parameter prediction method and system based on Bayesian classification

Abstract

The invention discloses a reservoir parameter prediction method and system based on Bayesian classification. The method comprises the following steps of: establishing a first inversion equation of reservoir physical property parameters based on known reservoir elastic parameters; based on the Bayesian formula, transforming the first inversion equation into the second inversion equation about the likelihood function of the reservoir elastic parameter; based on a Bayesian-sequential Gaussian method, solving the likelihood function of reservoir elastic parameters; obtaining reservoir physical parameters by solving the second inversion equation based on the likelihood function of the obtained reservoir elastic parameters. The invention does not need to carry out complex model initialization, fully considers the advantage of stochastic characteristics of geological and geophysical characteristics, and makes the inversion results have more practical geological significance, and it can also solve the problem that the variogram function is not reliable in the case of less well data, or even in the case of uneven distribution of wells in the work area in the geostatistical stochastic simulation method.

Description

technical field [0001] The invention belongs to the field of geophysical exploration in the petrochemical industry, and more specifically relates to a method and system for predicting reservoir parameters based on Bayesian classification. Background technique [0002] At present, the two most commonly used inversion methods of reservoir physical parameters based on conventional P-wave reflection seismic data in China are nonlinear inversion of reservoir physical parameters based on neural network and stochastic simulation method based on geostatistics. Both methods belong to the category of nonlinear inversion. The probabilistic inversion method of reservoir physical parameters based on Bayesian theory does not require complex model initialization, but establishes the probability and statistical relationship between reservoir physical parameters and reservoir elastic parameters based on the reservoir physical model, and then integrates with the stacked The combination of pr...

AI Technical Summary

Problems solved by technology

However, the current probabilistic inversion method of reservoir physical parameters based on Bayesian theory is difficult to obtain the likelihood function accurately during the calculation process. Therefore, it is necessary to develop a reservoir parameter prediction method based on Bayesian classification, which can obtain Stable and accurate likelihood function

Images