A Seismic Porosity Prediction Method Based on Well Log Support Vector Machine Modeling

Abstract

The invention discloses a seismic porosity prediction method based on logging curve support vector machine modeling. The support vector machine learning method is used to establish a nonlinear mapping model of uphole elastic parameters, gamma and porosity. It is transformed into a data-driven statistical learning problem, which avoids complex petrophysical modeling and parameter adjustment processes; secondly, the gamma curve of the reservoir division function is involved in statistical modeling, which makes the porosity prediction results interpretable and reliable. Rationality; finally, the support vector machine learning method is suitable for training sample data with small samples and low feature dimensions. The support vector machine can achieve nonlinear mapping through the kernel function, and the sequence minimum optimization algorithm with high computational efficiency is used to solve the objective function. The purpose is to solve the technical problem that the prediction of seismic porosity in the prior art relies on a probability model and has low applicability.

Description

technical field [0001] The invention relates to the technical field of oil and gas exploration, in particular to a seismic porosity prediction method based on logging curve support vector machine modeling. Background technique [0002] In oil and gas geophysics research, porosity is an important parameter to describe the characteristics of oil and gas reservoirs, and it is of great significance for reservoir prediction. [0003] The main methods of seismic porosity inversion are linear regression modeling prediction, geostatistical inversion and stochastic simulation. Among them, linear regression modeling prediction method is the most commonly used seismic porosity calculation method. The porosity curve on the well and the elastic parameter curve establish a linear regression equation, and the linear regression model is applied to the same elastic parameter attribute of the earthquake to predict the seismic porosity. This method has a faster calculation speed, but there is ...

AI Technical Summary

Problems solved by technology

[0005] The main purpose of the present invention is to provide a seismic porosity prediction method based on logging curve support vector machine modeling, aiming to solve the technical problem that the seismic porosity prediction depends on the probability model and the applicability is not high in the prior art

Images