Multiscale geologic feature detection fusing method based on deep learning and evolutionary learning

Abstract

The invention discloses a multiscale geologic feature detection fusing method based on deep learning and evolutionary learning, and relates to the technical field of geologic feature detection and fusion. The method comprises the following steps: S1, outputting a reservoir assessment rating parameter; and S1. 01, generating a monomer and multimer combined feature data according to the multidata volume, earthquake data and well logging data, and marking the feature data. Through adoption of the multiscale geologic feature detection fusing method based on deep learning and evolutionary learning,respective learning methods and data structures are applied specific to different data; different evolution and learning are realized according to different data; prediction and assessment can be implemented by different models specific to different geological conditions; self-correction, self-improvement and parameter assessment of the model can be realized in the learning process; no constant model and mode are included in the learning process; the finally-generated reservoir parameters have geological significance; and the prediction on the reservoir parameter is closer to the practical condition.

Description

technical field [0001] The invention relates to the technical field of geological feature detection and fusion, in particular to a multi-scale geological feature detection and fusion method based on deep learning and evolutionary learning. Background technique [0002] In recent years, some companies and research institutions at home and abroad have made some new progress in the research of reservoir description. In 2017, Cao Junxing and others patented a reservoir detection method based on deep learning of seismic data; in 2016, Cheng Guojian et al. applied deep learning algorithms to study rock image processing; in 2016, Duan Youxiang et al. applied convolutional neural networks to predict geological reservoir parameters; 2018 In 2017, Zheng Yuzhe applied deep learning to study reservoir physical parameters; in 2017, Smith et al. applied deep learning to data fusion research, etc. [0003] Reservoir description technology is a technology applied to seismic data interpreta...

AI Technical Summary

Problems solved by technology

[0005] Aiming at the deficiencies of the existing technology, the present invention provides a multi-scale geological feature detection and fusion method based on deep learning and evolutionary learning, which solves the problem that the current analysis method based on logging data and seismic data inversion cannot fully and accurately identify reservoirs Porosity, fluid composition and oil and gas content, and the current application research is mainly to apply the existing technology of deep learning to the oil field for reservoir prediction, and there is no in-depth research on the essence of intelligent learning.

Images