High Precision and Fast Inversion of Array Lateral Logging Based on BAS-LM Algorithm
Year: 2024
Type: article
Abstract: The high-resolution array lateral is affected by environmental factors, and the measured apparent resistivity often cannot reflect the true resistivity information of the reservoir, which needs to be inverted. To solve the problem that Levenberg-Marquarat (LM) algorithm used in geophysical inversion is highly dependent on the initial value and easily trapped in a local minimum, a strategy for optimizing the initial value selection of the basic LM algorithm using Beetle Antennae Search (BAS) algorithm is proposed to enhance the global optimization ability of the LM algorithm. Based on the Oklahoma classic reservoir model, the Particle Swarm Optimization (PSO) algorithm, the Genetic Algorithm (GA), BAS algorithm, LM algorithm, and BAS-LM algorithm are used to perform the inversion of the reservoir layer resistivity, intrusion zone radius, and intrusion zone resistivity (three parameters) and compare and verify the results. The research results show that the BAS-LM inversion algorithm has good adaptability to complex reservoir models, and the three-parameter inversion errors can be controlled below 10%, effectively recovering the mud invasion profile information; after optimizing the initial value by the BAS algorithm 10 times, it can meet the inversion requirements of the LM algorithm to a certain extent, and to some extent overcome the problems of low inversion accuracy of the BAS algorithm and high dependence of the LM algorithm on the initial value; compared with classic intelligent optimization algorithms such as PSO and GA, the BAS-LM algorithm has a faster convergence rate and inversion time, which is only 1/6 to 1/11 of the classic algorithms. This method is of great significance to the field real-time inversion of HRLA.
Authors: NI Xiaowei, HUAN Zhipeng, XU Sihui, ZHAO Zhaoxia, CHEN Zhanglong, CHEN Jisheng
Language: English
Subfield: Geophysics
Field: Earth and Planetary Sciences
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W6887478563
FWCI:
Cites:
Cited by:
Related to: