Transient electromagnetic forward modeling method and system considering induced polarization effect based on deep learning

By using a CNN-LSTM hybrid network model based on deep learning, combined with the Cole-Cole complex resistivity model and the piecewise weighted mean square error loss function, the computational efficiency and accuracy problems of traditional transient electromagnetic methods in simulating excited polarization effects are solved, realizing fast and accurate transient electromagnetic response simulation, which is suitable for the detection of complex geoelectric structures.

CN122065690AActive Publication Date: 2026-05-19CHINA UNIV OF GEOSCIENCES (WUHAN)
4 Cites 0 Cited by

Patent Information

Application Number
CN202610518685.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-04-20
Publication Date
2026-05-19
Estimated Expiration
2046-04-20

AI Technical Summary

Technical Problem

Traditional transient electromagnetic methods struggle to accurately represent the complex resistivity dispersion characteristics when simulating responses with induced polarization effects, resulting in insufficient explanatory power for sign inversion phenomena and high computational costs.

Method used

A CNN-LSTM hybrid network model based on deep learning is adopted. The Cole-Cole complex resistivity model is used to replace the actual resistivity of the formation. The network is trained by combining a piecewise weighted mean square error loss function. The key parameters of the induced polarization effect are explicitly introduced to construct an end-to-end mapping model from underground medium parameters to transient electromagnetic response.

Benefits of technology

It significantly improves computational efficiency, enables fast and accurate transient electromagnetic response simulation, is effectively applicable to complex geoelectric structures, ensures the reliability of response prediction throughout the entire time period, and lowers the computational threshold.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention belongs to the crossing field of artificial intelligence and geophysics, and particularly discloses a transient electromagnetic forward modeling method and system considering an induced polarization effect based on deep learning, and the method comprises the steps: inputting Cole-Cole key parameters representing the induced polarization effect into a CNN-LSTM mixed deep learning network model for forward modeling, and obtaining transient electromagnetic response data; wherein Cole-Cole key parameters representing the induced polarization effect serve as multi-dimensional input features of the CNN-LSTM mixed deep learning network model, transient electromagnetic response data of a corresponding time channel serve as output, and the CNN-LSTM mixed deep learning network model is trained by adopting a segmented weighted mean square error loss function. According to the method, the segmented weighted mean square error loss function is adopted in the training stage, higher weights are given to the symbol inversion points and the prediction errors of the adjacent time windows, and the reliability of all-time response prediction is guaranteed.
Need to check novelty before this filing date? Find Prior Art