Residual Neural Network for Transient Electromagnetic Probing Depth Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for calculating the probing depth in transient electromagnetic (TEM) geophysical detection are uncertain due to the lack of direct and precise algorithms, with existing methods either neglecting uneven resistivity structures and sampling points or relying on inversion models with multiple solutions and parameter dependencies.
Innovation Solution
A method using a residual neural network to predict the probing depth directly from observation data, specifically by simulating the transient electromagnetic field with one-dimensional underground resistivity models and establishing a rapid mapping between observation data and probing depth using the neural network.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If the first method (assuming plane electromagnetic wave incidence under uniform half space) is used to calculate probing depth, then the calculation process is simple, but the result has great uncertainty because it does not fully consider uneven resistivity structure and sampling points
Solution Approach 1:
The patent transforms the probing depth calculation from traditional physical parameter-based methods to a data-driven approach using neural networks. The system trains a neural network model on synthetic TEM data with various resistivity structures, sampling configurations, and noise levels. During inference, the model directly predicts probing depth from observed TEM responses, automatically adapting to different geological conditions without requiring explicit parameter adjustments or assumptions about uniform half-space conditions.
2Reliability
If the second method (calculating probing depth based on Jacobian matrix of inversion model) is used, then it can be applied to any model and considers noise and sampling points, but the result has certain errors because inversion has multiple solutions and depends on many parameters
Solution Approach 1:
The patent introduces a neural network model as an intermediary between TEM observation data and probing depth prediction. Instead of directly calculating probing depth from inversion models or using simplified analytical formulas, the system uses the trained neural network as a mediator that has learned the complex mapping relationship from synthetic training data. This intermediary approach avoids the multiple solutions problem of traditional inversion while maintaining applicability to various geological models.
3Adaptability or versatility
If traditional methods are used to obtain probing depth, then the process follows conventional workflows, but the calculation efficiency is low and inversion parameter settings significantly affect the results
Solution Approach 1:
The patent performs preliminary action by pre-training a neural network model on extensive synthetic TEM data covering a wide range of geological conditions, sampling configurations, and noise levels before actual field data analysis. This pre-computed knowledge is stored in the neural network's weights and biases, allowing rapid probing depth prediction during field surveys without requiring time-consuming iterative inversions or repeated analytical calculations for each new dataset.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach allows for direct and precise prediction of probing depth from observation data, reducing the influence of inversion parameter settings and improving calculation efficiency, while also providing uncertainty estimates for comprehensive geological and geophysical data interpretation.
Implementation Method 1
When it works, it introduces step current and transmits pulsed electromagnetic field. In the process of underground diffusion, a pulsed electromagnetic field will excite induced current in the electrically inhomogeneous body, also known as eddy current; by observing the primary diffusion field and the secondary electromagnetic field generated by the eddy current inside the anomalous body, the underground electrical structure can be inversion obtained.
Data Source
AI summary
The present disclosure belongs to the field of geophysical exploration, and discloses a method, a system, a medium, a device and a terminal for predicting the probing depth of transient electromagnetic. Based on the existing published resistivity model database, the transient electromagnetic field in layered medium is calculated, and the probing depth is calculated based on Jacobian matrix to establish a training data set. The simulated induced electromotive force is used as the input of the neural network, and the calculated probing depth is used as the output of the network. A rapid mapping between observation data and probing depth is established by using residual neural network.