An ore prospecting method for predicting migration of ore-forming fluid based on an AI model
By using AI models to predict the migration of ore-forming fluids, the problems of time-consuming, labor-intensive, and error-prone traditional methods have been solved. This has enabled efficient and accurate prediction of the migration paths of ore-forming fluids and the locations of mineral sedimentation, thereby improving the efficiency and accuracy of mineral resource exploration.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-16
- Publication Date
- 2026-03-17
AI Technical Summary
Traditional methods are time-consuming and labor-intensive in predicting the migration of ore-forming fluids and determining the location of mineral sedimentation. They are also prone to introducing human error and are difficult to deeply explore the intrinsic relationships in the data, resulting in insufficient accuracy and reliability of the prediction results.
AI models are used to predict the migration of ore-forming fluids. By acquiring geological and fluid data, performing correlation analysis and feature extraction, a prediction model for the migration of ore-forming fluids is constructed and trained. Combined with 3D model construction and visualization, the prediction accuracy is improved.
It improves the accuracy of predicting the migration of ore-forming fluids, enhances the efficiency and success rate of mineral resource exploration, and provides intuitive geological information to support exploration strategies.
Smart Images
Figure CN119691402B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of mineral resource exploration technology, and in particular to a mineral exploration method based on AI models for predicting the migration of ore-forming fluids. Background Technology
[0002] In current mineral exploration practices, the prediction of ore-forming fluid migration and the determination of mineral deposition locations traditionally rely on geologists' experience, geostatistical analysis, and simple mathematical models. However, these methods have many limitations.
[0003] First, the acquisition and processing of geological data often rely on manual operations, which are not only time-consuming and labor-intensive but also prone to human error. In particular, when faced with massive amounts of multi-dimensional geological data, the processing efficiency and accuracy of traditional methods are severely challenged.
[0004] Secondly, in terms of correlation analysis between geological data and fluid data, traditional methods can only perform simple statistical analysis, making it difficult to deeply explore the intrinsic connections and potential patterns between the data. This results in an insufficient understanding of the migration mechanism of ore-forming fluids, and the accuracy and reliability of the prediction results are greatly reduced.
[0005] Therefore, there is an urgent need for a mineral exploration method based on AI models to predict the migration of ore-forming fluids, which can overcome the shortcomings of existing technologies and provide strong support for the exploration and development of mineral resources. Summary of the Invention
[0006] The purpose of this invention is to provide a mineral exploration method based on an AI model for predicting the migration of ore-forming fluids, which can improve the accuracy of predicting the migration of ore-forming fluids.
[0007] To achieve the above objectives, the present invention provides the following solution:
[0008] A mineral exploration method based on an AI model for predicting the migration of ore-forming fluids includes:
[0009] Acquire geological and fluid data of the target mineralization; the geological data includes geological, geophysical, geochemical, and remote sensing data; the fluid data includes temperature, pressure, and fluid composition.
[0010] A correlation analysis is performed based on the geological data and the fluid data, and correlation features are extracted;
[0011] AI models are constructed and trained based on the geological data, fluid data, and related data. The trained model is then identified as the ore-forming fluid migration prediction model, and the migration path and mineral deposition location of the ore-forming fluid are determined using the ore-forming fluid migration prediction model.
[0012] A three-dimensional model is constructed based on the geological data, the migration path of the ore-forming fluid, and the location of mineral sedimentation, and the migration process of the ore-forming fluid based on the three-dimensional model is visualized.
[0013] Optionally, before constructing and training the AI model, the method further includes: preprocessing the geological data and the fluid data; the preprocessing employs a numerical resampling algorithm or a dynamic outlier removal algorithm.
[0014] Optionally, the processing procedure of the numerical resampling algorithm is as follows:
[0015] First, the validity and rationality of the geological data and the fluid data are judged: a range threshold for validity and rationality is set, and the next step is performed when the geological data and the fluid data meet the range threshold;
[0016] Then, within a set time period, the continuity of the geological data and the fluid data is checked respectively: when there are missing points in the geological data and the fluid data, they are supplemented by second-order Lagrange interpolation to obtain resampled data;
[0017] Finally, an FIR digital filter is constructed using the window function method, and the resampled data is filtered using the FIR digital filter to obtain effective data.
[0018] Optionally, the processing procedure of the dynamic outlier removal algorithm is as follows:
[0019] According to the set caching time, the geological data and the fluid data are temporarily stored in different buffers. For the buffers that have completed temporary storage, the data in the set buffers are popped out of the stack in sequence, and dynamic anomaly detection is performed against the set range thresholds for validity and rationality to filter out outliers and obtain valid data.
[0020] Optionally, the popping order is first-in, last-out.
[0021] Optionally, the construction and training process of the ore-forming fluid migration prediction model includes:
[0022] A pre-trained network for constructing an AI model based on attention mechanisms and generative adversarial networks; the pre-trained network includes an input layer, a hidden layer, and an output layer connected in sequence; the hidden layer includes a convolutional layer, a recurrent layer, and a fully connected layer connected in sequence.
[0023] The pre-trained network is trained using the geological data, the fluid data, and the associated data, and Dropout or L2 regularization is used to prevent overfitting. When the trained network meets the set iteration conditions or accuracy, the trained network is determined to be the ore-forming fluid migration prediction model.
[0024] Optionally, the training may employ RMSE loss and MAE loss as loss functions.
[0025] Optionally, it also includes: building a data monitoring interface to visualize the migration process of the ore-forming fluid, and using data visualization tools to dynamically display it on the data monitoring interface.
[0026] According to specific embodiments provided by the present invention, the present invention discloses the following technical effects:
[0027] This invention discloses a mineral exploration method for predicting the migration of ore-forming fluids based on an AI model. The method includes acquiring geological and fluid data of the target mineralization; performing correlation analysis on the geological and fluid data and extracting correlation features; constructing and training an AI model based on the geological, fluid, and correlation data, determining the trained model as the ore-forming fluid migration prediction model, and using the ore-forming fluid migration prediction model to determine the migration path of the ore-forming fluid and the location of mineral deposition; constructing a three-dimensional model based on the geological data, the migration path of the ore-forming fluid, and the location of mineral deposition, and visualizing the ore-forming fluid migration process based on the three-dimensional model. This invention can improve the accuracy of predicting the migration of ore-forming fluids. Attached Figure Description
[0028] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0029] Figure 1 This is a flowchart of the mineral exploration method based on an AI model for predicting the migration of ore-forming fluids, as described in this invention. Detailed Implementation
[0030] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0031] The purpose of this invention is to provide a mineral exploration method based on an AI model for predicting the migration of ore-forming fluids, which can improve the accuracy of predicting the migration of ore-forming fluids.
[0032] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.
[0033] like Figure 1 As shown, this invention provides a mineral exploration method based on an AI model for predicting the migration of ore-forming fluids, comprising:
[0034] A mineral exploration method based on an AI model for predicting the migration of ore-forming fluids includes:
[0035] Step 100: Obtain geological and fluid data of the target mineralization; the geological data includes geological, geophysical, geochemical and remote sensing data; the fluid data includes temperature, pressure and fluid composition.
[0036] Step 200: Perform correlation analysis based on the geological data and the fluid data, and extract correlation features.
[0037] Step 300: Based on the geological data, the fluid data, and the associated data, construct and train an AI model, determine the trained model as the ore-forming fluid migration prediction model, and use the ore-forming fluid migration prediction model to determine the migration path of the ore-forming fluid and the location of mineral deposition.
[0038] Step 400: Construct a three-dimensional model based on the geological data, the migration path of the ore-forming fluid, and the location of mineral sedimentation, and visualize the migration process of the ore-forming fluid based on the three-dimensional model.
[0039] Before constructing and training the AI model in step 300, the method further includes: preprocessing the geological data and the fluid data; the preprocessing adopts a numerical resampling algorithm or a dynamic outlier removal algorithm.
[0040] As a specific implementation method, the processing procedure of the numerical resampling algorithm is as follows:
[0041] First, the validity and rationality of the geological and fluid data are assessed: a range threshold for validity and rationality is set, and the process proceeds to the next step when the geological and fluid data meet the range threshold. Then, within a set time period, the continuity of the geological and fluid data is checked: if there are missing points in the geological and fluid data, they are supplemented using quadratic Lagrange interpolation to obtain resampled data. Finally, an FIR digital filter is constructed using the window function method, and the resampled data is filtered using the FIR digital filter to obtain valid data.
[0042] As a specific implementation method, the processing procedure of the dynamic outlier removal algorithm is as follows:
[0043] According to the set caching time, the geological data and the fluid data are temporarily stored in different buffers. For each buffer that has completed temporary storage, the data in the buffer is popped out sequentially and dynamically anomaly detected against set thresholds for validity and reasonableness. Outliers are then filtered out to obtain valid data. The popping order is first-in, last-out.
[0044] Before making predictions, the model needs to be trained. Therefore, a process for constructing and training the ore-forming fluid migration prediction model is provided, specifically including:
[0045] A pre-trained network for an AI model is constructed based on an attention mechanism and a generative adversarial network. The pre-trained network comprises an input layer, a hidden layer, and an output layer connected in sequence. The hidden layer comprises a convolutional layer, a recurrent layer, and a fully connected layer connected in sequence. The pre-trained network is trained using the geological data, the fluid data, and the associated data, and Dropout or L2 regularization is used to prevent overfitting. When the trained network meets the set iteration conditions or accuracy, the trained network is determined as the ore-forming fluid migration prediction model. The loss functions used in the training are RMSE loss and MAE loss.
[0046] As a specific implementation method, the visualization display in step 400 also includes: building a data monitoring interface to visualize the migration process of the ore-forming fluid, and using data visualization tools to dynamically display it on the data monitoring interface.
[0047] Based on the above technical solution, the following embodiments are provided.
[0048] Step 1: Data Collection
[0049] Geological data collection: Collect geological, geophysical, geochemical and remote sensing data of the target area.
[0050] Fluid data collection: Collect fluid temperature, pressure, and composition data for the target area.
[0051] Step 2: Data Preprocessing
[0052] Numerical resampling:
[0053] Determine the validity and reasonableness of the data, and set a threshold range.
[0054] Continuity checks are performed on the data, and quadratic Lagrange interpolation is used to supplement missing data points.
[0055] Construct an FIR digital filter to filter the resampled data and obtain effective data.
[0056] Dynamic outlier removal:
[0057] Data is temporarily stored in a buffer and popped from the stack in a last-in-last-out (LIFO) order.
[0058] Outliers are dynamically detected and filtered out to obtain valid data.
[0059] Step 3: Association Analysis and Feature Extraction
[0060] Correlation analysis: Correlation analysis is performed based on preprocessed geological and fluid data.
[0061] Feature extraction: Extracting features related to the migration of ore-forming fluids.
[0062] Step 4: AI Model Building and Training
[0063] Pre-trained network construction: A pre-trained network is constructed based on the attention mechanism and generative adversarial network, including an input layer, hidden layers and an output layer, wherein the hidden layers include convolutional layers, recurrent layers and fully connected layers.
[0064] Model training:
[0065] The pre-trained network is trained using geological data, fluid data, and correlation data.
[0066] Use Dropout or L2 regularization to prevent overfitting.
[0067] Use RMSE loss and MAE loss as the loss functions for training.
[0068] When the network meets the set iteration conditions or accuracy, it is determined to be a prediction model for the migration of ore-forming fluids.
[0069] Step 5: Prediction of ore-forming fluid migration paths and mineral sedimentation locations
[0070] Prediction model application: Use a trained ore-forming fluid migration prediction model to determine the migration path of ore-forming fluids and the location of mineral deposition.
[0071] Step 6: 3D Model Construction and Visualization
[0072] 3D model construction: A 3D model is constructed based on geological data, the migration path of ore-forming fluids, and the location of mineral sedimentation.
[0073] Visualization: The migration process of ore-forming fluids is dynamically displayed on the data monitoring interface using data visualization tools.
[0074] Optional step: Setting up the data monitoring interface
[0075] Interface setup: Build a data monitoring interface to display dynamic data on the migration process of ore-forming fluids.
[0076] Dynamic display: The migration process of ore-forming fluids is dynamically displayed on the interface using data visualization tools.
[0077] Implementation effect
[0078] By implementing the above methods, geological exploration teams can more accurately predict the migration paths of ore-forming fluids and the locations of mineral deposits, improving the efficiency and success rate of mineral resource exploration. Furthermore, the construction and visualization of 3D models provide the team with intuitive geological information, helping to better understand the mineralization process and formulate exploration strategies.
[0079] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.
[0080] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.
Claims
1. A method for ore prospecting based on AI model for ore-forming fluid migration prediction, characterized in that, The application relates to a method for predicting the migration of ore-forming fluid. The method comprises the following steps: obtaining geological data and fluid data of a target ore-forming area; the geological data comprises geological, geophysical, geochemical and remote sensing data; the fluid data comprises temperature, pressure and fluid composition; performing correlation analysis based on the geological data and the fluid data, and extracting correlation features; constructing and training an AI model based on the geological data, the fluid data and the correlation features, determining the trained model as an ore-forming fluid migration prediction model, and determining the migration path of the ore-forming fluid and the position of the ore-forming fluid by using the ore-forming fluid migration prediction model; constructing a three-dimensional model based on the geological data, the migration path of the ore-forming fluid and the position of the ore-forming fluid, and visually displaying the migration process of the ore-forming fluid based on the three-dimensional model; the construction and training process of the ore-forming fluid migration prediction model comprises the following steps: constructing a pre-training network of the AI model based on an attention mechanism and a generative adversarial network; the pre-training network comprises an input layer, a hidden layer and an output layer which are connected in sequence; the hidden layer comprises a convolution layer, a recurrent layer and a full connection layer which are connected in sequence; training the pre-training network by using the geological data, the fluid data and the correlation features, and preventing network overfitting by using Dropout or L2 regularization; when the trained network meets the set iteration condition or accuracy, the trained network is determined as the ore-forming fluid migration prediction model; before the AI model is constructed and trained, the geological data and the fluid data are preprocessed; the preprocessing adopts a numerical resampling algorithm or an abnormal value dynamic screening algorithm; the processing process of the numerical resampling algorithm comprises the following steps: firstly, judging the effectiveness and rationality of the geological data and the fluid data; a range threshold of the effectiveness and rationality is set, and when the geological data and the fluid data meet the range threshold, the next step is entered; then, continuously detecting the geological data and the fluid data within a set time length; when the geological data and the fluid data have missing points, the missing points are supplemented by using a quadratic Lagrange interpolation to obtain resampling data; finally, constructing an FIR digital filter by using a window function method, and filtering the resampling data by using the FIR digital filter to obtain effective data; the processing process of the abnormal value dynamic screening algorithm comprises the following steps:
2. The method of claim 1, wherein the AI model is trained using a plurality of training data sets, each of the training data sets including a plurality of mineralization data and a plurality of migration data of ore-forming fluid. storing the geological data and the fluid data in different buffer areas according to a set buffer time; for the buffer areas that have completed the storage, the data in the set buffer areas are popped out in sequence, and dynamic abnormality detection is performed on the data by using the set range threshold of the effectiveness and rationality to screen out abnormal values and obtain effective data.
3. The method of claim 1, wherein the AI model is trained using a plurality of training data sets, each of the training data sets including a plurality of mineralization data and a plurality of migration data of ore-forming fluid. The pop-out order is first-in last-out.
4. The method of claim 1, wherein the AI model is trained using a plurality of training data sets, each of the training data sets including a plurality of mineralization data and a plurality of migration data of ore-forming fluid. The loss function used in the training is an RMSE loss and an MAE loss. The application further comprises the following steps: building a data monitoring interface for visually displaying the migration process of the ore-forming fluid, and dynamically displaying the data on the data monitoring interface by using a data visualization tool.