Intelligent prediction method, device, equipment and storage medium for uranium concentration in in-situ leaching uranium

By training the acid amount and uranium concentration prediction model, the predicted acid amount and injection amount data are used to predict uranium concentration, which solves the problem of large uranium concentration prediction error in the prior art and improves the accuracy of the prediction.

CN116842842BActive Publication Date: 2024-12-06BEIJING RESEARCH INSTITUTE OF CHEMICAL ENGINEERING AND METALLURGY
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202310833710.1
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-07-07
Publication Date
2024-12-06
Estimated Expiration
2043-07-07

AI Technical Summary

Technical Problem

In the existing uranium leaching technology, since the ore layer is buried underground, the existing technology cannot accurately characterize the non-uniform distribution of minerals and permeability space, resulting in a large deviation between the numerical simulation prediction results and the measured data, especially due to the various notes. The recording data of the acid quantity of water wells is limited, resulting in a large deviation from the real value of uranium concentration prediction.

Method used

Using an intelligent prediction method for uranium concentration in ground-leached uranium, the acid amount prediction model and uranium concentration prediction model are trained and verified by obtaining the injection volume data, acid amount data and uranium concentration data of the mining area, and the acid amount prediction model is used to predict the acid amount. The predicted acid quantity data and injection quantity data output from the model are used as inputs to predict uranium concentration to reduce the prediction error.

Benefits of technology

By training and verifying the prediction model of acid amount and injection amount, we obtain a prediction value that is closer to the real uranium concentration, which reduces prediction errors and improves the accuracy of prediction.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN116842842B_ABST
    Figure CN116842842B_ABST
Patent Text Reader

Abstract

The present invention provides an intelligent prediction method, device, equipment and storage medium for uranium concentration in in-situ leaching uranium, which relates to the technical field of in-situ leaching uranium, including obtaining first information and obtaining second information; using the pumping liquid volume data and acid volume data in the first information as the first input data and the uranium concentration data as the first output data, training and verifying a preset first prediction model to obtain a uranium concentration prediction model; using the pumping liquid volume data in the first information as the second input data and the acid volume data as the second output data, training and verifying a preset second prediction model to obtain an acid volume prediction model; inputting the second information into the acid volume prediction model, outputting the predicted acid volume data of the date to be predicted, using the pumping liquid volume data and the predicted acid volume data of the date to be predicted as the input of the uranium concentration prediction model, and outputting the predicted uranium concentration of the date to be predicted. The present invention makes the predicted uranium concentration value closer to the predicted value of the actual uranium concentration.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of in-situ leaching uranium, and in particular to an intelligent prediction method, device, equipment and storage medium for uranium concentration in in-situ leaching uranium. Background Art

[0002] In in-situ uranium mining, the existing technology cannot accurately describe the spatial non-uniform distribution of minerals and permeability because the ore layer is buried underground and in a gray box state, resulting in a large deviation between the numerical simulation prediction results and the measured data of in-situ uranium mining. Constructing a prediction model for uranium concentration in the mining well with self-regulation ability in the in-situ uranium mining process is at the forefront of in-situ uranium mining simulation research. However, when constructing the existing in-situ uranium mining process prediction model, the acid volume record data of each injection well is very limited, so the final predicted uranium concentration deviates greatly from the actual value. Summary of the invention

[0003] The purpose of the present invention is to provide a method, device, equipment and storage medium for intelligent prediction of uranium concentration in in-situ leaching uranium to improve the above problems. In order to achieve the above purpose, the technical solution adopted by the present invention is as follows:

[0004] In a first aspect, the present application provides a method for intelligently predicting uranium concentration in in-situ leaching uranium, comprising:

[0005] Acquire first information, the first information including pumped fluid volume data, acid volume data and uranium concentration data of a mining area, the mining area including a plurality of well locations, and the first information being time series data;

[0006] Acquiring second information, wherein the second information includes data on the amount of fluid drawn on the date to be predicted;

[0007] Using the pumping liquid volume data and the acid volume data in the first information as the first input data and the uranium concentration data as the first output data, training and verifying the preset first prediction model to obtain a uranium concentration prediction model;

[0008] Using the pumping liquid volume data in the first information as the second input data and the acid volume data as the second output data, the preset second prediction model is trained and verified to obtain an acid volume prediction model;

[0009] The second information is input into the acid quantity prediction model, and the predicted acid quantity data for the date to be predicted is output. The injection liquid quantity data for the date to be predicted and the predicted acid quantity data for the date to be predicted are used as inputs of the uranium concentration prediction model, and the predicted uranium concentration for the date to be predicted is output.

[0010] In a second aspect, the present application also provides an intelligent prediction device for uranium concentration in in-situ leaching uranium, comprising:

[0011] A first acquisition module, the first acquisition module is used to acquire first information, the first information includes pumped liquid volume data, acid volume data and uranium concentration data of the mining area, the mining area includes a plurality of well locations, and the first information is time series data;

[0012] A second acquisition module, the second acquisition module is used to acquire second information, the second information includes the data of the amount of liquid drawn on the date to be predicted;

[0013] A first training module, wherein the first training module is used to train and verify a preset first prediction model using the pumping liquid volume data and the acid volume data in the first information as first input data and the uranium concentration data as first output data, so as to obtain a uranium concentration prediction model;

[0014] A second training module, wherein the second training module is used to train and verify a preset second prediction model using the pumping liquid volume data in the first information as second input data and the acid volume data as second output data, so as to obtain an acid volume prediction model;

[0015] A prediction module, wherein the prediction module is used to input the second information into the acid quantity prediction model, output the predicted acid quantity data for the date to be predicted, use the pumping liquid quantity data for the date to be predicted and the predicted acid quantity data for the date to be predicted as inputs to the uranium concentration prediction model, and output the predicted uranium concentration for the date to be predicted.

[0016] In a third aspect, the present application also provides an intelligent prediction device for uranium concentration in in-situ leaching uranium, comprising:

[0017] Memory for storing computer programs;

[0018] A processor is used to implement the steps of the method for intelligently predicting the uranium concentration in in-situ leaching uranium when executing the computer program.

[0019] In a fourth aspect, the present application further provides a readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of the above-mentioned intelligent prediction method for uranium concentration in in-situ leaching uranium are implemented.

[0020] The beneficial effects of the present invention are as follows: the present invention predicts the acid amount on the date to be predicted by training and verifying a good acid amount prediction model and the amount of pumped fluid on the date to be predicted, and then the predicted acid amount on the date to be predicted and the obtained amount of pumped fluid on the date to be predicted are input into a trained and verified uranium concentration prediction model, thereby obtaining a predicted value that is closer to the actual uranium concentration.

[0021] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or be understood by implementing the embodiments of the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description, claims, and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0023] Figure 1 It is a schematic diagram of the process of the intelligent prediction method of uranium concentration in in-situ leaching uranium according to an embodiment of the present invention;

[0024] Figure 2 It is a schematic diagram of the structure of the intelligent prediction device for uranium concentration in in-situ leaching uranium according to an embodiment of the present invention;

[0025] Figure 3 It is a schematic diagram of the structure of the intelligent prediction device for uranium concentration in in-situ leaching uranium described in an embodiment of the present invention.

[0026] Figure 4 These are the monitoring data of the amount of pumped fluid, the amount of acid and the uranium concentration in the C7 mining area of ​​the Bayan-Ula uranium mine in the embodiment of the present invention.

[0027] Figure 5 This is the location of the pumping and injection wells and the structured treatment of the C7 mining area in the embodiment of the present invention.

[0028] Figure 6 It is a fitting effect diagram of the acid amount in the training stage and the verification stage of the acid amount prediction model in the embodiment of the present invention.

[0029] Figure 7 This is a diagram showing the effect of acid quantity prediction for each production well during the model training stage and the verification stage in an embodiment of the present invention.

[0030] Figure 8 It is a diagram showing the uranium concentration fitting effect in the training stage and the verification stage of the uranium concentration prediction model in an embodiment of the present invention.

[0031] Fig. 9 Graph showing the effect of uranium concentration prediction in the training and verification stages for each mining well in an embodiment of the present invention.

[0032] Markings in the figure: 610, first acquisition module; 620, second acquisition module; 630, first training module; 640, second training module; 650, prediction module; 711, first construction module; 712, second construction module; 713, third construction module; 714, fourth construction module; 715, fifth construction module; 716, sixth construction module; 717, seventh construction module; 718, collection module; 719, first processing module; 720, second processing module; 721, third processing module; 800, intelligent prediction device for uranium concentration in in-situ leaching uranium; 801, processor; 802, memory; 803, multimedia component; 804, I / O interface; 805, communication component. DETAILED DESCRIPTION

[0033] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. The components of the embodiments of the present invention generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0034] It should be noted that similar reference numerals and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further defined and explained in the subsequent drawings. At the same time, in the description of the present invention, the terms "first", "second", etc. are only used to distinguish the description and cannot be understood as indicating or implying relative importance.

[0035] Embodiment 1:

[0036] See also Figure 1 , the figure shows that the method includes step S100, step S200, step S300, step S400, and step S500.

[0037] This embodiment provides a method for intelligently predicting uranium concentration in in-situ leaching uranium, comprising:

[0038] Step S100: acquiring first information, wherein the first information includes pumped fluid volume data, acid volume data and uranium concentration data of a mining area, wherein the mining area includes a plurality of well locations, and the first information is time series data;

[0039] Among them, several well locations include several water injection wells and mining wells; the pumped liquid volume data includes the injection volume of each water injection well and the pumped liquid volume of each mining well; the acid volume data is the pH data of each water injection well; and the uranium concentration data is the uranium concentration data of each mining well.

[0040] The step S100 also includes obtaining the location data of each well in the mining area, wherein the location data of each well in the mining area is obtained:

[0041] Step S101: establishing coordinates for the mining area, dividing the mining area into a plurality of cells, each of which contains at most one well location;

[0042] Step S102: setting the input data format of the uranium concentration prediction model according to the time series length of the first information, the preset number of model channels and the number of cells in the horizontal and vertical directions within the coordinates, and setting the output data format of the uranium concentration prediction model according to the preset number of model channels and the number of cells in the horizontal and vertical directions within the coordinates.

[0043] The step 100 further includes:

[0044] Step S103: collecting the pumped fluid volume data, acid volume data and uranium concentration data of the mining area;

[0045] Step S104: performing denoising and supplementation processing on the collected data in sequence;

[0046] Step S105: normalizing the data obtained by data supplementation;

[0047] In order to eliminate the impact of the magnitude difference of input and output variables on the model training and prediction process, the data can be normalized through this program.

[0048] Step S106: Based on the linear interpolation method, the normalized data of the amount of fluid drawn at different time points are processed at a unified time interval to obtain the first information.

[0049] Step S200: obtaining second information, wherein the second information includes data on the amount of fluid drawn on the date to be predicted;

[0050] Step S300: using the pumping liquid volume data and the acid volume data in the first information as the first input data and the uranium concentration data as the first output data, training and verifying a preset first prediction model to obtain a uranium concentration prediction model;

[0051] The step S300 also includes a method for constructing the preset first prediction model:

[0052] Step S301: The method for constructing the preset first prediction model includes:

[0053] Step S302: constructing a first convolutional neural network layer, where the first convolutional neural network layer is used to perform spatial information fusion and dimension conversion processing on the first input data;

[0054] Step S303: constructing a first long short-term memory network layer, where the first long short-term memory network layer is used to perform time series information fusion processing on the output of the first convolutional neural network layer;

[0055] Step S304: construct a second convolutional neural network layer, where the second convolutional neural network layer is used to extract features from the output of the first long short-term memory network layer.

[0056] Convolutional neural networks adopt the idea of ​​shared weights. While expressing complex input-output relationships, they greatly reduce the number of weights in the neural network, thereby effectively alleviating the problem of model overfitting caused by the mismatch between the number of weights and the number of monitoring data during neural network training, and have good stability.

[0057] The memory unit in the long short-term memory network contains three gate structures, namely the forget gate, input gate and output gate. The forget gate selectively forgets the information in the memory unit and retains the useful information; the input gate controls the input of new information and selectively records it to the memory unit; the output gate ensures that the memory unit only outputs relevant information synchronized with the current time. The gate structure ensures that information passes selectively by discarding useless information and adding useful information, so that the state of the memory unit is constantly updated, achieving the purpose of controlling and protecting the state of the memory unit.

[0058] The step S300 also includes training and validating the first prediction model:

[0059] Step S305: Divide the first information into training data and verification data according to time periods;

[0060] Step S306: randomly selecting a number of days of injection volume data, acid volume data and uranium concentration data from the training data to train the first prediction model, thereby obtaining a uranium concentration prediction model;

[0061] Step S307: performing weight training on the uranium concentration prediction model based on the ADAM optimization algorithm to obtain a uranium concentration prediction model;

[0062] Step S308: randomly intercepting the injection volume data, acid volume data and uranium concentration data of the several days from the verification data to verify the uranium concentration prediction model;

[0063] Step S400: using the pumping liquid volume data in the first information as the second input data and the acid volume data as the second output data, training and verifying the preset second prediction model to obtain an acid volume prediction model;

[0064] The step S400 also includes the method for constructing the preset second prediction model including:

[0065] Step S401: constructing a residual connection layer, wherein the residual connection layer is used to add the input of the first convolutional neural network layer to the output of the second convolutional neural network layer;

[0066] The residual connection layer realizes the input of the predicted acid volume and injection volume into the uranium concentration prediction model. It realizes the input of the second information into the acid volume prediction model, outputs the predicted acid volume data for the date to be predicted, uses the injection volume data for the date to be predicted and the predicted acid volume data for the date to be predicted as the input of the uranium concentration prediction model, and outputs the predicted uranium concentration for the date to be predicted. The residual connection layer enhances the gradient propagation and solves the gradient dissipation, making the output of the acid volume data more stable, and will not cause the loss of acid volume data due to the structure of the neural network, so that the predicted value output by the subsequent uranium concentration prediction model is closer to the actual value.

[0067] Step S402: constructing a third convolutional neural network layer, wherein the third convolutional neural network layer is used to perform spatial information fusion and dimension conversion processing on the output of the residual connection layer;

[0068] Step S403: constructing a second long short-term memory network layer, where the second long short-term memory network layer is used to perform time series information fusion processing on the output of the first convolutional neural network layer;

[0069] Step S404: construct a fourth convolutional neural network layer, and the fourth convolutional neural network layer is used to extract features from the output of the second long short-term memory network layer.

[0070] Step S400: further comprising training and validating the second prediction model:

[0071] Step S405: training the preset prediction model according to the pumping liquid volume data and the acid volume data in the training data to obtain an acid volume prediction model;

[0072] Step S406: training the acid volume prediction model based on the ADAM optimization algorithm according to the single well acid volume data in the training data and the total acid volume data of the mining area to obtain the acid volume prediction model;

[0073] Step S407: verifying the acid quantity prediction model according to the verification data.

[0074] Step S500: input the second information into the acid quantity prediction model, output the predicted acid quantity data for the date to be predicted, use the pumping liquid quantity data for the date to be predicted and the predicted acid quantity data for the date to be predicted as inputs to the uranium concentration prediction model, and output the predicted uranium concentration for the date to be predicted.

[0075] Embodiment 2:

[0076] like Figure 2 As shown, this embodiment provides an intelligent prediction device for uranium concentration in in-situ leaching uranium, the device comprising:

[0077] A first acquisition module 610, the first acquisition module 610 is used to acquire first information, the first information includes pumped liquid volume data, acid volume data and uranium concentration data of the mining area, the mining area includes a plurality of well locations, and the first information is time series data;

[0078] The first acquisition module 610 is also used to establish coordinates for the mining area, divide the mining area into a number of cells, each of which contains at most one well location; set the input data format of the uranium concentration prediction model according to the time series length of the first information, the preset number of model channels, and the number of cells in the horizontal and vertical directions within the coordinates, and set the output data format of the uranium concentration prediction model according to the preset number of model channels and the number of cells in the horizontal and vertical directions within the coordinates.

[0079] A second acquisition module 620, the second acquisition module 620 is used to acquire second information, the second information including the pumping volume data of the to-be-predicted date;

[0080] A first training module 630, wherein the first training module 630 is used to train and verify a preset first prediction model using the pumping liquid volume data and the acid volume data in the first information as first input data and the uranium concentration data as first output data, so as to obtain a uranium concentration prediction model;

[0081] A second training module 640, the second training module 640 is used to train and verify a preset second prediction model using the pumping liquid volume data in the first information as second input data and the acid volume data as second output data, to obtain an acid volume prediction model;

[0082] Prediction module 650, the prediction module 650 is used to input the second information into the acid quantity prediction model, output the predicted acid quantity data for the date to be predicted, use the pumping liquid quantity data for the date to be predicted and the predicted acid quantity data for the date to be predicted as inputs to the uranium concentration prediction model, and output the predicted uranium concentration for the date to be predicted.

[0083] The first training module 630 includes:

[0084] A first construction module 710, wherein the first construction module 710 is used to construct a first convolutional neural network layer, and the first convolutional neural network layer is used to perform spatial information fusion and dimension conversion processing on the first input data;

[0085] A second construction module 711, wherein the second construction module 711 is used to construct a first long short-term memory network layer, and the first long short-term memory network layer is used to perform time series information fusion processing on the output of the first convolutional neural network layer;

[0086] The third construction module 713 is used to construct a second convolutional neural network layer, and the second convolutional neural network layer is used to extract features from the output of the first long short-term memory network layer.

[0087] The second training module 640 includes:

[0088] A fourth construction module 714, wherein the fourth construction module 714 is used to construct a residual connection layer, wherein the residual connection layer is used to add the input of the first convolutional neural network layer to the output of the second convolutional neural network layer;

[0089] A fifth construction module 715, wherein the fifth construction module 715 is used to construct a third convolutional neural network layer, and the third convolutional neural network layer is used to perform spatial information fusion and dimension conversion processing on the output of the residual connection layer;

[0090] A sixth construction module 716, wherein the sixth construction module 716 is used to construct a second long short-term memory network layer, and the second long short-term memory network layer is used to perform time series information fusion processing on the output of the first convolutional neural network layer;

[0091] A seventh construction module 717, wherein the seventh construction module 717 is used to construct a fourth convolutional neural network layer, and the fourth convolutional neural network layer is used to extract features from the output of the second long short-term memory network layer.

[0092] The first acquisition module 610 includes:

[0093] A collection module 718, the collection module 718 is used to collect the pumping fluid volume data, acid volume data and uranium concentration data of the mining area;

[0094] The first processing module 719 performs denoising and supplementation processing on the collected data in sequence;

[0095] A second processing module 720, the second processing module 720 is used to normalize the data obtained by data supplementation;

[0096] The third processing module 721 is used to perform unified time interval processing on the normalized data of the amount of liquid drawn at different time points based on a linear interpolation method to obtain first information.

[0097] It should be noted that, regarding the device in the above embodiment, the specific manner in which each module performs the operation has been described in detail in the embodiment of the method, and will not be elaborated here.

[0098] Embodiment 3:

[0099] Corresponding to the above method embodiment, this embodiment also provides an intelligent prediction device for uranium concentration in in-situ uranium leaching. The intelligent prediction device for uranium concentration in in-situ uranium leaching described below and the intelligent prediction method for uranium concentration in in-situ leaching uranium described above can refer to each other.

[0100] Figure 3 FIG. 8 is a block diagram of an intelligent prediction device 800 for uranium concentration in in-situ leaching uranium according to an exemplary embodiment. Figure 3 As shown, the intelligent prediction device 800 for uranium concentration in in-situ leached uranium may include: a processor 801 and a memory 802. The intelligent prediction device 800 for uranium concentration in in-situ leached uranium may also include one or more of a multimedia component 803, an I / O interface 804, and a communication component 805.

[0101] The processor 801 is used to control the overall operation of the intelligent prediction device 800 for uranium concentration in in-situ leaching uranium to complete all or part of the steps in the above-mentioned intelligent prediction method for uranium concentration in in-situ leaching uranium. The memory 802 is used to store various types of data to support the operation of the intelligent prediction device 800 for uranium concentration in in-situ leaching uranium. These data may include, for example, instructions for any application or method operating on the intelligent prediction device 800 for uranium concentration in in-situ leaching uranium, as well as application-related data, such as contact data, sent and received messages, pictures, audio, video, etc. The memory 802 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, disk or optical disk. The multimedia component 803 may include a screen and an audio component. The screen may be, for example, a touch screen, and the audio component is used to output and / or input audio signals. For example, the audio component may include a microphone for receiving external audio signals. The received audio signal may be further stored in the memory 802 or sent via the communication component 805. The audio component also includes at least one speaker for outputting audio signals. The I / O interface 804 provides an interface between the processor 801 and other interface modules, and the above-mentioned other interface modules can be keyboards, mice, buttons, etc. These buttons can be virtual buttons or physical buttons. The communication component 805 is used for wired or wireless communication between the intelligent prediction device 800 for uranium concentration in uranium leaching and other devices. Wireless communication, such as Wi-Fi, Bluetooth, Near Field Communication (NFC), 2G, 3G or 4G, or a combination of one or more of them, so the corresponding communication component 805 can include: Wi-Fi module, Bluetooth module, NFC module.

[0102] In an exemplary embodiment, the intelligent prediction device 800 for uranium concentration in in-situ leaching uranium can be implemented by one or more application specific integrated circuits (Application Specific Integrated Circuit, referred to as ASIC), digital signal processors (Digital Signal Processor, referred to as DSP), digital signal processing devices (Digital Signal Processing Device, referred to as DSPD), programmable logic devices (Programmable Logic Device, referred to as PLD), field programmable gate arrays (Field Programmable Gate Array, referred to as FPGA), controllers, microcontrollers, microprocessors or other electronic components to execute the above-mentioned intelligent prediction method for uranium concentration in in-situ leaching uranium.

[0103] In another exemplary embodiment, a computer-readable storage medium including program instructions is also provided, and when the program instructions are executed by a processor, the steps of the above-mentioned intelligent prediction method for uranium concentration in in-situ leaching uranium are implemented. For example, the computer-readable storage medium can be the above-mentioned memory 802 including program instructions, and the above-mentioned program instructions can be executed by the processor 801 of the intelligent prediction device 800 for uranium concentration in in-situ leaching uranium to complete the above-mentioned intelligent prediction method for uranium concentration in in-situ leaching uranium.

[0104] Embodiment 4:

[0105] The type of mining method of the present invention was applied to the Bayanwula C7 mining area, and the data of injection fluid volume, uranium concentration and acid concentration monitored in the mining area from 2018 / 08 / 30 to 2019 / 08 / 29 were selected as the basis for forward simulation of uranium concentration, analysis of main controlling factors and verification of inversion control model.

[0106] like Figure 4 As shown, the developed uranium concentration prediction model is applied to the Bayanwula C7 mining area, which monitors the liquid extraction volume data from 2017 / 09 / 21 to 2020 / 07 / 24, the liquid injection volume data from 2018 / 08 / 30 to 2019 / 08 / 29 and 2019 / 10 / 21 to 2020 / 07 / 20, and the total acid and total uranium volume data from 2017 / 09 / 22 to 2020 / 07 / 24. Compared with the time span of the existing data, considering the completeness of different types of data, from the perspective of model verification, this embodiment selects the liquid extraction and injection volume, uranium concentration and acid concentration data from 2018 / 08 / 30 to 2019 / 08 / 29.

[0107] The first 300 days of data from 2018 / 08 / 30 to 2019 / 08 / 29 were used as training data for the first prediction model, and the data from the last 60 days were used as verification data for the preset first prediction model. In the 300 days of verification data, the volume of injection fluid for 21 days (3 weeks) was randomly intercepted as input data for the uranium concentration prediction model to predict the uranium concentration on the 21st day, forming a total of 280 sets of training data. In the verification process, the volume of injection fluid for 21 days was also used as input data to predict the uranium concentration on the 21st day. By comparing with the measured data, the accuracy and stability of the uranium concentration prediction model were tested.

[0108] like Figure 5 As shown, in order to facilitate the convolutional neural network to identify the location and data of each well, this embodiment divides the C7 mining area of ​​about 300m×300m into 10×10 cells, and each cell contains at most 1 mining well or 1 injection well. In the neural network model, the input data format containing injection and production wells is T×C×W×L: 21×3×10×10, and the output data format is C×W×L: 1×10×10, where T represents the time series length; C is the number of channels, the input data includes time, pumping fluid volume and injection acid volume, and the output data is the output uranium concentration; W×L represents the number of horizontal and vertical cells. Considering that different types of input data have significant differences in order of magnitude, in order to prevent the difference from affecting the skewness of the model results, this embodiment normalizes the time item in the input data to make it obey the 0-1 distribution, and standardizes the pumping fluid volume and the injection acid volume to make their values ​​obey the global standard normal distribution.

[0109] like Figure 4 and Figure 5 As shown, the main function of the uranium concentration prediction model constructed in this embodiment is to use the amount of pumped fluid and the amount of acid as input data to predict the uranium concentration and total uranium concentration value of each mining well. However, the acid volume record data of each pumping well is very limited (there are only 4 to 7 groups of acid concentration record values ​​for different injection wells in a year), and the total acid volume record data cannot be directly used as input data for uranium concentration prediction based on the in situ leaching uranium records. Therefore, firstly, the constructed acid volume prediction model is used, with the amount of pumped fluid as input data and the acid volume as output data, so as to realize the prediction of the acid volume, and the preset second prediction model weights are trained through the single well acid volume record values ​​and total acid volume record values ​​within a 300-day period. The training results are shown in the figure below. Figure 6As shown in the figure, the maximum error of the calculated value of total acid concentration within 300 days compared with the measured value is about 1.25g / L, the average error is 0.13g / L, and the relative error is 2.6%. In the verification stage, the average error of the total acid concentration prediction within 301-361 days is about 0.26g / L, and the relative error is 5.2%. From the statistical analysis results of the acid calculation errors at each time point in the two stages, the difference in errors between the two is not significant, indicating that after the model training, there is no serious overfitting problem, and the model has good prediction ability.

[0110] like Figure 7 As shown in the figure, in addition to having a good prediction effect on total acid concentration, the trained model can also control the maximum error of the prediction effect of acid concentration in a single mining well to 1.5g / L, and the model also shows good prediction ability at different mining well locations. Therefore, based on this model, the daily injection acid volume data of each injection well is supplemented as the input data for subsequent uranium concentration prediction.

[0111] like Figure 8 As shown in the figure, the flow rate, acid volume and time are used as known data to input the uranium concentration prediction model. After the first convolution layer performs spatial information fusion on the input data, the dimension is converted and then input into the LSTM layer for time series information fusion and mining. Finally, the convolution layer performs feature extraction to obtain the uranium concentration prediction value. Comparing it with the measured data, it can be seen that the maximum error of the total uranium concentration prediction in the training stage is 1.0 mg / L, the average error is 0.24 mg / L, and the relative error is 6.0%; the maximum error of the total uranium concentration in the verification stage is 0.85 mg / L, the average error is 0.29 mg / L, and the average relative error is 7.2%.

[0112] like Fig. 9 As shown, the maximum error of the uranium concentration prediction model in predicting the uranium concentration of a single well is about 1 mg / L, and the relative error can be controlled within 25%.

[0113] Embodiment 5:

[0114] Corresponding to the above method embodiment, a readable storage medium is also provided in this embodiment. The readable storage medium described below and the intelligent prediction method for uranium concentration in in-situ leaching uranium described above can refer to each other.

[0115] A readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the method for intelligently predicting uranium concentration in in-situ leaching uranium of the above method embodiment.

[0116] The readable storage medium may specifically be a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, or other readable storage medium that can store program codes.

[0117] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

[0118] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. An intelligent prediction method for uranium concentration in in-situ leaching uranium, characterized in that: include: Acquire first information, the first information including pumped fluid volume data, acid volume data and uranium concentration data of a mining area, the mining area including a plurality of well locations, and the first information being time series data; Acquiring second information, wherein the second information includes data on the amount of fluid drawn on the date to be predicted; Using the pumping liquid volume data and the acid volume data in the first information as the first input data and the uranium concentration data as the first output data, training and verifying the preset first prediction model to obtain a uranium concentration prediction model; Using the pumping liquid volume data in the first information as the second input data and the acid volume data as the second output data, the preset second prediction model is trained and verified to obtain an acid volume prediction model; The second information is input into the acid quantity prediction model, and the predicted acid quantity data for the date to be predicted is output. The injection liquid quantity data for the date to be predicted and the predicted acid quantity data for the date to be predicted are used as inputs of the uranium concentration prediction model, and the predicted uranium concentration for the date to be predicted is output.

2. The intelligent prediction method for uranium concentration in in-situ leaching uranium according to claim 1, characterized in that ,The method for constructing the preset first prediction model includes: Constructing a first convolutional neural network layer, wherein the first convolutional neural network layer is used to perform spatial information fusion and dimension conversion processing on the first input data; Constructing a first long short-term memory network layer, wherein the first long short-term memory network layer is used to perform time series information fusion processing on the output of the first convolutional neural network layer; Construct a second convolutional neural network layer, wherein the second convolutional neural network layer is used to extract features from the output of the first long short-term memory network layer.

3. The intelligent prediction method for uranium concentration in in-situ leaching uranium according to claim 2, characterized in that ,The method for constructing the preset first prediction model includes: Constructing a residual connection layer, wherein the residual connection layer is used to add the input of the first convolutional neural network layer to the output of the second convolutional neural network layer; Constructing a third convolutional neural network layer, wherein the third convolutional neural network layer is used to perform spatial information fusion and dimension conversion processing on the output of the residual connection layer; Constructing a second long short-term memory network layer, wherein the second long short-term memory network layer is used to perform time series information fusion processing on the output of the first convolutional neural network layer; Construct a fourth convolutional neural network layer, wherein the fourth convolutional neural network layer is used to extract features from the output of the second long short-term memory network layer.

4. The intelligent prediction method for uranium concentration in in-situ leaching uranium according to claim 1, characterized in that , obtaining the first information also includes obtaining the location data of each well in the mining area, and the obtaining the location data of each well in the mining area includes: Establishing coordinates for the mining area, dividing the mining area into a plurality of cells, each of which contains at most one well location; The input data format of the uranium concentration prediction model is set according to the time series length of the first information, the preset number of model channels, and the number of cells in the horizontal and vertical directions within the coordinates; and the output data format of the uranium concentration prediction model is set according to the preset number of model channels and the number of cells in the horizontal and vertical directions within the coordinates.

5. An intelligent prediction device for uranium concentration in in-situ leaching uranium, characterized in that: include: A first acquisition module, the first acquisition module is used to acquire first information, the first information includes pumped liquid volume data, acid volume data and uranium concentration data of the mining area, the mining area includes a plurality of well locations, and the first information is time series data; A second acquisition module, the second acquisition module is used to acquire second information, the second information includes the data of the amount of liquid drawn on the date to be predicted; A first training module, wherein the first training module is used to train and verify a preset first prediction model using the pumping liquid volume data and the acid volume data in the first information as first input data and the uranium concentration data as first output data, so as to obtain a uranium concentration prediction model; A second training module, wherein the second training module is used to train and verify a preset second prediction model using the pumping liquid volume data in the first information as second input data and the acid volume data as second output data, so as to obtain an acid volume prediction model; A prediction module, wherein the prediction module is used to input the second information into the acid quantity prediction model, output the predicted acid quantity data for the date to be predicted, use the pumping liquid quantity data for the date to be predicted and the predicted acid quantity data for the date to be predicted as inputs to the uranium concentration prediction model, and output the predicted uranium concentration for the date to be predicted.

6. The intelligent prediction device for uranium concentration in in-situ leaching uranium according to claim 5, characterized in that: The first training module includes: A first building module, wherein the first building module is used to build a first convolutional neural network layer, and the first convolutional neural network layer is used to perform spatial information fusion and dimension conversion processing on the first input data; A second construction module, wherein the second construction module is used to construct a first long short-term memory network layer, and the first long short-term memory network layer is used to perform time series information fusion processing on the output of the first convolutional neural network layer; A third building module, wherein the third building module is used to construct a second convolutional neural network layer, and the second convolutional neural network layer is used to extract features from the output of the first long short-term memory network layer.

7. The intelligent prediction device for uranium concentration in in-situ leaching uranium according to claim 6, characterized in that: The second training module comprises: A fourth building block, the fourth building block is used to build a residual connection layer, the residual connection layer is used to add the input of the first convolutional neural network layer to the output of the second convolutional neural network layer; A fifth building block, the fifth building block is used to build a third convolutional neural network layer, and the third convolutional neural network layer is used to perform spatial information fusion and dimension conversion processing on the output of the residual connection layer; A sixth construction module, the sixth construction module is used to construct a second long short-term memory network layer, and the second long short-term memory network layer is used to perform time series information fusion processing on the output of the first convolutional neural network layer; A seventh building block, the seventh building block is used to construct a fourth convolutional neural network layer, and the fourth convolutional neural network layer is used to extract features from the output of the second long short-term memory network layer.

8. The intelligent prediction device for uranium concentration in in-situ leaching uranium according to claim 5, characterized in that: The first acquisition module includes: A collection module, the collection module is used to collect the pumped fluid volume data, acid volume data and uranium concentration data of the mining area; The first processing module performs denoising and supplementation processing on the collected data in turn; A second processing module, the second processing module is used to normalize the data obtained by data supplementation; The third processing module is used to perform unified time interval processing on the normalized data of the amount of liquid drawn at different time points based on a linear interpolation method to obtain the first information.

9. An intelligent prediction device for uranium concentration in in-situ leaching uranium, characterized in that: include: Memory for storing computer programs; A processor is used to implement the steps of the method for intelligently predicting uranium concentration in in-situ leaching uranium as claimed in any one of claims 1 to 4 when executing the computer program.

10. A readable storage medium, characterized in that: The readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the method for intelligently predicting uranium concentration in in-situ leaching uranium as claimed in any one of claims 1 to 4.

Citation Information

Patent Citations

  • Independent acidification method for acid method-based in-situ leaching uranium mining single-mining area

    CN110331281A

  • Leached uranium concentration prediction method and device, electronic equipment and storage medium

    CN115440317A