A method and device for constructing and applying a rare earth extraction simulation model, and a medium

Abstract

The application discloses a rare earth extraction simulation model construction method, application method, device and medium, relates to the technical field of chemical process simulation, and comprises the following steps: acquiring a rare earth extraction historical data set under different working conditions; each sample in the rare earth extraction historical data set comprises historical production process characteristic parameters of rare earth extraction and actual values of component content of each separation stage; a multi-branch generative adversarial network is trained by using the rare earth extraction historical data set to obtain a rare earth extraction simulation model; the multi-branch generative adversarial network comprises a main network and a plurality of branch networks; each separation stage in the rare earth extraction process is provided with a branch network; the application constructs a multi-branch generative adversarial network for different numbers of extraction tanks, improves the prediction precision and accuracy of the rare earth extraction simulation, and more accurately reflects the hierarchical output characteristics of the component content in actual rare earth production.

Description

Technical Field

[0001] This application relates to the field of chemical process simulation technology, and in particular to the construction, application method, apparatus and medium of a rare earth extraction simulation model. Background Technology

[0002] Rare earth elements comprise 17 elements, including the lanthanides (Sc) and yttrium (Y) from the periodic table. In the actual rare earth extraction process, the traditional method for detecting component content involves manual quantitative sampling in a specific extraction tank followed by offline analysis. Staff then adjust on-site production process parameters based on the analysis results. This method is not very timely. Therefore, it is essential to accurately predict the rare earth element composition in the extraction tank during the extraction process, allowing for timely adjustments to operating variables and improving product quality.

[0003] Research on chemical process simulation methods mainly falls into two categories: one is static modeling based on in-depth mechanistic understanding, which constructs mathematical models through detailed analysis of the physicochemical characteristics of the extraction process. This process highly relies on the participation of experts with extensive domain knowledge to ensure that the model can reflect the actual process. The other is data-driven dynamic modeling, which uses a large amount of process data to train the model to achieve real-time simulation and prediction of the process. However, due to the complex physicochemical changes involved in rare earth extraction, characterized by complex mechanisms, multiple stages, and strong inter-stage coupling, mechanistic modeling neglects the influence of many factors. Therefore, the established mechanistic models cannot accurately predict the changes in the rare earth element content during the extraction process. To date, no mechanistic model can comprehensively and accurately depict all the details of the rare earth extraction and separation process; all of them contain a certain degree of idealized assumptions. At the same time, with the continuous advancement of data processing technology and data acquisition systems, data-driven modeling methods trained based on historical data are increasingly favored. This method does not rely on extensive domain expert experience; sufficient historical production data is sufficient to effectively train the model.

[0004] In data-driven modeling methods, due to the superior performance of neural network algorithms, there are already numerous research cases applying neural networks to rare earth extraction modeling. However, rare earth extraction involves complex physicochemical changes, characterized by complex mechanisms, numerous stages, and strong inter-stage coupling. Furthermore, some mechanisms remain unclear. Existing data-driven methods typically employ shallow neural networks—black-box models—to model extraction and separation processes involving dozens or even hundreds of stages. The prediction results are output in a single dimension at the final network output layer. Moreover, in complex industrial scenarios with varying operating conditions, rare earth extraction suffers from data scarcity. This leads to neural networks being unable to handle the complex nonlinear relationships due to sample scarcity, resulting in insufficient learning ability and an inability to accurately describe the hierarchical output of component content. Furthermore, these methods fail to consider the different mapping relationships between component content at each stage of the extraction tank and the input features, as well as the coupling between adjacent extraction tanks caused by component transfer, exchange, and cascade effects. This results in component content values ​​that cannot simulate the actual changes in component content across different extraction tanks. Therefore, for the simulation of multi-stage complex rare earth extraction processes, existing methods using shallow neural networks have limitations and insufficient prediction accuracy. Summary of the Invention

[0005] The purpose of this application is to provide a method, apparatus and medium for constructing and applying a rare earth extraction simulation model, which can construct a rare earth extraction simulation model that can accurately simulate the actual changes in the component content in different extraction tanks, thereby improving the accuracy of rare earth extraction simulation.

[0006] To achieve the above objectives, this application provides the following solution:

[0007] Firstly, this application provides a method for constructing a rare earth extraction simulation model, comprising:

[0008] Obtain historical datasets of rare earth extraction under different operating conditions; each sample in the historical dataset of rare earth extraction includes historical production process characteristic parameters of rare earth extraction and the actual values ​​of component content of each separation stage.

[0009] A multi-branch generative adversarial network (GAN) is trained using a historical dataset of rare earth extraction to obtain a rare earth extraction simulation model. The GAN includes a main network and multiple first branch networks. Each separation stage in the extraction process corresponds to a first branch network. The main network includes a series of input layers and multiple hidden layers. The output of each hidden layer is connected to the input of each first branch network. Each first branch network includes a generator, a discriminator, and a regressor. The input of the generator is connected to the output of the corresponding hidden layer. The input of the regressor is connected to the output of the generator, and the output of the regressor is the simulated result of the component content of the corresponding separation stage in the extraction process. During training, the input of the regressor also includes the actual value of the component content of the corresponding separation stage.

[0010] Secondly, this application provides a method for applying a rare earth extraction simulation model, including:

[0011] Obtain the characteristic parameters of the rare earth extraction production process to be simulated;

[0012] The characteristic parameters of the rare earth extraction production process to be simulated are input into the rare earth extraction simulation model to obtain the simulation results of the content of each component in the rare earth extraction process. The rare earth extraction simulation model is constructed based on the above-mentioned construction method of the rare earth extraction simulation model. The rare earth extraction simulation model includes the main network and the second branch network in the multi-branch generative adversarial network. The second branch network includes the generator and the regressor of the first branch network in the multi-branch generative adversarial network.

[0013] Thirdly, this application provides an apparatus for constructing a rare earth extraction simulation model, comprising:

[0014] The sample acquisition module is used to acquire historical datasets of rare earth extraction under different operating conditions; each sample in the historical dataset of rare earth extraction includes historical production process characteristic parameters of rare earth extraction and the actual values ​​of component content of each separation stage.

[0015] The model building module is used to train a multi-branch generative adversarial network (GAN) using a historical dataset of rare earth extraction to obtain a rare earth extraction simulation model. The multi-branch GAN includes a main network and multiple first branch networks. Each separation stage in the extraction process corresponds to a first branch network. The main network includes a series of input layers and multiple hidden layers. The output of each hidden layer is connected to the input of each first branch network. Each first branch network includes a generator, a discriminator, and a regressor. The input of the generator is connected to the output of the corresponding hidden layer. The input of the regressor is connected to the output of the generator. The output of the regressor is the simulated result of the component content of the corresponding separation stage in the extraction process. During training, the input of the regressor also includes the actual value of the component content of the corresponding separation stage.

[0016] Fourthly, this application provides an application device for a rare earth extraction simulation model, comprising:

[0017] The parameter acquisition module is used to acquire the characteristic parameters of the rare earth extraction production process to be simulated.

[0018] The rare earth extraction simulation module is used to input the characteristic parameters of the rare earth extraction production process to be simulated into the rare earth extraction simulation model to obtain the simulation results of the content of each component in the rare earth extraction process. The rare earth extraction simulation model is constructed based on the above-mentioned construction method of the rare earth extraction simulation model. The rare earth extraction simulation model includes a main network and a second branch network in a multi-branch generative adversarial network. The second branch network includes the generator and regressor of the first branch network in the multi-branch generative adversarial network.

[0019] Fifthly, this application provides a computer device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-described method for constructing a rare earth extraction simulation model, or the above-described method for applying a rare earth extraction simulation model.

[0020] Sixthly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described method for constructing a rare earth extraction simulation model or the above-described method for applying a rare earth extraction simulation model.

[0021] According to the specific embodiments provided in this application, the following technical effects are disclosed:

[0022] This application provides a method, apparatus, and medium for constructing and applying a rare earth extraction simulation model. A multi-branch generative adversarial network (GAN) is designed, with each branch corresponding to a first-level extraction cell. The branches are connected in series through hidden layers in the main network to jointly model extraction cells of different levels. The main network uses multiple hidden layers to learn the original features level by level. The main network contains hidden layers of the same level as the extraction stage, reflecting the hierarchical output of component content in rare earth production. Multiple independent branch networks are introduced under the main network. Each branch network first introduces a generator that is mapped level by level to the latent space to generate samples, enabling the generator of each branch network to extract the features learned by the hidden layer of the previous branch network. The multi-branch GAN designed in this application considers the different mapping relationships between component content at each level of the extraction cell and the input features, as well as the actual characteristics that component transfer, exchange, and cascade effects can cause coupling between adjacent extraction cells. Therefore, applying the multi-branch GAN designed in this application can improve the accuracy of rare earth extraction simulation. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of this application 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 this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 This is an application environment diagram of a method for constructing a rare earth extraction simulation model according to an embodiment of this application;

[0025] Figure 2 A flowchart illustrating a method for constructing a rare earth extraction simulation model according to an embodiment of this application;

[0026] Figure 3 A schematic diagram illustrating the technical concept of a method for constructing a rare earth extraction simulation model according to an embodiment of this application;

[0027] Figure 4 This is a schematic diagram of the structure of a multi-branch generative adversarial network provided in an embodiment of this application;

[0028] Figure 5 This is a schematic diagram of the main network structure provided in an embodiment of this application;

[0029] Figure 6 This is a schematic diagram of the structure of a discriminator and regressor network provided in an embodiment of this application;

[0030] Figure 7 A flow chart of a rare earth extraction process provided in an embodiment of this application;

[0031] Figure 8 A multi-branch generative adversarial network provided in an embodiment of this application provides a simulated prediction curve of the organic phase content at level 25 in the test set;

[0032] Figure 9 A multi-branch generative adversarial network provided in an embodiment of this application provides a simulated prediction curve of the water phase content at level 25 in the test set;

[0033] Figure 10 A comparison chart of the Pr component content distribution and model prediction provided in the 50th example of an embodiment of this application;

[0034] Figure 11 A simulation result of the rare earth extraction process for the 36th test set sample provided in an embodiment of this application;

[0035] Figure 12 A simulation result of the rare earth extraction process for the 50th test set sample provided in an embodiment of this application;

[0036] Figure 13 A simulation result of the rare earth extraction process for the 142nd test set sample provided in an embodiment of this application;

[0037] Figure 14 A schematic diagram of the functional modules of a device for constructing a rare earth extraction simulation model according to an embodiment of this application;

[0038] Figure 15 A flowchart illustrating an application method of a rare earth extraction simulation model provided in an embodiment of this application;

[0039] Figure 16 A schematic diagram of the functional modules of an application device for a rare earth extraction simulation model provided in an embodiment of this application;

[0040] Figure 17 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0041] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0042] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0043] The method for constructing a rare earth extraction simulation model provided in this application embodiment can be applied to, for example... Figure 1The application environment shown depicts a scenario where the terminal communicates with the server via a network. The data storage system stores the data the server needs to process. This data storage system can be configured independently, integrated into the server, or located in the cloud or on another server. The terminal can send historical datasets of rare earth extraction under different operating conditions (each sample in the historical dataset includes historical production process characteristic parameters of rare earth extraction and the actual values ​​of component content for each separation stage) to the server. After receiving the historical datasets of rare earth extraction under different operating conditions, the server uses the datasets to train a multi-branch generative adversarial network to obtain a rare earth extraction simulation model. The multi-branch generative adversarial network includes a main network and multiple first branch networks. Each separation stage in the extraction process corresponds to a first branch network. The main network includes a sequentially connected input layer and multiple hidden layers. The output of each hidden layer is connected to the input of each first branch network. Each first branch network includes a generator, a discriminator, and a regressor. The input of the generator is connected to the output of the corresponding hidden layer. The input of the regressor is connected to the output of the generator, and the output of the regressor is the simulation result of the component content for the corresponding separation stage in the extraction process. During training, the input of the regressor also includes the actual values ​​of component content for the corresponding separation stage. In addition, in some embodiments, the method for constructing the rare earth extraction simulation model can also be implemented by the server or the terminal alone. For example, the terminal can directly construct the model based on the historical dataset of rare earth extraction under different working conditions, or the server can obtain the historical dataset of rare earth extraction under different working conditions from the data storage system and construct the model.

[0044] The terminal can be, but is not limited to, various desktop computers, laptops, smartphones, tablets, IoT devices, and portable wearable devices. The server can be a standalone server or a server cluster composed of multiple servers, or it can be a cloud server.

[0045] In one exemplary embodiment, such as Figure 2 and Figure 3 As shown, a method for constructing a rare earth extraction simulation model is provided. This method is executed by a computer device, specifically by a terminal or server alone, or by both a terminal and a server. In this embodiment, the method is applied to... Figure 1 The following steps, 101 to 102, are used as an example to illustrate the process of using a server in the example.

[0046] Step 101: Obtain historical datasets of rare earth extraction under different working conditions (i.e., Figure 3 The original rare earth dataset); each sample in the rare earth extraction history dataset includes historical production process characteristic parameters of rare earth extraction and the actual values ​​of component content for each separation stage.

[0047] For the actual rare earth extraction production process, based on production indicators and process requirements, characteristics such as feed liquid component content, relative separation coefficients between components, detergent flow rate, extractant flow rate, feeding method, outlet indicators, and outlet separation coefficient are obtained for each production condition. Then, through sampling and analysis, the component content values ​​at each stage when extraction and separation reach dynamic equilibrium are obtained. This leads to the construction of a historical rare earth extraction dataset, where each sample includes a set of production process characteristic parameters and the corresponding actual component content values ​​for each separation stage. The historical rare earth extraction dataset contains sufficient extraction data under different production conditions.

[0048] Step 102: Train a multi-branch generative adversarial network using a historical dataset of rare earth extraction to obtain a rare earth extraction simulation model; such as... Figure 4 and Figure 5 As shown, the multi-branch generative adversarial network includes a main network and multiple first branch networks; each separation stage in the extraction process corresponds to a first branch network; the main network includes an input layer and multiple hidden layers connected in series; the output of each hidden layer is connected to the input of each first branch network in a one-to-one correspondence; each first branch network includes a generator, a discriminator, and a regressor; the input of the generator is connected to the output of the corresponding hidden layer, the output of the generator is connected to the input of the discriminator and the input of the regressor respectively, the input of the discriminator also includes the actual value of the component content of the corresponding separation stage, the output of the discriminator is connected to the input of the generator, and the output of the regressor is the simulated result of the component content of the corresponding separation stage in the extraction process. During training, the input of the regressor also includes the actual value of the component content of the corresponding separation stage.

[0049] like Figure 6 As shown, the input of discriminator D is processed through three convolutional layers (with channels of 32, 64, and 128 respectively) and the LeakyReLU activation function to extract features step by step, and then undergoes flattening processing (corresponding to...). Figure 6 After passing through the FLATTEN layer, the input is fed into a fully connected layer (220-dimensional and 1-dimensional), and then through a Sigmoid activation function to output the probability (corresponding to...). Figure 6 In The input to the regressor R is flattened after passing through a convolutional layer (32 output channels) and LeakyReLU activation, and then through two fully connected layers (1024-dimensional and 2-dimensional) to output the simulation results of the component content corresponding to the separation level (corresponding to...). Figure 6 In ). Figure 6In this context, "Bs" represents the batch size, and "Dense" represents the fully connected layer. The parameters of the first convolutional layer of the regressor R are the same as those of the first convolutional layer of the discriminator D, i.e., they share convolutional layers. Therefore, the discriminator D consists of a first convolutional layer, a second convolutional layer, a third convolutional layer, a first fully connected layer, and a second fully connected layer connected in sequence. The regressor R consists of a fourth convolutional layer, a third fully connected layer, and a fourth fully connected layer connected in sequence.

[0050] like Figure 4 , Figure 5 , Figure 6 and Figure 7 As shown, for the rare earth extraction process, based on the process design, the number of stages for the extraction section and the washing section are determined to be as follows: and Then the total number of separation levels is , No. Each stage is a feed stage. Each stage corresponds to one branch of a multi-branch generative adversarial network (GAN). Each stage contains an organic phase and an aqueous phase. Since the sum of the contents of the difficult-to-extract and easy-to-extract components in the same phase is 1, this application simplifies the structure of the GAN by considering only the easy-to-extract component A in the organic phase and the difficult-to-extract component B in the aqueous phase as the outputs of the GAN. Each first branch network has two outputs: the easy-to-extract component value in the organic phase and the difficult-to-extract component value in the aqueous phase. Considering the actual characteristics of rare earth extraction sites, component transfer and cascade effects can cause coupling between adjacent extraction cells. Therefore, a multi-branch GAN is designed, where each branch network corresponds to one extraction cell. The branch networks are connected in series through the hidden layers in the main network to jointly model extraction cells of different stages. The main network uses multiple hidden layers to learn the original features step by step. A hidden layer, the process closely mirroring the hierarchical output of component content in rare earth production. An additional layer is introduced under the main network. Each first branch network (first branch network) first introduces a generator G structure (e.g., a fully connected layer) to progressively map to the latent space to generate samples. This allows the generator of each first branch network to extract features learned from the hidden layers of the previous first branch network. An independent discriminator D is then introduced under each first branch network. Each discriminator, constrained by its own loss function, forces the generator to learn different patterns in the dataset. A regressor R is added after the discriminator in each first branch network. The discriminator and regressor share a shallow network; that is, the parameters of the first convolutional layer of the regressor and the discriminator are shared. Low-level features of each first branch network can be captured and shared by the shallow network. The regressor of each first branch network has two outputs: easily extractable component A in the organic phase and difficult-to-extract component B in the aqueous phase. Assume that a total of [number missing] samples are extracted from the historical rare earth extraction dataset. The input features, i.e., the input dimensions of the rare earth extraction historical dataset, are: The specific formula for a multi-branch generative adversarial network, analogous to a multi-input multi-output model, is as follows.

[0051]

[0052] Among them, the extraction process Each production process characteristic parameter is used as a condition variable, where the condition variable is represented as follows: Condition variables and noise spliced ​​together As input to a multi-branch generative adversarial network, The dependent variable is Real data, They represent levels 1 to 12 respectively. The actual value of the component content in the first stage of extraction; Refers to the easily extractable component A in the organic phase output by branch network j; Refers to the difficult-to-extract component B in the aqueous phase output of branch network j. Refers to the j-th stage rare earth extraction simulation process. .

[0053] The specific formulas for the main network and the generator G are as follows:

[0054]

[0055] in, Indicates the main network Dimensional input, including ; This represents the feature transfer process from the input layer to the first hidden layer in the main network; Indicates the first in the main network Hidden layer to the first The feature transfer process of each hidden layer; among which... This represents the nonlinear activation function of the first hidden layer; This is the bias vector for the first hidden layer; This represents the weight parameter matrix of the first hidden layer; Indicates the first Nonlinear activation functions for each hidden layer; For the first The bias vectors of each hidden layer; Indicates the first The weight parameter matrix of each hidden layer; For the first The output of each hidden layer Indicates the number of neurons in the hidden layer; Represents the first branch network Generator Activation function; and Represents the first branch network Generator The weight matrix and bias vector are used. To stabilize features and prevent shift, the hidden layer and generator are selected... The function is used as an activation function, assuming the weighted input of feature propagation is... The activation function is ,in express function.

[0056] Among them, the discriminator in the first branch network j and regressor The specific formulas are as follows:

[0057]

[0058]

[0059] in, Represents the generator in the first branch network j. The generated data and the corresponding real data A one-dimensional sequence, This represents the actual content of the component in stage j during the extraction process; , , , , They represent the discriminator. The weight matrices for the first to fifth layers; Represents the generator in branch j of the network. The generated data and input feature variables The lower bound of mutual information between them; , , , , Discriminator The biases of the first to fifth layers; RELU() and LeakyRELU() represent activation functions; BatchNorm() represents batch normalization; TANH() represents the tanh activation function; , , Representing the regressor The weight matrices of the first and third layers; , , Representing the regressor The bias from the first to the third layer.

[0060] In another exemplary embodiment of this application, in step 102, considering that the original feature parameters in the original rare earth extraction historical dataset have different dimensions and varying values, which is not conducive to training the multi-branch generative adversarial network using the stochastic gradient descent algorithm, it is necessary to eliminate the limitation of different dimensions and remove unrealistic sample data from the original dataset through data cleaning and other methods. Specifically, the production process feature parameters in the original rare earth extraction historical dataset are standardized to ensure that the original dataset conforms to a standard normal distribution, i.e., a mean of 0. The features in the original dataset are standardized, where, These are the production process characteristic parameters from the original rare earth extraction historical dataset. and These are the corresponding production process characteristic parameters. The mean and variance, These are the standardized production process characteristic parameters. Therefore, in step 102, a multi-branch generative adversarial network is trained using the rare earth extraction historical dataset to obtain a rare earth extraction simulation model, including:

[0061] (a1) Perform data cleaning and standardization preprocessing on the historical dataset of rare earth extraction to obtain the preprocessed historical dataset of rare earth extraction.

[0062] (a2) A multi-branch generative adversarial network was trained using the preprocessed rare earth extraction historical dataset to obtain a rare earth extraction simulation model.

[0063] In another exemplary embodiment of this application, in step 102, during the training of the multi-branch generative adversarial network, backpropagation and stochastic gradient descent are used to iteratively update the target loss function defined by each first branch network. Training continues until the target loss function of each branch meets the preset convergence criterion, at which point training is considered complete, and the optimized multi-branch generative adversarial network is obtained. To avoid potential overfitting risks and low training efficiency caused by fluctuations in the learning curve of the target loss function during training, this application adopts a strategic adjustment after the training iterations reach a certain number, reducing the initial learning rate by a predetermined proportion. The training termination condition is set to reach the preset maximum number of iterations or the target loss function of each branch converges to a specified threshold. Therefore, in step 102, the multi-branch generative adversarial network is trained using a rare earth extraction historical dataset to obtain a rare earth extraction simulation model, including:

[0064] (b1) Construct training and test sets based on the historical dataset of rare earth extraction and set the initial learning rate.

[0065] (b2) Input the training set into a multi-branch generative adversarial network and output the simulation results of the component content of the corresponding separation stage during the extraction process.

[0066] (b3) Calculate the loss function gradient of each hidden layer and the corresponding first branch network based on the simulation results of the component content of each separation stage and the actual values ​​of the component content of the corresponding separation stage.

[0067] (b4) Adjust the model parameters of each hidden layer and the corresponding first branch network according to the backpropagation of the loss function gradient.

[0068] (b5) Determine whether the current iteration number has reached the preset iteration number. If so, reduce the initial learning rate according to the preset ratio to obtain the reduced learning rate.

[0069] (b6) Replace the initial learning rate with the reduced learning rate and return to the step "Input the training set into the multi-branch generative adversarial network" until the gradient of each loss function converges, thus obtaining the trained multi-branch generative adversarial network, i.e., the rare earth extraction simulation model.

[0070] Before training the multi-branch generative adversarial network, the objective optimization function of the multi-branch generative adversarial network is first determined, as shown in the following formula:

[0071]

[0072] in, Expressing expectations; Discriminator Weight parameters; Represents generator and the weight parameters of the j-th hidden layer; Input data for generating adversarial networks; These are the actual label values, i.e., the actual content of each component. For generator The generated samples, and These represent the output scores of the discriminator for the generated samples and the real samples, respectively. ,in, Indicates the distribution of interpolated sampled real data. and the distribution of generated data A new numerical point is generated from the points between them as follows: ,have . Is it the discriminator in The norm of the gradient at that point; This represents the data distribution consisting of noise z and condition variable C, where z is random Gaussian noise; hyperparameters This represents the weights of the constraints imposed on the generator G by the balanced discriminator D and the regressor R. This indicates that when its input is a real sample At that time, the predicted value made by the regression model R; Regressor Network weight parameters; This indicates that when the input to the regressor R is the generated data... At that time, the variational method is used to estimate With input feature variables Mutual information between The lower bound of mutual information is determined by minimizing... The negative log-likelihood is obtained, and the specific process is as follows: formula.

[0073]

[0074] in, Indicates the total number of samples; This represents the input feature variable (production process feature parameter) corresponding to the h-th sample.

[0075] The specific formulas for the loss optimization functions of each first branch network are as follows.

[0076]

[0077]

[0078] in, This represents the loss optimization function of the discriminator D. This represents the loss optimization function of the discriminator G; This represents the actual value of the component content in stage j of the extraction process. ; Indicates the first The predicted component content of the h-th sample output by the first branch network; and These represent the discriminator generating the corresponding samples. and real samples The output score.

[0079] Training is performed using backpropagation and stochastic gradient descent algorithms. During backpropagation, the neural network models of the discriminator D and generator G are trained and parameterized. According to the chain rule The gradient can be propagated from the output layer to the input layer layer by layer through a chain-like computation, and correspondingly there are... Indicates the first The cumulative weight matrix and bias vector of the generator and its corresponding hidden layer in the first branch network. Indicates the first The weight matrix and bias vector accumulated from the convolutional and fully connected layers of the discriminator in the first branch network. Therefore, the loss functions of the discriminator D and the generator G can be calculated. and Relative to parameters and parameters gradient value and Finally, gradient descent is used to update the value, and the formula for calculating the gradient is shown below.

[0080]

[0081] in, and Let represent the gradient of the loss function of the discriminator D and the gradient of the loss function of the generator G, respectively. and The learning rate; This represents the weight parameters of the discriminator; The weight parameters represent the generators (including the hidden layers of the main network and the generators of the first branch network); and This represents the gradient.

[0082] Maximize the first branch network through supervised learning. The loss function of the discriminator Minimize the first branch network The loss function of the generator The gradient descent algorithm is used to update the weight parameters of the discriminator. and generator weight parameters When the first The weight matrix of the hidden layer corresponding to the first branch network and the weight matrix of the first branch network. When the bias vectors of the hidden layers corresponding to the first branch networks all converge to the set threshold, the th... The first branch of the network stops training. (This is a judgment / determination / initiative). Is it greater than or equal to? ,if Greater than or equal to This indicates that the target loss of each first branch network has converged, and the trained multi-branch generative adversarial network is output; if Less than Then let Continue training.

[0083] In another exemplary embodiment of this application, after performing the step "training a multi-branch generative adversarial network using a rare earth extraction historical dataset to obtain a rare earth extraction simulation model", the method for constructing the rare earth extraction simulation model further includes:

[0084] The rare earth extraction simulation model was tested using a test set to obtain test evaluation metrics. The simulation was stopped when the accuracy reached the set threshold, yielding the final multi-branch generative adversarial network.

[0085] Rare earth extraction involves complex physicochemical changes, characterized by complex mechanisms, multiple levels, and strong coupling between levels. Some of these mechanisms remain unclear. For data-driven modeling, existing shallow neural network modeling methods are essentially black-box models, with predictions output in a single dimension at the final output layer. Furthermore, in complex industrial scenarios with varying operating conditions, rare earth extraction suffers from data scarcity. This leads to neural networks being unable to handle the complex nonlinear relationships due to the scarcity of samples, resulting in insufficient learning ability and an inability to accurately describe the hierarchical output of component content. This invention addresses the problem of insufficient prediction accuracy in existing neural network modeling methods by proposing a multi-branch generative adversarial network (GAN). It considers the different mapping relationships between the component content at each level of the extraction tank and the input features, as well as the actual characteristics of coupling between adjacent extraction tanks due to component transfer, exchange, and cascade effects. It also considers the insufficient learning ability of shallow neural networks to learn complex nonlinear relationships due to the scarcity of on-site data. By connecting first-branch networks in series in each hidden layer, a multi-branch GAN is constructed to jointly model extraction tanks of different levels. This multi-branch GAN includes a main network (containing multiple series-connected hidden layers) and multiple first-branch networks (each branch network includes a generator, discriminator, and regressor). A target loss optimization function is established for each first-branch network and its corresponding hidden layer. Then, using backpropagation and stochastic gradient descent, the target loss function defined by each first-branch network in the model is iteratively updated to obtain the trained multi-branch GAN. Ultimately, the simulated component content at each level is closer to the actual values ​​in the extraction production site, and the model process closely reflects the hierarchical output of component content in actual rare earth production.

[0086] In this application, the extraction process comprises: Level, multi-branch generative adversarial networks correspond to A first branch network is introduced sequentially from the first to the last level of the main network, introducing a first branch network every other hidden layer. This is done considering the different mapping relationships between the component content at each level of the extraction tank and the input features, the coupling between adjacent extraction tanks due to component transfer and cascade effects, and the insufficient learning ability caused by scarce field data. A shallow feature sharing mechanism (i.e., sharing the parameters of the first convolutional layer) is proposed between the discriminator and the regressor to improve the regressor's performance while reducing model redundancy. The hidden layers corresponding to each first branch network form a sub-generative adversarial model, preventing the generator from receiving fixed feedback from the discriminator. This allows the generator to utilize its deep feature learning ability to generate training samples containing feature information from each branch sequentially. Each branch's independent regressor can predict component content, conforming to the complexity distribution of the extraction and separation reaction. A target loss optimization function is established for each branch for training. The multi-branch generative adversarial network outputs the content values ​​of easily extractable organic phase components and difficult-to-extract aqueous phase components for each level corresponding to the branch. During the training of generative adversarial networks, the generator can expand the data by generating new samples based on the original scarce dataset through adversarial training. This is more reasonable than other networks that only use the original scarce dataset during training, and solves the problem of insufficient learning ability caused by the scarcity of on-site data.

[0087] The following example verifies the effectiveness of the method in this application. The example involves the separation of four elements: lanthanum, cerium, praseodymium, and neodymium (La, Ce, Pr, Nd), with a feed rate MF of 1.00 mol. Analysis yields the mole fractions fa, fb, fc, and fd of each element in the feed solution, under the condition fa + fb + fc + fd = MF. The separation process needs to consider the first stage, where the component is cut into cerium, with lanthanum-cerium and praseodymium-neodymium products obtained at the two outlets respectively. For example... Figure 10 As shown, the separation process has 58 stages, with the 39th stage being the feed stage. The feed liquid component distribution is [0.2543, 0.4912, 0.0486, 0.2012]. Furthermore, it can be seen that the predicted component distribution values ​​basically fit the actual value curve, proving the effectiveness of the method in this application.

[0088] like Figure 8 and Figure 9 As shown, in the extraction process, the fluctuations and changes in component content in the middle extraction tank are more complex and frequent compared to those at the two ends. A multi-branch generative adversarial network (GAN) was used to compare the predicted values ​​of easily extractable organic phase component A and difficult-to-extract aqueous phase component B in 200 test samples. The GAN showed the best curve fit to the actual values ​​for both the organic and aqueous phases in the test set, demonstrating high accuracy. This indicates that the proposed GAN can learn the complex nonlinear relationship between the operating parameters of the rare earth extraction process and the content of components at each stage with high accuracy. Figure 8 and Figure 9 In this context, TRUE refers to the actual true value of the component content data; BPNN, CGAN, and MBRDN represent BP neural network, conditional generative adversarial network, and multi-branch residual deep network, respectively; MB-RGAN represents the multi-branch generative adversarial network proposed in this application.

[0089] Figure 11 , Figure 12 , Figure 13 The output simulated curves of the component content of each branch of the multi-branch generative adversarial network are shown, corresponding to the feed stage in the example sample. The values ​​are 36, 105, and 142. From Figure 8 , Figure 9 , Figure 10 It can be seen from this that the easily extracted component A is in The content of the organic phase outlet component is close to 1, and the content of the difficult-to-extract component B at the first-stage aqueous phase outlet component is close to 1. This simulates the actual changes in component content in different extraction tanks, and the numerical error is within an acceptable range. The simulation reflects the changes in component content at each stage during the actual extraction process.

[0090] Based on the above, the content of each component obtained by simulating the rare earth extraction process using the method proposed in this application is close to the actual value on the extraction production site. It can truly reflect the rare earth extraction process and enable production technicians to judge the production situation and adjust the production process characteristic parameters in a timely manner.

[0091] In one exemplary embodiment, such as Figure 14 As shown, a device for constructing a rare earth extraction simulation model is proposed, comprising:

[0092] The sample acquisition module M1 is used to acquire historical datasets of rare earth extraction under different operating conditions; each sample in the historical dataset of rare earth extraction includes historical production process characteristic parameters of rare earth extraction and the actual values ​​of component content of each separation stage.

[0093] The model building module M2 is used to train a multi-branch generative adversarial network (GAN) using a historical dataset of rare earth extraction to obtain a rare earth extraction simulation model. The multi-branch GAN includes a main network and multiple first branch networks. Each separation stage in the extraction process corresponds to a first branch network. The main network includes a series of input layers and multiple hidden layers. The output of each hidden layer is connected to the input of each first branch network. Each first branch network includes a generator, a discriminator, and a regressor. The input of the generator is connected to the output of the corresponding hidden layer. The output of the generator is connected to the input of the discriminator and the input of the regressor. The input of the discriminator also includes the actual value of the component content of the corresponding separation stage. The output of the discriminator is connected to the input of the generator. The output of the regressor is the simulated result of the component content of the corresponding separation stage in the extraction process. During training, the input of the regressor also includes the actual value of the component content of the corresponding separation stage.

[0094] In another exemplary embodiment of this application, a method for applying a rare earth extraction simulation model is proposed, such as... Figure 15 As shown, it includes:

[0095] Step 201: Obtain the characteristic parameters of the rare earth extraction production process to be simulated.

[0096] Step 202: Input the characteristic parameters of the rare earth extraction production process to be simulated into the rare earth extraction simulation model to obtain the simulation results of the content of each component in the rare earth extraction process; the rare earth extraction simulation model is constructed based on the above-mentioned construction method of the rare earth extraction simulation model; the rare earth extraction simulation model includes the main network and the second branch network in the multi-branch generative adversarial network; the second branch network includes the generator and regressor of the first branch network in the multi-branch generative adversarial network.

[0097] Based on the same inventive concept, this application also provides an application device for a rare earth extraction simulation model to implement the application method of the rare earth extraction simulation model described above. The solution provided by this device is similar to the solution described in the above method. Therefore, the specific limitations of the application device embodiments of one or more rare earth extraction simulation models provided below can be found in the limitations of the application method of the rare earth extraction simulation model above, and will not be repeated here.

[0098] In one exemplary embodiment, such as Figure 16 As shown, an application device for a rare earth extraction simulation model is proposed, comprising:

[0099] The parameter acquisition module T1 is used to acquire the characteristic parameters of the rare earth extraction production process to be simulated.

[0100] The rare earth extraction simulation module T2 is used to input the characteristic parameters of the rare earth extraction production process to be simulated into the rare earth extraction simulation model to obtain the simulation results of the content of each component in the rare earth extraction process. The rare earth extraction simulation model is constructed based on the above-mentioned construction method of the rare earth extraction simulation model. The rare earth extraction simulation model includes the main network and the second branch network in the multi-branch generative adversarial network. The second branch network includes the generator and regressor of the first branch network in the multi-branch generative adversarial network.

[0101] This application also provides an application scenario in which the above-described rare earth extraction simulation model application method is applied. Specifically, the rare earth extraction simulation model application method provided in this embodiment can be applied in a rare earth extraction simulation scenario. This scenario includes a data acquisition stage, a rare earth extraction simulation stage, and a rare earth extraction process control stage; the data acquisition stage is used to collect the production process characteristic parameters of rare earth extraction production; the rare earth extraction simulation stage is used to perform rare earth extraction simulation based on the collected production process characteristic parameters; and the rare earth extraction process control stage is used to adjust the production process characteristic parameters in the rare earth extraction production process based on the rare earth extraction simulation results. The rare earth extraction simulation model application method provided in this embodiment belongs to the rare earth extraction simulation stage.

[0102] In one exemplary embodiment, a computer device is provided, which may be a server or a terminal, and its internal structure diagram may be as follows. Figure 17 As shown. This computer device (i.e., computer equipment) includes a processor, memory, input / output interfaces (I / O), and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operation of the operating system and computer programs in the non-volatile storage media. The database stores video tag processing data. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communicating with external terminals via a network connection. When the computer program is executed by the processor, it implements a method for constructing a rare earth extraction simulation model or a method for applying a rare earth extraction simulation model.

[0103] Those skilled in the art will understand that Figure 17The structures shown are merely block diagrams of some structures related to the present application and do not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than shown in the figures, or combine certain components, or have different component arrangements. In an exemplary embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0104] In one exemplary embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0105] It should be noted that the data involved in this application (including but not limited to data used for analysis, data stored, data displayed, etc.) are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0106] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0107] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, 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 this application.

Claims

1. A method for constructing a rare earth extraction simulation model, characterized in that, The method for constructing the rare earth extraction simulation model includes: Obtain historical datasets of rare earth extraction under different operating conditions; each sample in the historical dataset of rare earth extraction includes historical production process characteristic parameters of rare earth extraction and the actual values ​​of component content of each separation stage. A multi-branch generative adversarial network (GAN) was trained using a historical dataset of rare earth extraction to obtain a rare earth extraction simulation model. The GAN includes a main network and multiple first branch networks. Each separation stage in the extraction process corresponds to a first branch network. The main network includes a series of input layers and multiple hidden layers. The output of each hidden layer is connected to the input of each first branch network. Each first branch network includes a generator, a discriminator, and a regressor. The input of the generator is connected to the output of the corresponding hidden layer. The input of the regressor is connected to the output of the generator, and the output of the regressor is the simulated result of the component content of the corresponding separation stage in the extraction process. During training, the input of the regressor also includes the actual value of the component content of the corresponding separation stage. The specific formula for multi-branch generative adversarial networks is as follows: In the formula, the extraction process is used as the basis for the expression. Each production process characteristic parameter is used as a condition variable, where the condition variable is represented as follows: Condition variables and noise spliced ​​together As input to a multi-branch generative adversarial network, The dependent variable is Real data, They represent levels 1 to 12 respectively. The actual value of the component content in the first stage of extraction; Refers to the easily extractable component A in the organic phase output by branch network j; Refers to the difficult-to-extract component B in the aqueous phase output of branch network j; Refers to the j-th stage rare earth extraction simulation process. .

2. The method for constructing a rare earth extraction simulation model according to claim 1, characterized in that, The output of the generator is also connected to the input of the discriminator, whose input includes the actual values ​​of the component content corresponding to the separation stage. The output of the discriminator is connected to the input of the generator.

3. The method for constructing a rare earth extraction simulation model according to claim 1, characterized in that, A multi-branch generative adversarial network was trained using a historical dataset of rare earth extraction to obtain a rare earth extraction simulation model, including: Training and test sets were constructed based on historical datasets of rare earth extraction, and an initial learning rate was set. The training set is input into a multi-branch generative adversarial network, which outputs simulated results of the component content at the corresponding separation stage during the extraction process. The loss function gradient of each hidden layer and the corresponding first branch network is calculated based on the simulation results of the component content of each separation stage and the actual values ​​of the component content of the corresponding separation stage. The model parameters of each hidden layer and the corresponding first branch network are adjusted by backpropagation based on the gradient of the loss function. Determine whether the current iteration count has reached the preset iteration count. If so, reduce the initial learning rate according to the preset ratio to obtain the reduced learning rate. The initial learning rate is replaced with the reduced learning rate, and the process returns to the step "input the training set into the multi-branch generative adversarial network" until the gradient of each loss function converges or the maximum number of iterations is reached, thus obtaining the trained multi-branch generative adversarial network, i.e., the rare earth extraction simulation model.

4. The method for constructing a rare earth extraction simulation model according to claim 3, characterized in that, After performing the step "training a multi-branch generative adversarial network using a historical dataset of rare earth extraction to obtain a rare earth extraction simulation model", the method for constructing the rare earth extraction simulation model further includes: The rare earth extraction simulation model was tested using a test set, and test evaluation indicators were obtained.

5. The method for constructing a rare earth extraction simulation model according to claim 1, characterized in that, A multi-branch generative adversarial network was trained using a historical dataset of rare earth extraction to obtain a rare earth extraction simulation model, including: Data cleaning and standardization preprocessing were performed on the historical dataset of rare earth extraction to obtain the preprocessed historical dataset of rare earth extraction. A multi-branch generative adversarial network was trained using a preprocessed historical dataset of rare earth extraction to obtain a rare earth extraction simulation model.

6. A method for applying a rare earth extraction simulation model, characterized in that, The application methods of the rare earth extraction simulation model include: Obtain the characteristic parameters of the rare earth extraction production process to be simulated; The characteristic parameters of the rare earth extraction production process to be simulated are input into the rare earth extraction simulation model to obtain the simulation results of the content of each component in the rare earth extraction process; the rare earth extraction simulation model is constructed based on the construction method of the rare earth extraction simulation model according to any one of claims 1 to 5; the rare earth extraction simulation model includes a main network and a second branch network in a multi-branch generative adversarial network; the second branch network includes the generator and regressor of the first branch network in the multi-branch generative adversarial network.

7. A device for constructing a rare earth extraction simulation model, characterized in that, The apparatus for constructing the rare earth extraction simulation model includes: The sample acquisition module is used to acquire historical datasets of rare earth extraction under different operating conditions; each sample in the historical dataset of rare earth extraction includes historical production process characteristic parameters of rare earth extraction and the actual values ​​of component content of each separation stage. The model building module is used to train a multi-branch generative adversarial network (GAN) using a historical dataset of rare earth extraction to obtain a rare earth extraction simulation model. The multi-branch GAN includes a main network and multiple first branch networks. Each separation stage in the extraction process corresponds to a first branch network. The main network includes a series of input layers and multiple hidden layers. The output of each hidden layer is connected to the input of each first branch network. Each first branch network includes a generator, a discriminator, and a regressor. The input of the generator is connected to the output of the corresponding hidden layer. The input of the regressor is connected to the output of the generator, and the output of the regressor is the simulated result of the component content of the corresponding separation stage in the extraction process. During training, the input of the regressor also includes the actual value of the component content of the corresponding separation stage. The specific formula for multi-branch generative adversarial networks is as follows: In the formula, the extraction process is used as the basis for the expression. Each production process characteristic parameter is used as a condition variable, where the condition variable is represented as follows: Condition variables and noise spliced ​​together As input to a multi-branch generative adversarial network, The dependent variable is Real data, They represent levels 1 to 12 respectively. The actual value of the component content in the first stage of extraction; Refers to the easily extractable component A in the organic phase output by branch network j; Refers to the difficult-to-extract component B in the aqueous phase output of branch network j; Refers to the j-th stage rare earth extraction simulation process. .

8. An application device for a rare earth extraction simulation model, characterized in that, The application device for the rare earth extraction simulation model includes: The parameter acquisition module is used to acquire the characteristic parameters of the rare earth extraction production process to be simulated. A rare earth extraction simulation module is used to input the characteristic parameters of the rare earth extraction production process to be simulated into a rare earth extraction simulation model to obtain the simulation results of the content of each component in the rare earth extraction process; the rare earth extraction simulation model is constructed based on the construction method of the rare earth extraction simulation model according to any one of claims 1 to 5; the rare earth extraction simulation model includes a main network and a second branch network in a multi-branch generative adversarial network; the second branch network includes the generator and regressor of the first branch network in the multi-branch generative adversarial network.

9. A computer device, comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the computer program to implement the method for constructing a rare earth extraction simulation model according to any one of claims 1-5, or the method for applying the rare earth extraction simulation model according to claim 6.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the method for constructing the rare earth extraction simulation model according to any one of claims 1-5, or the method for applying the rare earth extraction simulation model according to claim 6.

Images