Method, device, storage medium and product for dynamically predicting aluminum electrolysis cell sidewall

Abstract

The application discloses a kind of aluminum electrolytic cell cell help dynamic prediction method, equipment, storage medium and product, the prediction method includes according to structural parameter and process parameter constructs aluminum electrolytic cell multi-physical field simulation model, and then obtains the cell help thickness and cell shell temperature under different process parameters, and constructs sample data set;Sample data set is used to train cell help size prediction model, and target prediction model is obtained;Real-time process parameters of the aluminum electrolytic cell to be predicted are predicted using the target prediction model, and cell help thickness prediction value and cell shell temperature prediction value are obtained;Whether target prediction model meets prediction accuracy requirement is judged according to cell help thickness measurement value and cell help thickness prediction value, and aluminum electrolytic cell multi-physical field simulation model is adjusted and target prediction model is retrained according to the judgment result, until meeting prediction accuracy requirement.The application improves prediction accuracy and robustness, and improves the adaptability to complex and changeable working conditions.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of aluminum electrolysis cell, and particularly relates to an aluminum electrolysis cell side wall dynamic prediction method, device, storage medium and product. BACKGROUND

[0002] The aluminum electrolysis cell is a key electrolysis device in the aluminum metallurgy industry, which is used to extract aluminum from aluminum oxide through an electro-metallurgical process. The main components of the aluminum electrolysis cell include a cell shell, a cell cavity, an anode, a cathode, an inner lining, and an electrolyte, etc. The performance and service life of the aluminum electrolysis cell are the most important in the electrolysis process, which are affected by multiple factors, among which the shape, thickness and thermal stability of the side wall directly determine the operation efficiency and safety of the aluminum electrolysis cell. The side wall mainly refers to the two side regions of the aluminum electrolysis cell, which is the most concentrated part of the aluminum electrolysis cell during operation, and is subjected to heat, stress and corrosion.

[0003] With the development of aluminum electrolysis technology, the size of the aluminum electrolysis cell gradually increases (the current maximum reaches 600kA), and the working conditions become more complex and variable, which puts higher requirements on the thermal behavior, structural change and service life of the side wall. Therefore, how to accurately predict and monitor the dynamic change of the side wall, timely find the abnormal state of the side wall and perform fault diagnosis, and then provide feedback for the control of the aluminum electrolysis cell, has become an important technical problem in the aluminum electrolysis production process.

[0004] The existing technology mainly focuses on estimating the thickness of the side wall at the point of the side cell shell temperature measurement, and is not related to the big data of the aluminum electrolysis cell, that is, it is not suitable for prediction under variable working conditions, and can only simply perform damage warning, and cannot perform real-time high-precision calculation on the thickness of the side wall of the aluminum electrolysis cell, so the practical value is poor.

[0005] A Chinese patent document with the authorized publication number CN106709149B discloses a kind of aluminum electrolysis cell three-dimensional hearth shape real-time prediction method and system based on neural network, which is mainly based on BP neural network model to calculate three-dimensional side wall, the method emphasizes the fitting of neural network model to part process parameters and finite element simulation model, can only cover a few application scenarios. SUMMARY

[0006] The present application aims to provide an aluminum electrolysis cell side wall dynamic prediction method, device, storage medium and product, to solve at least one of the problems that the traditional method cannot predict the thickness of the side wall in real time, the prediction accuracy is poor, and it cannot be applied to complex and variable working conditions.

[0007] The present application solves the above technical problems by the following technical solutions: an aluminum electrolysis cell side wall dynamic prediction method, comprising:

[0008] Acquire structural parameters and process parameters of the aluminum electrolysis cell, and construct a multi-physical field simulation model of the aluminum electrolysis cell according to the structural parameters and the process parameters;

[0009] Based on the multi-physical field simulation model of the aluminum electrolysis cell, obtain the cell wall thickness and the cell shell temperature under different process parameters;

[0010] Construct a sample data set according to the cell wall thickness and the cell shell temperature under different process parameters;

[0011] Construct a cell wall size prediction model, train the cell wall size prediction model using the sample data set, and obtain a target prediction model;

[0012] Acquire real-time process parameters of the aluminum electrolysis cell to be predicted, predict the real-time process parameters using the target prediction model, and obtain a cell wall thickness prediction value and a cell shell temperature prediction value;

[0013] Acquire a cell wall thickness measurement value of an anode of the aluminum electrolysis cell to be predicted;

[0014] Determine whether the target prediction model meets the prediction accuracy requirement according to the cell wall thickness measurement value and the cell wall thickness prediction value; if yes, output the cell wall thickness prediction value and the cell shell temperature prediction value under the real-time process parameters; if no, adjust the physical performance parameters of the multi-physical field simulation model of the aluminum electrolysis cell, and repeat the steps of constructing a sample data set, retraining a target prediction model, predicting real-time process parameters, and determining the prediction accuracy of the target prediction model based on the adjusted multi-physical field simulation model of the aluminum electrolysis cell.

[0015] Further, the structural parameters include geometric dimensions and physical performance parameters, the physical performance parameters include the electrical conductivity of each part of the material and the thermal conductivity under different temperature and different service life stage working conditions; the process parameters include temperature field, current distribution, aluminum level and electrolyte level.

[0016] Further, the electrical conductivity and the thermal conductivity of the material are acquired by using a nanoCT method, and the specific implementation process is as follows:

[0017] The acquired CT material data is processed by using Avizo software to obtain a three-dimensional sampling model;

[0018] The three-dimensional sampling model is repaired in a model processing software to obtain a repaired model;

[0019] The repaired model is imported into a mesh drawing software, and the repaired model is meshed to obtain a material mesoscopic model;

[0020] The material mesoscopic model is introduced into a finite element simulation software, and a temperature boundary condition and a thermal conductivity boundary condition are loaded into the material mesoscopic model to calculate comprehensive electrical conductivity and thermal conductivity of the material.

[0021] Further, based on the aluminum electrolysis cell multi-physical field simulation model, the cell wall thickness and the cell shell temperature under different process parameters are obtained, specifically including:

[0022] The process parameters, physical boundary conditions and parameter constraint conditions are loaded on the aluminum electrolysis cell multi-physical field simulation model, and iterative solution is performed to obtain the cell wall thickness and the cell shell temperature under the process parameters.

[0023] Further, the physical boundary conditions and the parameter constraint conditions are optimized by using a multi-scale simulation method.

[0024] Further, a sample data set is constructed according to the cell wall thickness and the cell shell temperature under different process parameters, including:

[0025] The process parameters are denoised and missing data are filled in;

[0026] The processed process parameters are standardized and normalized;

[0027] The normalized process parameters are integrated and format-converted;

[0028] Key features are extracted from the format-converted process parameters by using statistical analysis and machine learning techniques, and the correlation of each key feature with the cell wall thickness and the cell shell temperature is evaluated by combining a deep learning algorithm;

[0029] According to the evaluation results, relevant features are determined, and a sample data set is constructed according to the determined relevant features and the corresponding cell wall thickness and cell shell temperature.

[0030] Further, in the training process of the cell wall size prediction model or the target prediction model, an automatic parameter optimization algorithm based on Bayesian optimization is used to optimize the model parameters.

[0031] Based on the same concept, the present application also provides an electronic device comprising a memory, a processor and a computer program / instruction stored in the memory, wherein the processor executes the computer program / instruction to implement the aluminum electrolysis cell wall dynamic prediction method as described above.

[0032] Based on the same concept, the present application also provides a computer readable storage medium having a computer program / instruction stored thereon, wherein the computer program / instruction is executed by a processor to implement the aluminum electrolysis cell wall dynamic prediction method as described above.

[0033] Based on the same concept, the application also provides a computer program product comprising computer programs / instructions which, when executed by a processor, implement the aluminum reduction cell sidewall dynamic prediction method as described above.

[0034] Advantages

[0035] Compared with the prior art, the application has the advantages that:

[0036] The application is based on real-time process parameters and deep learning technology to perform real-time prediction of sidewall thickness and shell temperature, and judges whether the target prediction model is suitable for the current working condition based on the measured sidewall thickness value, adjusts the multi-physical field simulation model of the aluminum reduction cell and re-trains the target prediction model to make the target prediction model suitable for the prediction of sidewall thickness and shell temperature under complex and variable working conditions, thereby improving the prediction accuracy and robustness and improving the adaptability to complex and variable working conditions. BRIEF DESCRIPTION OF DRAWINGS

[0037] In order to more clearly illustrate the technical solutions of the application, the following will briefly introduce the drawings needed to be used in the embodiments. Obviously, the drawings in the following description are only a part of the application, and other drawings can be obtained by those skilled in the art without any creative effort.

[0038] Figure 1 is a flow chart of the aluminum reduction cell sidewall dynamic prediction method in the embodiment of the application;

[0039] Figure 2 is a CFD simulation result graph of the sidewall inner shape under different pole distances in the embodiment of the application;

[0040] Figure 3 is a visualization schematic diagram of the sidewall dynamic prediction result in the embodiment of the application. DETAILED DESCRIPTION

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

[0042] The technical solutions of the application will be described in detail in the following with specific embodiments. The following specific embodiments can be combined with each other, and some embodiments can not be described again for the same or similar concepts or processes.

[0043] Embodiment 1

[0044] Figure 1A flow chart of the aluminum electrolysis cell sidewall dynamic prediction method provided by the present application is shown in FIG. 1. Figure 1 As shown in FIG. 1, the aluminum electrolysis cell sidewall dynamic prediction method comprises the following steps:

[0045] Step 1: Obtain the structural parameters and process parameters of the aluminum electrolysis cell.

[0046] In this embodiment, the structural parameters include geometric dimensions and physical property parameters, and the physical property parameters include the electrical conductivity of the materials of each part and the thermal conductivity under different temperature and different service life stage working conditions; the process parameters include temperature field, current distribution, aluminum level and electrolyte level, and the temperature field includes side cell shell temperature, bottom cell shell temperature and cathode steel rod temperature, and the current distribution includes anode current. The process parameters of different aluminum electrolysis cells can be obtained by collecting through corresponding sensors (for example, collecting temperature data every 5 seconds by using distributed electric thermocouple sensors), and / or calling historical working condition data stored by using SQL database or HDF5, to provide multi-working condition input for the sidewall size prediction model. The process parameters can also be obtained by manual collection, for example, aluminum level and electrolyte level, and when obtained by manual collection, the manually collected data is automatically corrected based on a machine learning model to eliminate subjective errors. The sensor collected data and the historical working condition data can be combined and stored in HDF5 format with gzip compression to reduce data volume.

[0047] Taking a certain 420kA aluminum electrolysis cell as an example, the structural parameters thereof are shown in Table 1.

[0048] Table 1: Part of the structural parameters of a certain 420kA aluminum electrolysis cell

[0049] Parameter Value Remark Series current / kA 420 Adjustable range: 400 kA ~ 430 kA Anode carbon block size / mm 1700×660×540 Per carbon block size Cathode carbon block size / mm 3680×665×485 Per carbon block size Anode carbon block quantity 48 Uniform distribution Cathode carbon block quantity 24 Uniform distribution Cathode steel rod size / mm 2200×100×200 Single steel rod size Steel rod quantity 96 Arrangement position of each steel rod Aluminum level / mm 220 Dynamic change in operation, real-time calculation required Electrolyte level / mm 180 Dynamic change in operation, real-time calculation required Pole pitch / mm 45 Dynamic change range: 20 ~ 50 mm

[0050] In another specific embodiment of the present application, the operating parameters of the aluminum electrolysis cell can also be obtained (collected every 10 minutes), specifically including set voltage, real-time voltage, bus displacement, furnace bottom pressure drop, discharging frequency, fluorine salt addition amount, temperature and needle vibration, etc., and the basic working state of the aluminum electrolysis cell is obtained through these information. Part of the operating parameters of the aluminum electrolysis cell are shown in Table 2.

[0051] Table 2: Part of the operating parameters of the aluminum electrolysis cell

[0052] Molecular ratio Fe content Si content Furnace bottom pressure drop Alumina concentration … Swing Needle vibration Mean value 2.46 0.10 0.04 315.57 2.96 … 1.23 4.68 Standard deviation 0.09 0.01 0.00 37.76 0.77 … 0.51 2.13 Minimum value 2.17 0.07 0.02 216.00 0.49 … 1.00 1.00 25％ 2.41 0.10 0.03 289.00 2.40 … 1.00 3.00 50％ 2.45 0.10 0.04 310.00 2.85 … 1.00 4.00 75％ 2.50 0.11 0.04 337.00 3.44 … 1.00 6.00 Maximum value 3.12 0.29 0.12 495.00 6.22 … 6.00 23.00 Coefficient of variation 0.04 0.18 0.21 0.12 0.26 … 0.41 0.45

[0053] In order to facilitate calling and checking, all the parameters obtained in step 1 are stored by using SQL database.

[0054] Step 2: Construct a multi-physical field simulation model of the aluminum electrolysis cell according to the structural parameters and process parameters obtained in step 1.

[0055] In the specific embodiment of the application, the specific implementation process of constructing a multi-physical field simulation model of an aluminum electrolysis cell includes:

[0056] A three-dimensional model of the aluminum electrolysis cell is constructed according to the structural parameters of the aluminum electrolysis cell, then the three-dimensional model is imported into the Openfoam platform for meshing, and the model after meshing is given material properties according to the physical performance parameters of different parts.

[0057] In order to improve the accuracy of the physical performance parameters of each part of the material, the nanoCT method is used to obtain the electrical conductivity of the material, and the specific implementation process is as follows:

[0058] Step 2.1: The CT material data (i.e. material data obtained by computer tomography) is processed by using Avizo software to obtain a three-dimensional sampling model;

[0059] Step 2.2: Repair the three-dimensional sampling model in the model processing software (such as Geomagic Studio) to process unreasonable holes, abnormal surfaces and noise points to obtain a repaired model;

[0060] Step 2.3: Import the repaired model into the mesh drawing software (such as fluent meshing), and mesh the repaired model to obtain a material mesoscopic model;

[0061] Step 2.4: Import the material mesoscopic model into the finite element simulation software, and load the temperature boundary condition and thermal conductivity boundary condition into the material mesoscopic model to simulate the heat transfer and loss mechanism of different parts, and calculate the comprehensive electrical conductivity and thermal conductivity of the material.

[0062] The comprehensive electrical conductivity of the material is used as the electrical conductivity parameter of the material, and the nanoCT method is used to describe the dynamic evolution process of the material property parameters affecting the prediction of the cell wall thickness and the cell shell temperature, which more detailedly describes the mesoscopic material characteristics and improves the calculation accuracy of the cell wall thickness and the cell shell temperature under non-standard working conditions.

[0063] The nanoCT method is a prior art, which can be found in the literature: Wang, Gang, et al. "Quantitative analysis of microscopic structure and gas seepage characteristics of low-rank coal based on CT three-dimensional reconstruction of CT images and fractal theory." Fuel 256 (2019): 115900.

[0064] Step 3: Based on the aluminum electrolytic cell multi-physical field simulation model constructed in step 2, the cell wall thickness and cell shell temperature under different process parameters are obtained.

[0065] In the specific embodiment of the present application, based on the aluminum electrolytic cell multi-physical field simulation model, the cell wall thickness and cell shell temperature under different process parameters are obtained, specifically including:

[0066] Load the process parameters, physical boundary conditions and parameter constraint conditions on the aluminum electrolytic cell multi-physical field simulation model, and perform iterative solution to obtain the cell wall thickness and cell shell temperature under the process parameters. The multi-physical field includes electric field, thermal field and flow field, the electric field and thermal field dynamically interact, and the heat conduction equation and electromagnetic field equation are combined to simulate the heat loss and cell wall erosion process of the aluminum electrolytic cell sidewall, and calculate the cell wall thickness and cell shell temperature.

[0067] Generally, the physical boundary conditions and parameter constraint conditions are set according to experience, and the present application adopts a multi-scale simulation method to optimize the physical boundary conditions and parameter constraint conditions set according to experience, calculates the interfacial behavior of the electrolyte and the inner village material by analyzing the physical and chemical properties of the microstructure, and provides more optimal physical boundary conditions and parameter constraint conditions for the aluminum electrolytic cell multi-physical field simulation model. The multi-scale simulation method is prior art.

[0068] The aluminum electrolytic cell multi-physical field simulation model of the present application is obtained by micro-interface behavior, mesoscopic material properties and macro multi-scale modeling method, which can comprehensively describe the dynamic evolution process of several important parameters affecting the prediction of the cell wall, and provide analysis on the problems such as cell wall solidification and melting, multi-layer heat transfer and electrolyte component full life cycle parameter coverage.

[0069] Figure 2 The CFD simulation result graph of the cell wall inner shape under different pole distances is shown, and the numbers on the color band represent the liquid volume fraction, which distinguishes the solid-liquid phase of the electrolyte through the liquid volume fraction.

[0070] Step 4: Construct a sample data set according to the cell wall thickness and cell shell temperature under different process parameters obtained in step 3.

[0071] In the specific embodiment of the present application, the sample data set is constructed according to the cell wall thickness and cell shell temperature under different process parameters, including:

[0072] Step 4.1: Denoising and missing value filling processing are performed on the process parameters to ensure the integrity and accuracy of the data.

[0073] For missing or abnormal values, a interpolation method based on variational autoencoder (VAE) is used for repair, and the interpolation formula is:

[0074]

[0075] wherein x i represents the inserted data, x k , x j represents two known data.

[0076] Step 4.2: Standardization and normalization processing of the processed process parameters are performed to eliminate the dimension effect and facilitate subsequent analysis.

[0077] Step 4.3: Integration and format conversion processing of the normalized process parameters; wherein ETL (i.e. extraction, transformation and loading) process is used for integration to provide a global perspective for subsequent analysis; different sources of data are converted to a unified format (including but not limited to.npy, HDF5), and built-in compression (such as gzip or zlib) is used to reduce storage space.

[0078] According to the type and source of data, the data is stored as multiple files or stored in different data sets in the same file; wherein for unstructured data, image format.png,.jpg or text format.csv is used for storage, and compression technology is combined, which can improve storage efficiency, reduce space occupation, and optimize data access and processing performance.

[0079] Step 4.4: Using statistical analysis and machine learning techniques, key features are extracted from the process parameters after format conversion processing, and deep learning algorithms are used to evaluate the correlation of each key feature with the thickness of the tank and the shell temperature (for example, calculating the correlation coefficient).

[0080] Machine learning techniques include but are not limited to principal component analysis, feature importance evaluation, etc., and deep learning algorithms include but are not limited to convolutional neural networks and graph neural networks.

[0081] Step 4.5: According to the evaluation results, determine the relevant features, and according to the determined relevant features and the corresponding tank thickness and shell temperature, construct a sample data set. That is, according to the correlation size, select features with strong correlation with the tank thickness and shell temperature to improve the effectiveness and reliability of the prediction model. Each sample in the sample data set includes an input sample and an output sample, the input sample is the relevant feature corresponding to the process parameter, and the output sample is the tank thickness and shell temperature.

[0082] In order to cover more complex working conditions, the distributed current intensity of the aluminum electrolysis tank, the cover material thickness, the electrolyte primary crystal temperature, the pole distance and the temperature gradient are also introduced in the input sample.

[0083] Step 5: Constructing a tank size prediction model.

[0084] In this embodiment, the slot side size prediction model selects a convolutional neural network model or a recurrent neural network model.

[0085] Step 6: Train the slot side size prediction model constructed in step 5 using the sample data set constructed in step 4 to obtain a target prediction model.

[0086] During the training process of the slot side size prediction model, an Adam optimizer and an automatic parameter tuning algorithm based on Bayesian optimization are used for model parameter optimization.

[0087] Step 7: Obtain real-time process parameters of the aluminum electrolysis cell to be predicted, and use the target prediction model obtained in step 6 to predict the real-time process parameters to obtain a slot side thickness prediction value and a slot shell temperature prediction value.

[0088] Step 8: Obtain a slot side thickness measurement value of a certain anode of the aluminum electrolysis cell to be predicted.

[0089] The service life of the anode of the aluminum electrolysis cell is limited, and the service life is determined according to the anode residual amount and the service condition. When the service life is reached, the anode needs to be replaced. When the anode is manually replaced, the slot side thickness can be measured, i.e., a slot side thickness measurement value is obtained, so as to determine whether the target prediction model is suitable for the current working condition of the aluminum electrolysis cell to be predicted according to the slot side thickness measurement value and the slot side thickness prediction value.

[0090] Step 9: Determine whether the target prediction model meets the prediction accuracy requirement according to the slot side thickness measurement value obtained in step 8 and the slot side thickness prediction value obtained in step 7; if yes, output the slot side thickness prediction value and the slot shell temperature prediction value under the real-time process parameters; if no, adjust the physical performance parameters of the aluminum electrolysis cell multi-physics field simulation model, and based on the adjusted aluminum electrolysis cell multi-physics field simulation model, repeat steps 3, 4, 6 (retrain the target prediction model), 7 and 9 until the target prediction model meets the prediction accuracy requirement, indicating that the target prediction model can adapt to the current working condition.

[0091] When the target prediction model cannot meet the prediction accuracy requirement, the influence of the heat transfer of the electrolysis cell side, the material performance attenuation and the working condition change on the slot side thickness is comprehensively considered, the physical performance parameters of the aluminum electrolysis cell multi-physics field simulation model are adjusted, the slot side thickness and the slot shell temperature under different process parameters are obtained again based on the adjusted aluminum electrolysis cell multi-physics field simulation model, and then the sample data set is constructed again, the target prediction model is retrained using the sample data set, and then the slot side thickness and the slot shell temperature are predicted using the retrained target prediction model, so as to ensure that the target prediction model can adapt to different working conditions.

[0092] In the specific embodiments of the present application, the present application also develops an open API interface, supports real-time integration of the target prediction model with the industrial control system, builds a real-time feedback closed loop between dynamic prediction and the industrial control system, dynamically optimizes the target prediction model based on the dynamic prediction data of the cell side and the cell side measurement data, expands the application range of the target prediction model, and realizes prediction across specifications and across working conditions. By introducing the open API interface, the target prediction model supports rapid adaptation in different aluminum electrolytic cell specifications and complex working conditions, improving the versatility and flexibility of the prediction model.

[0093] The application of the present application to a certain 420kA aluminum electrolytic cell can predict the thickness of the cell side and the change of the cell shell temperature in real time, with a prediction error of less than 3% for the thickness of the cell side and a prediction accuracy of more than 95% for the change trend of the cell shell temperature. Table 3 shows the prediction results for different numbered anodes.

[0094] Table 3 Prediction results corresponding to different numbered anodes

[0095]

[0096]

[0097] The coordinate positions in Table 3 are based on a three-dimensional coordinate system, which takes the starting point of the intersection of the aluminum electrolytic cell flue end center line, the cell shell and the cell bottom lining as the coordinate origin, takes the upward direction as the X axis, takes the left and right directions as the Y axis, and takes the direction towards the aluminum outlet as the Z axis.

[0098] Figure 3 A visualization diagram of the dynamic prediction results of the cell side is shown, which shows the real-time cell side thickness of the aluminum electrolytic cell obtained by the dynamic prediction method of the present application and visualizes it, wherein the color band represents the cell side thickness, and the subgraph is the cell side interface graph of the corresponding position. The prediction time of the present application is 1 min, and the prediction accuracy is 95.31% of the offline prediction, but the offline prediction time is 10-20h.

[0099] Traditional cell side calculation is usually based on fixed physical equations, empirical formulas or simple neural network algorithms, without fully considering the dynamic changes of material properties, resulting in a significant decrease in calculation accuracy under non-standard working conditions. The present application improves the adaptability and prediction accuracy of complex working conditions by fusing structure parameters, process parameters and multi-scale simulation methods, and dynamically adjusting the parameters of the prediction model combined with deep learning technology.

[0100] Compared with the traditional method, the dynamic iterative optimization mechanism is generally based on a one-time model construction, the accuracy and robustness of the prediction model are poor, and the parameters are designed for a specific electrolytic cell, which is difficult to apply in different specifications and working conditions, the adaptability of the target prediction model under the current working condition is evaluated according to the measured cell wall thickness and the cell wall thickness prediction value when the anode is replaced, the aluminum electrolytic cell multi-physical field simulation model is adjusted when the prediction accuracy requirement is not met, and the target prediction model is retrained, so that the prediction result is highly consistent with the actual working condition.

[0101] Embodiment 2

[0102] The embodiment of the present application also provides an electronic device, which comprises a memory, a processor and a computer program / instruction stored in the memory, and the processor executes the computer program / instruction to realize the aluminum electrolytic cell cell wall dynamic prediction method in the embodiment of the present application.

[0103] Although not shown, the electronic device includes a processor, which can perform various appropriate operations and processes according to programs and / or data stored in a read-only memory (ROM) or programs and / or data loaded from a storage part into a random access memory (RAM). The processor can be a multi-core processor, or can include multiple processors. In some embodiments, the processor can include a general-purpose main processor and one or more special-purpose coprocessors, such as a central processing unit, a graphics processing unit (GPU), a neural network processing unit (NPU), a digital signal processor (DSP), etc. In the RAM, various programs and data required for device operation are also stored. The processor, ROM and RAM are connected to each other through a bus. An input / output (I / O) interface is also connected to the bus.

[0104] The above processor and memory are used together to execute programs / instructions stored in the memory, which can realize the methods, steps or functions described in the above embodiments when executed by a computer.

[0105] Although not shown, the embodiment of the present application also provides a computer readable storage medium having a computer program / instruction stored thereon, which realizes the aluminum electrolytic cell cell wall dynamic prediction method in the embodiment of the present application when executed by a processor.

[0106] Read only memory (ROM), electrically programmable read only memory (EPROM), electrically erasable programmable read only memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), digital versatile discs (DVD), or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other medium which can be used to store information which can be accessed by a computer, such as cache memories, etc., which can be accessed by a computer. In this context, a "computer readable medium" can include a transitory medium such as a modulated data signal, carrier wave, or other transport mechanism, and can include a non-transitory medium such as a storage device, memory, or other storage medium. The application is not limited to any particular type of computer readable medium.

[0107] Although not shown, the embodiments of the present application further provide a computer program product, comprising: computer programs / instructions, which, when executed by a processor, implement the aluminum reduction cell sidewall dynamic prediction method in the embodiments of the present application.

[0108] The above disclosure is merely a specific implementation of the present application, but the protection scope of the present application is not limited thereto, and any person skilled in the art can easily think of changes or modifications within the technical scope disclosed in the present application, which shall be covered within the protection scope of the present application.

Claims

1. A method of dynamic prediction of an aluminium electrolysis cell sidewall, characterized by, The prediction method comprises: obtaining structural parameters and process parameters of an aluminum electrolysis cell, and constructing a multi-physical field simulation model of the aluminum electrolysis cell according to the structural parameters and process parameters; obtaining the thickness of the cell wall and the temperature of the cell shell under different process parameters based on the multi-physical field simulation model of the aluminum electrolysis cell; constructing a sample data set according to the thickness of the cell wall and the temperature of the cell shell under different process parameters; constructing a cell wall size prediction model, training the cell wall size prediction model using the sample data set, and obtaining a target prediction model; obtaining real-time process parameters of an aluminum electrolysis cell to be predicted, predicting the real-time process parameters using the target prediction model, and obtaining a predicted value of the thickness of the cell wall and a predicted value of the temperature of the cell shell; obtaining a measured value of the thickness of the cell wall of an anode of the aluminum electrolysis cell to be predicted; determining whether the target prediction model meets the prediction accuracy requirement according to the measured value of the thickness of the cell wall and the predicted value of the thickness of the cell wall; if yes, outputting the predicted value of the thickness of the cell wall and the predicted value of the temperature of the cell shell under the real-time process parameters; if no, adjusting the physical performance parameters of the multi-physical field simulation model of the aluminum electrolysis cell, and repeating the steps of constructing a sample data set, retraining a target prediction model, predicting real-time process parameters, and determining the prediction accuracy of the target prediction model based on the adjusted multi-physical field simulation model of the aluminum electrolysis cell; the structural parameters include geometric dimensions and physical performance parameters, and the physical performance parameters include the electrical conductivity of each part of the material and the thermal conductivity under different temperature and service life conditions; the process parameters include temperature field, current distribution, aluminum level and electrolyte level; the nanoCT method is used to obtain the electrical conductivity and thermal conductivity of the material, and the specific implementation process is as follows: the CT material data obtained is processed by using Avizo software to obtain a three-dimensional sampling model; the three-dimensional sampling model is repaired in a model processing software to obtain a repaired model; the repaired model is imported into a mesh drawing software, and the repaired model is meshed to obtain a material mesoscopic model; the material mesoscopic model is imported into a finite element simulation software, and temperature boundary conditions and thermal conductivity boundary conditions are loaded into the material mesoscopic model to calculate the comprehensive electrical conductivity and thermal conductivity of the material; based on the multi-physical field simulation model of the aluminum electrolysis cell, the thickness of the cell wall and the temperature of the cell shell under different process parameters are obtained, which specifically comprises: loading process parameters, physical boundary conditions and parameter constraint conditions on the multi-physical field simulation model of the aluminum electrolysis cell, and iteratively solving to obtain the thickness of the cell wall and the temperature of the cell shell under the process parameters.

2. The aluminium electrolysis cell cell side dynamic prediction method according to claim 1, characterized in that, The physical boundary conditions and parameter constraint conditions are obtained by using a multi-scale simulation method.

3. The aluminium electrolysis cell cell side dynamic prediction method according to claim 1, characterized in that, The sample data set is constructed according to the thickness of the cell wall and the temperature of the cell shell under different process parameters, which comprises: performing denoising and missing data filling processing on the process parameters; performing standardization and normalization processing on the processed process parameters; performing integration and format conversion processing on the normalized process parameters; extracting key features from the format-converted process parameters using statistical analysis and machine learning techniques, and evaluating the correlation of each key feature with the thickness of the cell wall and the temperature of the cell shell by combining a deep learning algorithm; According to the evaluation result, relevant features are determined, and a sample data set is constructed according to the determined relevant features and corresponding sidewall thickness and shell temperature.

4. The aluminium reduction cell cell side dynamic prediction method of any one of claims 1 to 3, characterised in that, In the training process of the sidewall size prediction model or the target prediction model, an automatic parameter tuning algorithm based on Bayesian optimization is used to optimize the model parameters.

5. An electronic device comprising a memory, a processor, and a computer program / instructions stored on the memory, wherein, The processor executes the computer program / instructions to implement the aluminum electrolysis cell sidewall dynamic prediction method of any one of claims 1-4.

6. A computer readable storage medium having stored thereon computer programs / instructions, characterized in that, The computer program / instructions, when executed by the processor, implement the aluminum electrolysis cell sidewall dynamic prediction method of any one of claims 1-4.

7. A computer program product comprising computer programs / instructions, characterized in that, The computer program / instructions, when executed by the processor, implement the aluminum electrolysis cell sidewall dynamic prediction method of any one of claims 1-4.

Images