A method and system for evaluating the environmental impact during the recycling stage of lithium batteries based on big data

Abstract

The present invention relates to the field of big data technology, and specifically, to a method and system for evaluating the environmental impact during the lithium battery recycling stage based on big data. It includes the following steps: using an Internet of Things sensor network to collect environmental status data and battery disassembly data in real time, and screening the data into relevant data and irrelevant data; calculating the electromagnetic field intensity of the disassembly equipment through an interpolation method, and using a graph neural network combined with a spatio-temporal convolutional neural network to construct a data association knowledge graph; introducing an electromagnetic driving force term to construct an electromagnetic-driven heavy metal diffusion equation, and using a Fourier neural operator to solve the electromagnetic-driven heavy metal diffusion equation; using a graph neural network to construct a dynamic Fourier neural operator. The method and system for evaluating the environmental impact during the lithium battery recycling stage based on big data construct and use a Fourier neural operator to solve the electromagnetic-driven heavy metal diffusion equation, effectively evaluating the impact of the equipment magnetic field on heavy metal diffusion and thus on the environment.

Description

Technical Field

[0001] The present invention relates to the technical field of big data, and specifically, to a method and system for evaluating the environmental impact during the lithium battery recycling stage based on big data. Background Art

[0002] The method and system for evaluating the environmental impact during the lithium battery recycling stage based on big data aims to analyze and evaluate the diffusion impact of the electromagnetic field intensity of the environment on heavy metals during the lithium battery disassembly process. By combining the electromagnetic field intensity and the environmental concentration of heavy metals, an electromagnetic-driven heavy metal diffusion equation is constructed and solved to analyze the diffusion distribution of heavy metals during the lithium battery disassembly process, so as to realize the evaluation of the environmental impact during the lithium battery recycling and disassembly stage.

[0003] The existing methods and systems for evaluating the environmental impact during the lithium battery recycling stage easily ignore the interaction between the electromagnetic field intensity of the equipment in the disassembly workshop and the diffusion of heavy metals. Moreover, due to the different electromagnetic field intensities generated by the equipment in different disassembly workshops and the different electromagnetic forces received by heavy metals, it will lead to deviations in the evaluation of the environmental impact of the interaction between electromagnetic interference and the diffusion of different heavy metals during the disassembly and recycling process. Therefore, a method and system for evaluating the environmental impact during the lithium battery recycling stage based on big data are provided. Summary of the Invention

[0004] The purpose of the present invention is to provide a method and system for evaluating the environmental impact during the lithium battery recycling stage based on big data, so as to solve the deviation in the evaluation of the environmental impact of the interaction between electromagnetic interference and the diffusion of different heavy metals during the disassembly and recycling process, which is caused by the different electromagnetic field intensities generated by the equipment in different disassembly workshops and the different electromagnetic forces received by heavy metals as mentioned in the above background art.

[0005] To achieve the above purpose, the present invention provides a method for evaluating the environmental impact during the lithium battery recycling stage based on big data, including:

[0006] S1. Use the Internet of Things sensor network to collect environmental status data and battery disassembly data in real time, and use the random forest algorithm based on SHAP values to screen the environmental status data and battery disassembly data into relevant data and irrelevant data;

[0007] S2. Based on the relevant data of the environmental status data and battery disassembly data, calculate the electromagnetic field intensity of the disassembly equipment through an interpolation method, and use a graph neural network combined with a spatio-temporal convolutional neural network to construct a data association knowledge graph;

[0008] S3. Based on the electromagnetic field intensity of the disassembly equipment, introduce an electromagnetic driving force term to construct an electromagnetic-driven heavy metal diffusion equation, and use a Fourier neural operator to solve the electromagnetic-driven heavy metal diffusion equation to evaluate the diffusion impact of electromagnetic interference on heavy metals during the lithium battery disassembly process;

[0009] S4. Use a graph neural network to optimize the Fourier neural operator and construct a dynamic Fourier neural operator for dynamically evaluating the environmental impact of heavy metal diffusion according to different recycling workshop environments.

[0010] As a further improvement of this technical solution, the environmental state data includes: environmental temperature, heavy metal concentration, and time series of the electromagnetic field intensity in the disassembly workshop;

[0011] The battery disassembly data includes: battery specifications and disassembly speed.

[0012] As a further improvement of this technical solution, in S2, based on the relevant data of the environmental state data and the battery disassembly data, calculate the electromagnetic field intensity of the disassembly equipment through an interpolation method, and use a graph neural network combined with a spatio-temporal convolutional neural network to construct a data association knowledge graph. The specific method steps are as follows:

[0013] S2.1. Calculate the electromagnetic field intensity of the disassembly equipment based on the relevant data of the environmental state data and the battery disassembly data;

[0014] S2.2. Use a graph neural network combined with a spatio-temporal convolutional neural network to fuse the relevant data of the environmental state data and the battery disassembly data, and define and construct a data association knowledge graph.

[0015] As a further improvement of this technical solution, in S2.1, based on the relevant data of the environmental state data and the battery disassembly data, calculate the electromagnetic field intensity of the disassembly equipment through an interpolation method. The specific method is as follows:

[0016] ;

[0017] Wherein, is time; is the environmental location; is the electromagnetic field intensity at time and location ; is the interpolation method; is the time series of the electromagnetic field intensity;

[0018] In S2.2, use a graph neural network combined with a spatio-temporal convolutional neural network to fuse the relevant data of the environmental state data and the battery disassembly data, and define and construct a data association knowledge graph. The specific method steps are as follows:

[0019] S2.2.1. Use a graph neural network to analyze the relationships among sensors, battery disassembly data, and environmental state data, generate node feature representations, and define them as the nodes of the data association knowledge graph;

[0020] S2.2.2. Analyze the spatio-temporal variation characteristics of battery disassembly data and environmental status data using a spatio-temporal convolutional neural network to obtain spatio-temporal variation dependencies, which are defined as the edges of the data association knowledge graph.

[0021] S2.2.3. Construct a data association knowledge graph based on the nodes and edges of the data association knowledge graph.

[0022] As a further improvement of this technical solution, in S3, based on the electromagnetic field intensity of the disassembly equipment and introducing an electromagnetic driving force term, construct an electromagnetic-driven heavy metal diffusion equation, and use the Fourier neural operator to solve the electromagnetic-driven heavy metal diffusion equation to evaluate the influence of electromagnetic interference on the diffusion of heavy metals during the disassembly of lithium batteries. The specific method steps are as follows:

[0023] S3.1. Based on the electromagnetic field intensity, magnetic field intensity, and heavy metal concentration, use spatio-temporal data fusion technology to construct a spatio-temporal joint input matrix and calculate the electromagnetic driving force term.

[0024] S3.2. Based on the electromagnetic driving force term, construct an electromagnetic-driven heavy metal diffusion equation, and use the Fourier neural operator to solve the electromagnetic-driven heavy metal diffusion equation to evaluate the heavy metal concentration at the next time.

[0025] As a further improvement of this technical solution, in S3.1, based on the electromagnetic field intensity, magnetic field intensity, and heavy metal concentration, use spatio-temporal data fusion technology to construct a spatio-temporal joint input matrix and calculate the electromagnetic driving force term. The specific method steps are as follows:

[0026] S3.1.1. Based on the electromagnetic field intensity, magnetic field intensity, and heavy metal concentration, use spatio-temporal data fusion technology to construct a spatio-temporal joint input matrix:

[0027] ;

[0028] ;

[0029] Among them, is the spatio-temporal joint input matrix; is the electromagnetic field intensity at time and position ; is the magnetic field intensity at time and position ; is the heavy metal concentration at time and position ; is the boundary condition dynamic mask; is the equipment position;

[0030] S3.1.2. Based on the time and position The electromagnetic field strength and magnetic field strength are used to calculate the electromagnetic driving force term:

[0031] ;

[0032] Among them, is the heavy metal ion index; is the th kind of heavy metal ion's equivalent acceleration at time and position ; is the th kind of heavy metal ion's electric charge; is the th kind of heavy metal ion's mass; is the th kind of heavy metal ion's velocity at time and position ;

[0033] Among them, the velocity of the th kind of heavy metal ion at time and position is used as the electromagnetic driving force term.

[0034] As a further improvement of this technical solution, in S3.2, an electromagnetic-driven heavy metal diffusion equation is constructed based on the electromagnetic driving force term, and the Fourier neural operator is used to solve the electromagnetic-driven heavy metal diffusion equation to evaluate the heavy metal concentration at the next time. The specific method steps are as follows:

[0035] S3.2.1. Construct an electromagnetic-driven heavy metal diffusion equation based on the electromagnetic driving force term:

[0036] ;

[0037] Among them, is the rate of change of the heavy metal concentration with respect to time at time and position ; is the gradient operator; is the diffusion coefficient of the th kind of heavy metal; is the influence of the acceleration effect of the electromagnetic field on heavy metals on the diffusion process; is the real-time pollution source term of the th kind of heavy metal;

[0038] S3.2.2. Represent the spatio-temporal joint input matrix in the frequency domain:

[0039] ;

[0040] Among them, is the frequency-domain representation of the spatio-temporal joint input matrix, i.e., the Fourier neural operator; is the Fourier transform operation; is the element-wise multiplication operation; is the index of the Fourier transform frequency;

[0041] S3.2.3. Convert the electromagnetic-driven heavy metal diffusion equation into the frequency domain, use the Fourier neural operator to solve the electromagnetic-driven heavy metal diffusion equation, and through the inverse Fourier transform, convert the solution in the frequency domain back to the spatio-temporal domain to obtain the heavy metal concentration at the next time step.

[0042] As a further improvement of this technical solution, in S4, a graph neural network is used to optimize the Fourier neural operator to construct a dynamic Fourier neural operator for dynamically evaluating the environmental impact of heavy metal diffusion according to different recycling workshop environments, specifically as follows:

[0043] ;

[0044] where is the operation of the graph neural network; is the optimized Fourier neural operator; is the dynamic weight learned by the graph neural network; is the graph node index; is the total number of graph nodes;

[0045] ;

[0046] where is the dynamic Fourier neural operator; is the adaptive weight of the dynamic Fourier neural operator.

[0047] As a further improvement of this technical solution, the dynamic Fourier neural operator is used to adaptively solve the electromagnetic-driven heavy metal diffusion equation according to the changes in the workshop environment during the lithium battery recycling stage.

[0048] On the other hand, the present invention provides a big data-based environmental impact assessment system for the lithium battery recycling stage, including a memory, a processor, and a computer program stored in the memory and executable on the processor. The processor executes the computer program to implement the big data-based environmental impact assessment method for the lithium battery recycling stage described in any one of the above.

[0049] Compared with the prior art, the beneficial effects of the present invention:

[0050] 1. In the method and system for evaluating the environmental impact during the lithium - battery recycling stage based on big data, through the spatio - temporal data analysis of electromagnetic - field intensity and heavy - metal concentration, the diffusion impact of the electromagnetic field on different types of heavy metals during the disassembly process can be evaluated. With the assistance of big - data technology, the accuracy and real - time performance of the evaluation are improved.

[0051] 2. In the method and system for evaluating the environmental impact during the lithium - battery recycling stage based on big data, by using a graph neural network to construct a dynamic Fourier neural operator, the interaction between the electromagnetic - field intensity generated by different devices and the diffusion characteristics of different heavy metals is evaluated. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 It is the overall method flow chart of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0053] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0054] Embodiment 1: Please refer to Figure 1 As shown, this embodiment provides a method for evaluating the environmental impact during the lithium - battery recycling stage based on big data, including the following steps:

[0055] S1. Use the Internet - of - Things sensor network to collect environmental - state data and battery - disassembly data in real time, and use the random - forest algorithm based on SHAP values to screen the environmental - state data and battery - disassembly data into relevant data and irrelevant data;

[0056] The environmental - state data includes: environmental temperature, heavy - metal concentration, and the time series of the electromagnetic - field intensity in the disassembly workshop;

[0057] The battery - disassembly data includes: battery specifications and disassembly speed.

[0058] In this embodiment, temperature has an important impact on the chemical reaction and disassembly process of lithium batteries. Lithium batteries may experience thermal runaway under high - temperature conditions, resulting in the leakage of harmful gases and heavy metals. The increase in temperature may also accelerate the diffusion of heavy metals in the environment; lithium batteries contain a certain amount of harmful heavy metals (such as nickel, cobalt, lithium, lead, etc.). During the battery - disassembly process, heavy metals may enter the soil environment through cracking, dissolution, or gas diffusion. This method focuses on the situation of heavy - metal - contaminated soil.

[0059] In this embodiment, the time series of the electromagnetic field intensity in the disassembly workshop is collected because when the disassembly equipment is working, it may generate a high-intensity electromagnetic field, which will change the distribution of the electric and magnetic fields in the surrounding environment, and then affect the diffusion mode of heavy metals;

[0060] In step S1 of this embodiment, the random forest algorithm based on SHAP values is used to screen the environmental status data and battery disassembly data into relevant data and irrelevant data. The specific method is as follows:

[0061] Use the complete data set of environmental status data and battery disassembly data as the training set; for supervised learning problems, the target variable can be the label related to the environmental impact during the disassembly process;

[0062] Construct multiple decision tree models, each tree is trained on different subsets of data, and prediction is made through majority voting;

[0063] After training by the random forest model, the contribution of each feature to the prediction result can be quantified through feature importance;

[0064] Calculate the SHAP value of each feature using the trained random forest model; the SHAP value is achieved by allocating the contribution of each feature to the model prediction;

[0065] By calculating the SHAP value of each feature, the features that have a greater impact on the model prediction result (i.e., the features with larger SHAP values) can be screened out; these features are considered relevant because they have a greater impact on the prediction result; for those features with SHAP values close to zero or with small changes, it means that their contribution to the prediction result is negligible, so these features can be considered irrelevant;

[0066] The relevant data screened based on the SHAP value will be used for subsequent modeling and analysis, while the irrelevant data can be removed or ignored during model training, which helps to improve the prediction accuracy of the model, reduce redundant features, and enhance the efficiency of model training and prediction.

[0067] S2. Calculate the electromagnetic field intensity of the disassembly equipment through the interpolation method based on the relevant data of the environmental status data and battery disassembly data, and use the graph neural network combined with the spatio-temporal convolutional neural network to construct a data association knowledge graph;

[0068] In step S2 of this embodiment, based on the relevant data of the environmental status data and battery disassembly data, calculate the electromagnetic field intensity of the disassembly equipment through the interpolation method, and use the graph neural network combined with the spatio-temporal convolutional neural network to construct a data association knowledge graph. The specific method steps are as follows:

[0069] S2.1. Calculate the electromagnetic field intensity of the disassembly equipment based on the relevant data of the environmental status data and battery disassembly data;

[0070] S2.2. Use a graph neural network combined with a spatio-temporal convolutional neural network to fuse the relevant data of the environmental state data and the battery disassembly data, and define and construct a data association knowledge graph.

[0071] In the present embodiment S2.1, the electromagnetic field intensity of the disassembly equipment is that the electromagnetic field during the operation of the disassembly equipment will affect the diffusion of surrounding heavy metals. By calculating the electromagnetic field intensity, detailed electromagnetic environment data can be provided for subsequent analysis; the electromagnetic field intensity data is crucial for predicting the diffusion path of heavy metals because the electromagnetic field may accelerate or inhibit the migration, aggregation or diffusion of heavy metals.

[0072] In the present embodiment S2.2, the graph neural network is a deep learning method for processing graph-structured data; the core advantage of the graph neural network lies in its ability to capture the complex relationships between devices, sensors, and pollutant concentrations; the graph neural network aggregates based on the features of adjacent nodes through an information transfer mechanism between nodes to learn the hidden features of the nodes; the spatio-temporal convolutional neural network is a class of methods for processing data with both spatial and temporal dependencies; the data such as the electromagnetic field and heavy metal concentration in the disassembly workshop have spatial and temporal dependencies; the intensity of the electromagnetic field has a spatial distribution in the workshop and changes over time; while the heavy metal concentration also varies at different positions and different times; using GNN and spatio-temporal modeling techniques can simultaneously consider the spatial and temporal relationships.

[0073] The knowledge graph is a graphical structure used to represent the internal relationships between different data. The knowledge graph represents the relationship between the environmental state data and the battery disassembly data as a graph structure, which is convenient for analyzing their interactions and dependencies.

[0074] In the present embodiment S2.1, based on the relevant data of the environmental state data and the battery disassembly data, the electromagnetic field intensity of the disassembly equipment is calculated by an interpolation method. The specific method is as follows:

[0075] ;

[0076] where is time; is the environmental location; is the electromagnetic field intensity at time and location ; is the interpolation method; is the electromagnetic field intensity time series;

[0077] In this embodiment S2.2, a graph neural network is used in combination with a spatio-temporal convolutional neural network to fuse the relevant data of the environmental state data and the battery disassembly data, and a data association knowledge graph is defined and constructed. The specific method steps are as follows:

[0078] S2.2.1. Use a graph neural network to analyze the relationships among sensors, battery disassembly data, and environmental state data, generate node feature representations, and define them as the nodes of the data association knowledge graph;

[0079] S2.2.2. Use a spatio-temporal convolutional neural network to analyze the spatio-temporal change characteristics of the battery disassembly data and the environmental state data, obtain the spatio-temporal change dependence relationship, and define it as the edge of the data association knowledge graph;

[0080] S2.2.3. Based on the nodes and edges of the data association knowledge graph, construct the data association knowledge graph.

[0081] In this embodiment, the data relationship between heavy metal lead and the data of sensors and electromagnetic fields is represented in the data association knowledge graph. The heavy metal lead node: represents the concentration, diffusion coefficient, and acceleration effect information of lead; the features in the node include the physical and chemical properties of lead, electromagnetic drive characteristics, concentration changes, etc.;

[0082] The sensor node records the electromagnetic field intensity and its relationship with the environment; the electromagnetic field node records the electromagnetic field intensity at a certain time and position and its influence on the diffusion of heavy metal lead.

[0083] S3. Based on the electromagnetic field intensity of the disassembly equipment, introduce an electromagnetic driving force term to construct an electromagnetic-driven heavy metal diffusion equation, and use the Fourier neural operator to solve the electromagnetic-driven heavy metal diffusion equation to evaluate the influence of electromagnetic interference on the diffusion of heavy metals during the lithium battery disassembly process;

[0084] In this embodiment S3, based on the electromagnetic field intensity of the disassembly equipment, introduce an electromagnetic driving force term to construct an electromagnetic-driven heavy metal diffusion equation, and use the Fourier neural operator to solve the electromagnetic-driven heavy metal diffusion equation to evaluate the influence of electromagnetic interference on the diffusion of heavy metals during the lithium battery disassembly process. The specific method steps are as follows:

[0085] S3.1. Based on the electromagnetic field intensity, magnetic field intensity, and heavy metal concentration, use spatio-temporal data fusion technology to construct a spatio-temporal joint input matrix and calculate the electromagnetic driving force term;

[0086] S3.2. Based on the electromagnetic driving force term, construct an electromagnetic-driven heavy metal diffusion equation, and use the Fourier neural operator to solve the electromagnetic-driven heavy metal diffusion equation to evaluate the heavy metal concentration at the next time.

[0087] In this embodiment, the influence of the electromagnetic field on the migration rate and diffusion coefficient of metal ions is generated by the electromagnetic force (Lorentz force). The electromagnetic field will cause the acceleration of charged particles, thereby affecting their motion in the medium. For charged particles, the electric field provides a direct force, and the magnetic field generates a force perpendicular to the velocity and magnetic field directions through the cross-action with the particle velocity, which affects the migration rate of the particles.

[0088] The diffusion coefficient describes the diffusion ability of particles in the medium. Under normal circumstances, the diffusion coefficient is determined by factors such as temperature and concentration gradient. However, the action of the electromagnetic field will change the path and velocity of the metal ion movement, so it will also affect the diffusion coefficient. The influence of the electric field and magnetic field makes the ions may undergo more complex trajectory movements in some cases, such as deflecting along the electric field or magnetic field directions, resulting in the actual diffusion process being different from the classical diffusion equation. By introducing the electromagnetic driving force term, the influence of the electromagnetic field can be reflected in the diffusion equation, modifying the diffusion coefficient, so as to obtain a diffusion equation with electromagnetic field effects. The electromagnetic field usually changes the equivalent diffusion coefficient of the diffusion process, making it not only depend on temperature and concentration gradient, but also be corrected by the influence of the electric field and magnetic field.

[0089] The Fourier neural operator is a mathematical method that combines the Fourier transform and neural networks, mainly used for solving partial differential equations. When dealing with the electromagnetic-driven heavy metal diffusion equation, the FNO can approximate the equation in the frequency domain, greatly accelerating the calculation process. The core idea of the Fourier neural operator is to transform the traditional PDE solving method into a frequency domain problem, directly approximating the solution in the frequency domain. Through the Fourier transform, the heavy metal diffusion equation can be transformed into an algebraic problem in the frequency domain, reducing the spatial and temporal dimensions, and thus accelerating the solving process. Through the frequency domain approximation method, variables such as the electromagnetic field intensity and magnetic field intensity can be directly embedded into the diffusion equation without explicitly performing complex numerical simulations in the spatio-temporal domain, solving PDEs containing nonlinear terms and complex boundary conditions, thereby accelerating the solution of the heavy metal diffusion equation.

[0090] In this embodiment S3.1, based on the electromagnetic field intensity, magnetic field intensity, and heavy metal concentration, the spatio-temporal data fusion technology is used to construct a spatio-temporal joint input matrix and calculate the electromagnetic driving force term. The specific method steps are as follows:

[0091] S3.1.1. Based on the electromagnetic field intensity, magnetic field intensity, and heavy metal concentration, use the spatio-temporal data fusion technology to construct a spatio-temporal joint input matrix:

[0092] ;

[0093] ;

[0094] Among them, is a spatio-temporal joint input matrix; is the electromagnetic field strength at time and position ; is the magnetic field strength at time and position ; is the heavy metal concentration at time and position ; is the boundary condition dynamic mask; is the device position;

[0095] S3.1.2. Calculate the electromagnetic driving force term based on the electromagnetic field strength and magnetic field strength at time and position :

[0096] ;

[0097] where is the heavy metal ion index; is the equivalent acceleration of the th heavy metal ion at time and position ; is the electric charge of the th heavy metal ion; is the mass of the th heavy metal ion; is the velocity of the th heavy metal ion at time and position ;

[0098] where the velocity of the th heavy metal ion at time and position is used as the electromagnetic driving force term.

[0099] In this embodiment, the electromagnetic driving force term is calculated based on the electromagnetic field strength and magnetic field strength, and then the equivalent acceleration of the heavy metal ions is obtained: The electromagnetic force acts on charged particles, especially under the combined action of an electric field and a magnetic field; for charged heavy metal ions, the specific manifestation of their force can be described by the Lorentz force formula, indicating that the motion of charged particles in an electric field and a magnetic field is affected by both; by combining the Lorentz force formula with Newton's second law, we can obtain the acceleration of heavy metal ions under the action of an electromagnetic field.

[0100] In this embodiment S3.2, an electromagnetic-driven heavy metal diffusion equation is constructed based on the electromagnetic driving force term, and the Fourier neural operator is used to solve the electromagnetic-driven heavy metal diffusion equation to evaluate the heavy metal concentration at the next time. The specific method steps are as follows:

[0101] S3.2.1. Construct an electromagnetic-driven heavy metal diffusion equation based on the electromagnetic driving force term:

[0102] ;

[0103] where, is the rate of change of the heavy metal concentration with respect to time at time and position ; is the gradient operator; is the th diffusion coefficient of the heavy metal; is the influence of the acceleration effect of the electromagnetic field on the heavy metal on the diffusion process; is the th real-time pollution source term of the heavy metal;

[0104] S3.2.2. Represent the spatio-temporal joint input matrix in the frequency domain:

[0105] ;

[0106] where, is the frequency-domain representation of the spatio-temporal joint input matrix, i.e., the Fourier neural operator; is the Fourier transform operation; is the element-wise multiplication operation; is the index of the Fourier transform frequency;

[0107] S3.2.3. Convert the electromagnetic-driven heavy metal diffusion equation into the frequency domain form, use the Fourier neural operator to solve the electromagnetic-driven heavy metal diffusion equation, and through the inverse Fourier transform, convert the solution in the frequency domain back to the spatio-temporal domain to obtain the heavy metal concentration at the next time step.

[0108] In this embodiment S3.2.3, the electromagnetic-driven heavy metal diffusion equation is converted into the frequency domain form as follows:

[0109] The original electromagnetic-driven heavy metal diffusion equation is:

[0110] ;

[0111] is the diffusion term, and by performing the Fourier transform on it, it can be converted into a calculation in the frequency domain: ; is the square of the spatial gradient in the frequency domain;

[0112] is the electromagnetic driving term, representing the acceleration effect of the electromagnetic field on heavy metals. This term is solved through Fourier transform: Regarding as the electromagnetic acceleration term related to the position which is a modulation factor affecting the concentration change. In the frequency domain, the change of the electromagnetic acceleration term is represented by multiplying by the heavy metal concentration That is, the acceleration term after Fourier transform is multiplied by the frequency domain representation of the heavy metal concentration: ; is the factor related to the spatial gradient in the frequency domain;

[0113] is the real-time pollution source term of the th kind of heavy metal, which directly provides the contribution of the external source to the diffusion of heavy metals. The Fourier transform of the source term is: ;

[0114] Through Fourier transform, we transform each term of the electromagnetic-driven heavy metal diffusion equation into the form in the frequency domain: ;

[0115] Obtain the frequency domain representation of the electromagnetic-driven heavy metal diffusion equation;

[0116] The Fourier neural operator is used to efficiently solve this equation in the frequency domain. The heavy metal diffusion equation is quickly solved by the Fourier neural operator to obtain the heavy metal concentration at the next time step; this process dynamically adjusts the Fourier coefficients to adapt to environmental changes and the influence of the electromagnetic field on heavy metal diffusion, and transforms the result back to the spatio-temporal domain through inverse Fourier transform, so as to predict the position and concentration of heavy metals at the next time step;

[0117] In this embodiment, lead (Pb) is a common and important heavy metal, and its impact on the environment and human health is very large. By combining the electromagnetic-driven heavy metal diffusion equation and the Fourier neural operator to analyze it, we can specifically evaluate the diffusion process of lead during the recycling stage and evaluate environmental pollution. After obtaining the spatio-temporal evolution results of lead concentration, we can evaluate the environmental impact of the electromagnetic field on lead diffusion based on the following points:

[0118] Analyze the trend of lead concentration changing with time, especially the expansion of high-concentration areas. If, under the influence of the electromagnetic field, the lead concentration increases significantly in some areas, this means that the electromagnetic field may accelerate the migration of lead;

[0119] By analyzing the spatial distribution of the solution of the diffusion equation, evaluate the areas that may be polluted by lead during the recycling process. If the electromagnetic field source configuration during the recycling process is improper, it may cause lead to diffuse from the inside of the equipment to the surrounding air, soil or water body;

[0120] If the electromagnetic field significantly accelerates the lead diffusion, it is recommended to regulate the electromagnetic field intensity in the recycling equipment to avoid the excessive electromagnetic field exacerbating the migration of lead.

[0121] According to the spatial distribution results of lead diffusion, optimize the layout of the recycling equipment and the position of the electromagnetic field source to reduce the expansion of the lead pollution area.

[0122] S4. Use a graph neural network to optimize the Fourier neural operator and construct a dynamic Fourier neural operator to dynamically evaluate the environmental impact of heavy metal diffusion according to different recycling workshop environments.

[0123] In this embodiment S4, a graph neural network is used to optimize the Fourier neural operator and construct a dynamic Fourier neural operator to dynamically evaluate the environmental impact of heavy metal diffusion according to different recycling workshop environments, as follows:

[0124] ;

[0125] Among them, is the operation of the graph neural network; is the optimized Fourier neural operator; is the dynamic weight learned by the graph neural network; is the graph node index; is the total number of graph nodes;

[0126] ;

[0127] Among them, is the dynamic Fourier neural operator; is the adaptive weight of the dynamic Fourier neural operator.

[0128] The dynamic Fourier neural operator is used to adaptively solve the electromagnetic drive heavy metal diffusion equation according to the changes in the workshop environment during the lithium battery recycling stage.

[0129] In this embodiment, the combination of the graph neural network and the Fourier neural operator forms a dynamic Fourier neural operator, which adaptively solves the electromagnetic-driven heavy metal diffusion equation according to the changes in the workshop environment during the lithium battery recycling stage. The graph neural network is a neural network architecture capable of processing graph-structured data, used to model and learn the spatio-temporal variation relationship between the environmental state data of the lithium battery recycling workshop and the battery disassembly process. Through the graph neural network, the complex spatio-temporal dependencies between the environmental state data and the battery disassembly data can be captured, so as to perform adaptive modeling in a dynamically changing environment. The spatio-temporal graph embedding obtained through the graph neural network can be combined with the electromagnetic field strength, magnetic field strength, and heavy metal concentration to generate a new spatio-temporal joint input matrix. By performing frequency domain transformation on these input data through Fourier transform and conducting training and optimization, the dynamic Fourier neural operator can be obtained.

[0130] Embodiment 2: This embodiment provides an environment impact assessment system for the lithium battery recycling stage based on big data, including a memory, a processor, and a computer program stored in the memory and executable on the processor. The processor executes the computer program to implement the environment impact assessment method for the lithium battery recycling stage based on big data described in any one of the above.

[0131] The above shows and describes the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and the descriptions in the specification are only preferred examples of the present invention and are not used to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed.

Claims

1. A method for environmental impact assessment of lithium battery recycling based on big data, characterized in that: The following steps are involved: S1. Use the IoT sensor network to collect environmental status data and battery disassembly data in real time, and use the random forest algorithm based on SHAP value to filter the environmental status data and battery disassembly data into relevant data and irrelevant data; S2. Based on the relevant data of environmental status data and battery disassembly data, the electromagnetic field strength of the disassembly equipment is calculated by interpolation method, and a data association knowledge graph is constructed using a graph neural network combined with a spatiotemporal convolutional neural network; S3. Based on the electromagnetic field strength of the disassembly equipment, the electromagnetic driving force term is introduced to construct the electromagnetic driven heavy metal diffusion equation, and the electromagnetic driven heavy metal diffusion equation is solved using the Fourier neural operator to evaluate the impact of electromagnetic interference on the diffusion of heavy metals during the disassembly of lithium batteries; S4. Use graph neural networks to optimize the Fourier neural operator and construct a dynamic Fourier neural operator to dynamically evaluate the environmental impact of heavy metal diffusion according to different recycling workshop environments.

2. The environmental impact assessment method for lithium battery recycling based on big data according to claim 1 is characterized by: The environmental status data include: ambient temperature, heavy metal concentration, and electromagnetic field intensity time series of the disassembly workshop; Battery disassembly data includes: battery specifications and disassembly speed.

3. The environmental impact assessment method for lithium battery recycling based on big data according to claim 2 is characterized in that: In S2, based on the relevant data of the environmental status data and the battery disassembly data, the electromagnetic field strength of the disassembly equipment is calculated by the interpolation method, and the data association knowledge graph is constructed by using the graph neural network combined with the spatiotemporal convolutional neural network. The specific method steps are as follows: S2.

1. Calculate the electromagnetic field strength of the dismantling equipment based on the environmental status data and the relevant data of the battery dismantling data; S2.

2. Use graph neural network combined with spatiotemporal convolutional neural network to fuse environmental status data with relevant data of battery disassembly data, define and construct data association knowledge graph.

4. The environmental impact assessment method for lithium battery recycling based on big data according to claim 3 is characterized by: In S2.1, based on the relevant data of the environmental status data and the battery disassembly data, the electromagnetic field strength of the disassembly equipment is calculated by the interpolation method, and the specific method is as follows: ； in, For time; for environmental location; For in time and location The electromagnetic field strength; is the interpolation method; is the electromagnetic field intensity time series; In S2.2, a graph neural network is used in combination with a spatiotemporal convolutional neural network to fuse the environmental status data with the relevant data of the battery disassembly data, and a data association knowledge graph is defined and constructed. The specific method steps are as follows: S2.2.

1. Use graph neural network to analyze the relationship between sensor, battery disassembly data and environmental status data, generate node feature representation, and define it as a node of the data association knowledge graph; S2.2.

2. Use spatiotemporal convolutional neural network to analyze the spatiotemporal variation characteristics of battery disassembly data and environmental status data, and obtain spatiotemporal variation dependencies, which are defined as the edges of the data association knowledge graph; S2.2.

3. Construct a data association knowledge graph based on the nodes and edges of the data association knowledge graph.

5. The environmental impact assessment method for lithium battery recycling based on big data according to claim 4 is characterized in that: In S3, based on the electromagnetic field strength of the disassembly equipment, the electromagnetic driving force term is introduced to construct the electromagnetic driven heavy metal diffusion equation, and the electromagnetic driven heavy metal diffusion equation is solved using the Fourier neural operator to evaluate the influence of electromagnetic interference on the diffusion of heavy metals during the disassembly of lithium batteries. The specific method steps are as follows: S3.

1. Based on the electromagnetic field intensity, magnetic field intensity and heavy metal concentration, the spatiotemporal data fusion technology is used to construct the spatiotemporal joint input matrix and calculate the electromagnetic driving force term; S3.

2. Construct an electromagnetically driven heavy metal diffusion equation based on the electromagnetic driving force term, use the Fourier neural operator to solve the electromagnetically driven heavy metal diffusion equation, and evaluate the heavy metal concentration at the next time.

6. The environmental impact assessment method for lithium battery recycling based on big data according to claim 5 is characterized by: In S3.1, based on the electromagnetic field intensity, magnetic field intensity and heavy metal concentration, the spatiotemporal data fusion technology is used to construct a spatiotemporal joint input matrix and calculate the electromagnetic driving force term. The specific method steps are as follows: S3.1.

1. Based on electromagnetic field intensity, magnetic field intensity and heavy metal concentration, a spatiotemporal joint input matrix is ​​constructed using spatiotemporal data fusion technology: ； ； in, is the spatiotemporal joint input matrix; For in time and location The electromagnetic field strength; For in time and location The magnetic field strength; For in time and location Heavy metal concentrations; Dynamic masking for boundary conditions; is the device location; S3.1.2, based on time and location The electromagnetic field strength and magnetic field strength are calculated to calculate the electromagnetic driving force: ； in, Index for heavy metal ions; For the Heavy metal ions in time and location The equivalent acceleration of For the The charge of the heavy metal ion; For the The mass of heavy metal ions; For the Heavy metal ions in time and location speed; Among them, Heavy metal ions in time and location Speed That is, it serves as the electromagnetic driving force term.

7. The environmental impact assessment method for lithium battery recycling based on big data according to claim 6 is characterized by: In S3.2, an electromagnetically driven heavy metal diffusion equation is constructed based on the electromagnetic driving force term, and the electromagnetically driven heavy metal diffusion equation is solved using a Fourier neural operator to evaluate the heavy metal concentration at the next time. The specific method steps are as follows: S3.2.

1. Construct the electromagnetic driven heavy metal diffusion equation based on the electromagnetic driving force term: ； in, For time and location The rate of change of heavy metal concentration over time; is the gradient operator; For the Diffusion coefficient of heavy metals; The influence of the acceleration effect of electromagnetic field on the diffusion process of heavy metals; For the Real-time pollution source items of various heavy metals; S3.2.

2. Express the spatiotemporal joint input matrix in the frequency domain: ； in, It is the frequency domain representation of the spatiotemporal joint input matrix, i.e., the Fourier neural operator; is the Fourier transform operation; Element-wise multiplication operation; is the index of the Fourier transform frequency; S3.2.

3. Convert the electromagnetic driven heavy metal diffusion equation into frequency domain form, use Fourier neural operator to solve the electromagnetic driven heavy metal diffusion equation, and convert the solution in the frequency domain back to the space-time domain through inverse Fourier transform to obtain the heavy metal concentration in the next time step.

8. The method for environmental impact assessment of lithium battery recycling based on big data according to claim 7 is characterized in that: In S4, a graph neural network is used to optimize the Fourier neural operator and construct a dynamic Fourier neural operator to dynamically evaluate the environmental impact of heavy metal diffusion according to different recycling workshop environments, as follows: ； in, Operations of graph neural networks; is the optimized Fourier neural operator; The dynamic weights learned by the graph neural network; is the graph node index; is the total number of graph nodes; ； in, is a dynamic Fourier neural operator; is the adaptive weight of the dynamic Fourier neural operator.

9. The environmental impact assessment method for lithium battery recycling based on big data according to claim 8 is characterized by: The dynamic Fourier neural operator is used to adaptively solve the electromagnetic driven heavy metal diffusion equation according to the changes in the workshop environment during the lithium battery recycling stage.

10. A lithium battery recycling environmental impact assessment system based on big data, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: The processor executes a computer program to implement the environmental impact assessment method for the lithium battery recycling stage based on big data as described in any one of claims 1 to 9.

Images