Artificial intelligence simulation method and device for atmospheric aerosol chemical process

Abstract

The embodiment of the invention discloses an artificial intelligence simulation method and equipment for an atmospheric aerosol chemical process. The method comprises the following steps: acquiring three-dimensional initial concentrations and environmental factors of various aerosols and related gases from a non-static atmospheric numerical mode; wherein the environmental factors comprise temperature, pressure, air density and water vapor mixing ratio, and the plurality of aerosols comprise inorganic aerosols; according to the three-dimensional initial concentration and the environmental factors, adopting a pre-trained target model to predict and obtain predicted concentrations of the inorganic aerosol and the related gas after a specified step length; and the rest part iterates the environmental factors and other processes according to the predicted concentrations of the inorganic aerosol and the related gas, so that alternative calculation of an aerosol chemical process in a numerical mode is realized, and accurate simulation of the predicted concentrations of the inorganic aerosol and the related gas is realized. And therefore, the global high-fidelity simulation precision is realized, and the calculation efficiency is remarkably improved.

Description

Technical Field

[0001] This specification relates to the field of atmospheric numerical simulation, and in particular to an artificial intelligence simulation method and equipment for atmospheric aerosol chemical processes. Background Art

[0002] Atmospheric aerosols are solid or liquid particles suspended in the air that have a profound impact on the Earth's climate system and air quality. Numerical models are indispensable analytical tools and are key to understanding these phenomena, playing an important role in air quality management and developing climate change mitigation strategies, however, coupling aerosol chemistry and interactions into these models poses a major computational challenge.

[0003] This is mainly due to the need to solve a complex set of rigid nonlinear differential equations that control aerosol chemical processes, coupled with the need to use implicit integration schemes to ensure numerical stability; in addition, in order to adapt to various methods for describing the evolution of particle size distribution (PSD), some aerosol chemical processes may need to be repeatedly calculated. In addition, unlike atmospheric photochemical processes that only involve chemical reactions between species, the complete aerosol chemical and interaction processes include many other complex processes, such as nucleation, condensation, and thermodynamic processes. Therefore, the computational burden is greatly increased. This computational intensity often creates a dilemma because it conflicts with many other methods in numerical simulations to improve simulation accuracy, such as increasing spatial resolution to reduce uncertainties in numerical models.

[0004] Based on this, a fast and accurate simulation scheme is needed to predict the concentration of inorganic aerosol in numerical models. Summary of the invention

[0005] The embodiments of this specification provide an artificial intelligence simulation method, device and storage medium for atmospheric aerosol chemical processes to solve the following technical problems: a fast and accurate simulation solution for predicting the concentration of inorganic aerosols in a numerical model is needed.

[0006] To solve the above technical problems, one or more embodiments of this specification are implemented as follows:

[0007] In a first aspect, an embodiment of the present specification provides an artificial intelligence simulation method for atmospheric aerosol chemical processes, which is applied to a non-hydrostatic atmospheric numerical model including aerosol chemistry, wherein the non-hydrostatic atmospheric numerical model includes a target model and other parts, and the method includes:

[0008] The three-dimensional initial concentrations and environmental factors of various aerosols and related gases are obtained from the non-hydrostatic atmospheric numerical model; wherein the environmental factors include temperature, pressure, air density and water vapor mixing ratio, the various aerosols include inorganic aerosols, and the related gases are gases that have ion exchange relationships with the inorganic aerosols; based on the three-dimensional initial concentrations and environmental factors, the predicted concentrations of the inorganic aerosols and related gases after a specified step length are predicted using a pre-trained target model; the remaining part iterates the environmental factors and other processes based on the predicted concentrations of the inorganic aerosols and related gases.

[0009] In a second aspect, one or more embodiments of this specification provide an electronic device, including:

[0010] at least one processor; and,

[0011] a memory communicatively connected to the at least one processor; wherein,

[0012] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to the first aspect.

[0013] In a third aspect, an embodiment of the present specification provides a non-volatile computer storage medium storing computer executable instructions. When a computer reads the computer executable instructions in the storage medium, the instructions enable one or more processors to execute the method described in the first aspect.

[0014] At least one of the above-mentioned technical solutions adopted in one or more embodiments of the present specification can achieve the following beneficial effects: by obtaining the three-dimensional initial concentrations and environmental factors of various aerosols and related gases from the non-hydrostatic atmospheric numerical model; wherein the environmental factors include temperature, pressure, air density and water vapor mixing ratio, and the various aerosols include inorganic aerosols; according to the three-dimensional initial concentrations and environmental factors, the predicted concentrations of the inorganic aerosols and related gases after a specified step length are predicted using a pre-trained target model; the remaining part iterates the environmental factors and other processes according to the predicted concentrations of the inorganic aerosols and related gases, thereby realizing alternative calculations of aerosol chemical processes in the numerical model, realizing accurate simulation of the predicted concentrations of inorganic aerosols and related gases, and further realizing global high-fidelity simulation accuracy and significantly improved computational efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the embodiments of this specification or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this specification. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0016] Figure 1 A schematic flow chart of an artificial intelligence simulation of an atmospheric aerosol chemical process provided for one or more embodiments of this specification;

[0017] Figure 2 A schematic diagram of a model architecture used in an artificial intelligence model of aerosol chemistry and interactions (AIMACI) provided in an embodiment of this specification;

[0018] Figure 3 A schematic diagram of the data density and distribution of the column concentrations of the four main aerosols provided for the examples of this specification;

[0019] Figure 4 A schematic diagram of the frequency distribution of four key aerosols provided for the embodiments of this specification;

[0020] Figure 5 A schematic diagram of the time series of total surface nitrate nitrogen concentration under different environmental conditions in four seasons in the Yangtze River Delta region simulated by the present application scheme provided in the embodiments of this specification;

[0021] Figure 6 A schematic diagram comparing the computing speed of the numerical computing solution under different computing configurations provided in the embodiments of the present application and the artificial intelligence solution of the present application;

[0022] Figure 7 A schematic diagram of the structure of an electronic device provided in an embodiment of this specification. DETAILED DESCRIPTION

[0023] The embodiments of this specification provide an artificial intelligence simulation method, device, equipment and storage medium for atmospheric aerosol chemical processes.

[0024] In order to enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below in conjunction with the drawings in the embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this application, not all of them. Based on the embodiments of this specification, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of this application.

[0025] like Figure 1As shown, Figure 1 A flow chart of an artificial intelligence simulation method for atmospheric aerosol chemical processes provided for one or more embodiments of this specification is applied to a non-hydrostatic atmospheric numerical model including aerosol chemistry, wherein the non-hydrostatic atmospheric numerical model includes a target model and other parts.

[0026] Figure 1 The process in may include the following steps:

[0027] S102: Obtaining three-dimensional initial concentrations and environmental factors of various aerosols and related gases from the non-hydrostatic atmospheric numerical model; wherein the environmental factors include temperature, pressure, air density and water vapor mixing ratio, and the various aerosols include inorganic aerosols.

[0028] In conventional refined non-hydrostatic atmospheric numerical models, atmospheric chemical processes are coupled online, that is, non-hydrostatic atmospheric numerical models can simulate and calculate the concentration changes of chemical substances after a certain time step based on environmental factors (including temperature, pressure, humidity, etc.) and a given number of equations.

[0029] While in this application, we only use the AI ​​mode (i.e. the target model part) to simulate aerosol interactions and chemistry, and the rest of the modes (i.e. the rest part) keep the original numerical models.

[0030] like Figure 2 As shown, Figure 2 A schematic diagram of the model architecture used in an artificial intelligence model of aerosol chemistry and interactions (AIMACI) provided in an embodiment of this specification.

[0031] Environmental factors mainly include key meteorological variables that affect chemistry: temperature, pressure, air density and water vapor mixing ratio.

[0032] Since the scheme of this application can handle all important major aerosol species at urban, regional and global scales, including sulfate (SO4 2- ), nitrate (NO3 - ), chloride (Cl - ), carbonate (CO3 2- ), ammonium (NH4 + ), sodium (Na+), calcium (Ca 2+ ), black carbon (BC), organic carbon (OC), other inorganic matter (OIN), mineral dust (DUST), methane sulfonic acid (MSA) and liquefied water content of aerosol (Water), etc.

[0033] Among them, organic aerosols also play a vital role in numerical simulations, but due to their inherent chemical complexity, they are not simulated and predicted in this application.

[0034] In addition, there are some inorganic aerosol species that are not output variables. However, they play an important role in affecting the acidity or alkalinity of the atmospheric environment, which in turn affects the formation of overall aerosols. Therefore, these inorganic aerosol species are necessary input variables to ensure that the model accurately reflects the conditions that affect aerosol production. Specifically, they include mineral dust, black carbon (BC), organic carbon (OC), calcium salts and carbonates. In addition, some related gases have a direct ion exchange relationship with the inorganic aerosols, and these gases are also needed as necessary inputs and outputs. For example, the related gases (Gas species) include H2SO4, HNO3, NH3, HCL and methanesulfonic acid, etc.

[0035] That is, in this application, the inorganic aerosols for concentration prediction mainly include: SO4 2- 、NO3 - NH4 + 、Na + , Cl - , MSA|, aerosol water content and aerosol number concentration, and the related gas (Gas species). As shown in Table 1, Table 1 is a schematic diagram of the input and output variables provided in the embodiments of the present application.

[0036] Table 1 Input and output variables of the aerosol chemistry and interaction artificial intelligence model

[0037]

[0038] After the initial time point and boundary conditions are given, the three-dimensional initial concentration of aerosol and the three-dimensional initial concentration of related gases can be given by the non-hydrostatic atmospheric numerical model.

[0039] S104, according to the three-dimensional initial concentration and environmental factors, using a pre-trained target model to predict the predicted concentration of the inorganic aerosol and related gases after a specified step length.

[0040] The target model is Figure 2 The AIMACI part is shown in Figure 1. It is trained in advance with data generated by a non-hydrostatic atmospheric numerical model.

[0041] For example, to generate training, validation, and test datasets, we performed a WRF-Chem simulation over eastern China from 2019-03-01 00:00 to 2019-03-19 23:00 UTC. The simulation results were split into: the first 16 days starting at 2019-03-02 00:00 UTC were used as the training set, the penultimate day was used as the validation set, and the last day was used as the test set.

[0042] The simulation has a horizontal resolution of 0.2° and covers 140 × 105 grid cells within the geographical range of 107.1°E–127.9°E and 19.7°N–47.5°N, with 49 vertical layers extending to 50 hPa.

[0043] A 2-minute dynamic time step and a 1-hour chemical time step were used during the simulation. The concentrations of aerosol species and gas species related to air-particle partitioning were recorded every hour, as well as key meteorological variables (i.e., environmental factors) that affect chemistry: temperature, pressure, air density, and water vapor mixing ratio.

[0044] Each training sample includes 65 input features (4 meteorological variables, 5 gases and 14 aerosols with 4 different particle sizes) and 37 output targets (5 gases and 8 aerosols with 4 different particle sizes). The 4 different particle sizes are 0.039 microns to 0.156 microns, 0.156 microns to 0.625 microns, 0.625 microns to 2.5 microns, and 2.5 microns to 10.0 microns.

[0045] The target model includes an embedding layer, an integrator, and an output representation layer. The input embedding layer receives meteorological variables and initial concentrations of various aerosols and related gases as input features, and maps the input features into a potential high-dimensional subspace, wherein the various aerosols include inorganic aerosols; the integrator is used to learn the nonlinear process of the predicted concentrations of the meteorological variables and various aerosols and related gases from the latent subspace, and integrate them over time; the output representation layer maps the integration results of the latent high-dimensional subspace back to the predicted concentrations of the various aerosols and related gases;

[0046] like Figure 2 As shown in , the integrator contains a stack of multiple identical blocks, each block contains a multi-head attention layer and a feedforward layer, the multi-head attention layer learns the reaction relationship between different chemicals from multiple angles, the head attention layer and the feedforward layer are residually connected, and layer normalization is performed.

[0047] Furthermore, the validation data subset is used to fine-tune the hyperparameters and configuration of the trained model; the test data subset is used to evaluate the generalization ability of the trained model; and the model that has undergone parameter training, hyperparameter fine-tuning, and meets the generalization ability evaluation is determined as the target model.

[0048] When evaluating the model, the following indicators are used to evaluate the generalization ability of the trained model:

[0049]

[0050] in, represents the predicted concentration of the trained model, C represents the predicted concentration of the original numerical algorithm, Represents the average value of the predicted concentration of the original numerical algorithm, N lat Indicates the number of grids in the width direction, N lon represents the number of grids in the length direction, represents the predicted concentration of the artificial intelligence method, Represents the average of the predicted concentrations by the artificial intelligence algorithm.

[0051] After the test was passed, in order to facilitate the coupling of the AIMACI scheme written in Python with the numerical model based on Fortran, we used PyTorch's TorchScript and LibTorch tools to encapsulate the AIMACI scheme into a static library that can be called by the numerical model. That is, the target model can be used as a static library or dynamic library in a packaged form, and coupled with the rest of the non-hydrostatic atmospheric numerical model through a C++ function interface to obtain a coupled model. Coupling artificial intelligence algorithms of different complexity with a series of numerical models provides a highly flexible and easy plug-and-play solution.

[0052] S106, the remaining part iterates the environmental factors and other processes according to the predicted concentration of the inorganic aerosol.

[0053] like Figure 2 As shown in the figure, after the AIMACI part simulates the interaction and chemical process of some inorganic aerosols, the rest of the part maintains the original numerical mode and iterates again in units of step size based on the results of the AIMACI part. The results of the iteration are used as the input of the next iteration until the iteration termination condition is reached.

[0054] By obtaining the three-dimensional initial concentrations and environmental factors of various aerosols and related gases from the non-hydrostatic atmospheric numerical model; wherein the environmental factors include temperature, pressure, air density and water vapor mixing ratio, and the various aerosols include inorganic aerosols; according to the three-dimensional initial concentrations and environmental factors, the predicted concentrations of the inorganic aerosols and related gases after a specified step length are predicted using a pre-trained target model; the remaining part iterates the environmental factors and other processes according to the predicted concentrations of the inorganic aerosols and related gases, thereby realizing alternative calculations of aerosol chemical processes in the numerical model, realizing accurate simulation of the predicted concentrations of inorganic aerosols and related gases, and further realizing global high-fidelity simulation accuracy and significantly improved computational efficiency.

[0055] In order to verify the solution of the present application, the applicant also conducted the following experiments.

[0056] Four key aerosol types were selected, including chemical type, liquid water content in aerosol and number concentration of aerosol. The simulation was performed using the scheme of this application, and the following results were obtained: Figure 3 Results shown. Figure 3 This is a schematic diagram of the data density and distribution of the column concentrations of the four main aerosols provided in the embodiments of this specification, specifically including sulfate, nitrate, liquid water content in aerosols, and number concentration of aerosols.

[0057] Among them, sulfate mainly comes from fossil fuel combustion emissions and plays a vital role in acid rain, aerosol formation and the interaction between aerosols and clouds; nitrate formed by oxidation of nitrogen oxides is the main inorganic component of aerosols that affects air quality and ecosystem health; the liquid water content in aerosols is crucial for understanding the hygroscopicity of particles and their ability to serve as cloud condensation nuclei, thereby affecting cloud formation and precipitation processes; the number concentration of aerosols is a key indicator for assessing the overall aerosol load and its direct impact on visibility, radiation balance and climate feedback mechanisms. The evaluation of these four aerosols can fully reflect the effectiveness of the scheme in simulating aerosol chemistry.

[0058] The numerical simulation results show that the column concentrations of sulfate, nitrate and liquid water in the aerosol are high in the range of 0.156-0.625 μm, while the column concentrations are significantly larger in the range of 0.039-0.156 μm. This indicates that although the number of small-sized particles is large; their overall contribution to the total mass is not significant because they have a smaller mass compared to large-sized particles.

[0059] In addition, this application also comprehensively evaluates the scheme of this application, including: (1) Stable and accurate simulation capability: The AIMACI scheme should accurately reproduce the spatiotemporal and size distributions of various aerosols without rapidly accumulating errors during the simulation process. (2) Robust generalization capability: The AIMACI scheme should be applicable to scenarios other than the training data, such as different seasons, showing its robustness under various environmental conditions. (3) High computational efficiency: Compared with traditional numerical schemes, the AIMACI scheme should provide higher computational efficiency, which is critical for high-resolution, long-term simulations.

[0060] like Figure 4 As shown, Figure 4 Schematic diagram of the frequency distribution of the four key aerosols provided for the examples of this specification, and the last column shows the predicted concentrations of the surface concentrations of these key aerosol types. The calculation results cover the area from 109.1°E to 125.9°E and 22.1°N to 360 44.9°N, and the data results are averaged over the entire simulation period.

[0061] It can be seen that the frequency distributions of sulfate and nitrate surface concentrations show relatively uniform patterns, while the liquid water content and aerosol number concentration show extreme values, resulting in significantly skewed distributions. The AIMACI scheme accurately captures these distributions. It is worth noting that there are significant differences in the PSDs of different aerosol types. Sulfate and nitrate concentrations peak within 0.156-0.625 μm, while the liquid-liquid water content of aerosols is mainly concentrated in 2.5-10.0 μm, and the aerosol concentration is mainly concentrated in 0.039-0.156 μm. These findings highlight the complexity of accurately simulating PSD, and the near-perfect performance of the scheme of the present application in reproducing these complex patterns, with stable and accurate simulation capabilities.

[0062] like Figure 5 As shown, Figure 5 A schematic diagram of the time series of total surface nitrate nitrogen concentration under different environmental conditions in four seasons in the Yangtze River Delta region simulated by the scheme of the present application provided in the embodiments of this specification. Although there were occasional significant deviations during the simulation in July, such as the significant overestimation on July 20, this anomaly did not lead to a rapid escalation of the error or interruption of subsequent simulation results. On the contrary, as the simulation proceeded, the error gradually decreased, making the results of the scheme of the present application close to those of the numerical scheme. It can be seen that the scheme of the present application maintains a high consistency with the numerical format, showing its strong robust generalization ability.

[0063] like Figure 6 As shown, Figure 6 Schematic diagram comparing the calculation speed of the numerical mode and the scheme of the present application under different computing configurations provided by the embodiments of the present application. When using a single CPU core, the scheme of the present application achieves a calculation acceleration of about 5 times compared to the conventional numerical scheme, with a time cost of 48.51 seconds, while the time cost of the numerical scheme is 229.74 seconds. When using a single GPU, this progress is further expanded, and the scheme of the present application completes the calculation in only 0.83 seconds, which is about 277 times faster than the numerical scheme running on a single CPU core. In addition, the simulation speed of testing the artificial intelligence scheme on the GPU shows that the future implementation of heterogeneous computing platforms integrating CPUs and GPUs will produce significant enhancements in computing efficiency.

[0064] In summary, in conventional numerical models, solving the rigid differential equations that control aerosol chemistry is significantly time-consuming, and the strict time integration scheme required for numerical stability leads to limited breakthroughs in simulation speed using available numerical techniques. This study developed and evaluated a new artificial intelligence model of aerosol chemistry and interactions, called AIMACI, to achieve fast and accurate simulation of aerosol chemical processes.

[0065] In the second aspect, based on the same idea, one or more embodiments of this specification also provide a device corresponding to the above method, such as Figure 7 As shown, Figure 7 A schematic diagram of the structure of an electronic device provided in an embodiment of this specification, the device includes:

[0066] at least one processor; and,

[0067] a memory communicatively connected to the at least one processor; wherein,

[0068] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to the first aspect.

[0069] In the third aspect, based on the same idea, the embodiment of this specification also provides a non-volatile computer storage medium corresponding to the above method, storing computer executable instructions. When the computer reads the computer executable instructions in the storage medium, the instructions enable one or more processors to execute the method described in the first aspect.

[0070] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device, equipment, and non-volatile computer storage medium embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0071] The above is a description of a specific embodiment of the specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0072] The above description is only one or more embodiments of this specification and is not intended to limit this specification. For those skilled in the art, one or more embodiments of this specification may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of one or more embodiments of this specification shall be included in the scope of the claims of this specification.

Claims

1. An artificial intelligence simulation method for atmospheric aerosol chemical processes, applied to a non-hydrostatic atmospheric numerical model including aerosol chemistry, wherein the non-hydrostatic atmospheric numerical model includes a target model and other parts, the method comprising: Obtaining three-dimensional initial concentrations and environmental factors of multiple aerosols and related gases from the non-hydrostatic atmospheric numerical model; wherein the environmental factors include temperature, pressure, air density and water vapor mixing ratio, the multiple aerosols include inorganic aerosols, and the related gases are gases that have an ion exchange relationship with the inorganic aerosols; According to the three-dimensional initial concentration and environmental factors, a pre-trained target model is used to predict the predicted concentration of the inorganic aerosol and related gases after a specified step length; The remainder of the process iterates the environmental factors and other processes based on the predicted concentrations of the inorganic aerosols and related gases.

2. The method of claim 1, wherein: The target model is pre-trained in the following way: A conventional non-hydrostatic atmospheric numerical model is used to generate a training data subset, a validation data subset and a test data subset; The training data subset is used to perform parameter training on the embedding layer, the integrator and the output representation layer included in the initial model to generate a trained model, wherein the input embedding layer receives the meteorological variables, the predicted concentrations of multiple aerosols and related gases as input features, and maps the input features into a potential high-dimensional subspace, wherein the multiple aerosols include inorganic aerosols; the integrator is used to learn the nonlinear process of the predicted concentrations of the meteorological variables, multiple aerosols and related gases from the potential subspace, and integrate them over time; the output representation layer maps the integration result of the potential high-dimensional subspace back to the predicted concentrations of the multiple aerosols and related gases; Using the validation data subset to fine-tune hyperparameters and configurations of the trained model; The test data subset is used to evaluate the generalization ability of the trained model; The model that has undergone parameter training, hyperparameter fine-tuning, and met the generalization ability evaluation is determined as the target model.

3. The method of claim 2, wherein: The integrator contains a stack of multiple identical blocks, each block contains a multi-head attention layer and a feedforward layer, the multi-head attention layer learns the reaction relationship between different chemical substances from multiple angles, the head attention layer and the feedforward layer are residually connected, and layer normalization is performed.

4. The method of claim 1, wherein: The inorganic aerosol includes: SO4 2- 、NO3 - NH4 + 、Na + , Cl - , MSA, aerosol water content and aerosol number concentration; the predicted concentration includes the predicted concentrations of four particle sizes from 0.039 microns to 10 microns, and the related gases include H2SO4, HNO3, NH3, HCL and methane sulfonic acid.

5. The method of claim 2, wherein: The test data subset is used to evaluate the generalization ability of the trained model, including: The following indicators are used to evaluate the generalization ability of the trained model: in, represents the predicted concentration of the trained model, C represents the predicted concentration of the original numerical algorithm, Represents the average value of the predicted concentration of the original numerical algorithm, N lat Indicates the number of grids in the width direction, N lon represents the number of grids in the length direction, represents the predicted concentration of the artificial intelligence method, Represents the average of the predicted concentrations by the artificial intelligence algorithm.

6. The method of claim 1, wherein: In the non-hydrostatic atmosphere numerical model, the target model is a static library or a dynamic library in a packaged form, and is coupled with the remaining parts through a C++ function interface to obtain a coupled model.

7. An electronic device comprising: at least one processor; as well as, a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 6.

Images