Industrial chain electrical load sensitive factor extraction method based on deep learning

Abstract

The invention discloses an industrial chain electrical load sensitive factor extraction method based on deep learning, and the method comprises the steps: obtaining historical electricity price, meteorological data, industry classification and power consumption data, and carrying out the preprocessing of the data; constructing a graph neural network according to the preprocessed data; performing model training based on the constructed graph neural network; performing hierarchical attribution and sensitivity analysis; and analyzing and applying a result. According to the method, the load relation between different industries is analyzed by constructing a graph neural network model, the accuracy of load sensitivity analysis is improved by utilizing a self-supervised learning and hierarchical attribution technology, especially in a rapidly fluctuating market environment, the influence degree of external factors such as economy and meteorology on the industry chain electrical load is automatically extracted and quantified, and the load sensitivity analysis accuracy is improved. And the accuracy and sensitivity of electrical load prediction are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of industrial chain electricity load management, and in particular to a method for extracting sensitive factors of industrial chain electricity load based on deep learning. Background Art

[0002] Most of the existing power load analysis methods rely on traditional statistical models (such as regression analysis, time series analysis, etc.). These models are usually based on historical data and focus on load forecasting by observing the correlation between data. However, with the continuous changes in external factors such as economy, meteorology, and policies, especially in a complex industrial chain environment, the load analysis method that relies solely on statistical models has shown significant limitations, which are specifically manifested in the following points. First, the existing statistical models do not analyze economic, meteorological and other factors in depth, and it is difficult to flexibly adapt to the complex external environment of the industrial chain, especially the differentiated responses of different user categories, such as the impact of residential users depends on temperature, while the impact of industrial users depends on production processes; second, traditional load forecasting methods focus on a single industry and ignore the mutual influence of upstream and downstream power demand in the industrial chain, especially when the industrial chain fluctuates or sudden demand changes, the prediction accuracy is low; third, different industries and users have different sensitivities to external factors, but existing models often use general processing, resulting in insufficient prediction accuracy, especially in a rapidly fluctuating market environment, the model response is insensitive; fourth, traditional methods are difficult to capture the load change characteristics of emergencies (such as extreme weather), lack a detection mechanism to deal with load mutations, and limit the stability and dispatching accuracy of the power system.

[0003] In summary, the current power load analysis method has significant deficiencies in adapting to the complex industrial chain environment and the interaction of multiple factors, and cannot meet the increasingly complex reality of the power demand structure. Therefore, there is an urgent need for a new method that can automatically identify and quantify the impact of various external factors on the load, especially in dynamic and sudden fluctuation scenarios to achieve more accurate load forecasting. Summary of the invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides a method for extracting sensitive factors of electricity load in an industrial chain based on deep learning. It analyzes the load relationship between different industries by constructing a graph neural network model, and uses self-supervised learning and hierarchical attribution technology to improve the accuracy of load sensitivity analysis, especially in a rapidly fluctuating market environment. It automatically extracts and quantifies the impact of external factors such as economy and meteorology on the electricity load of the industrial chain, thereby improving the accuracy and sensitivity of electricity load forecasting.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: a method for extracting sensitive factors of power load in an industrial chain based on deep learning, comprising:

[0007] Obtain historical electricity prices, meteorological data, industry classifications and electricity consumption data, and pre-process the data;

[0008] Build a graph neural network based on the preprocessed data;

[0009] Perform model training based on the constructed graph neural network;

[0010] Conduct stratified attribution and sensitivity analyses;

[0011] Result analysis and application.

[0012] As a preferred solution of the method for extracting sensitive factors of power load in the industrial chain based on deep learning described in the present invention, the preprocessing includes that for any feature x, the normalization process can be expressed as:

[0013]

[0014] Among them, x min and x max are the minimum and maximum values ​​of feature x respectively. The processed data D = {x1, x2, ..., x n} will serve as the basic data for the input of the graph neural network GNN.

[0015] As a preferred solution of the method for extracting sensitive factors of power load in the industrial chain based on deep learning described in the present invention, the preprocessing includes, after normalization, using the random forest algorithm to screen the features, and the importance score of the random forest is calculated as follows:

[0016]

[0017] Where T is the number of decision trees, ΔG t Represents feature x j The reduction of the Gini coefficient in tree t, let the selected features be F = {f1, f2, ..., f k}, the feature will be used as the node feature in the graph neural network, representing the input information of the nodes in each industry. In the GNN construction, the node feature h v That is, it comes from this screening step.

[0018] As a preferred solution of the method for extracting sensitive factors of power load in the industrial chain based on deep learning described in the present invention, the construction of the graph neural network includes constructing a graph neural network according to the feature set F obtained by screening in the first step, and treating each industry in the industrial chain as a node V = {v1, v2, ..., vm}, the feature vector of the node is the key factor h obtained by the first step screening v =F, the connection weight w calculated from the data set uv Representing the correlation of electricity demand between industries, the constructed graph G=(V, E) will represent the power load relationship between the upstream and downstream of the industrial chain.

[0019] As a preferred solution of the method for extracting sensitive factors of power load in the industrial chain based on deep learning described in the present invention, wherein: the construction of the graph neural network includes, in the GNN, the feature vector h generated in the first step v As the initial node feature, the message passing mechanism is applied to simulate the correlation influence between different industries. Each node aggregates and updates the information of neighbor nodes at each layer through the following formula:

[0020]

[0021] in, is the hidden state of node v at layer k, is the neighbor set of node v, W (k) is the weight matrix of the kth layer, σ is the activation function, The initial features are obtained from data preprocessing and feature screening. The mechanism accumulates the information of upstream and downstream nodes at each layer to simulate how different external factors transmit and affect load fluctuations.

[0022] As a preferred solution of the method for extracting sensitive factors of power load in the industrial chain based on deep learning described in the present invention, the construction of the graph neural network includes the edge weights w between nodes in different industries. uv It represents the strength of association between the upstream and downstream of the industrial chain. The weight value is calculated by the covariance of the historical load data between the nodes:

[0023]

[0024] Among them, L u and L v are the historical load data of nodes u and v respectively, and The calculated w is its standard deviation. uv Will affect the strength of graph neural network message passing.

[0025] As a preferred solution of the method for extracting sensitive factors of power load in the industrial chain based on deep learning described in the present invention, the model training based on the constructed graph neural network includes self-supervised learning through a masked autoencoder, using the graph structure G constructed in the second step as the input of the model, selecting some node features for masking, and reconstructing the masked features using MAE. The loss function of the model is:

[0026]

[0027] in, represents the set of obscured nodes, h i is the real node feature, To predict features, the pre-training process uses self-supervised learning on the graph structure to improve the generalization ability of the model;

[0028] Based on the graph structure, the model is fine-tuned in a supervised manner using data with large load fluctuations. The training goal is to predict the load value of each industry through the model. The loss function of the model is defined as:

[0029]

[0030] Among them, y i is the actual load value, To predict the load value, λ is the regularization coefficient, which is used to control the model complexity and prevent overfitting. This training step ensures that the model can learn the load change pattern under the influence of different industrial chain associations and external factors.

[0031] As a preferred solution of the method for extracting sensitive factors of power load in the industrial chain based on deep learning described in the present invention, the hierarchical attribution and sensitivity analysis includes hierarchical attribution of the trained model using layer-by-layer correlation propagation, calculating the sensitivity of each external factor to load fluctuation layer by layer, and LRP assigning the model output to the input features layer by layer through back propagation. The calculation formula is:

[0032]

[0033] Among them, R j is the correlation of feature j, w jk is the connection weight, a j is the feature activation value.

[0034] A computer device includes a memory and a processor, wherein the memory stores a computer program, and is characterized in that when the processor executes the computer program, it implements the steps of a method for extracting sensitive factors of electricity load in an industrial chain based on deep learning.

[0035] A computer-readable storage medium having a computer program stored thereon, characterized in that when the computer program is executed by a processor, the steps of a method for extracting sensitive factors of electricity load in an industrial chain based on deep learning are implemented.

[0036] Beneficial effects of the present invention: (1) In-depth analysis of the impact of economic and meteorological factors, and adaptation to the complex environment of the industrial chain. This patent introduces masked autoencoders (MAE) and graph neural networks (GNNs), which can use unlabeled data in the absence of labels, and automatically capture the differentiated impact of external factors such as economy and meteorology on different industries through self-supervised learning. For example, the present invention can identify the high sensitivity of residential users to temperature, or the dependence of industrial users on production processes, thereby improving the adaptability and accuracy of load forecasting.

[0037] (2) Consider the correlation of electricity demand in the upstream and downstream of the industrial chain to improve prediction accuracy. The present invention constructs an upstream and downstream correlation model based on a graph neural network, which can dynamically capture the supply and demand relationship of different industries in the industrial chain, ensuring that the impact of industrial chain fluctuations on electricity demand is fully considered during load forecasting. Especially when demand changes suddenly or the industrial chain adjusts, the present invention can provide more accurate prediction support.

[0038] (3) Differentiate the sensitivity of users in different industries and improve the response flexibility of the model. The present invention combines the layer-by-layer relevance propagation (LRP) in the layered attribution technology to identify the sensitivity of each external factor to the load, thereby refining the differentiated needs of different industries. Through sensitivity analysis, the present invention can effectively cope with the rapidly fluctuating market environment and improve the response speed and flexibility of the model in the prediction process.

[0039] (4) Enhanced adaptability to emergencies and ensuring the stability of the power system. The MAE model improves the robustness of the model during self-supervised learning, so that it still has strong prediction stability under emergencies (such as extreme weather). Even if sudden load changes occur, the present invention can quickly identify and adjust the prediction results, thereby supporting the stability and efficient scheduling of the power system. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0041] Figure 1 A schematic flow chart of a method for extracting sensitive factors of electricity load in an industrial chain based on deep learning is provided as an embodiment of the present invention.

[0042] Figure 2 A schematic diagram of the structure of a graph neural network industry chain model for a method for extracting sensitive factors of electricity load in an industry chain based on deep learning is provided as an embodiment of the present invention.

[0043] Figure 3 A schematic diagram of the sensitivity analysis process of a hierarchical attribution technology for a method for extracting sensitive factors of electricity load in an industrial chain based on deep learning is provided as an embodiment of the present invention. DETAILED DESCRIPTION

[0044] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0045] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0046] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0047] The present invention is described in detail with reference to schematic diagrams. When describing the embodiments of the present invention, for the sake of convenience, the cross-sectional diagrams showing the device structure will not be partially enlarged according to the general scale, and the schematic diagrams are only examples, which should not limit the scope of protection of the present invention. In addition, in actual production, the three-dimensional dimensions of length, width and depth should be included.

[0048] At the same time, in the description of the present invention, it should be noted that the directions or positional relationships indicated by the terms "upper, lower, inner and outer" are based on the directions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore cannot be understood as limiting the present invention. In addition, the terms "first, second or third" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0049] In the present invention, unless otherwise clearly specified and limited, the terms "install, connect, connect" should be understood in a broad sense, for example: it can be a fixed connection, a detachable connection or an integral connection; it can also be a mechanical connection, an electrical connection or a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0050] Example 1, reference Figure 1-Figure 3 , which is the first embodiment of the present invention, and provides a method for extracting sensitive factors of power load in an industrial chain based on deep learning, comprising:

[0051] The first step is data preprocessing. Obtain historical electricity prices, meteorological data (such as temperature and humidity), industry classification and electricity consumption data. All data need to be normalized to eliminate differences in different dimensions. For any feature x, the normalization process can be expressed as:

[0052]

[0053] Among them, x min and x max are the minimum and maximum values ​​of feature x respectively. These processed data D = {x1, x2, ..., x n} will serve as the basic data for the input of the Graph Neural Network (GNN).

[0054] After normalization, the random forest algorithm is used to screen the features and retain the features that have a greater impact on load fluctuations. The importance score of the random forest is calculated as follows:

[0055]

[0056] Where T is the number of decision trees, ΔG t Represents feature x j The reduction of Gini coefficient in tree t. Let the selected features be F = {f1, f2, ..., f k}, these features will be used as node features in the graph neural network, representing the input information of nodes in each industry. In the subsequent GNN construction, the node feature h v This screening step ensures that the graph neural network focuses on the factors that have the most significant impact on load fluctuations.

[0057] The second step is model construction. Based on the feature set F obtained in the first step, a graph neural network is constructed. Each industry in the industrial chain is regarded as a node V = {v1, v2, ..., v m}, the feature vector of the node is the key factor h obtained by the first step screening v=F, the connection weight w calculated from the data set uv The constructed graph G = (V, E) represents the power load relationship between the upstream and downstream of the industrial chain.

[0058] In GNN, the feature vector h generated by the first step v As the initial node feature, the message passing mechanism is applied to simulate the correlation impact between different industries. Each node aggregates and updates the information of neighboring nodes at each layer through the following formula:

[0059]

[0060] in, is the hidden state of node v at layer k, is the neighbor set of node v, W (k) is the weight matrix of the kth layer, σ is the activation function (such as ReLU), is the initial feature obtained from data preprocessing and feature screening. This mechanism simulates how different external factors affect load fluctuations by accumulating information from upstream and downstream nodes at each layer.

[0061] In the constructed graph structure, the edge weights w between nodes in different industries are uv Indicates the strength of association between the upstream and downstream of the industrial chain. The weight value is calculated by the covariance of historical load data between nodes:

[0062]

[0063] Among them, L u and L v are the historical load data of nodes u and v respectively, and is its standard deviation. The calculated w uv Will affect the strength of graph neural network message passing, Figure 2 The structure diagram of the graph neural network industry chain model is shown.

[0064] The third step is to train the model based on the constructed graph neural network. Self-supervised learning is performed through Masked Autoencoders (MAE). The graph structure G constructed in the second step is used as the input of the model, some node features are selected for masking, and the masked features are reconstructed using MAE. The loss function of the model is:

[0065]

[0066] in, represents the set of obscured nodes, h i is the real node feature, The pre-training process uses self-supervised learning to improve the generalization ability of the model on the graph structure, providing a more reliable basis for subsequent load fluctuation prediction.

[0067] Based on the graph structure in the second step, the model is fine-tuned using data with large load fluctuations. The training goal is to predict the load value of each industry through the model. The loss function of the model is defined as:

[0068]

[0069] Among them, y i is the actual load value, To predict the load value, λ is the regularization coefficient, which is used to control the model complexity and prevent overfitting. This training step ensures that the model can learn the load change pattern under the influence of different industrial chain associations and external factors.

[0070] Step 4: Hierarchical attribution and sensitivity analysis. The model trained in step 3 is hierarchically attributed using layer-wise relevance propagation (LRP), and the sensitivity of each external factor to load fluctuation is calculated layer by layer. LRP assigns the model output to the input features layer by layer through back propagation. The calculation formula is:

[0071]

[0072] Among them, R j is the correlation of feature j, w jk is the connection weight, a j is the feature activation value. Figure 3 A process diagram of sensitivity analysis of hierarchical attribution technique is shown.

[0073] The LRP method can identify and quantify the impact of each external factor on the load, providing explanatory support for the model. Based on the LRP results, the sensitivity of each feature to load fluctuations is calculated, and the ranking of key factors affecting load changes is obtained. These sensitivity analysis results provide data support for power companies to optimize power dispatch and load forecasting, helping to achieve more accurate power load forecasting.

[0074] Step 5: Result analysis and application. The output of the overall solution is a ranking table of the sensitivity of loads in different industries to external factors, and the load transmission mechanism in the upstream and downstream relationships of the industrial chain. Through this analysis result, power companies can identify highly sensitive factors and optimize power dispatching, thereby improving the accuracy of power load forecasting. Table 1 shows a comparison table between the present invention and the prior art.

[0075] Table 1 Comparison between the present invention and the prior art

[0076]

[0077]

[0078] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should be included in the scope of the claims of the present invention.

[0079] Example 3

[0080] The third embodiment of the present invention is different from the first two embodiments in that:

[0081] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program codes.

[0082] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0083] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0084] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0085] Although the preferred embodiments of the present application have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0086] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A method for extracting sensitive factors of power load in an industrial chain based on deep learning, characterized by: include, Obtain historical electricity prices, meteorological data, industry classifications and electricity consumption data, and pre-process the data; Build a graph neural network based on the preprocessed data; Perform model training based on the constructed graph neural network; Conduct stratified attribution and sensitivity analyses; Result analysis and application.

2. The method for extracting sensitive factors of power load in an industrial chain based on deep learning as claimed in claim 1, characterized in that: The preprocessing includes, for any feature x, the normalization process can be expressed as: Among them, x min and x max are the minimum and maximum values ​​of feature x respectively. The processed data D = {x1, x2, ..., x n } will serve as the basic data for the input of the graph neural network GNN.

3. The method for extracting sensitive factors of power load in an industrial chain based on deep learning as claimed in claim 2, characterized in that: The preprocessing includes, after normalization, using the random forest algorithm to screen the features, and the importance score of the random forest is calculated as follows: Where T is the number of decision trees, ΔG t Represents feature x j The reduction of the Gini coefficient in tree t, let the selected features be F = {f1, f2, ..., f k }, the feature will be used as the node feature in the graph neural network, representing the input information of the nodes in each industry. In the GNN construction, the node feature h v That is, it comes from this screening step.

4. The method for extracting sensitive factors of power load in an industrial chain based on deep learning as claimed in claim 3, characterized in that: The construction of the graph neural network includes constructing a graph neural network according to the feature set F obtained by the first step, treating each industry in the industrial chain as a node V = {v1, v2, ..., v m }, the feature vector of the node is the key factor h obtained by the first step screening v =F, the connection weight w calculated from the data set uv Representing the correlation of electricity demand between industries, the constructed graph G=(V, E) will represent the power load relationship between the upstream and downstream of the industrial chain.

5. The method for extracting sensitive factors of power load in an industrial chain based on deep learning as claimed in claim 4, characterized in that: The construction of the graph neural network includes, in the GNN, the feature vector h generated in the first step v As the initial node feature, the message passing mechanism is applied to simulate the correlation influence between different industries. Each node aggregates and updates the information of neighbor nodes at each layer through the following formula: in, is the hidden state of node v at layer k, is the neighbor set of node v, W (k) is the weight matrix of the kth layer, σ is the activation function, The initial features are obtained from data preprocessing and feature screening. The mechanism accumulates the information of upstream and downstream nodes at each layer to simulate how different external factors transmit and affect load fluctuations.

6. The method for extracting sensitive factors of power load in an industrial chain based on deep learning as claimed in claim 5, characterized in that: The construction of the graph neural network includes the edge weights w between nodes in different industries. uv It represents the strength of association between the upstream and downstream of the industrial chain. The weight value is calculated by the covariance of the historical load data between nodes: Among them, L u and L v are the historical load data of nodes u and v respectively, and The calculated w is its standard deviation. uv Will affect the strength of graph neural network message passing.

7. The method for extracting sensitive factors of power load in an industrial chain based on deep learning as claimed in claim 6, characterized in that: The model training based on the constructed graph neural network includes self-supervised learning through a masked autoencoder, using the graph structure G constructed in the second step as the input of the model, selecting some node features for masking, and reconstructing the masked features using MAE. The loss function of the model is: in, represents the set of obscured nodes, h i is the real node feature, To predict features, the pre-training process uses self-supervised learning on the graph structure to improve the generalization ability of the model; Based on the graph structure, the model is fine-tuned in a supervised manner using data with large load fluctuations. The training goal is to predict the load value of each industry through the model. The loss function of the model is defined as: Among them, y i is the actual load value, To predict the load value, λ is the regularization coefficient, which is used to control the model complexity and prevent overfitting. This training step ensures that the model can learn the load change pattern under the influence of different industrial chain associations and external factors.

8. The method for extracting sensitive factors of power load in an industrial chain based on deep learning as claimed in claim 7, characterized in that: The hierarchical attribution and sensitivity analysis includes hierarchical attribution of the trained model using layer-by-layer correlation propagation, calculating the sensitivity of each external factor to load fluctuation layer by layer, and LRP assigning the model output to the input features layer by layer through back propagation. The calculation formula is: Among them, R j is the correlation of feature j, w jk is the connection weight, a j is the feature activation value.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 8 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

Images