Power grid risk data processing method and device, equipment, storage medium and program product

Abstract

The embodiment of the invention provides a power grid risk data processing method and device, equipment, a storage medium and a program product. The method comprises the following steps: acquiring login roles of login users, and determining a first role feature of the login role of each login user from a pre-constructed role multi-dimensional matrix; determining a user feature vector of the login user according to the user data of the login user and the first role feature; extracting risk elements from the multi-source power grid risk data; determining association weights of the risk elements and the user feature vectors, and determining a dynamic content weight matrix based on the association weights; the dynamic content weight matrix comprises a weight corresponding to a risk early warning index of the login user; and according to the weight corresponding to the risk early warning index of each login user, determining a target risk factor corresponding to each login user, and according to the target risk factor, generating a power grid risk early warning content sent to the terminal of each login user. The method is used for improving the power grid risk early warning effect.

Description

Technical Field

[0001] The present application relates to the field of electric power technology, and in particular to a method, device, equipment, storage medium and program product for processing power grid risk data. Background Art

[0002] As the core of the power system, the safe and stable operation of the power grid is crucial to ensuring its safe and stable operation. Power grid risk warnings identify potential risks and issue warnings in advance by monitoring and analyzing real-time data on factors such as the power system's operating status, the health of power equipment, and the external environment, thereby preventing power grid safety incidents.

[0003] In related technologies, risk management systems analyze risk monitoring data related to the user's management scope to generate and push grid risk warning information to the user, allowing them to take timely measures to address grid risks. The data content and timing of monitoring data used to generate grid risk warning information often vary from user to user, leading to discrepancies in the risk warning information received.

[0004] Based on the above analysis, we can see that in related technologies, due to the discrepancies in the risk warning information obtained by each user, it is difficult for users at different management levels to collaborate, thus affecting the effectiveness of power grid risk warning. Therefore, how to improve the effectiveness of power grid risk warning has become an urgent problem to be solved. Summary of the Invention

[0005] The embodiments of the present application provide a power grid risk data processing method, apparatus, equipment, storage medium and program product to improve the effectiveness of power grid risk warning.

[0006] In a first aspect, an embodiment of the present application provides a method for processing power grid risk data, comprising:

[0007] Obtaining a login role of at least one login user, and determining a first role characteristic of each login user's login role from a pre-constructed multi-dimensional role matrix; wherein the first role characteristic of each login user includes the grid risk management level, jurisdiction, and grid risk data access rights to which the login user belongs;

[0008] For each logged-in user, determining a user feature vector of the logged-in user based on the user data of the logged-in user and the first role feature, wherein the user data includes historical operation data, risk management tasks, jurisdiction, and historical power grid risk information of the jurisdiction;

[0009] extracting at least one risk factor from multi-source power grid risk data;

[0010] Determining the association weights of each risk factor and each of the user feature vectors, and determining a dynamic content weight matrix based on the association weights; the dynamic content weight matrix includes weights corresponding to each logged-in user and at least one risk warning indicator configured for each logged-in user;

[0011] According to the weight corresponding to the risk warning indicator of each logged-in user, the target risk factor corresponding to each logged-in user is determined, and the power grid risk warning content sent to the terminal of each logged-in user is generated according to the target risk factor; wherein, the power grid risk warning content for different logged-in users is different.

[0012] In a possible implementation, the role multidimensional matrix is ​​obtained based on the following construction steps:

[0013] Obtaining job descriptions corresponding to a plurality of risk management roles in the risk management architecture of the power grid, extracting responsibility keywords from each job description, and constructing a role-responsibility mapping relationship table based on the responsibility keywords corresponding to each risk management role; wherein the plurality of risk management roles belong to different power grid risk management levels;

[0014] Based on the geographic information system, determine the topological relationship of power equipment within the respective jurisdictions of multiple risk management roles;

[0015] A dynamic authority adjustment factor is set for each risk management role. The dynamic authority adjustment factor is used to adjust the power grid risk data access permission threshold according to the power grid risk event level.

[0016] In a possible implementation, extracting at least one risk factor from multi-source power grid risk data includes:

[0017] Perform semantic analysis on the multi-source power grid risk data, and determine the at least one risk factor at multiple moments according to the semantic analysis results.

[0018] In a possible implementation, determining the association weights between each risk factor and each user feature vector, and determining a dynamic content weight matrix based on the association weights, includes:

[0019] The information of each risk factor and each user feature vector at multiple moments is input into a pre-built risk association model, the risk association model is used to calculate the association weight of each risk factor and each user feature vector, and the dynamic content weight matrix is ​​generated according to the association weight.

[0020] In one possible implementation, the risk association model includes a multi-head attention mechanism and a bidirectional long short-term memory sub-model; calculating the association weights between each risk factor and each user feature vector using the risk association model, and generating the dynamic content weight matrix based on the association weights, includes:

[0021] Utilizing the bidirectional long-short-term memory sub-model in the risk association model to extract the temporal characteristics of each risk factor from the information of each risk factor at multiple moments;

[0022] Utilize the multi-head attention mechanism to calculate the association weight between each risk factor and each user feature vector based on the temporal characteristics of each risk factor;

[0023] For each logged-in user's risk warning indicator, according to the associated weight corresponding to the logged-in user, the softmax function in the risk association model is used to determine the dynamic content weight matrix corresponding to the logged-in user.

[0024] In one possible implementation, the multi-head attention mechanism includes a first group of attention heads, a second group of attention heads, a third group of attention heads, and a fourth group of attention heads;

[0025] The first group of attention heads is used to calculate the association weight between the job responsibilities and the grid risk type in each user feature vector;

[0026] The second set of attention heads is used to calculate the spatial correlation weights between the jurisdiction and the risk propagation path in each user feature vector;

[0027] The third group of attention heads is used to calculate the compliance of the power grid risk data access rights and the power grid risk data in each user feature vector;

[0028] The fourth group of attention heads is used to calculate the correlation between the historical decision-making pattern and the risk factor information in each user feature vector.

[0029] In a possible implementation, the power grid risk management level includes a first management level, a second management level, and a third management level; wherein,

[0030] The first management level corresponds to the first jurisdiction and the first power grid risk event level; the second management level corresponds to the second jurisdiction and the second power grid risk event level; the third management level corresponds to the third jurisdiction and the third power grid risk event level; wherein,

[0031] The first jurisdiction is larger than the second jurisdiction; the second jurisdiction is larger than the third jurisdiction;

[0032] The first power grid risk event level is higher than the second power grid risk event level, and the second power grid risk event level is higher than the third power grid risk event level.

[0033] In a possible implementation, generating power grid risk warning content sent to terminals of respective logged-in users based on the target risk factors includes:

[0034] For logged-in users belonging to the first management level, an overview of the power grid risk situation and trend forecast are generated based on the target risk factors corresponding to the logged-in users;

[0035] For login users belonging to the second management level, operational-level risk control indicators and grid risk disposal progress are generated based on the target risk factors corresponding to the login users;

[0036] For login users belonging to the third management level, equipment-level inspection tasks and power grid risk emergency response information are generated based on the target risk factors corresponding to the login users.

[0037] In a second aspect, an embodiment of the present application provides a power grid risk data processing device, comprising:

[0038] A first acquisition module is configured to acquire a login role of at least one logged-in user and determine a first role characteristic of each logged-in user's login role from a pre-constructed multi-dimensional role matrix; wherein the first role characteristic of each logged-in user includes the grid risk management level, jurisdiction, and grid risk data access rights of the logged-in user;

[0039] a first determining module configured to determine, for each logged-in user, a user feature vector of the logged-in user based on user data of the logged-in user and the first role feature, wherein the user data includes historical operation data, risk management tasks, jurisdiction, and historical power grid risk information of the jurisdiction;

[0040] A second acquisition module is used to extract at least one risk factor from multi-source power grid risk data;

[0041] a second determination module, configured to determine an association weight between each risk factor and each user feature vector, and determine a dynamic content weight matrix based on the association weight; the dynamic content weight matrix includes a weight corresponding to each logged-in user and at least one risk warning indicator configured for each logged-in user;

[0042] A generation module is used to determine the target risk factor corresponding to each logged-in user based on the weight corresponding to the risk warning indicator of each logged-in user, and generate the power grid risk warning content sent to the terminal of each logged-in user based on the target risk factor; wherein the power grid risk warning content of different logged-in users is different.

[0043] In a third aspect, an embodiment of the present application provides a power grid risk data processing device, comprising: a memory, a processor;

[0044] The memory stores computer-executable instructions;

[0045] The processor executes the computer-executable instructions stored in the memory, so that the processor executes the above first aspect and / or various possible implementations of the first aspect.

[0046] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, in which computer-executable instructions are stored. When the computer-executable instructions are executed by a processor, they are used to implement the first aspect above and / or various possible implementation methods of the first aspect.

[0047] In a fifth aspect, an embodiment of the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the above first aspect and / or various possible implementation methods of the first aspect.

[0048] The power grid risk data processing method, device, equipment, storage medium and program product provided in the embodiments of the present application obtain the login role of the logged-in user, and determine the user feature vector of the logged-in user by using a pre-built role multi-dimensional matrix and the user data of the logged-in user; extract risk factors from multi-source power grid risk data, and determine the association weights between the risk factors and the user feature vectors; determine a dynamic content weight matrix corresponding to the risk warning indicators of the logged-in user based on the association weights; determine the target risk factors corresponding to the logged-in user according to the weights corresponding to the risk warning indicators of the logged-in user; generate power grid risk warning content sent to the terminals of each logged-in user according to the target risk factors; use unified multi-source power grid risk data to extract risk factors, and use the association relationship between the feature vectors of each logged-in user and the risk factors to generate power grid risk warning content for each logged-in user; each logged-in user collaborates on the basis of their respective power grid risk warning content uniformly generated using the same analysis data, reducing the problem of difficulty in collaborative work caused by information asymmetry, thereby improving the effect of power grid risk warning. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0050] Figure 1 Schematic diagram of the application scenario of the power grid risk data processing method provided in this application;

[0051] Figure 2A flowchart of a method for processing power grid risk data provided in an embodiment of the present application;

[0052] Figure 3 A schematic diagram of the process of constructing a multi-dimensional matrix of roles provided in an embodiment of the present application;

[0053] Figure 4 A schematic diagram of the structure of a power grid risk data processing device provided in an embodiment of the present application;

[0054] Figure 5 A schematic diagram of the structure of the power grid risk data processing device provided in an embodiment of the present application.

[0055] The above drawings illustrate specific embodiments of the present application, which will be described in more detail below. These drawings and the textual description are not intended to limit the scope of the present application in any way, but rather to illustrate the concepts of the present application to those skilled in the art by reference to specific embodiments. DETAILED DESCRIPTION

[0056] Exemplary embodiments will be described in detail herein, with examples illustrated in the accompanying drawings. In the following description, when referring to the drawings, identical numerals in different figures represent identical or similar elements, unless otherwise indicated. The embodiments described in the following exemplary embodiments are not intended to represent all embodiments consistent with the present application. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present application, as detailed in the appended claims.

[0057] Figure 1 The application scenario diagram of the power grid risk data processing method provided in this application is as follows: Figure 1 As shown, multiple users at different management levels use risk management system 10 to manage power grid risks. The multiple users include logged-in user 101, logged-in user 102, and logged-in user 103. When the power grid may encounter risks, for example, based on weather forecast monitoring data, it is determined that the power grid is about to be affected by a typhoon, the multiple logged-in users at different management levels need to obtain risk warning information within the scope of the power grid they manage and perform risk management based on the risk warning information.

[0058] In the related technology, for the logged-in users 101, 102 and 103 who log in to the risk management system 10, the risk management system obtains the monitoring data related to the management scope of the logged-in users from the power grid risk monitoring data according to the management information of different logged-in users, performs data analysis, and displays the power grid risk warning information to the logged-in users based on the analysis results, so that the logged-in users can take corresponding risk warning actions according to the risk warning information.

[0059] Based on the above scenarios, we can see that because different logged-in users often obtain inconsistent data content and time for monitoring data used to generate power grid risk warning information, the risk warning information obtained by different logged-in users may be biased. This makes it difficult for logged-in users at different management levels to collaborate, affecting the effectiveness of power grid risk warnings. Therefore, using related technologies for risk management has the technical problem of poor effectiveness of power grid risk warnings.

[0060] The power grid risk data processing method, device, equipment, storage medium and program product provided in this application are used to solve the above-mentioned technical problems.

[0061] The following specific embodiments describe in detail the technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0062] Figure 2 A flow chart of a method for processing power grid risk data provided in an embodiment of the present application is shown in FIG. Figure 2 As shown, the method includes:

[0063] S201. Obtain the login role of at least one login user, and determine the first role characteristics of the login role of each login user from a pre-built role multi-dimensional matrix; wherein the first role characteristics of each login user include the power grid risk management level, jurisdiction and power grid risk data access rights to which the login user belongs.

[0064] It should be noted that multiple managers perform risk management on the power grid at the same time, and managers log in to the risk management system to become login users of the risk management system.

[0065] S202. For each logged-in user, determine a user feature vector of the logged-in user based on the user data and first role feature of the logged-in user, wherein the user data includes historical operation data, risk management tasks, jurisdiction, and historical power grid risk information of the jurisdiction.

[0066] It should be noted that the first role characteristics include role characteristics that can be determined before the user logs into the risk management system; and the user data is data determined after the user logs into the risk management system.

[0067] For example, the division of the jurisdiction of the power grid has been determined before the user logs into the risk management system. Therefore, the correspondence information between the role and the jurisdiction can be used as information in the first role feature vector.

[0068] Exemplarily, the function modules in the risk management system that the logged-in user frequently visits need to be determined after the user logs in to the risk management system. Therefore, the function module data in the risk management system that the logged-in user frequently visits can be used as user data to supplement the feature vector of the logged-in user.

[0069] For each logged-in user, the logged-in user corresponds to one of the multiple roles in the pre-constructed role multi-dimensional feature matrix.

[0070] For example, a unique identifier can be set for each role. A logged-in user enters user login information to log into the risk management system, where the user login information includes the user's unique identifier. When the user's unique identifier matches the role's unique identifier, a user feature vector for the logged-in user can be generated using at least one role feature of the matching role in a pre-built multi-dimensional role matrix and the user data of the logged-in user.

[0071] It should be noted that the historical operation data of the logged-in user includes the operation data of the historical operation risk management system of the logged-in user.

[0072] As an example, a logged-in user database can be established to store the user data of logged-in users. After the logged-in user logs into the risk management system, the user data is obtained by accessing the user database. The user database is updated in real time based on the user's historical operations.

[0073] The generated user feature vector of the logged-in user may be standard data after data processing such as data cleaning, denoising, and normalization. Furthermore, the user feature vector may be updated in real time as the role information and the user data of the logged-in user are updated.

[0074] For example, the user feature vector corresponding to the login user with the work number of 2013 includes the weight coefficient of working years, the professional field matching degree, the historical decision preference value, and the geographical area risk characteristic value. The user feature vector of the login user can be recorded as R_2103=[0.85, 0.92, 0.78, 0.67], where R_2103 is used to uniquely mark the login user, 0.85 corresponds to the weight coefficient of working years after data processing, 0.92 corresponds to the professional field matching degree after data processing, 0.78 corresponds to the historical decision preference value after data processing, and 0.67 corresponds to the geographical area risk characteristic value after data processing.

[0075] S203: Extract at least one risk factor from multi-source power grid risk data.

[0076] It should be noted that the multi-source power grid risk data may be monitoring data for monitoring possible risks in the power grid, including multiple types of monitoring data at multiple moments.

[0077] For example, multi-source grid risk data can include grid SCADA (Supervisory Control and Data Acquisition) monitoring data. Multi-source grid risk data can also include grid power equipment status monitoring data, status monitoring data of power equipment within the grid, geographic grid meteorological warning monitoring data, and maintenance work order data for grid power equipment.

[0078] Semantic analysis may be performed on the multi-source power grid risk data, and at least one risk factor at multiple moments may be determined based on the semantic analysis results.

[0079] S204. Determine the association weights between each risk factor and each user feature vector, and determine a dynamic content weight matrix based on the association weights; the dynamic content weight matrix includes the weights corresponding to each logged-in user and at least one risk warning indicator configured for each logged-in user.

[0080] It should be noted that the association weights of each risk factor and each user feature vector include: the association weights of each risk factor and each user feature in the user feature vector of each logged-in user, wherein the larger the association weight value, the greater the association relationship between the user feature vector of the logged-in user and the risk factor.

[0081] Different risk warning indicators correspond to different risk warning contents. The configured risk warning indicators for logged-in users are associated with the user feature vector of the logged-in user. Risk factors and user feature vectors have associated weights. Therefore, a dynamic content weight matrix can be determined based on the associated weights.

[0082] The configured risk warning indicator for the logged-in user can be determined based on the user data of the logged-in user.

[0083] As an example, risk warning indicators can be configured according to the risk warning tags selected by the logged-in user on the risk management system. The risk warning tags can be pre-configured in the risk management system by dragging and dropping components. The risk warning tags may include "equipment overload hotspots", "network-wide power outage losses", and "anchor risk warnings". The logged-in user can configure their respective risk warning indicators by dragging and dropping.

[0084] As another example, the user's historical operation data in the user data of the logged-in user can be accessed to determine the historical decision information of the logged-in user. The risk warning indicator of the logged-in user can be determined based on the historical decision information.

[0085] S205. Determine the target risk factor corresponding to each logged-in user based on the weight corresponding to the risk warning indicator of each logged-in user, and generate power grid risk warning content sent to the terminal of each logged-in user based on the target risk factor; wherein, the power grid risk warning content for different logged-in users is different.

[0086] It should be noted that different logged-in users correspond to different risk warning indicators, and different risk warning indicators often correspond to different target risk factors. The power grid risk warning content is generated based on the target risk factors. Therefore, the power grid risk warning content for different logged-in users is different.

[0087] The method provided in the embodiment of the present application obtains the login role of the logged-in user, and determines the user feature vector of the logged-in user by using a pre-built role multi-dimensional matrix and the user data of the logged-in user; extracts risk factors from multi-source power grid risk data, and determines the association weights between the risk factors and the user feature vectors; determines a dynamic content weight matrix corresponding to the risk warning indicators of the logged-in user based on the association weights; determines the target risk factors corresponding to the logged-in user according to the weights corresponding to the risk warning indicators of the logged-in user; generates power grid risk warning content sent to the terminals of each logged-in user according to the target risk factors; uses unified multi-source power grid risk data to extract risk factors, and uses the association relationship between the feature vectors of each logged-in user and the risk factors to generate power grid risk warning content for each logged-in user; each logged-in user collaborates on the basis of their respective power grid risk warning content uniformly generated using the same analysis data, thereby reducing the problem of difficulty in collaborative work caused by information asymmetry, thereby improving the effect of power grid risk warning.

[0088] In some specific embodiments, different logged-in users have different grid risk management levels, each corresponding to a different scope of jurisdiction. Furthermore, each level manages risk events of different severity levels. Based on each logged-in user's management level, the risk management system sets different database access permissions for each logged-in user and generates different grid risk warning content.

[0089] Specifically, the power grid risk management level in S201 includes a first management level, a second management level, and a third management level.

[0090] The first management level corresponds to the first jurisdiction and the first grid risk event level; the second management level corresponds to the second jurisdiction and the second grid risk event level; and the third management level corresponds to the third jurisdiction and the third grid risk event level. The first jurisdiction is larger than the second jurisdiction; the second jurisdiction is larger than the third jurisdiction; the first grid risk event level is higher than the second grid risk event level, and the second grid risk event level is higher than the third grid risk event level.

[0091] Exemplarily, the first management level may be the leadership level of the power grid. As the leadership level of the power grid, the first jurisdiction corresponding to the first management level includes the entire power grid.

[0092] For the first management level, the first power grid risk event corresponding to the first management level may include a fault linkage event across power grid geographical areas. Specifically, in S205, power grid risk warning content is generated and sent to the terminals of each logged-in user according to the target risk factor, including:

[0093] For logged-in users belonging to the first management level, an overview of the power grid risk situation and trend forecast are generated based on the target risk factors corresponding to the logged-in users.

[0094] The second management level may be a management level of the power grid. As the management level of the power grid, the second jurisdiction corresponding to the second management level includes a portion of the power grid.

[0095] For the second management level, the second power grid risk event corresponding to the second management level may include risk tasks within the jurisdiction. Specifically, in S205, power grid risk warning content is generated and sent to the terminals of each logged-in user according to the target risk factor, including:

[0096] For login users belonging to the second management level, operational-level risk management and control indicators and grid risk disposal progress are generated based on the target risk factors corresponding to the login users.

[0097] As another example, the second grid risk event may include the distribution of overloaded electrical equipment within the jurisdiction.

[0098] The third management level may be the execution level of the power grid. As the execution level of the power grid, the third jurisdiction corresponding to the third management level is a partial power grid scope of the second jurisdiction.

[0099] For the third management level, the third power grid risk event corresponding to the third management level may include an alarm event of a certain power equipment in the power grid. Specifically, in S205, the power grid risk warning content sent to each logged-in user terminal based on the target risk factor includes:

[0100] For login users belonging to the third management level, equipment-level inspection tasks and power grid risk emergency response information are generated based on the target risk factors corresponding to the login users.

[0101] It should be noted that different management levels correspond to at least one login user. By setting different management levels, login users at different management levels can pay attention to different levels of power grid risk events, so as to deploy risk warning work in a progressive manner, thereby achieving efficient power grid risk management.

[0102] The pre-built role multi-dimensional matrix used to construct the user feature vector of the logged-in user includes multiple role features of multiple roles, and each role has at least one role feature among the multiple role features; for any role, the role has a power grid risk management level, a certain jurisdiction in the power grid, and power grid risk data access rights, wherein the power grid risk data access rights are used to set different database access rights for different roles according to the role's power grid risk management level and jurisdiction, so as to prevent the logged-in user from accessing cross-level power grid risk data.

[0103] Figure 3 A schematic diagram of a process for constructing a multi-dimensional matrix of roles provided in an embodiment of the present application is shown in FIG. Figure 3 As shown, the role multidimensional matrix is ​​obtained based on the following construction steps:

[0104] S301. Obtain job descriptions corresponding to multiple risk management roles in the risk management architecture of the power grid, extract responsibility keywords from each job description, and construct a role-responsibility mapping relationship table based on the responsibility keywords corresponding to each risk management role; wherein the multiple risk management roles belong to different power grid risk management levels.

[0105] For example, for a multi-level management role structure consisting of a first management level, a second management level, and a third management level, a multi-dimensional role matrix can be generated based on the job hierarchy and functional attributes in the power grid organization. Each role is associated with characteristic parameters such as job responsibilities, jurisdiction, and security permissions.

[0106] Specifically, we can extract responsibilities keywords such as "strategic decision-making," "emergency planning," and "equipment inspection" from job descriptions organized into three levels: the first, second, and third management levels. "Strategic decision-making" and "emergency planning" correspond to leadership-level roles, while "equipment inspection" corresponds to executive-level roles. This mapping between roles and responsibilities can then serve as the role characteristics in a multidimensional role matrix.

[0107] S302: Based on the geographic information system, determine the topological relationship of the power equipment within the respective jurisdictions of multiple risk management roles.

[0108] For example, a geographic information system (GIS) can be used to annotate the density of multiple power devices within a grid's geographic area, along with information about the climate characteristics of each region and historical disasters that have occurred within each region (e.g., marking a region as a high-risk volcanic area or a lightning strike zone). Based on this GIS-annotated data, a topological relationship diagram for multiple power devices can be established. This topological relationship diagram includes the geographic region information for the power devices, as well as information about the climate characteristics of the region and historical disasters that have occurred within the region. Based on the jurisdiction of a risk management role, a sub-topological relationship diagram for that jurisdiction within the topological relationship diagram is obtained. This sub-topological relationship diagram can serve as a role feature in the multidimensional role matrix.

[0109] S303: Set dynamic authority adjustment factors for different risk management roles. The dynamic authority adjustment factors are used to adjust the threshold of grid risk data access rights according to the grid risk event level.

[0110] It should be noted that for the three-level management structure of the power grid, the grid risk event levels are high, medium, and low. The roles of the first management level correspond to high-level grid risk events, the roles of the second management level correspond to medium-level grid risk events, and the roles of the third management level correspond to low-level grid risk events. The dynamic permission adjustment factors corresponding to the high, medium, and low levels can be used to control the database access rights that the first management level can access for making strategic decisions, the database access rights that the second management level can access for determining risk command tasks, and the database access rights that the third management level can access for formulating equipment disposal instructions.

[0111] For example, the initial value of the dynamic permission adjustment factor can be set based on the management level and jurisdiction of the management role, and the management role is set to access the power grid risk data within the jurisdiction. After the user logs in to the system, the dynamic permission adjustment factor of the logged-in user can be dynamically adjusted based on the logged-in user's specific database access rights.

[0112] It should be noted that the steps in constructing the multidimensional role matrix described above are not necessarily sequential, and can be performed individually or in combination. The constructed multidimensional role matrix includes information about the role's job responsibilities, the topological relationships of the power equipment within its jurisdiction, and information about access rights to grid risk data. This data provides the foundation for the subsequent generation of grid risk warning content, enabling accurate grid risk warnings to be generated for logged-in users.

[0113] In some embodiments, determining the association weights between each risk factor and each user feature vector in S204, and determining a dynamic content weight matrix based on the association weights, includes:

[0114] The information of each risk factor and each user feature vector at multiple times is input into a pre-built risk association model, the risk association model is used to calculate the association weight of each risk factor and each user feature vector, and a dynamic content weight matrix is ​​generated based on the association weight.

[0115] It should be noted that the pre-built risk association model is a model trained using historical data. The historical data used for model training can be the historical data of the power grid for the previous N years, where N is an integer greater than 1.

[0116] For example, historical data from the power grid over the past seven years is obtained and divided into a training set, a validation set, and a test set in a ratio of 7:2:1 to train the initial model. The historical data includes various risks and corresponding risk warning data.

[0117] For example, it may include fire risk and its warning data, flood risk and its warning data, typhoon risk and its warning data, and anchor risk and its warning data.

[0118] The model training process includes using risk propagation factors to modify the association weights of cascading failure scenarios across power regions. The larger the risk propagation factor, the more the value of the association weight is affected by the data processing results of other regions.

[0119] The model training process also includes using residual connection technology to process data to solve the problem of gradient disappearance during model training.

[0120] In some implementations of these embodiments, the risk association model includes a multi-head attention mechanism and a bidirectional long short-term memory sub-model; the risk association model is used to calculate the association weights of each risk factor and each user feature vector, and a dynamic content weight matrix is ​​generated based on the association weights, including:

[0121] First, the bidirectional long-short-term memory sub-model in the risk association model is used to extract the temporal characteristics of each risk factor from the information of each risk factor at multiple moments.

[0122] Second, the multi-head attention mechanism is used to calculate the association weight between each risk factor and each user feature vector based on the temporal characteristics of each risk factor.

[0123] Third, for each logged-in user's risk warning indicator, according to the associated weight corresponding to the logged-in user, the softmax function in the risk association model is used to determine the dynamic content weight matrix corresponding to the logged-in user.

[0124] Regarding the first step above, it should be noted that the bidirectional long short-term memory sub-model consists of forward and backward long short-term memory neural networks. The forward long short-term memory neural network is used to capture the temporal forward evolution law of risk factors; the backward long short-term memory neural network is used to capture the temporal reverse evolution law of risk factors.

[0125] For example, a bidirectional long short-term memory sub-model can be used to analyze the time series characteristics of various risk factors in the power grid within 72 hours, thereby providing logged-in users with trend data on risk changes over time and then conducting risk trend prediction.

[0126] For the second step above, the multi-head attention mechanism can include the first group of attention heads, the second group of attention heads, the third group of attention heads, and the fourth group of attention heads;

[0127] The first group of attention heads is used to calculate the association weights between job responsibilities and grid risk types in each user's feature vector;

[0128] The second set of attention heads is used to calculate the spatial correlation weights between the jurisdiction and the risk propagation path in each user’s feature vector;

[0129] The third group of attention heads is used to calculate the compliance of the grid risk data access rights and grid risk data in each user's feature vector;

[0130] The fourth group of attention heads is used to calculate the correlation between the historical decision-making pattern and risk factor information in each user's feature vector.

[0131] For example, the first group of attention heads includes the 1st and 2nd attention heads, the second group of attention heads includes the 3rd and 4th attention heads, the third group of attention heads includes the 5th and 6th attention heads, and the fourth group of attention heads can be the 7th and 8th attention heads. The eight attention heads run in parallel, focusing on different dimensions of user features.

[0132] The multi-head attention mechanism in the model can focus on key time steps, thereby analyzing the risk characteristics of key time steps.

[0133] For the third step above, the softmax function is used to normalize the association weights of the target risk factors corresponding to the risk warning indicators of the logged-in users and generate probability outputs.

[0134] For example, regarding the risk warning indicators of meteorological risk levels, the softmax function outputs the dynamic content weight of the first-level warning as 0.7, the dynamic content weight of the second-level warning as 0.2, and the dynamic content weight of the third-level warning as 0.1. Therefore, the power grid risk warning content sent to the logged-in user includes the dynamic content of the first-level warning.

[0135] The following is an example of a power grid facing the risk of a typhoon passing through to further illustrate this technical solution.

[0136] Step 1: Obtain the corresponding login roles of leadership-level users, management-level users, and execution-level users, and extract the responsibilities keywords of each role according to the job description of the three-level management structure; use the annotation of the geographic information system to determine the topological relationship diagram of the power equipment corresponding to each role according to the jurisdiction of each role; set the leadership-level role to have access to the database permissions of the entire power grid, and the management-level and execution-level roles to have access to the database permissions corresponding to their respective jurisdictions.

[0137] Step 2. Leadership-level users log in to the risk management system through facial recognition, management-level users log in to the risk management system using their work ID and password, and executive-level users log in to the risk management system through mobile terminals. Based on the role characteristics and user data of the leadership-level users, a user feature vector for the leadership-level users is generated. Based on the role characteristics and user data of the management-level users, a user feature vector for the management-level users is generated. Based on the role characteristics and user data of the executive-level users, a user feature vector for the executive-level users is generated. Based on the user data of the leadership-level users, the database access rights of the leadership-level users are adjusted, and the "network-wide economic loss prediction module" of the risk management system is loaded. Based on the user data of the management-level users, the database access rights of the management-level users are adjusted, and the "geographical area equipment overload heat map module" of the risk management system is loaded. Based on the user data of the executive-level users, the database access rights of the executive-level users are adjusted, and the "power equipment real-time alarm module" of the risk management system is loaded.

[0138] Step 3: Obtain meteorological data for the past 72 hours through the data interface provided by the Meteorological Bureau. Use the data acquisition and monitoring system to obtain the load factor of the five 220 kV power lines within the power grid, monitoring data for each power device, and maintenance work order data for each power device. Extract multiple risk factors from this acquired data.

[0139] Step 4: Input the role feature vectors and risk factor information of the leadership-level login users, management-level login users, and executive-level login users into the pre-trained risk association model.

[0140] Step 5. For users logged in at the leadership level, the risk management system displays the following information at the top of the "Network-wide Economic Loss Forecast Module": "The typhoon may cause a 23% loss in network load. It is recommended to immediately initiate a backup line call agreement with a certain geographic region." It also provides decision-making assistance information: "Three cross-regional support plans are available. Each plan is expected to reduce economic losses by 180 million yuan, 150 million yuan, and 120 million yuan, respectively." For users logged in at the management level, the risk management system displays a heat map of overloaded power equipment within its jurisdiction through the "Geographical Area Equipment Overload Heat Map Module" and generates a "Pre-Typhoon Special Patrol Work Order." The "Pre-Typhoon Special Patrol Work Order" requires the inspection of 76 towers within 12 hours. The work order includes the following information: "Currently, 12 emergency repair teams are available. It is recommended that geographic regions 1 and 2 be prioritized for deployment." For users logged in at the executive level, the risk management system issues a strong reminder through the "Power Equipment Real-time Alarm Module": "There is an unreinforced color room 200 meters around the 45th tower of the 220 kV Binhai Line. Cleaning is required within 6 hours (with geolocation coordinates)."

[0141] After the typhoon passes, the data from this risk warning can also be used to optimize and adjust the parameters of the risk association model.

[0142] Figure 4 A schematic diagram of the structure of the power grid risk data processing device provided in an embodiment of the present application is shown in FIG. Figure 4 As shown, the power grid risk data processing 40 provided in this embodiment includes:

[0143] A first acquisition module 401 is configured to acquire a login role of at least one logged-in user and determine a first role characteristic of each logged-in user's login role from a pre-constructed multi-dimensional role matrix; wherein the first role characteristic of each logged-in user includes the grid risk management level, jurisdiction, and grid risk data access rights of the logged-in user;

[0144] A first determining module 402 is configured to determine, for each logged-in user, a user feature vector of the logged-in user based on the user data of the logged-in user and a first role feature, wherein the user data includes historical operation data, risk management tasks, jurisdiction, and historical power grid risk information of the jurisdiction;

[0145] A second acquisition module 403 is configured to extract at least one risk factor from the multi-source power grid risk data;

[0146] The second determination module 404 is configured to determine the association weights between each risk factor and each user feature vector, and determine a dynamic content weight matrix based on the association weights; the dynamic content weight matrix includes the weights corresponding to each logged-in user and at least one risk warning indicator configured for each logged-in user;

[0147] Generation module 405 is used to determine the target risk factor corresponding to each logged-in user based on the weight corresponding to the risk warning indicator of each logged-in user, and generate the power grid risk warning content sent to the terminal of each logged-in user based on the target risk factor; wherein, the power grid risk warning content of different logged-in users is different.

[0148] In one possible implementation, the first acquisition module 401 is further configured to obtain job descriptions corresponding to multiple risk management roles in the risk management architecture of the power grid, extract responsibility keywords from each job description, and construct a role-responsibility mapping relationship table based on the responsibility keywords corresponding to each risk management role; wherein the multiple risk management roles belong to different power grid risk management levels;

[0149] Based on the geographic information system, determine the topological relationship of power equipment within the respective jurisdictions of multiple risk management roles;

[0150] Set dynamic permission adjustment factors for different risk management roles. The dynamic permission adjustment factors are used to adjust the grid risk data access permission threshold according to the grid risk event level.

[0151] In a possible implementation, the second acquisition module 403 is further configured to perform semantic analysis on the multi-source power grid risk data, and determine at least one risk factor at multiple moments according to the semantic analysis results.

[0152] In one possible implementation, the second determination module 404 is further used to input the information of each risk factor and each user feature vector at multiple moments into a pre-built risk association model, use the risk association model to calculate the association weights of each risk factor and each user feature vector, and generate a dynamic content weight matrix based on the association weights.

[0153] In one possible implementation, the risk association model includes a multi-head attention mechanism and a bidirectional long short-term memory sub-model. The second determination module 404 is further configured to utilize the bidirectional long short-term memory sub-model in the risk association model to extract temporal features of each risk factor from information on each risk factor at multiple moments.

[0154] The multi-head attention mechanism is used to calculate the association weight between each risk factor and each user feature vector based on the temporal characteristics of each risk factor;

[0155] For each logged-in user's risk warning indicator, according to the associated weight corresponding to the logged-in user, the softmax function in the risk association model is used to determine the dynamic content weight matrix corresponding to the logged-in user.

[0156] In one possible implementation, the multi-head attention mechanism includes a first group of attention heads, a second group of attention heads, a third group of attention heads, and a fourth group of attention heads;

[0157] The first group of attention heads is used to calculate the association weights between job responsibilities and grid risk types in each user's feature vector;

[0158] The second set of attention heads is used to calculate the spatial correlation weights between the jurisdiction and the risk propagation path in each user’s feature vector;

[0159] The third group of attention heads is used to calculate the compliance of the grid risk data access rights and grid risk data in each user's feature vector;

[0160] The fourth group of attention heads is used to calculate the correlation between the historical decision-making pattern and risk factor information in each user's feature vector.

[0161] In a possible implementation, the grid risk management level includes a first management level, a second management level, and a third management level; wherein,

[0162] The first management level corresponds to the first jurisdiction and the first grid risk event level; the second management level corresponds to the second jurisdiction and the second grid risk event level; the third management level corresponds to the third jurisdiction and the third grid risk event level; among them,

[0163] The first jurisdiction is larger than the second jurisdiction; the second jurisdiction is larger than the third jurisdiction;

[0164] The first power grid risk event level is higher than the second power grid risk event level, and the second power grid risk event level is higher than the third power grid risk event level.

[0165] In a possible implementation, the generating module 405 is further configured to generate a power grid risk situation overview and trend forecast for a logged-in user belonging to the first management level according to the target risk factor corresponding to the logged-in user;

[0166] For login users belonging to the second management level, operational-level risk control indicators and grid risk disposal progress are generated based on the target risk factors corresponding to the login users;

[0167] For login users belonging to the third management level, equipment-level inspection tasks and power grid risk emergency response information are generated based on the target risk factors corresponding to the login users.

[0168] The power grid risk data processing device provided in this embodiment can execute the method provided in the above method embodiment. Its implementation principle and technical effects are similar, and are not described in detail in this embodiment.

[0169] Figure 5 This is a schematic diagram of the structure of the power grid risk data processing equipment provided in this application. Figure 5As shown, the power grid risk data processing device 50 provided in this embodiment includes: at least one processor 501 and a memory 502. Optionally, the device 50 also includes a communication component 503. The processor 501, the memory 502, and the communication component 503 are connected via a bus.

[0170] In a specific implementation process, at least one processor 501 executes the computer-executable instructions stored in the memory 502, so that the at least one processor 501 performs the above method.

[0171] The specific implementation process of the processor 501 can be found in the above method embodiment. Its implementation principle and technical effects are similar and will not be repeated here in this embodiment.

[0172] In the above embodiments, it should be understood that the processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASICs), etc. A general-purpose processor may be a microprocessor or any conventional processor. The steps of the method disclosed in the present invention may be directly executed by a hardware processor or by a combination of hardware and software modules within the processor.

[0173] The memory may include random access memory (RAM) and may also include non-volatile memory (NVM), such as at least one disk storage.

[0174] A bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus. Buses can be categorized as address buses, data buses, and control buses. For ease of illustration, the buses in the drawings of this application are not limited to just one bus or just one type of bus.

[0175] The present application also provides a computer program product, including a computer program, which implements the above method when executed by a processor.

[0176] The present application also provides a computer-readable storage medium, in which computer-executable instructions are stored. When a processor executes the computer-executable instructions, the above method is implemented.

[0177] The readable storage medium may be implemented by any type of volatile or non-volatile memory device, or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The readable storage medium may be any available medium that can be accessed by a general-purpose or special-purpose computer.

[0178] An exemplary readable storage medium is coupled to a processor so that the processor can read information from the readable storage medium and write information to the readable storage medium. Of course, the readable storage medium can also be an integral part of the processor. The processor and the readable storage medium can be located in an application specific integrated circuit (ASIC). Of course, the processor and the readable storage medium can also exist in the device as discrete components.

[0179] The division of units is merely a logical functional division; actual implementations may employ alternative divisions, such as combining or integrating multiple units or components into another system, or omitting or disabling certain features. Furthermore, any direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection between devices or units, either through an interface, electrical, mechanical, or other means.

[0180] Units described as separate components may or may not be physically separate, and components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0181] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0182] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the various embodiments of the method of the present invention. The aforementioned storage medium includes various media that can store program code, such as USB flash drives, mobile hard drives, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical disks.

[0183] Those skilled in the art will appreciate that all or part of the steps in the above-described method embodiments can be implemented using hardware associated with program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0184] Finally, it should be noted that those skilled in the art will readily identify other embodiments of the present invention after considering the specification and practicing the invention disclosed herein. The present invention is intended to cover any variations, uses, or adaptations of the present invention that follow the general principles of the present invention and include common knowledge or customary techniques in the art not disclosed herein. The present invention is not limited to the precise structure described above and illustrated in the accompanying drawings, and various modifications and variations may be made without departing from the scope thereof. The scope of the present invention is limited solely by the appended claims.

Claims

1. A method for processing power grid risk data, characterized in that: The method comprises: Obtaining a login role of at least one login user, and determining a first role characteristic of each login user's login role from a pre-constructed multi-dimensional role matrix; wherein the first role characteristic of each login user includes the grid risk management level, jurisdiction, and grid risk data access rights to which the login user belongs; For each logged-in user, determining a user feature vector of the logged-in user based on the user data of the logged-in user and the first role feature, wherein the user data includes historical operation data, risk management tasks, jurisdiction, and historical power grid risk information of the jurisdiction; extracting at least one risk factor from multi-source power grid risk data; Determining the association weights of each risk factor and each of the user feature vectors, and determining a dynamic content weight matrix based on the association weights; the dynamic content weight matrix includes weights corresponding to each logged-in user and at least one risk warning indicator configured for each logged-in user; According to the weight corresponding to the risk warning indicator of each logged-in user, the target risk factor corresponding to each logged-in user is determined, and the power grid risk warning content sent to the terminal of each logged-in user is generated according to the target risk factor; wherein, the power grid risk warning content for different logged-in users is different.

2. The method according to claim 1, characterized in that The role multidimensional matrix is ​​obtained based on the following construction steps: Obtaining job descriptions corresponding to a plurality of risk management roles in the risk management architecture of the power grid, extracting responsibility keywords from each job description, and constructing a role-responsibility mapping relationship table based on the responsibility keywords corresponding to each risk management role; wherein the plurality of risk management roles belong to different power grid risk management levels; Based on the geographic information system, determine the topological relationship of power equipment within the respective jurisdictions of multiple risk management roles; A dynamic authority adjustment factor is set for each risk management role. The dynamic authority adjustment factor is used to adjust the power grid risk data access permission threshold according to the power grid risk event level.

3. The method according to claim 1, characterized in that The step of extracting at least one risk factor from the multi-source power grid risk data includes: Perform semantic analysis on the multi-source power grid risk data, and determine the at least one risk factor at multiple moments according to the semantic analysis results.

4. The method according to any one of claims 1 to 3, characterized in that Determining the association weights between each risk factor and each user feature vector, and determining a dynamic content weight matrix based on the association weights, includes: The information of each risk factor and each user feature vector at multiple moments is input into a pre-built risk association model, the risk association model is used to calculate the association weight of each risk factor and each user feature vector, and the dynamic content weight matrix is ​​generated according to the association weight.

5. The method according to claim 4, characterized in that The risk association model includes a multi-head attention mechanism and a bidirectional long short-term memory sub-model; the risk association model is used to calculate the association weights of each risk factor and each user feature vector, and the dynamic content weight matrix is ​​generated based on the association weights, including: Utilizing the bidirectional long-short-term memory sub-model in the risk association model to extract the temporal characteristics of each risk factor from the information of each risk factor at multiple moments; Utilize the multi-head attention mechanism to calculate the association weight between each risk factor and each user feature vector based on the temporal characteristics of each risk factor; For each logged-in user's risk warning indicator, according to the associated weight corresponding to the logged-in user, the softmax function in the risk association model is used to determine the dynamic content weight matrix corresponding to the logged-in user.

6. The method according to claim 5, characterized in that The multi-head attention mechanism includes a first group of attention heads, a second group of attention heads, a third group of attention heads, and a fourth group of attention heads; The first group of attention heads is used to calculate the association weight between the job responsibilities and the grid risk type in each user feature vector; The second set of attention heads is used to calculate the spatial correlation weights between the jurisdiction and the risk propagation path in each user feature vector; The third group of attention heads is used to calculate the compliance of the power grid risk data access rights and the power grid risk data in each user feature vector; The fourth group of attention heads is used to calculate the correlation between the historical decision-making pattern and the risk factor information in each user feature vector.

7. The method according to claim 1, characterized in that The grid risk management level includes the first management level, the second management level and the third management level; wherein, The first management level corresponds to the first jurisdiction and the first power grid risk event level; the second management level corresponds to the second jurisdiction and the second power grid risk event level; the third management level corresponds to the third jurisdiction and the third power grid risk event level; wherein, The first jurisdiction is larger than the second jurisdiction; the second jurisdiction is larger than the third jurisdiction; The first power grid risk event level is higher than the second power grid risk event level, and the second power grid risk event level is higher than the third power grid risk event level.

8. The method according to claim 7, characterized in that The generating of the power grid risk warning content sent to the terminal of each logged-in user according to the target risk factor includes: For logged-in users belonging to the first management level, an overview of the power grid risk situation and trend forecast are generated based on the target risk factors corresponding to the logged-in users; For login users belonging to the second management level, operational-level risk control indicators and grid risk disposal progress are generated based on the target risk factors corresponding to the login users; For login users belonging to the third management level, equipment-level inspection tasks and power grid risk emergency response information are generated based on the target risk factors corresponding to the login users.

9. A power grid risk data processing device, characterized in that: include: A first acquisition module is configured to acquire a login role of at least one logged-in user and determine a first role characteristic of each logged-in user's login role from a pre-constructed multi-dimensional role matrix; wherein the first role characteristic of each logged-in user includes the grid risk management level, jurisdiction, and grid risk data access rights of the logged-in user; a first determining module configured to determine, for each logged-in user, a user feature vector of the logged-in user based on user data of the logged-in user and the first role feature, wherein the user data includes historical operation data, risk management tasks, jurisdiction, and historical power grid risk information of the jurisdiction; A second acquisition module is used to extract at least one risk factor from multi-source power grid risk data; a second determination module, configured to determine an association weight between each risk factor and each user feature vector, and determine a dynamic content weight matrix based on the association weight; the dynamic content weight matrix includes a weight corresponding to each logged-in user and at least one risk warning indicator configured for each logged-in user; A generation module is used to determine the target risk factor corresponding to each logged-in user based on the weight corresponding to the risk warning indicator of each logged-in user, and generate the power grid risk warning content sent to the terminal of each logged-in user based on the target risk factor; wherein the power grid risk warning content of different logged-in users is different.

10. A power grid risk data processing device, characterized in that: include: Memory, processor; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory, so that the processor performs the method according to any one of claims 1 to 8.

11. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer-executable instructions, which are used to implement the method according to any one of claims 1 to 8 when executed by a processor.

12. A computer program product, characterized in that The invention comprises a computer program, which implements the method according to any one of claims 1 to 8 when executed by a processor.

Images