Power grid operation risk prediction and early warning method and system

Abstract

The invention relates to the technical field of power grid operation safety, in particular to a power grid operation risk prediction and early warning method and system. The method comprises the following steps: acquiring multi-source data for power grid operation risk prediction and early warning; performing data normalization operation on the multi-source data; when normalization preprocessing is carried out on the multi-source data, normalization preprocessing is carried out on the main network model and the distribution network model separately, and the normalization preprocessing is further separated from operation data, so that the accuracy of data normalization is improved; particularly, when the distribution network model is normalized, distance differences in different scenes can be better adapted by dynamically adjusting a threshold value, and the accuracy of comparison and checking is improved; the improved Euclidean distance calculation method can more accurately reflect the difference between the nodes and improve the normalization accuracy by introducing a weighting factor and considering the weight of the equipment type and the connection relationship.

Description

Technical Field

[0001] The present invention relates to the technical field of power grid operation safety, and in particular to a power grid operation risk prediction and early warning method and system. Background Art

[0002] As a crucial infrastructure for energy transmission in modern society, the safe and stable operation of power grids is crucial. However, power grid operations are affected by a variety of factors. For example, long-term operation can cause equipment in the grid to malfunction due to natural aging and wear. For example, the aging of equipment like transformers and circuit breakers can lead to degraded insulation and poor contact, potentially causing short circuits and power outages. Severe weather conditions can significantly impact power grid operations. For example, lightning strikes can damage transmission line insulators, causing short circuits. Strong winds can cause conductors to vibrate and collide, resulting in line breakage. Heavy rains and floods can destroy towers and substations, disrupting the grid's normal operation. With socioeconomic development and improved living standards, power grid loads are constantly changing. During peak hours, grid loads increase dramatically, potentially leading to equipment overloads and voltage drops. During off-peak hours, sudden drops in load can cause grid frequency fluctuations, impacting grid stability.

[0003] Traditional power grid risk assessment methods rely heavily on empirical judgment and post-analysis, making it difficult to accurately predict and warn of risks in real time. This makes it difficult to effectively implement risk mitigation measures and ensure safe and stable grid operation. Therefore, a system and method that can accurately predict power grid operational risks in real time and issue timely warnings is urgently needed. Furthermore, existing preprocessing methods for multi-source data lack sufficient accuracy. Summary of the Invention

[0004] In order to solve the above technical problems, the present invention provides a power grid operation risk prediction and early warning method and system for solving the problems existing in the prior art.

[0005] The present invention provides a method for predicting and warning of power grid operation risks, comprising the following steps: S1: Acquire multi-source data for power grid operation risk prediction and early warning; S2: performing data normalization on the multi-source data; The data normalization operation includes model data normalization and operation data normalization; The model data standardization operation is specifically as follows: for the main network model, meta-model conversion is achieved by building a semantic mapping rule library; for the distribution network model, the connection relationship between the nodes of the reference relationship group is obtained through the single line diagram, and the connection relationship is compared and verified with the topological relationship in the distribution network model database, and then normalized and archived; S3: Build a risk prediction and early warning model; S4: Inputting the normalized multi-source data into the risk prediction and early warning model to obtain risk prediction and early warning results.

[0006] Preferably, the multi-source data includes a main network model, a distribution network model, main network operation data, distribution network operation data, and user-related data.

[0007] Preferably, the main network model includes the main network's topology, equipment parameters, and line connection relationship information; the distribution network model includes the distribution network's network structure, equipment configuration, and user access information; the distribution network operation data includes the distribution network's voltage, current, load, and fault information; the user-related data includes the user's electricity usage behavior, electricity-using equipment, and geographic location information.

[0008] Preferably, the metamodel conversion is achieved for the main network model by building a semantic mapping rule base, specifically: Performing ontology analysis on the main network model to extract core classes and their attribute relationships; Establish a mapping rule table between the target system metamodel and CIM elements; Convert the attribute relationships in the main network model so that they conform to the data format and semantic requirements of the target system; Reconstructing the association relationship in the main network model to adapt to the topology requirements of the target system; Build the target-library table mapping to search the belonging library table and determine the normalized storage location.

[0009] Preferably, for the distribution network model, the connection relationship between the nodes of the control relationship group is obtained through the single line diagram, and the connection relationship is compared and checked with the topological relationship in the distribution network model database, and then normalized and archived; Specifically: Obtaining single-line diagram data of the distribution network model; Identifying nodes of a distribution network model in the one-line diagram data; Extracting connection relationships between nodes in the single-line graph data to form a comparison relationship group; The control relationship group is compared and checked with the topological relationship in the distribution network model database.

[0010] Preferably, the process of calculating the inconsistency between the topological relationship between the control relationship group and the distribution network model database by improving the Euclidean distance is specifically as follows: Calculate the Euclidean distance between the coordinates of the inconsistent nodes in the topological relationship diagram in the distribution network model database and in the single-line diagram according to the device type and the node connection relationship; Among them, the improved Euclidean distance calculation formula is used to calculate the Euclidean distance. In the calculation process, not only the coordinate difference of the node in the single-line diagram is considered, but also the influence of the equipment type and connection relationship on the distance; Specifically, the calculation formula of the improved Euclidean distance d is: ; Wherein, (x1, y1) and (x2, y2) are the coordinates of the inconsistent nodes in the topological relationship diagram in the distribution network model database and the coordinates in the single-line diagram respectively; w x and w y are weight factors in the x-coordinate and y-coordinate directions, respectively; t1 and t2 are the device type codes of the inconsistent nodes in the topological relationship in the distribution network model database and the device type codes in the reference relationship group, respectively; c1 and c2 are the connection relationship codes of the inconsistent nodes in the topological relationship in the distribution network model database and the connection relationship codes in the reference relationship group, respectively; w t and w c are the weight factors of device type and connection relationship respectively; Dynamically adjust the threshold θ according to the calculation result of the Euclidean distance; The threshold value is dynamically adjusted according to the calculation result of the Euclidean distance as follows: Initial threshold setting: set an initial threshold θ0; Dynamically adjust the threshold θ according to the distribution of the Euclidean distance d; The calculation formula of the threshold is: ; Where σ is the standard deviation of the Euclidean distance of all inconsistent nodes, and μ is the mean of the Euclidean distance of all inconsistent nodes; The Euclidean distance between the topological relationship diagram of the inconsistent node in the distribution network model database and the coordinates in the single-line diagram is compared with the threshold θ, thereby achieving normalized archiving.

[0011] Preferably, an LSTM model is selected as a power grid operation risk prediction model; the LSTM model is trained using historical power grid operation data and external environment data; the data is divided into a training set and a validation set, and the model is trained and evaluated using a cross-validation method; during the training process, the learning rate is set to 0.01 and the number of training rounds is 100.

[0012] Preferably, the multi-source data is normalized and then input into a trained LSTM risk prediction model; The model outputs the risk level prediction results of the power grid in the next 24 hours, and divides the risk level into three levels: low risk, medium risk and high risk according to the set threshold; Generate detailed early warning information based on the predicted risk level; Early warning information is promptly released to power grid operation managers, relevant departments and users through multiple channels such as power grid monitoring system, SMS platform and email; medium-risk early warning information is pushed via SMS and email.

[0013] According to another aspect of the present invention, a power grid operation risk prediction and early warning system is provided, which adopts the above-mentioned power grid operation risk prediction and early warning method, and the system includes: Data acquisition module, which acquires multi-source data for power grid operation risk prediction and early warning; A preprocessing module, configured to perform a data normalization operation on the multi-source data; Model building module, used to build risk prediction and early warning models; The prediction and warning module is used to input the normalized multi-source data into the risk prediction and warning model to obtain risk prediction and warning results.

[0014] The embodiments of the present invention have the following technical effects: When preprocessing multi-source data, the present invention performs normalization preprocessing on the main network model and the distribution network model separately, and further separates them from the operating data, thereby improving the accuracy of data normalization; in particular, when normalizing the distribution network model, by dynamically adjusting the threshold, it can better adapt to the distance differences in different scenarios and improve the accuracy of comparison and verification; by introducing weighting factors and considering the weights of device types and connection relationships, the improved Euclidean distance calculation method can more accurately reflect the differences between nodes and improve the accuracy of normalization. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the specific embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0016] Figure 1 This is a flow chart of a method for predicting and warning power grid operation risks provided by an embodiment of the present invention; Figure 2 This is a flowchart of the inconsistency between the topological relationship in the comparison relationship group and the distribution network model database by improving the Euclidean distance calculation provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0017] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention are described clearly and completely below. Obviously, the embodiments described are only some of the embodiments of the present invention, not all of them. All other embodiments derived by persons of ordinary skill in the art based on the embodiments of the present invention without inventive effort are also within the scope of protection of the present invention.

[0018] Example 1, Figure 1 The flowchart of the method for predicting and warning the risk of power grid operation is shown, which includes the following steps: S1: Acquire multi-source data for power grid operation risk prediction and early warning; The multi-source data includes the main network model, the distribution network model, the main network operation data, the distribution network operation data, and user-related data.

[0019] Among them, the main network model includes information such as the main network's topological structure, equipment parameters, line connection relationships, etc.; the main network model can be obtained from the control cloud, which is a cloud computing platform that integrates power grid dispatching, control and management functions. It stores detailed model data of the power grid and can provide accurate main network structure information for risk prediction.

[0020] Among them, the main network operation data refers to various parameters and states of the main network during actual operation, including voltage, current, power, frequency, etc.; these data can reflect the operation status of the main network in real time and are an important basis for risk prediction; the main network operation data can also be obtained from the control cloud. The control cloud collects the main network operation data through a real-time monitoring system, and stores and manages it, providing rich real-time data resources for risk prediction.

[0021] Among them, the distribution network model includes information such as the network structure, equipment configuration, and user access of the distribution network; as the terminal part of the power grid, the operating status of the distribution network directly affects the power supply quality and reliability of the user; the distribution network model can be obtained from the distribution automation system, which realizes the automatic monitoring and control of the distribution network, and can provide detailed distribution network model data, providing a structural basis for risk prediction of the distribution network.

[0022] Among them, the distribution network operation data includes the voltage, current, load, fault information, etc. of the distribution network. These data can reflect the actual operation status of the distribution network and are of great significance for assessing the risk of the distribution network; the distribution network operation data can also be obtained from the distribution automation system. The distribution automation system collects the operation data of the distribution network in real time through various sensors and monitoring equipment installed in the distribution network, and transmits it to the system, providing real-time distribution network operation information for risk prediction.

[0023] Among them, the user-related data includes information such as the user's electricity usage behavior, electricity-using equipment, and geographic location. These data can help analyze the impact of users on power grid operation, as well as the impact of power grid operation on users, thereby more comprehensively assessing power grid operation risks; the user-related data can be obtained through the REST interface, which is a network application program interface based on the HTTP protocol that can easily interact with other systems.

[0024] S2: performing data normalization on the multi-source data; The data normalization operation includes model data normalization and operation data normalization; Among them, the model data standardization operation is specifically as follows: for the main network model, meta-model conversion is realized by constructing a semantic mapping rule library; for the distribution network model, the connection relationship between the nodes of the control relationship group is obtained through the single-line diagram, and compared and verified with the topological relationship in the distribution network model database, and then normalized and archived.

[0025] Among them, for the main network model, the metamodel conversion is realized by building a semantic mapping rule base as follows: Performing ontology analysis on the main network model to extract core classes and their attribute relationships; The core classes include VoltageLevel, Breaker, Feeder, etc.

[0026] Establish a mapping rule table between the target system metamodel and CIM elements; The mapping rule table defines how to convert the elements in the main network model into a meta-model that the target system can understand and process. Specifically, the mapping is a class name mapping, including: Substation→Substation; Breaker → switch; Feeder→feeder; Through this mapping, the names of the core classes in the main network model are converted into corresponding terms in the target system, so that data between different systems can be effectively mapped and interacted.

[0027] Convert the attribute relationships in the main network model so that they conform to the data format and semantic requirements of the target system; For example, the cim: Identified Object.name attribute is converted into the device name. In the main network model, Identified Object is a general class used to identify various objects in the power grid, and the name attribute is used to name these objects. During the data standardization process, the value of this name attribute is extracted as the name of the device in the target system, thereby realizing the conversion and unification of attributes.

[0028] Reconstructing the association relationship in the main network model to adapt to the topology requirements of the target system; For example, the cim: Connectivity Node is converted into a topological connection point. In the main network model, Connectivity Node is used to represent the connection relationship between devices in the power grid. It defines how the devices are connected to form the topology of the power grid. During the data standardization process, these Connectivity Nodes are reconstructed into topological connection points in the target system, thereby accurately representing the connection relationship and topology of the power grid.

[0029] Build a target-library table mapping to search the belonging library table and determine the normalized storage location; In this step, you can determine in which database table of the target system the standardized data should be stored for subsequent data query and processing; by searching the database structure of the target system, find the library table corresponding to the current data, and store the data in the appropriate location to achieve normalized data storage.

[0030] Among them, for the distribution network model, the connection relationship between the nodes of the reference relationship group is obtained through the single line diagram, and compared and checked with the topological relationship in the distribution network model database, and then normalized and archived. Specifically: Obtaining single-line diagram data of the distribution network model; In this step, the single-line diagram data is exported from the distribution network automation system. The single-line diagram data graphically presents the topology of the distribution network, including substations, lines, switches, transformers and other equipment and the connection relationships between them.

[0031] Identifying nodes of a distribution network model in the one-line diagram data; In this step, graph parsing techniques are used to identify nodes in the single-line diagram data. Each node represents a device, such as a switch, transformer, or segment point. The node's identification information (such as device name and number) and location coordinates are extracted.

[0032] Extracting connection relationships between nodes in the single-line graph data to form a comparison relationship group; Based on the connections between the nodes, the connection relationship between the nodes is determined; specifically, for each pair of connected nodes, their connection order and connection type (such as cable connection, overhead line connection, etc.) are recorded to form a comparison relationship group; for example, node A is connected to node B, and node B is connected to node C, then a comparison relationship group (AB, BC) is formed.

[0033] Comparing and verifying the control relationship group with the topological relationship in the distribution network model database; The topological relationship of the distribution network model is read from the database of the distribution network model. The model topological relationship is stored in the form of a data table, including the connection information of the devices, such as device 1 connected to device 2, device 2 connected to device 3, etc. The comparison relationship group extracted from the single-line diagram is compared one by one with the topological relationship in the database of the distribution network model. For each pair of node connection relationships in the comparison relationship group, it is checked whether the topological relationship in the database of the distribution network model has the same connection relationship.

[0034] During the comparison and verification process, there will generally be inconsistencies in the comparison and verification. In this case, this embodiment provides an improved solution for calculating the inconsistencies in the comparison and verification by using Euclidean distance, so as to improve the accuracy of the verification result.

[0035] Specifically, if Figure 2 As shown, the process of calculating the inconsistency between the topological relationship between the control relationship group and the distribution network model database by improving the Euclidean distance is specifically as follows: Calculate the Euclidean distance between the coordinates of the inconsistent nodes in the topological relationship diagram in the distribution network model database and in the single-line diagram according to the device type and the node connection relationship; Among them, the improved Euclidean distance calculation formula is used to calculate the Euclidean distance. In the calculation process, not only the coordinate difference of the node in the single-line diagram is considered, but also the influence of the equipment type and connection relationship on the distance; Specifically, the calculation formula of the improved Euclidean distance d is: ; Wherein, (x1, y1) and (x2, y2) are the coordinates of the inconsistent nodes in the topological relationship diagram in the distribution network model database and the coordinates in the single-line diagram respectively; w x and w y The weight factors in the x-coordinate and y-coordinate directions are used to adjust the influence of coordinate differences in different directions on distance calculation. The weight factors can be adjusted according to the actual scenario. For example, if the horizontal layout change has a greater impact on the topological relationship than the vertical layout change in the distribution network, you can set w x >w y ; t 1 and t2 are the device type codes of the inconsistent nodes in the topological relationship in the distribution network model database and the device type codes in the comparison relationship group, respectively. c 1 and c 2 are the connection relationship codes of the inconsistent nodes in the topological relationship in the distribution network model database and the connection relationship codes in the comparison relationship group, respectively. w t and w c They are the weight factors of device type and connection relationship, respectively, used to adjust the impact of differences in device type and connection relationship on distance calculation.

[0036] For example, the equipment type code can be classified according to the type of equipment, for example, the switch equipment code is 1, the transformer equipment code is 2, the segmentation point equipment code is 3, etc. The connection relationship code can be classified according to the connection type, for example, the cable connection code is 1, the overhead line connection code is 2, etc.

[0037] Dynamically adjust the threshold θ according to the calculation result of the Euclidean distance; The threshold value is dynamically adjusted according to the calculation result of the Euclidean distance as follows: Initial threshold setting: set an initial threshold θ0; Exemplarily, the initial threshold is 0.1.

[0038] Dynamically adjust the threshold θ according to the distribution of the Euclidean distance d; The calculation formula of the threshold is: ; Among them, σ is the standard deviation of the Euclidean distance, and μ is the mean of the Euclidean distance; Comparing the Euclidean distance between the topological relationship diagram of the inconsistent node in the distribution network model database and the coordinates in the single-line diagram with the threshold θ, thereby achieving normalized archiving; Specifically, if the calculated Euclidean distance is less than the threshold θ, it means that although the topological relationship in the distribution network model database and the connection relationship of the single-line diagram are not completely consistent, the difference is small and can be considered to be within a reasonable error range; in this case, normalization and archiving are performed according to the comparison relationship group; If the calculated Euclidean distance is greater than the threshold θ, it means that there is a large difference between the topological relationship in the distribution network model database and the connection relationship of the single-line diagram. It may be that the distribution network model is wrong or the single-line diagram is not synchronized to the distribution network model after being updated; at this time, the relevant verification task is automatically generated by the topology management tool. The verification task includes a detailed task description, such as the device name, location coordinates, connection relationship in the comparison relationship group and other information of the inconsistent node. The verification task is assigned to manual operation and maintenance personnel. The operation and maintenance personnel go to the site to verify the problem according to the task information, check the actual connection status of the equipment, determine whether it is a model error or a single-line diagram error, and correct the model or single-line diagram. Finally, the corrected results are normalized and archived.

[0039] In this step, by dynamically adjusting the threshold, we can better adapt to the distance differences in different scenarios and improve the accuracy of comparison and verification; by introducing weighting factors and considering the weights of device types and connection relationships, the improved Euclidean distance calculation method can more accurately reflect the differences between nodes and improve the accuracy of comparison and verification.

[0040] The normalization of the operating data is specifically performed by using a minimum-maximum normalization method to map the operating data to the interval [0, 1]. For example, for voltage data, the maximum and minimum values are calculated and then normalized according to the formula.

[0041] S3: Build a risk prediction and early warning model; In this step, we select the LSTM model as the grid operation risk prediction model and determine the model's structural parameters, such as the number of hidden layer units and time step. Based on the characteristics of grid operation data, we set the number of hidden layer units to 50 and the time step to 24 hours.

[0042] The LSTM model was trained using historical power grid operation data and external environmental data. The data was divided into training and validation sets, and cross-validation was used for model training and evaluation. During training, the learning rate was set to 0.01 and the number of training rounds was set to 100. By adjusting the model's hyperparameters, the model achieved an accuracy of over 90% on the validation set.

[0043] Regularization techniques and early stopping methods are used to optimize the model to prevent overfitting.

[0044] S4: Inputting the normalized multi-source data into the risk prediction and early warning model to obtain risk prediction and early warning results.

[0045] After normalization, the multi-source data is input into the trained LSTM risk prediction model.

[0046] The model outputs a 24-hour forecast of the grid's risk level. Based on predefined thresholds, the risk level is categorized as low, medium, or high. For example, if the model predicts a 25% probability of failure in a specific area of the grid within the next 24 hours, it would be considered medium risk.

[0047] Generate detailed warning information based on the predicted risk level. For example, the warning information may be: "Warning Level: Medium Risk; Risk Type: Power Outage Due to Equipment Failure; Potentially Affected Area: Grid Area A; Expected Time of Occurrence: Within the Next 24 Hours." Early warning information is promptly disseminated to grid operation managers, relevant departments, and users through various channels, including the grid monitoring system, SMS platforms, and email. Medium-risk warnings are pushed via SMS and email to ensure that relevant personnel receive them promptly and can take appropriate countermeasures.

[0048] In Example 2, the present invention further provides a power grid operation risk prediction and early warning system, which adopts the power grid operation risk prediction and early warning method of Example 1, and includes: Data acquisition module, which acquires multi-source data for power grid operation risk prediction and early warning; A preprocessing module, configured to perform a data normalization operation on the multi-source data; Model building module, used to build risk prediction and early warning models; The prediction and warning module is used to input the normalized multi-source data into the risk prediction and warning model to obtain risk prediction and warning results.

[0049] Example 3: The present invention also provides an electronic device, including one or more processors and a memory.

[0050] The processor may be a central processing unit (CPU) or other forms of processing units having data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device to perform desired functions.

[0051] The memory may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may include, for example, random access memory (RAM) and / or cache memory. The non-volatile memory may include, for example, read-only memory (ROM), a hard disk, flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium, and the processor may execute the program instructions to implement a power grid operation risk prediction and early warning method of any embodiment of the present application described above and / or other desired functions. Various contents such as initial external parameters, thresholds, etc. may also be stored in the computer-readable storage medium.

[0052] In one example, the electronic device may further include an input device and an output device, these components interconnected via a bus system and / or other connection mechanisms (not shown). The input device may include, for example, a keyboard, a mouse, etc. The output device may output various information to the outside, including warning information, braking force, etc. The output device may include, for example, a display, a speaker, a printer, a communication network, and remote output devices connected thereto.

[0053] Of course, for the sake of simplicity, components such as buses, input / output interfaces, etc. are omitted. In addition, the electronic device may further include any other appropriate components according to specific application scenarios.

[0054] In addition to the above-mentioned methods and devices, an embodiment of the present application may also be a computer program product, which includes computer program instructions, which, when executed by a processor, enable the processor to implement the functions of a power grid operation risk prediction and early warning method provided by any embodiment of the present application.

[0055] The computer program product may be written in any combination of one or more programming languages to implement the program code for performing the operations of the embodiments of the present application, including object-oriented programming languages such as Java, C++, and conventional procedural programming languages such as "C" or similar programming languages. The program code may be executed entirely on the user's computing device, partially on the user's computing device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0056] In addition, an embodiment of the present application may also be a computer-readable storage medium having computer program instructions stored thereon. When the computer program instructions are executed by a processor, the processor enables the processor to implement a power grid operation risk prediction and early warning method provided by any embodiment of the present application.

[0057] The computer-readable storage medium may be any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. The readable storage medium may include, for example, but is not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or device, or any combination thereof. More specific examples (a non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof.

[0058] Finally, 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 above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the technical solutions of the embodiments of the present invention.

Claims

1. A method for predicting and warning of power grid operation risks, characterized in that: The following steps are involved: S1: Acquire multi-source data for power grid operation risk prediction and early warning; S2: performing data normalization on the multi-source data; The data normalization operation includes model data normalization and operation data normalization; The model data standardization operation is specifically as follows: for the main network model, meta-model conversion is achieved by building a semantic mapping rule library; for the distribution network model, the connection relationship between the nodes of the reference relationship group is obtained through the single line diagram, and the connection relationship is compared and verified with the topological relationship in the distribution network model database, and then normalized and archived; S3: Build a risk prediction and early warning model; S4: Inputting the normalized multi-source data into the risk prediction and early warning model to obtain risk prediction and early warning results.

2. The method for predicting and warning power grid operation risks according to claim 1, characterized in that: The multi-source data includes a main network model, a distribution network model, main network operation data, distribution network operation data and user-related data.

3. The method for predicting and warning power grid operation risks according to claim 2, characterized in that: The main network model includes the main network's topology, equipment parameters, and line connection relationship information; the distribution network model includes the distribution network's network structure, equipment configuration, and user access information; the distribution network operation data includes the distribution network's voltage, current, load, and fault information; and the user-related data includes the user's electricity usage behavior, electricity-using equipment, and geographic location information.

4. The method for predicting and warning power grid operation risks according to claim 1, characterized in that: The meta-model conversion is achieved for the main network model by building a semantic mapping rule base, specifically: Perform ontology analysis on the main network model to extract core classes and their attribute relationships; Establish a mapping rule table between the target system metamodel and CIM elements; Convert the attribute relationships in the main network model so that they conform to the data format and semantic requirements of the target system; Reconstructing the association relationship in the main network model to adapt to the topology requirements of the target system; Build the target-library table mapping to search the belonging library table and determine the normalized storage location.

5. The method for predicting and warning power grid operation risks according to claim 4, characterized in that: For the distribution network model, the connection relationship between the nodes of the reference relationship group is obtained through the single line diagram, and the connection relationship is compared and verified with the topological relationship in the distribution network model database, and then normalized and archived; Specifically: Obtaining single-line diagram data of the distribution network model; Identifying nodes of a distribution network model in the one-line diagram data; Extracting connection relationships between nodes in the single-line graph data to form a comparison relationship group; The control relationship group is compared and checked with the topological relationship in the distribution network model database.

6. The method for predicting and warning power grid operation risks according to claim 5, characterized in that: The process of calculating the inconsistency between the topological relationship between the control relationship group and the distribution network model database by improving the Euclidean distance is specifically as follows: Calculate the Euclidean distance between the coordinates of the inconsistent nodes in the topological relationship diagram in the distribution network model database and the coordinates in the single-line diagram according to the device type and the node connection relationship; Among them, the Euclidean distance is calculated using the improved Euclidean distance calculation formula; Specifically, the calculation formula of the improved Euclidean distance d is: ； Wherein, (x1, y1) and (x2, y2) are the coordinates of the inconsistent nodes in the topological relationship diagram in the distribution network model database and in the single-line diagram respectively; w x and w y are weight factors in the x-coordinate and y-coordinate directions respectively; t1 and t2 are device type codes of inconsistent nodes in the topological relationship of the distribution network model database and the device type codes in the reference relationship group respectively; c1 and c2 are connection relationship codes of inconsistent nodes in the topological relationship of the distribution network model database and the connection relationship codes in the reference relationship group respectively; w t and w c are the weight factors of device type and connection relationship respectively; Dynamically adjust the threshold θ according to the calculation result of the Euclidean distance; The threshold value is dynamically adjusted according to the calculation result of the Euclidean distance as follows: Initial threshold setting: set an initial threshold θ0; Dynamically adjust the threshold θ according to the distribution of the Euclidean distance d; The calculation formula of the threshold is: ； Where σ is the standard deviation of the Euclidean distance of all inconsistent nodes, and μ is the mean of the Euclidean distance of all inconsistent nodes; The Euclidean distance between the topological relationship diagram of the inconsistent node in the distribution network model database and the coordinates in the single-line diagram is compared with the threshold θ, thereby achieving normalized archiving.

7. The method for predicting and warning power grid operation risks according to claim 6, characterized in that: Select LSTM model as the power grid operation risk prediction model; The LSTM model is trained using historical power grid operation data and external environmental data; the data is divided into a training set and a validation set, and the model is trained and evaluated using a cross-validation method; During the training process, the learning rate is set to 0.01 and the number of training rounds is 100.

8. The method for predicting and warning power grid operation risks according to claim 7, characterized in that: After normalizing the multi-source data, the data is input into the trained LSTM risk prediction model; The model outputs the prediction results of the power grid risk level, and divides the risk level into three levels: low risk, medium risk and high risk according to the set threshold; Generate detailed early warning information based on the predicted risk level; Early warning information is promptly released to power grid operation managers, relevant departments and users through multiple channels such as power grid monitoring system, SMS platform and email; medium-risk early warning information is pushed via SMS and email.

9. A power grid operation risk prediction and early warning system, using the power grid operation risk prediction and early warning method according to any one of claims 1 to 8, characterized in that: The system comprises: Data acquisition module, which acquires multi-source data for power grid operation risk prediction and early warning; A preprocessing module, configured to perform a data normalization operation on the multi-source data; Model building module, used to build risk prediction and early warning models; The prediction and warning module is used to input the normalized multi-source data into the risk prediction and warning model to obtain risk prediction and warning results.

Images