Power supply and distribution digital twinning application method and device for emergency deduction

Abstract

The invention relates to the technical field of data center intelligent operation and maintenance, and particularly provides a power supply and distribution digital twinning application method and device for emergency deduction, and the method comprises the following steps: S1, data collection and processing; s2, power supply and distribution twin modeling; s3, knowledge acquisition; s4, constructing a knowledge graph; and S5, knowledge reasoning. Compared with the prior art, the method can solve the current situation that the emergency drilling scene and the online application are relatively separated, and deeply integrates the power supply and distribution digital twinning and the emergency drilling scene.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent operation and maintenance of data centers, and specifically provides a digital twin application method and device for power supply and distribution for emergency simulation. Background Art

[0002] As the core of information storage, processing and transmission, the stability and reliability of the power supply and distribution system of the data center is directly related to the continuity of business and the security of data. Therefore, regular power supply and distribution emergency drills to test and improve emergency response capabilities are the key to ensuring that the data center can quickly restore power supply and ensure normal business operations in emergencies.

[0003] At present, emergency drills are mainly conducted offline. However, large-scale offline emergency drills require a lot of manpower, material and financial resources. Due to resource constraints, offline drills can only be conducted for specific scenarios, personnel and processes, and it is difficult to fully cover all possible emergency situations and participants. At the same time, the evaluation of offline drills often relies on the subjective judgment and experience of observers, which is difficult to be objective, comprehensive and quantitative.

[0004] At the same time, the current power supply and distribution system usually adopts static visualization means to display, which only presents the operating conditions such as the equipment's own alarms. There is no correlation analysis between equipment alarms, which cannot support the rapid location of the root cause of the alarm. It is completely separated from business demands such as fault impact analysis and emergency simulation, and cannot support actual production requirements. Summary of the invention

[0005] In view of the deficiencies of the above-mentioned prior art, the present invention provides a digital twin application method for power supply and distribution for emergency simulation with strong practicality.

[0006] A further technical task of the present invention is to provide a digital twin application device for power supply and distribution that is rationally designed, safe and applicable for emergency simulation.

[0007] The technical solution adopted by the present invention to solve its technical problem is:

[0008] A power supply and distribution digital twin application method for emergency simulation has the following steps:

[0009] S1, data collection and processing;

[0010] S2, power supply and distribution twin modeling;

[0011] S3, knowledge acquisition;

[0012] S4, construction of knowledge graph;

[0013] S5. Knowledge reasoning.

[0014] Furthermore, in step S1, it includes:

[0015] S1-1. Obtain real-time alarm data, operation data and resource data of power supply and distribution equipment;

[0016] S1-2, standardize, organize and clean the data collected from multiple parties;

[0017] S1-3. Store the alarm and operation data of power supply and distribution equipment over the past period of time as a data set for AI training.

[0018] Furthermore, in step S2, based on the configuration-based twin tool, the drawing and editing of the power supply and distribution dynamic topology diagram is completed, which specifically includes:

[0019] S2-1, equipment modeling, representing various equipment in the power supply and distribution system in the form of 2D graphics;

[0020] S2-2, connection relationship modeling, the connection element represents the upstream and downstream connection relationship between devices, defines the upstream and downstream devices of each topological connection, and the current flows directly from the upstream device to the downstream device;

[0021] S2-3. Bind graphic element resources to establish a mapping relationship between graphic elements and the real power supply and distribution equipment connected to the system, and realize the positioning and dynamic update of graphic element equipment alarms, resources, and operating data.

[0022] Furthermore, in step S3, the PrefixSpan algorithm is used to acquire knowledge for mining sequence frequency patterns. In the alarm fault correlation analysis, the PrefixSpan algorithm is used to mine the time correlation and patterns between different faults and automatically mine historical alarm data.

[0023] Furthermore, based on the PrefixSpan mining algorithm, historical alarm data mining is automatically performed to form a point-to-point alarm propagation relationship. The specific steps are as follows:

[0024] S3-1, data collection, collect historical alarm fault data, pre-process the data, and filter out noise data items;

[0025] S3-2. Define fault events and sequences, regard each alarm fault event as an item set, combine multiple alarm fault events into sequences in chronological order, and form an alarm fault sequence data set;

[0026] S3-3. Set the support threshold, which is used to define the minimum occurrence frequency of the sequence frequency;

[0027] S3-4. Use the PrefixSpan algorithm to mine the fault sequence data set to find the sequence frequency that meets the support threshold. The PrefixSpan algorithm will recursively mine the sequence frequency corresponding to each prefix until no longer sequence frequency can be found;

[0028] S3-5. Analyze the sequence frequencies mined to identify the correlation between different alarm faults. The weight of the relationship is calculated using conditional probability, that is, based on the sequence frequency given by the mining algorithm, the ratio of [the number of times the secondary alarm occurs under the condition that the primary alarm occurs] to [the total number of times the primary alarm occurs] is used as the weight of the alarm relationship.

[0029] Furthermore, in step S4, two graphs are constructed:

[0030] (1) Based on topological drawing information, resource data is used to form a resource map;

[0031] (2) Based on the PrefixSpan algorithm, point-to-point knowledge is used to form an alarm graph;

[0032] Based on the prefixspan mining algorithm, alarm association relationships are automatically mined and annotated and confirmed by business experts; based on the discovered alarm association relationships and resource topology analysis, alarm relationship maps and resource maps are automatically constructed.

[0033] Furthermore, in step S5, the fault alarm is superimposed on the resource map to generate an alarm fact map; and the root cause of the fault is located based on the Random Walk algorithm. At the same time, the path is verified and optimized in combination with the Dijkstra shortest path algorithm to determine the fault alarm propagation path and the root cause device.

[0034] Furthermore, the Random Walk algorithm is a strategy for traversing nodes in a graph and exploring their adjacent nodes. It is used to simulate random processes in the physical world or to simulate information propagation in a network. In the analysis of the fault propagation path, the components in the topology are regarded as nodes of the graph, and the fault propagation path is represented by the connections between the nodes. The specific steps are as follows:

[0035] (a) Build a graph model: Model the drawn power supply and distribution topology as a graph, where nodes represent power supply and distribution devices in the topology and edges represent power supply lines between devices;

[0036] (b) Define the initial fault condition: determine the node or set of nodes where the fault starts. These nodes are the starting points of the Random Walk.

[0037] (c) Set the rules of Random Walk: Select the next node. In Random Walk, the current node will randomly select an adjacent node as the next node to visit. This is achieved by randomly selecting one from all neighbors of the current node.

[0038] Define fault propagation probability: Define a fault propagation probability for each edge, which represents the probability of a fault propagating from one node to adjacent nodes. Consider these probabilities when randomly selecting the next node.

[0039] Define stopping conditions: define when to stop the Random Walk;

[0040] (d) Execute Random Walk: Starting from the fault starting node, execute Random Walk according to the defined rules and record the visited nodes and fault propagation paths;

[0041] (e) Result analysis: Count all possible fault propagation paths and their frequencies, and identify components that are critical to fault propagation by analyzing the nodes or edges that are frequently visited;

[0042] Finally, based on the probability and path of fault propagation, the system is risk assessed to determine the areas most susceptible to faults.

[0043] Furthermore, the Dijkstra algorithm is used to find the shortest path from a single source point to all other nodes in a weighted graph. A greedy strategy is adopted to select the node with the shortest distance to the starting point from the unprocessed nodes each time, and update the distance of the neighboring nodes of the node. The specific steps are as follows:

[0044] 1) Build a graph model: Consider the components in the power supply and distribution topology as nodes of the graph, the connections between components as edges, and assign a weight to each edge;

[0045] 2) Set the starting point, and set the initial node where the alarm fault occurs as the starting point;

[0046] 3) Run Dijkstra algorithm: starting from the starting point, gradually find the shortest path to all other nodes;

[0047] 4) Analyze results: Based on the algorithm results, identify the shortest paths along which the fault may propagate, as well as the key nodes and edges on these paths.

[0048] A power supply and distribution digital twin application device for emergency simulation, comprising: at least one memory and at least one processor;

[0049] The at least one memory is used to store a machine-readable program;

[0050] The at least one processor is used to call the machine-readable program to execute a digital twin application method for power supply and distribution for emergency simulation.

[0051] Compared with the prior art, the power supply and distribution digital twin application method and device for emergency simulation of the present invention has the following outstanding beneficial effects:

[0052] The present invention is a power supply and distribution digital twin application for data center emergency simulation. It solves the current situation of relative separation between emergency drill scenarios and online applications, and deeply integrates the power supply and distribution digital twin with the emergency drill scenario. By simulating real circuit fault scenarios, the possible impact range of the alarm fault is analyzed through algorithms. The root cause of the alarm is quickly locked, thereby greatly shortening the troubleshooting time.

[0053] When an actual fault occurs, relevant personnel can quickly and accurately determine the fault point and take corresponding emergency measures to reduce the scope of the fault. At the same time, operators can refer to the emergency plan to manually switch the opening and closing status of the equipment, simulate possible fault handling operations, check the emergency plan's handling effect on specific faults, and promptly discover deficiencies in the emergency plan and revise and improve it to ensure the scientificity, pertinence and operability of the emergency plan. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0055] Attached Figure 1 It is a flow chart of a method for applying a digital twin of power supply and distribution for emergency simulation;

[0056] Attached Figure 2 It is a schematic diagram of an alarm spectrum in a digital twin application method for power supply and distribution in an emergency simulation;

[0057] Attached Figure 3 This is a schematic diagram of fault location using the Random Walk algorithm in the digital twin application method for power supply and distribution in an emergency simulation. DETAILED DESCRIPTION

[0058] In order to enable those skilled in the art to better understand the solution of the present invention, the present invention is further described in detail below in conjunction with specific implementation methods. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0059] A best embodiment is given below:

[0060] like Figure 1 As shown, a power supply and distribution digital twin application method for emergency simulation in this embodiment has the following steps:

[0061] S1, data collection and processing;

[0062] include:

[0063] S1-1. Data access: Obtain real-time alarm data, operation data, and resource data of power supply and distribution equipment. This is usually achieved through the following methods: First, connect to equipment or FSU to directly connect to equipment to obtain operating parameters; second, connect to a third-party system (usually a dynamic environment system) to achieve data collection and monitoring.

[0064] S1-2. Data processing: standardize, organize and clean data collected from multiple parties to ensure data consistency, improve data quality, and promote data analysis and decision-making.

[0065] S1-3. Data storage: The system stores the alarm and operation data of power supply and distribution equipment over the past period of time. It can be used as a data set for AI training to support continuous optimization and iteration of the model.

[0066] S2, power supply and distribution twin modeling;

[0067] Based on the configuration-based twin tool, the dynamic topology diagram of power supply and distribution can be quickly drawn and edited.

[0068] The specific steps are as follows:

[0069] S2-1. Equipment modeling: Various equipment in the power supply and distribution system (such as transformers, UPS, high and low voltage distribution cabinets, etc.) are represented in the form of 2D graphics. It supports manual layout drawing by dragging and dropping, and supports custom configuration of graphics attributes (height, width, color, rotation angle, etc.).

[0070] S2-2. Connection relationship modeling: The upstream and downstream connection relationships between devices are represented by connecting elements, and the upstream and downstream devices of each topological connection are defined. The current can flow directly from the upstream device to the downstream device.

[0071] S2-3. Graphics element resource binding: Establish a mapping relationship between graphics elements and the actual power supply and distribution equipment connected to the system to achieve accurate positioning and dynamic updating of graphics element equipment alarms, resources, and operating data.

[0072] S3, knowledge acquisition;

[0073] The PrefixSpan algorithm is an algorithm used to mine sequence frequency patterns. It is effective when processing time-ordered data sets. In the analysis of alarm fault correlation relationships, the PrefixSpan algorithm can be used to mine the time correlation and patterns between different faults, automatically mine historical alarm data, and thus provide an important basis for fault diagnosis.

[0074] Based on the PrefixSpan mining algorithm, historical alarm data can be automatically mined to form a point-to-point alarm propagation relationship. The specific steps are as follows:

[0075] S3-1. Data collection: Collect historical alarm fault data, including the time, type, impact, etc. of the alarm fault. Pre-process the data and filter out noise data items to ensure the accuracy and completeness of the data.

[0076] S3-2. Define fault events and sequences: Each alarm fault event is considered as an item set, which contains the specific information of the alarm fault (such as time, type, severity, etc.). Multiple alarm fault events are combined into sequences in chronological order to form an alarm fault sequence data set.

[0077] S3-3. Set the support threshold: The support threshold is used to define the minimum occurrence frequency of the sequence frequency. Set an appropriate threshold according to actual needs.

[0078] S3-4, PrefixSpan algorithm processing: Use the PrefixSpan algorithm to mine the fault sequence data set to find the sequence frequency that meets the support threshold. The PrefixSpan algorithm will recursively mine the sequence frequency corresponding to each prefix until no longer sequence frequency can be found.

[0079] S3-5. Result analysis: Analyze the mined sequence frequencies and identify the correlations between different alarm faults.

[0080] The weight of the relationship is calculated using conditional probability, that is, based on the sequence frequency given by the mining algorithm, the ratio of [the number of times the secondary alarm occurs under the condition that the primary alarm occurs] to [the total number of times the primary alarm occurs] is used as the weight of the alarm relationship.

[0081] S4, construction of knowledge graph;

[0082] like Figure 2 As shown, it includes constructing two graphs:

[0083] (1) Based on topological drawing information, resource data is used to form a resource map;

[0084] (2) Based on the PrefixSpan algorithm, point-to-point knowledge is used to form an alarm graph;

[0085] Based on the prefixspan mining algorithm, alarm associations are automatically mined and annotated and confirmed by business experts. Based on the discovered alarm associations and resource topology analysis, alarm relationship maps and resource maps are automatically constructed.

[0086] S5, knowledge reasoning;

[0087] like Figure 3 As shown, the fault alarm is superimposed on the resource map to generate an alarm fact map; and the root cause of the fault is located based on the Random Walk algorithm.

[0088] At the same time, the Dijkstra shortest path algorithm is used to verify and optimize the path, and determine the fault alarm propagation path and root cause device.

[0089] The Random Walk algorithm is a strategy for traversing nodes in a graph and exploring their adjacent nodes. It is often used to simulate random processes in the physical world or to simulate information propagation in a network. In the analysis of fault propagation paths, we can regard the components in the topology as nodes of the graph, and the fault propagation path is represented by the connections (edges) between nodes.

[0090] The specific steps are as follows:

[0091] (1) Build a graph model: The drawn power supply and distribution topology is modeled as a graph, where nodes represent power supply and distribution devices in the topology and edges represent power supply lines between devices.

[0092] (2) Define the initial fault condition: determine the node or set of nodes where the fault starts. These nodes will be the starting point of the RandomWalk.

[0093] (3) Set the rules of Random Walk: Select the next node. In Random Walk, the current node will randomly select an adjacent node as the next node to visit. This is usually achieved by randomly selecting one from all neighbors of the current node.

[0094] Define fault propagation probabilities: You can define a fault propagation probability for each edge, which represents the probability of a fault propagating from one node to its neighboring nodes. These probabilities can be taken into account when randomly selecting the next node;

[0095] Define stopping conditions: Define when to stop the Random Walk. For example, you can set a maximum number of steps, visit a specific set of nodes, or reach a certain propagation depth.

[0096] (4) Execute Random Walk: Starting from the fault starting node, execute Random Walk according to the defined rules. Record the visited nodes and fault propagation paths.

[0097] (5) Result analysis: Count all possible fault propagation paths and their frequencies. By analyzing which nodes or edges are frequently visited, components that are critical to fault propagation can be identified.

[0098] Finally, based on the probability and path of fault propagation, a risk assessment can be performed on the system to determine which parts are most susceptible to failure.

[0099] Dijkstra's algorithm is an algorithm for finding the shortest path from a single source point to all other nodes in a weighted graph. It uses a greedy strategy, each time selecting the node with the shortest distance to the starting point from the unprocessed nodes, and updating the distance of the node's neighbor nodes.

[0100] The specific steps are as follows:

[0101] 1) Build a graph model: The components in the power supply and distribution topology are regarded as nodes of the graph, the connection relationship between the components is regarded as the edge, and each edge is assigned a weight (such as the probability or cost of fault propagation).

[0102] 2) Set the starting point: Set the initial node where the alarm fault occurs as the starting point.

[0103] 3) Run Dijkstra's algorithm: starting from the starting point, gradually find the shortest path to all other nodes.

[0104] 4) Analyze results: Based on the algorithm results, identify the shortest paths along which the fault may propagate, as well as the key nodes and edges on these paths.

[0105] It helps to identify weak links in the power supply and distribution system, assess the impact of faults, and develop effective fault prevention and recovery strategies.

[0106] Based on the above method, a power supply and distribution digital twin application device for emergency simulation in this embodiment includes: at least one memory and at least one processor;

[0107] The at least one memory is used to store a machine-readable program;

[0108] The at least one processor is used to call the machine-readable program to execute a digital twin application method for power supply and distribution for emergency simulation.

[0109] The above-mentioned specific implementations are only specific cases of the present invention. The patent protection scope of the present invention includes but is not limited to the above-mentioned specific implementations. Any technical solutions that conform to the above-mentioned specific implementations of the present invention and any appropriate changes or substitutions made by ordinary technicians in the relevant technical field shall fall within the patent protection scope of the present invention.

[0110] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A digital twin application method for power supply and distribution in emergency simulation, characterized in that: The steps are as follows: S1, data collection and processing; S2, power supply and distribution twin modeling; S3, knowledge acquisition; S4, construction of knowledge graph; S5. Knowledge reasoning.

2. According to the method for applying a digital twin of power supply and distribution for emergency simulation in claim 1, it is characterized in that: In step S1, it includes: S1-1. Obtain real-time alarm data, operation data and resource data of power supply and distribution equipment; S1-2, standardize, organize and clean the data collected from multiple parties; S1-3. Store the alarm and operation data of power supply and distribution equipment over the past period of time as a data set for AI training.

3. According to claim 2, a method for applying a digital twin of power supply and distribution for emergency simulation is characterized in that: In step S2, based on the configuration twin tool, the drawing and editing of the power supply and distribution dynamic topology diagram is completed, which specifically includes: S2-1, equipment modeling, representing various equipment in the power supply and distribution system in the form of 2D graphics; S2-2, connection relationship modeling, the connection element represents the upstream and downstream connection relationship between devices, defines the upstream and downstream devices of each topological connection, and the current flows directly from the upstream device to the downstream device; S2-3. Bind graphic element resources to establish a mapping relationship between graphic elements and the real power supply and distribution equipment connected to the system, and realize the positioning and dynamic update of graphic element equipment alarms, resources, and operating data.

4. The method for applying a digital twin of power supply and distribution for emergency simulation according to claim 3 is characterized in that: In step S3, the PrefixSpan algorithm is used to acquire knowledge for mining sequence frequency patterns. In the alarm fault correlation analysis, the PrefixSpan algorithm is used to mine the time association and patterns between different faults and automatically mine historical alarm data.

5. The method for applying a digital twin of power supply and distribution for emergency simulation according to claim 4 is characterized in that: Based on the PrefixSpan mining algorithm, historical alarm data mining is automatically performed to form a point-to-point alarm propagation relationship. The specific steps are as follows: S3-1, data collection, collect historical alarm fault data, pre-process the data, and filter out noise data items; S3-2. Define fault events and sequences, regard each alarm fault event as an item set, combine multiple alarm fault events into sequences in chronological order, and form an alarm fault sequence data set; S3-3. Set the support threshold, which is used to define the minimum occurrence frequency of the sequence frequency; S3-4. Use the PrefixSpan algorithm to mine the fault sequence data set to find the sequence frequency that meets the support threshold. The PrefixSpan algorithm will recursively mine the sequence frequency corresponding to each prefix until no longer sequence frequency can be found; S3-5. Analyze the sequence frequencies mined to identify the correlation between different alarm faults. The weight of the relationship is calculated using conditional probability, that is, based on the sequence frequency given by the mining algorithm, the ratio of [the number of times the secondary alarm occurs under the condition that the primary alarm occurs] to [the total number of times the primary alarm occurs] is used as the weight of the alarm relationship.

6. The method for applying a digital twin of power supply and distribution for emergency simulation according to claim 5 is characterized in that: In step S4, two maps are constructed: (1) Based on topological drawing information, resource data is used to form a resource map; (2) Based on the PrefixSpan algorithm, point-to-point knowledge is used to form an alarm graph; Based on the prefixspan mining algorithm, alarm association relationships are automatically mined and annotated and confirmed by business experts; based on the discovered alarm association relationships and resource topology analysis, alarm relationship maps and resource maps are automatically constructed.

7. The method for applying a digital twin of power supply and distribution for emergency simulation according to claim 6 is characterized in that: In step S5, the fault alarm is superimposed on the resource map to generate an alarm fact map; and the root cause of the fault is located based on the Random Walk algorithm. At the same time, the path is verified and optimized in combination with the Dijkstra shortest path algorithm to determine the fault alarm propagation path and the root cause device.

8. The method for applying digital twins of power supply and distribution for emergency simulation according to claim 7 is characterized in that: The Random Walk algorithm is a strategy for traversing nodes in a graph and exploring their adjacent nodes. It is used to simulate random processes in the physical world or to simulate information propagation in a network. In the analysis of fault propagation paths, the components in the topology are regarded as nodes of the graph, and the fault propagation path is represented by the connections between the nodes. The specific steps are as follows: (a) Build a graph model: Model the drawn power supply and distribution topology as a graph, where nodes represent power supply and distribution devices in the topology and edges represent power supply lines between devices; (b) Define the initial fault condition: determine the node or set of nodes where the fault starts. These nodes are the starting points of the Random Walk. (c) Set the rules of Random Walk: Select the next node. In Random Walk, the current node will randomly select an adjacent node as the next node to visit. This is achieved by randomly selecting one from all neighbors of the current node. Define fault propagation probability: Define a fault propagation probability for each edge, which represents the probability of a fault propagating from one node to adjacent nodes. Consider these probabilities when randomly selecting the next node. Define stopping conditions: define when to stop the Random Walk; (d) Execute Random Walk: Starting from the fault starting node, execute Random Walk according to the defined rules and record the visited nodes and fault propagation paths; (e) Result analysis: Count all possible fault propagation paths and their frequencies, and identify components that are critical to fault propagation by analyzing the nodes or edges that are frequently visited; Finally, based on the probability and path of fault propagation, the system is risk assessed to determine the areas most susceptible to faults.

9. The method for applying a digital twin of power supply and distribution for emergency simulation according to claim 8 is characterized in that: The Dijkstra algorithm is used to find the shortest path from a single source point to all other nodes in a weighted graph. It adopts a greedy strategy, selects the node with the shortest distance to the starting point from the unprocessed nodes each time, and updates the distance of the neighboring nodes of the node. The specific steps are as follows: 1) Build a graph model: Consider the components in the power supply and distribution topology as nodes of the graph, the connections between components as edges, and assign a weight to each edge; 2) Set the starting point, and set the initial node where the alarm fault occurs as the starting point; 3) Run Dijkstra algorithm: starting from the starting point, gradually find the shortest path to all other nodes; 4) Analyze results: Based on the algorithm results, identify the shortest paths along which the fault may propagate, as well as the key nodes and edges on these paths.

10. A digital twin application device for power supply and distribution for emergency simulation, characterized in that: include: at least one memory and at least one processor; The at least one memory is used to store a machine-readable program; The at least one processor is configured to call the machine-readable program to execute the method according to any one of claims 1 to 9.

Images