Power distribution network panoramic dynamic topology global computing service method, integration method, system, electronic equipment, computer readable storage medium and computer program product

Abstract

The invention discloses a power distribution network panoramic dynamic topology global calculation service method, an integration method, a system, electronic equipment, a computer readable storage medium and a computer program product. The service method comprises the following steps: receiving data from a measurement center; based on an event-driven mechanism, pulling data related to the event; inputting the event-related data into a deep neural network to obtain a topology transaction probability; the topology transaction probability is compared with a set transaction probability threshold value, if the topology transaction probability is not larger than the transaction probability threshold value, global calculation is started, and if the topology transaction probability is larger than the transaction probability threshold value, only local calculation of a transaction area is started; pushing a calculation result to a measurement center; the integration method comprises the following steps: deploying a global computing service system in a real-time measurement center intranet; the method comprises the following steps: registering atomic-scale service of a global computing service system to obtain a service calling address; directionally pushing the data to a global computing service system based on the service calling address; and receiving a calculation result from the global calculation service system.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of power grid digitization, and particularly relates to a power distribution network panoramic dynamic topology global calculation service method, an integration method, a system, an electronic device, a computer readable storage medium and a computer program product. BACKGROUND

[0002] With the promotion of new power system construction, the scale of power distribution network equipment continues to expand, and the equipment nodes of some provincial power distribution networks exceed one million. Enterprise-level real-time measurement centers need to real-time gather multi-source data of main networks, distribution networks and power consumption sides to provide dynamic topology support for power outage analysis, fault research and judgment, line loss calculation and other businesses. At present, power distribution network topology calculation mainly relies on two types of technical solutions: one type is a traditional calculation component based on a relational database such as MySQL, which realizes topology construction and calculation through multi-table association query; the other type is a lightweight component based on a simplified graph model, which only supports small and medium-sized topology analysis. The existing technology has the following problems:

[0003] 1. Low global calculation performance: the traditional calculation component adopts single-thread traversal or simple parallel mode, and when facing one million equipment nodes, the topology instantiation time is often more than 20 minutes, which cannot meet the business needs of the real-time measurement center for "second-level response and minute-level calculation". The core reason is that the calculation architecture is not optimized for the physical characteristics of the power distribution network "layered and partitioned", resulting in waste of calculation resources.

[0004] 2. Difficulty in integration with the real-time measurement center: when the existing component is connected with the enterprise-level real-time measurement center, it needs to be hard-coded to adapt to the data format of the measurement center, such as the CIME model and the Kafka message format, and the adaptation period is as long as 2-3 weeks. Moreover, it cannot be compatible with the subsequent data model upgrade of the measurement center, which has poor flexibility and high integration cost.

[0005] 3. Slow response of dynamic topology: the topology of the power distribution network frequently changes due to switch position changes and equipment changes, and the existing component needs to recalculate the global topology. Even if there is a local change, it also needs to traverse all nodes, and the invalid calculation overhead accounts for more than 60%, resulting in a topology update response time of more than 10 seconds, which affects the decision efficiency of real-time businesses such as fault research. SUMMARY

[0006] The application aims to solve the problems of low global calculation performance of the existing power distribution network topology calculation component when facing one million equipment nodes, high adaptation cost and poor flexibility of the component when integrated with the enterprise-level real-time measurement center, and slow response of the component when the topology of the power distribution network changes and high invalid calculation ratio. The application provides a power distribution network panoramic dynamic topology global calculation service method, an integration method, a system, an electronic device, a computer readable storage medium and a computer program product, which are suitable for high-performance topology calculation scenarios of service integration with enterprise-level real-time measurement centers.

[0007] Technical solution: In a first aspect, the present application provides a power distribution network panoramic dynamic topology global calculation service method, comprising:

[0008] receiving data from a measurement center;

[0009] pulling event-related data based on an event-driven mechanism;

[0010] inputting the event-related data into a deep neural network to obtain a topology anomaly probability;

[0011] comparing the topology anomaly probability with a set anomaly probability threshold value, if not greater than the anomaly probability threshold value, starting global calculation, if greater than the anomaly probability threshold value, only starting local calculation of the anomaly region;

[0012] pushing the calculation result to the measurement center.

[0013] Further, if not greater than the anomaly probability threshold value, starting global calculation, if greater than the anomaly probability threshold value, only starting local calculation of the anomaly region, comprising:

[0014] dividing the power distribution network topology structure into multiple layers according to voltage levels;

[0015] for each layer, dividing into multiple subgraphs according to geographical regions or power supply partitions; each layer and each subgraph is independently calculated in parallel;

[0016] when not greater than the anomaly probability threshold value, starting global calculation, comprising: all subgraphs are independently calculated in parallel;

[0017] when greater than the anomaly probability threshold value, only starting local calculation of the anomaly region, comprising: only starting local calculation of the anomaly region subgraph.

[0018] Further, the input of the deep neural network includes the number of device state changes in the last one hour, and the voltage fluctuation amplitude.

[0019] In a second aspect, the present application provides a power distribution network panoramic dynamic topology global calculation service system, comprising:

[0020] a data conversion module configured to receive data from a measurement center;

[0021] a service integration module configured to pull event-related data based on an event-driven mechanism, and push the calculation result to the measurement center;

[0022] The global calculation module is configured to input the event-related data into a deep neural network to obtain a topology anomaly probability, compare the topology anomaly probability with a set anomaly probability threshold, if the topology anomaly probability is not greater than the anomaly probability threshold, start global calculation, and if the topology anomaly probability is greater than the anomaly probability threshold, only start local calculation of an anomaly region.

[0023] Further, in the global calculation module, the topology anomaly probability is compared with the set anomaly probability threshold, if the topology anomaly probability is not greater than the anomaly probability threshold, global calculation is started, and if the topology anomaly probability is greater than the anomaly probability threshold, only local calculation of the anomaly region is started, and specifically includes:

[0024] The power distribution network topology structure is divided into multiple layers according to voltage levels;

[0025] For each layer, multiple subgraphs are divided according to geographical regions or power supply partitions; each layer and each subgraph is independently calculated in parallel;

[0026] When the topology anomaly probability is not greater than the anomaly probability threshold, global calculation is started, including: all subgraphs are independently calculated in parallel;

[0027] When the topology anomaly probability is greater than the anomaly probability threshold, only local calculation of the anomaly region is started, including: only local calculation of the anomaly region subgraph is started.

[0028] Further, the input of the deep neural network includes the number of device state changes in nearly 1 hour and the voltage fluctuation amplitude.

[0029] In a third aspect, the application provides an integration method of a global calculation service system and a real-time measurement center, including:

[0030] The global calculation service system is deployed in a real-time measurement center intranet, and hardware and software environments are configured; the global calculation service system is a power distribution network panoramic dynamic topology global calculation service system according to any one of claims 1 to 6;

[0031] The global calculation service system is registered in a real-time measurement center digital portal to obtain a service calling address;

[0032] A real-time measurement center message pushing rule is configured, and data is pushed to the global calculation service system based on the service calling address;

[0033] The calculation result from the global calculation service system is received.

[0034] In a fourth aspect, the application provides an electronic device, including:

[0035] At least one processor;

[0036] and a memory in communication connection with the at least one processor;

[0037] The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor to enable the at least one processor to perform the power distribution network panoramic dynamic topology global calculation service method.

[0038] In a fifth aspect, the present application provides a computer readable storage medium, which stores computer instructions for enabling a processor to implement the power distribution network panoramic dynamic topology global calculation service method when executed by the processor.

[0039] In a sixth aspect, the present application provides a computer program product, which comprises a computer program for implementing the power distribution network panoramic dynamic topology global calculation service method when executed by a processor.

[0040] In a seventh aspect, the present application provides an electronic device, which comprises:

[0041] at least one processor;

[0042] and a memory in communication connection with the at least one processor;

[0043] The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor to enable the at least one processor to perform the global calculation service system and real-time measurement center integration method.

[0044] In an eighth aspect, the present application provides a computer readable storage medium, which stores computer instructions for enabling a processor to implement the global calculation service system and real-time measurement center integration method when executed by the processor.

[0045] In a ninth aspect, the present application provides a computer program product, which comprises a computer program for implementing the global calculation service system and real-time measurement center integration method when executed by a processor.

[0046] Advantages: Compared with the prior art, the present application has the following advantages:

[0047] (1) Dynamic topology response speed is improved: through DNN state perception pre-judgment of abnormal area, only local calculation is triggered, response time is shortened, invalid calculation overhead is reduced, and decision timeliness of real-time business such as fault research and judgment is guaranteed;

[0048] (2) The global computing performance is significantly improved: the application realizes the "hierarchical-subgraph" hybrid parallel computing, combines the TSM-TREE storage optimization, shortens the instantiation time, and significantly improves the computing efficiency, so as to meet the minute-level computing demand of the enterprise-level real-time measurement center.

[0049] (3) The integration flexibility and economy are improved: through the standardized interface and the event-driven mechanism, the core code of the component does not need to be modified, the adaptation cycle is shortened, the integration cost is reduced, and the subsequent data model upgrade of the measurement center is supported. BRIEF DESCRIPTION OF DRAWINGS

[0050] Figure 1 The flow chart of the high-performance panoramic dynamic topology global computing service component of the power distribution network is provided in the application.

[0051] Figure 2 The architecture diagram of the high-performance panoramic dynamic topology global computing service component of the power distribution network is provided in the application.

[0052] Figure 3 The "hierarchical-subgraph" parallel computing flow chart of the global computing module is provided in the application. DETAILED DESCRIPTION

[0053] In order to make the purpose, technical scheme and advantages of the application clearer, the following will combine the drawings and examples in the application to further describe the power distribution network panoramic dynamic topology global computing service component and integration method provided in the application.

[0054] Example one:

[0055] The application provides a power distribution network panoramic dynamic topology global computing service method, which comprises the following steps:

[0056] Receiving data from the measurement center;

[0057] Pulling the event-related data based on the event-driven mechanism;

[0058] Inputting the event-related data into the deep neural network to obtain the topology anomaly probability;

[0059] Comparing the topology anomaly probability with the set anomaly probability threshold value, if the topology anomaly probability is not greater than the anomaly probability threshold value, starting the global computing, if the topology anomaly probability is greater than the anomaly probability threshold value, only starting the local computing of the anomaly region;

[0060] Pushing the computing result to the measurement center.

[0061] The application triggers the local computing by predicting the anomaly region through the deep neural network, shortens the response time, reduces the invalid computing cost, guarantees the decision timeliness of the real-time business such as fault analysis, and improves the dynamic topology response speed.

[0062] Embodiment two:

[0063] On the basis of embodiment one, the embodiment of the application makes further optimization, including: the embodiment of the application realizes high-performance global computing of power distribution network panoramic dynamic topology and rapid integration of enterprise-level real-time measurement center through the technical system of "component modular design + lightweight integration mechanism + pre-judgment calculation optimization". The following is described in detail in two parts:

[0064] As shown in Figure 1 and 2 , the high-performance power distribution network panoramic dynamic topology global computing service component proposed by the embodiment of the application adopts "data-computation-integration-monitoring" full-link modular design, each module is independent and cooperates, and the modules interact through standardized data interfaces. Specifically, it includes a data conversion module, a global computing module, a service integration module and a monitoring module.

[0065] Now the data conversion module is further described.

[0066] The data conversion module receives the multi-source heterogeneous data of the enterprise-level real-time measurement center, completes cleaning, standardized conversion, intermediate storage, and provides consistent data input for the global computing module. Specifically:

[0067] Kafka message queue is used as a data access channel to receive three types of core data output by the enterprise-level real-time measurement center: ① power distribution network model data, CIME format, including device account, static topology relationship; ② real-time measurement data, voltage, current, power and other electrical quantities, sampling period 15 seconds / time; ③ device state data, switch opening and closing state, device operation and maintenance state.

[0068] Built-in data cleaning rules are used for data cleaning. For example, "default value filling + log alarm" is used for missing fields; when the unique device ID is missing, a "NULL-device type-timestamp" temporary ID is filled; format abnormal data such as non-numeric electrical quantities are filtered and returned; time series inconsistent data such as timestamp deviation exceeding 5 seconds are corrected based on the measurement center clock synchronization service.

[0069] When the qualified rate of cleaned data is lower than the threshold, the monitoring module is triggered to alarm.

[0070] Data conversion and storage: the cleaned data is uniformly converted into standardized JSON format, including device ID, data type, timestamp, core attributes (such as voltage value, switch state); stored in the Redis intermediate library for subsequent module calling.

[0071] Now the global computing module is further described.

[0072] The global computing module adopts a hybrid computing architecture of "hierarchical parallel + subgraph parallel", combined with deep neural network (DNN) state-aware optimization, and refers to Figure 3 , and specifically includes the following contents:

[0073] (1) Hierarchical parallel computing: The topology is divided into high-voltage layer (110kV and above), medium-voltage layer (10kV), and low-voltage layer (0.4kV) according to the voltage level of the power distribution network. Each layer starts a computing thread independently, and the boundary data between layers is transmitted through shared memory, such as high-voltage-medium-voltage tie-in transformer parameters, to avoid cross-layer data interaction bottlenecks.

[0074] (2) Subgraph parallel computing: Within each layer, the topology is divided into several subgraphs according to geographical regions such as city-level administrative districts or power supply areas such as substation power supply ranges. Each subgraph is assigned an independent computing core, and data consistency between subgraphs is achieved through distributed locks.

[0075] (3) DNN state-aware optimization: A 3-layer DNN model is constructed, with 8 feature nodes in the input layer, including device state change frequency, measurement data fluctuation amplitude, and historical anomaly records within the last 1 hour. The hidden layer contains 16 neurons, and the output topology anomaly probability threshold is set to 0.7. When the deep neural network (DNN) predicts that the topology anomaly probability is ≤0.7, global computing is started; when the anomaly probability is >0.7, only the local computing of the anomaly area subgraph is started, without the need for global recalculation.

[0076] (4) Storage optimization: The TSM-TREE file structure is used to store graph data, and the node attributes are compressed according to "device type-voltage level" classification, reducing disk I / O overhead and improving data read / write speed.

[0077] Now the service integration module will be further described.

[0078] The service integration module realizes lightweight integration of components and enterprise-level real-time measurement centers, including service registration, permission authentication, event-driven data interaction, without modifying the existing architecture of the measurement center. Specifically, it includes:

[0079] Among them, service registration can be understood as: standardized registration based on digital portal. Through the RESTful interface of the digital portal of the measurement center, atomic-level services such as topology instantiation, topology update, and topology query are registered, and service description, input / output parameters, and call frequency limit are filled in.

[0080] Permission authentication: secure access based on OAuth2.0 protocol, specifically including: using OAuth2.0 protocol, the component as a client applies for a token to the measurement center authentication server, and the token validity period is set to 2 hours to avoid repeated authentication.

[0081] Data interaction: based on the event-driven mechanism, the data update of the measurement center triggers the "data ready" event, and the component automatically pulls the data after listening to the event; after the calculation is completed, the component pushes the "result ready" event to the measurement center, and the measurement center obtains the result through the standardized interface.

[0082] The monitoring module will be further described.

[0083] The monitoring module monitors the running state of each module in real time, timely alarms failure and records log, and guarantees high reliability of service.

[0084] Real-time acquisition of running parameters of each module: ① calculation performance such as node processing rate, thread occupancy rate; ② data processing such as access success rate, cleaning qualified rate; ③ service state such as call success rate, response delay.

[0085] When the parameter exceeds the preset threshold, trigger email + SMS alarm, and record fault log to the ELK log system, support fault tracing.

[0086] The embodiment of the application aims to solve the problem of low global calculation performance of the existing power distribution network topology calculation component when facing million-level device nodes, shorten the topology instantiation time, meet the real-time demand of enterprise-level real-time measurement center; solve the problem of high adaptation cost and poor flexibility when integrating the component and enterprise-level real-time measurement center, realize lightweight integration, reduce the difficulty of subsequent maintenance; solve the problem of slow response and high invalid calculation proportion of the component when the power distribution network topology changes, improve the dynamic topology update efficiency, and support real-time business decision.

[0087] Embodiment three

[0088] The embodiment of the application proposes an integration method of global calculation service system and real-time measurement center, comprising:

[0089] The global calculation service system is deployed in the intranet of the real-time measurement center, and the hardware environment and software environment are configured; the global calculation service system is the power distribution network panoramic dynamic topology global calculation service system proposed in any one of the above embodiments;

[0090] Register the atomic-level service of the global calculation service system through the digital portal of the real-time measurement center, and obtain the service call address;

[0091] Configure the real-time measurement center message push rule, and based on the service call address, push the data to the global calculation service system;

[0092] Receive the calculation result from the global calculation service system.

[0093] Embodiment four:

[0094] On the basis of embodiment three, the embodiment of the application proposes an integration method of the component proposed in embodiment two and the enterprise-level real-time measurement center, which is roughly divided into five steps of deployment, registration, docking, triggering and returning. The existing architecture of the measurement center does not need to be modified, and the specific steps are as follows:

[0095] Component deployment: deploy the component in the intranet of the enterprise-level real-time measurement center. The hardware environment is that two 8-core 16G application servers are deployed for data conversion, service integration and monitoring modules, and one 16-core 32G computing server is deployed for global computing module. The software environment is JDK 1.8, Neo4j 4.4.10 (graph database) and Spring Cloud Hoxton.SR12 (microservice framework).

[0096] Service registration: log in to the digital portal of the measurement center, submit the registration application of the component atomic service, and after the administrator's approval, obtain the service call address and OAuth2.0 authentication key.

[0097] Data docking: configure the Kafka message topic in the measurement center, and push the power distribution network model data, real-time measurement data and device state data to the component data conversion module.

[0098] Calculation triggering: the measurement center data is automatically triggered when the "data ready" event is triggered, the component listens to the event and pulls the data, starts global calculation or local calculation, or initiates a calculation task manually through the portal, such as manually triggering historical topology backtracking.

[0099] Result returning: after the component calculation is completed, the "result ready" event is pushed to the topology service layer of the measurement center, supporting the upper-layer business call. The topology result is in JSON format, including node list (device ID, voltage level, state), edge list (connection relationship, line loss parameter) and topology version number.

[0100] Embodiment five:

[0101] The following will be described in detail in combination with the application scene of the target electric power company enterprise-level real-time measurement center:

[0102] The implementation environment includes:

[0103] (1) Hardware environment:

[0104] Device type Number Specific configuration parameters Deployment module Key role (distinguish from prior art) Computing server 1 Intel Xeon Gold 6248 (16 cores / 32 threads, base frequency 2.5GHz), 32GB DDR4 (2666MHz, dual channel), 2TB SSD (NVMe protocol, read / write speed 3500MB / s) Global computing module Provide 16-core parallel computing resources to support "hierarchical-subgraph" parallel computing (most prior art is below 8 cores, which cannot meet the requirements of million-level node parallelism) Application server 2 Intel Xeon E5-2680 v4 (8 cores / 16 threads, base frequency 2.4GHz), 16GB DDR4, 1TB HDD (7200rpm) Data conversion, service integration, monitoring module Dual-machine hot backup ensures data access stability and avoids loss of measurement data due to single-node failure (most prior art is single-node deployment) Kafka cluster server 3 Intel Xeon E5-2680 v4 (8 cores), 16GB DDR4, 1TB HDD Data access channel 3-node cluster configuration (1 controller + 2 replicas) to ensure real-time measurement data (15 seconds / second) without loss (most prior art is a single Kafka node) Graph database server 1 Same as computing server (share hardware resources) Neo4j graph database Configure 16GB heap memory + 8GB page cache to optimize graph data read / write (most prior art has 8GB heap memory, which is prone to OOM) Monitoring server 1 Intel Xeon E5-2680 v4 (8 cores), 16GB DDR4, 1TB HDD ELK log system Real-time storage of 6 months of fault logs, supporting module / time retrieval (most prior art has no long-term log storage)

[0105] (2) Software environment: the operating system is CentOS 7.9, the graph database is Neo4j 4.4.10, the distributed mode is started, the micro-service framework is Spring Cloud Hoxton.SR12, the DNN model training framework is TensorFlow 2.8, and the log system is ELK Stack (Elasticsearch 7.14, Logstash 7.14, Kibana 7.14).

[0106] The implementation steps include:

[0107] Step 1: Component deployment and initialization

[0108] (1) Deploy the global computing module on the computing server: install the Neo4j graph database, create the "power distribution network topology" database, and configure memory allocation (16 GB of heap memory); deploy the DNN state-aware sub-module, import the near-1-year target power company power distribution network topology change data training model, which contains 2000 switch position changes and 500 device operation changes, and the model accuracy is optimized to 92%.

[0109] (2) Deploy the data conversion module on the application server: configure the Kafka consumer, subscribe to the measurement center model data, real-time measurement, and device state three topics, set the consumption group to "topo-calc-group", and the message pulling frequency to 15 seconds / second; configure the data cleaning rules, such as "filling 'NULL-device type-time stamp' temporary ID when device ID is missing" and "marking as abnormal and filtering when voltage measurement value exceeds 380V".

[0110] (3) Deploy the service integration module and the monitoring module: configure the digital portal service registration parameters, fill in the service name "power distribution network topology global computing service", input parameters "topology version number, computing range (province / city)", and output parameters "topology JSON, computing time consumption"; configure the monitoring threshold, such as triggering an alarm when the call success rate is lower than 99.5% and the computing time consumption is more than 8 minutes.

[0111] Step 2: Service registration and permission configuration

[0112] (1) Log in to the target power company digital portal and enter the "service management" page, submit the component atomic service registration application, and upload the service interface document (including request method, parameter example).

[0113] (2) After the measurement center administrator approves, the component obtains the client ID and key, applies for an access token through the OAuth2.0 authentication server, sets the token validity period to 2 hours, and the component has built-in token automatic refreshing logic to avoid service interruption caused by token expiration.

[0114] Step 3: Data docking and calculation test

[0115] (1) The measurement center configures Kafka message push rules: CIME model data (full push at 3 am every day), real-time measurement data (15-second incremental push), and device status data (real-time push when the state changes) are pushed to the component.

[0116] (2) Initiate the provincial distribution network topology instantiation test: Call the "topology instantiation" service through the portal, input parameters "version number V20250601, calculation range of the whole province", and the component triggers global calculation: ① The data conversion module receives CIME model data (including 1.2 million device nodes), and the standardization rate after cleaning reaches 99.8%; ② The global calculation module is layered according to voltage level (12,000 nodes for high voltage, 450,000 nodes for medium voltage, and 738,000 nodes for low voltage), and 48 parallel threads are started for calculation according to 13 cities; ③ The DNN module predicts that the topology change probability of Wuxi and Suzhou is less than 0.3, and local recalculation is not needed; the final calculation time is 4 minutes and 20 seconds, and the consistency of the topology result with the measurement center static topology is 98.7%.

[0117] Step 4: Dynamic topology update test

[0118] (1) Simulate the 10kV switch position change in a certain city area under the jurisdiction of the target power company: The measurement center pushes the device status data of "switch ID=KG10086, state=off", and the component DNN module predicts that the topology change probability of the subgraph (5.2 million nodes in the city area medium voltage subgraph) is 0.85, triggering local calculation.

[0119] (2) Local calculation process: Only load the city area medium voltage subgraph data, recalculate the topology connection relationship after the switch is off, and the calculation time is 2 seconds and 15 milliseconds. The calculation result is pushed to the measurement center, supporting the fault research and judgment business to quickly locate the outage range.

[0120] Typical scenario application

[0121] (1) Time and section line loss analysis scenario: The component calculates the topology of each section line at a 15-minute interval, and outputs "line ID-section ID-topology relationship" data. The measurement center calculates the line loss combined with real-time power data, and the line loss calculation accuracy reaches 97%, which is 8% higher than traditional methods.

[0122] (2) Fault research and judgment scenario: The component updates the topology of the fault area in real time, and outputs the "fault device-associated user" relationship, supporting the power supply service command center to quickly dispatch repair work orders, and shortening the repair response time by 15%.

[0123] Through 6 months of trial operation, the component has completed topology calculation task 1440 times (10 times per day) in total, the service call success rate is 99.92%, the average calculation time is 4 minutes and 35 seconds, the average response time of dynamic topology update is 2.8 seconds, which fully meets the performance and reliability requirements of enterprise-level real-time measurement center on power distribution network topology calculation.

[0124] Embodiment six:

[0125] The embodiment discloses a computer device, comprising a memory, a processor and a computer program stored in the memory and executable on the processor, and the processor implements the steps disclosed in any one of the above embodiments when executing the computer program.

[0126] Embodiment seven:

[0127] The embodiment discloses a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the steps disclosed in any one of the above embodiments.

[0128] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by a computer program instructing related hardware, and the computer program can be stored in a non-volatile computer readable storage medium. When the computer program is executed, it can include the processes of the above-mentioned embodiments. In the embodiments provided in the present application, any reference to memory, storage, database or other medium can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory or optical memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration but not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0129] The technical features of the above embodiments can be combined in any way. In order to make the description concise, not all possible combinations of the technical features in the above embodiments are described, but as long as the combination of the technical features does not exist contradictory, it should be considered as the scope of the present application.

[0130] The above-described embodiments are merely illustrative of several embodiments of the present application, which are described in more detail and in a specific and detailed manner, but should not be construed as limiting the scope of the patent. It should be noted that for those skilled in the art, several modifications and improvements can be made without departing from the concept of the present application, and these are all within the scope of the present application. Therefore, the scope of protection of the patent of the present application should be subject to the appended claims.

Claims

1. A method for providing a panoramic dynamic topology calculation service for a distribution network, characterized in that: include: Receive data from the measurement center; Based on an event-driven mechanism, retrieve event-related data; The event-related data is input into a deep neural network to obtain the topological anomaly probability; The probability of topological anomaly is compared with a set anomaly probability threshold. If it is not greater than the anomaly probability threshold, global calculation is initiated. If it is greater than the anomaly probability threshold, local calculation of the anomaly region is initiated only. The calculation results are sent to the measurement center.

2. The method for providing a panoramic dynamic topology full-domain calculation service for a distribution network according to claim 1, characterized in that: If the probability of anomalies is not greater than the anomaly probability threshold, then global calculation is initiated; if the probability of anomalies is greater than the anomaly probability threshold, then only local calculation of the anomaly region is initiated, including: The distribution network topology is divided into multiple layers according to voltage level; For each layer, it is divided into multiple sub-maps according to geographical region or power supply zone; each layer and each sub-map is calculated independently and in parallel. When the probability of anomalies is not greater than the threshold, global computation is initiated, including independent parallel computation of all subgraphs. When the probability of anomalies exceeds the threshold, only local calculations of the anomaly region are initiated, including only local calculations of the subgraph of the anomaly region.

3. The method for providing a panoramic dynamic topology full-domain calculation service for a distribution network according to claim 1, characterized in that: The inputs to the deep neural network include the number of times the device status has changed in the past hour and the voltage fluctuation amplitude.

4. A panoramic dynamic topology full-domain computing service system for power distribution networks, characterized in that: include: The data conversion module is configured to receive data from the measurement center; The service integration module is configured to use an event-driven mechanism to pull event-related data and push the calculation results to the measurement center. The global calculation module is configured to input the event-related data into a deep neural network to obtain the topological anomaly probability; compare the topological anomaly probability with a set anomaly probability threshold; if it is not greater than the anomaly probability threshold, then global calculation is started; if it is greater than the anomaly probability threshold, then only local calculation of the anomaly region is started.

5. The distribution network panoramic dynamic topology full-domain calculation service system according to claim 4, characterized in that: In the global calculation module, the comparison of the topological anomaly probability with a set anomaly probability threshold is performed. If the probability is not greater than the threshold, global calculation is initiated; if it is greater than the threshold, local calculation of the anomaly region is initiated only. This specifically includes: The distribution network topology is divided into multiple layers according to voltage level; For each layer, it is divided into multiple sub-maps according to geographical region or power supply zone; each layer and each sub-map is calculated independently and in parallel. When the probability of anomalies is not greater than the threshold, global computation is initiated, including independent parallel computation of all subgraphs. When the probability of anomalies exceeds the threshold, only local calculations of the anomaly region are initiated, including only local calculations of the subgraph of the anomaly region.

6. The distribution network panoramic dynamic topology full-domain calculation service system according to claim 4, characterized in that: The inputs to the deep neural network include the number of times the device status has changed in the past hour and the voltage fluctuation amplitude.

7. A method for integrating a global computing service system with a real-time measurement center, characterized in that: include: A full-domain computing service system is deployed on the intranet of the real-time measurement center, and the hardware and software environments are configured; the full-domain computing service system is a full-domain computing service system for a distribution network panoramic dynamic topology as described in any one of claims 1 to 6; Register atomic-level services of the full-domain computing service system through the real-time measurement center's digital portal to obtain the service call address; Configure real-time measurement center message push rules to push data to the global computing service system based on the service call address; Receive computation results from the global computing service system.

8. An electronic device, characterized in that, The electronic device includes: At least one processor; and a memory communicatively connected to the at least one processor; The memory stores a computer program that can be executed by the at least one processor, which is then executed by the at least one processor to enable the at least one processor to perform the power distribution network panoramic dynamic topology full-domain calculation service method according to any one of claims 1-3.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that cause a processor to execute the power distribution network panoramic dynamic topology full-domain calculation service method as described in any one of claims 1-3.

10. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the power distribution network panoramic dynamic topology full-domain calculation service method as described in any one of claims 1-3.

11. An electronic device, characterized in that, The electronic device includes: At least one processor; and a memory communicatively connected to the at least one processor; The memory stores a computer program that can be executed by the at least one processor, which is then executed by the at least one processor to enable the at least one processor to perform the integration method of the global computing service system and the real-time measurement center as described in claim 7.

12. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that, when executed by a processor, implement the integration method of the global computing service system and the real-time measurement center as described in any one of claims 7.

13. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the integration method of the global computing service system and the real-time measurement center as described in any one of claims 7.

Images