A smart grid system and dispatching method based on cloud-edge fusion architecture

Abstract

The present invention discloses a smart grid system and dispatching method based on a cloud-edge fusion architecture, including a cloud system module, an edge system module, an edge gateway module and corresponding functional designs, as well as a smart grid dispatching method based on a cloud-edge fusion architecture, wherein the cloud constructs a cloud system for grid-wide services based on computing resources provided by cloud platform technology; the edge includes two types of edge systems, an edge cluster and an edge gateway; the edge cluster constructs a grid dispatching and operation monitoring system with elastic expansion capabilities based on a dynamic container strategy, supporting elastic access and real-time monitoring of controlled objects such as traditional plants, centralized new energy, and massive emerging grid-connected entities; the edge gateway exchanges data with controlled objects such as traditional plants, centralized new energy, and various emerging grid-connected entities in a standardized manner. As a smart grid system and dispatching method based on a cloud-edge fusion architecture, the present invention can be widely used in the field of smart grid technology.

Description

Technical Field

[0001] The present invention relates to the field of smart grid technology, and in particular to a smart grid system and scheduling method based on a cloud-edge fusion architecture. Background Art

[0002] With the construction of new power systems, distributed new energy, microgrids, and power market reforms are developing rapidly. Massive emerging grid-connected entities such as large-scale distributed energy, virtual power plants, and integrated parks will participate in the operation and regulation of the power grid. The grid structure and system characteristics will change from being planned and centralized to open, shared, and intelligently interactive. This has brought new challenges to the operation control and management of the power grid and the operation of the power market, and has put forward higher requirements for the platform openness, system capacity, access capability, and intelligence level of the power grid dispatching and operation monitoring system. The current grid adopts a hierarchical control system with obvious business boundaries and weak coordination capabilities, and lacks a top-level unified monitoring, prediction, optimization, and analysis support system for the overall situation. Support services have failed to effectively solve the support needs of interactive operation of sources, grids, loads and storage in the context of the development of new power systems. In addition, the controlled objects of traditional power grid dispatching, operation and monitoring systems are all limited and tightly coupled entities (power plants, substations, etc.). In the access scenario of massive emerging grid-connected entities, traditional power grid dispatching and operation technical systems are difficult to effectively support the massive access of emerging grid-connected entities, service elastic expansion, intelligent response decision-making and other requirements. However, the common way to solve the above problems at this stage is to improve the system's access capabilities and business capabilities by expanding operating resources on the basis of traditional power grid dispatching, operation and monitoring systems. However, this method is becoming increasingly difficult to support massive object access and complex and changing business needs. Summary of the Invention

[0003] In order to solve the above technical problems, the purpose of the present invention is to provide a smart grid system and dispatching method based on a cloud-edge fusion architecture, and propose a "cloud + edge" two-level integrated smart grid dispatching and operation platform to provide dispatching and operation support for various businesses such as comprehensive monitoring, analysis and prediction, collaborative control, and market transactions of new power systems.

[0004] The first technical solution adopted by the present invention is: a smart grid system based on a cloud-edge fusion architecture, comprising:

[0005] The cloud system module is used to connect network, provincial, and local edge clusters, collect network-wide operation data, implement global optimization analysis, network-wide scheduling decisions, and panoramic information display, supporting network-wide sharing and service sharing;

[0006] The edge system module is used for flexible access, real-time monitoring, and collaborative control of controlled objects such as traditional plants and stations, centralized new energy sites, and a large number of emerging market entities within the scope of regulation;

[0007] The edge gateway module is responsible for collecting local power grid operation data and marketing data, and sending them to the edge cluster to support grid operation mode switching, frequency regulation and peak regulation and other command operations, as well as support grid dispatching operation and power market operation.

[0008] The second technical solution adopted by the present invention is: a smart grid scheduling method based on cloud-edge fusion architecture, comprising:

[0009] Construct a smart grid system based on the cloud-edge fusion architecture as described in claim 1;

[0010] Collect demand data of various control units of source, grid, load and storage under the smart grid system;

[0011] The various control units send computing task requests to the edge system module and to the cloud system module through the edge system module according to local computing capabilities;

[0012] In the edge system module, under the planned rolling control mode, the edge cluster further decomposes the total edge cluster regulation demand or market bidding results issued by the cloud system into rolling plans for various control units of source, grid, load and storage under its jurisdiction based on the local grid model and operation data, and sends them to the regulated objects for execution;

[0013] In the cloud system module, under the real-time control mode, the edge cluster receives the real-time active power regulation requirements issued by the cloud unified frequency modulation, decomposes them into real-time control instructions for various control units of source, grid, load and storage under its jurisdiction, and sends them to the control objects for execution.

[0014] Furthermore, the edge system module provides flexible access to controlled objects such as traditional plants, centralized new energy sites, and a large number of emerging market entities within the scope of regulation and management, which specifically includes:

[0015] The edge cluster provides on-demand resource allocation capabilities based on container technology, enabling elastic expansion of edge cluster business applications. Based on the standard interfaces provided by the distributed service bus and message queue, it enables "plug and play" of application modules.

[0016] (1) Edge cluster container operating environment;

[0017] Security Zone I and the Security Zone I secure access zone use a customized container operating environment that meets the security protection requirements of the power monitoring system control area. Security Zone II and Security Zone III use general container service technologies to build a container operating environment that is compatible with cloud system application download and deployment.

[0018] (2) Data storage for edge cluster application modules;

[0019] Data is stored in the storage volume mounted on the container and managed by the application module itself. The format can be customized. Data is stored in the edge cluster database and storage service mounted on the container.

[0020] (3) Data interaction between edge cluster application modules;

[0021] Message transmission between application modules supported by different container groups is supported through a message bus. Application modules supported by the same container group have strong functional dependencies and can exchange data through shared memory, database, and storage.

[0022] (4) Service calls of edge cluster application modules;

[0023] A distributed service proxy is configured in the application module container group. The application module provides functional services to the outside world through the distributed service proxy. Each application module calls related services through the distributed service proxy on demand.

[0024] Furthermore, the edge system module further includes:

[0025] Build elastically scalable front-end collection based on container and microservice technologies to support dynamic and elastic access to controlled objects such as traditional power plants, centralized new energy stations, and a large number of emerging grid-connected entities;

[0026] Based on container and message bus technologies, the parallel processing mechanism for front-end data supports monitoring and control of large-scale access objects.

[0027] Based on the cloud-edge integrated collaborative control system, it supports real-time collaborative control of controlled objects such as traditional plants and stations, centralized new energy stations, and massive emerging grid-connected entities.

[0028] Furthermore, the dynamic and flexible access of the edge system module to the controlled object also includes designing a data preprocessing method for a class of dynamic containers, specifically:

[0029] Collect edge cluster pre-data, including remote control data S1, manual operation data S2, real-time statistical data S3 and related historical data S4;

[0030] At the beginning of each time slot, data preprocessing is performed on remote control data S1, manual operation data S2, real-time statistical data S3 and related historical data S4, abnormal values ​​are proposed, and related missing values ​​are filled to obtain the corresponding data volume value.

[0031] The speed of data generation is expressed as [V1, V2, V3, V4]. When the data generation speed is lower than the processing speed, that is, At this time, the container capacity R iNo processing is required. When the data generation rate is higher than the processing rate, the capacity of the container is increased;

[0032] We plan to jointly optimize the computing resources [C1, C2, C3, C4] allocated to the four types of data in parallel on the edge cluster, as well as the corresponding container capacity [R1, R2, R3, R4] allocated to the four types of data, to minimize the task preprocessing latency. The optimization problem is expressed as:

[0033]

[0034]

[0035]

[0036] In the above formula, δ represents the duration of the task processing time slot, constraint (1) indicates that the container capacity allocated to each type of data must be greater than the rate at which the task is generated minus the data processing rate to ensure that the relevant data cannot be discarded, and constraint (2) represents the capacity size limit and computing resource limit;

[0037] The optimization problem is optimized and solved through genetic algorithms. After the dynamic container calculation is completed, the calculated results are automatically calculated based on the real-time operation data and sent to the message queue for use by other businesses.

[0038] Furthermore, the edge cluster supports two ways to perform operational predictions:

[0039] The first method is to remotely call the intelligent prediction module of the cloud-edge fusion cloud system to perform load forecasting and new energy forecasting through a two-level cloud-edge fusion method and obtain the forecast results directly from the cloud system;

[0040] The second method is to carry out model training for bus load forecasting, new energy power forecasting, system load forecasting and distributed energy forecasting within the jurisdiction of the edge cluster in the cloud system based on cloud artificial intelligence technology, and download the trained model to the edge cluster through the cloud-edge two-level fusion method. After the edge cluster is deployed, prediction is carried out based on the operating data of the edge cluster to obtain the prediction results.

[0041] Furthermore, the cloud system adopts a "cloud native" solution, based on a container operating environment, to support application "microservices" and elastic system expansion and upgrades; based on the open and standard platform interface provided by the service bus and message queue, it builds a "plug-and-play" application module to support the joint construction of cloud applications across the entire network, specifically:

[0042] (1) Application module deployment in cloud systems;

[0043] Deployment is based on the cloud platform's container runtime environment, which provides services such as container cluster management, internal communication, scheduling, start / stop, and horizontal scaling.

[0044] (2) Communication between cloud system application modules;

[0045] Communication is based on the virtual IP provided by the cloud platform VPC network. Application module services composed of multiple container groups can be mounted under a balancing IP to achieve load balancing of the application module's external service traffic.

[0046] (3) Data storage of cloud system application modules;

[0047] Data is stored in the storage volume mounted on the container and managed by the application module itself. The format can be customized. Data is stored in the cloud platform database and storage service mounted on the container.

[0048] (4) Data interaction between cloud system application modules;

[0049] Message transmission between application modules supported by different container groups is supported through a message bus. Application modules supported by the same container group have strong functional dependencies and can exchange data through shared memory, databases, and storage.

[0050] (5) Service calls of cloud system application modules;

[0051] It supports encapsulating all or a certain sub-functional logic of the application module and mounting it on the cloud service bus to provide functional services to the outside world. Each application module calls the services mounted on the cloud service bus through the cloud service bus on demand.

[0052] The beneficial effects of the method and system of the present invention are as follows: the present invention proposes a "cloud + edge" two-level integrated smart grid dispatching and operation platform to provide dispatching and operation support for various businesses such as comprehensive monitoring of source, grid, load and storage of new power systems, analysis and prediction, collaborative control, market transactions, etc., by adopting advanced computer technologies such as large-scale computing and storage of cloud platforms to build a cloud system, as a unified "brain" shared by the entire power grid, collecting the operation data of the entire power grid, carrying out optimization calculation and analysis of the entire power grid, and supporting the operation of the entire power grid system and power market operation. By adopting advanced computer technologies such as containers, microservices, and message queues, an elastically scalable edge cluster is constructed, which serves as a real-time operation control system for power grid dispatching agencies at the network, provincial, and local levels, supporting the elastic access and real-time monitoring of controlled objects such as traditional power plants, centralized new energy, and massive emerging grid-connected entities by power grid dispatching agencies at all levels. By adopting computer technologies such as service agents to build a cloud-edge integrated data interaction system, a flat, standardized, and highly secure cloud-edge interaction channel is established at the cloud and edge levels to support the efficient and collaborative operation of cloud-edge business integrated interaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 This is a structural block diagram of a smart grid system based on a cloud-edge fusion architecture of the present invention;

[0054] Figure 2 This is a step flow chart of a smart grid scheduling method based on a cloud-edge fusion architecture of the present invention. DETAILED DESCRIPTION

[0055] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. The step numbers in the following embodiments are provided for ease of description only and do not limit the order of the steps. The order of execution of the steps in the embodiments can be adaptively adjusted based on the understanding of those skilled in the art.

[0056] Reference Figure 1 and Figure 2 The present invention provides a smart grid system and scheduling method based on a cloud-edge fusion architecture, which adopts a "cloud + edge" system architecture for platform construction, including a cloud system and an edge system. The edge system is composed of several edge computing nodes, including two types of edge systems: edge clusters and edge gateways, including:

[0057] The cloud system is built on a unified network based on cloud platform technology, adopting a hierarchical design. It consists of three layers from bottom to top, including the public resource layer, the business platform layer, and the business application layer.

[0058] Specifically, the cloud system is built based on the cloud platform's IaaS, PaaS, and DaaS cloud service components. As the intelligent brain of the power grid, it connects network, provincial, and local edge clusters, aggregates network-wide operational data, and enables global optimization analysis, network-wide dispatching decisions, and panoramic information display, supporting network-wide sharing and service sharing. The cloud system consists of three parts: the public resource layer, the business platform layer, and the business application layer. The public resource layer is provided with basic operating resources by the cloud platform. The business platform layer provides standardized and unified public services for upper-level business applications. The business application layer includes advanced applications such as power grid operation monitoring, computational analysis, operation prediction, operation optimization, evaluation analysis, and intelligent interaction. The cloud system interacts with edge clusters for data and application collaboration.

[0059] (1) In terms of technology, the cloud system adopts a "cloud native" solution, based on the container operating environment, to support the "microserviceization" of applications and the elastic expansion and upgrade of the system; based on the open and standard platform interface provided by the service bus and message queue, it builds a "plug and play" application module to support the joint construction of cloud applications across the entire network.

[0060] 1) Application module deployment in the cloud system: Deployment is based on the cloud platform's container runtime environment, and the cloud platform implements services such as container cluster management, internal communication, scheduling, start and stop, and horizontal scaling;

[0061] 2) Communication between cloud system application modules: Communication is based on the virtual IP provided by the cloud platform's VPC network. Application module services composed of multiple container groups can be mounted under a single balancing IP address to achieve load balancing of the application module's external service traffic.

[0062] 3) Data storage of cloud system application modules: Data is stored in the storage volume mounted on the container and managed by the application module itself. The format can be customized. Data is stored in the cloud platform database and storage service mounted on the container.

[0063] 4) Data exchange between cloud system application modules: Messages between application modules supported by different container groups are transmitted via a message bus. Application modules supported by the same container group have strong functional dependencies and can exchange data through shared memory, databases, and storage.

[0064] 5) Service invocation of cloud system application modules: This allows the encapsulation of all or a specific sub-function logic of an application module and its mounting on the cloud service bus, providing external functional services. Each application module can call the services mounted on the cloud service bus on demand.

[0065] (2) In terms of business, the cloud system is based on a panoramic large model of the entire network of 10kV and above (within the wall) including emerging grid-connected entities, realizing comprehensive perception, intelligent prediction and global optimization of the source, network, load and storage of the large power grid. Based on the cloud-edge intelligent operation and maintenance system, it realizes the intensive operation and maintenance of the secondary system, and improves the "observability, measurability and controllability" level of the new power system through the cloud system.

[0066] 1) Currently, power grid models are constructed according to the power grid regulation and management scope at the network, provincial, and local levels, serving as the basis for business applications such as power grid operation and analysis at this level. The cloud system breaks through the hierarchical power grid model and unifies the source, grid, load, and storage models of the entire network to build a panoramic power grid model covering voltage levels of 10kV and above, which is used to support services such as comprehensive perception of the source, grid, load, and storage of the entire grid, intelligent prediction, and global optimization. Model splicing consists of two parts: the splicing of traditional power grid models and the splicing of emerging grid-connected subject models. For the splicing of traditional power grid models, in order to improve the efficiency of model splicing, a unified model coding method is used to standardize the model, and a cloud-edge two-level model splicing method is used to implement the splicing of traditional power grid models. That is, the edge clusters at the network, provincial, and local levels complete the generation of the traditional power grid model at this level according to the unified model coding and upload it to the cloud system. The cloud system completes the splicing of the traditional power grid model of the entire network based on the traditional power grid models and model boundaries at all levels. For the emerging grid-connected entity model, the network, provincial and local edge clusters generate the emerging grid-connected entity aggregation control unit according to the emerging grid-connected entity grid-connected node-user association relationship and send it to the cloud system. Based on the emerging grid-connected entity grid-connected node-user association relationship, the aggregation control unit and the measurement data, the cloud system equates the aggregation control unit to a virtual load / unit and connects it to the traditional power grid model of the entire network, eventually forming a panoramic model of the source, grid, load and storage of the entire network of 10kV and above for use by the cloud system business applications.

[0067] 2) The cloud system is based on a panoramic model of the power grid. Through a two-level cloud-edge integration approach, the cloud system obtains operational data on power sources, grids, loads, and storage at voltage levels of 10kV and above from edge clusters at the network, provincial, and local levels. This enables the cloud system to achieve comprehensive network-wide perception of voltage levels of 10kV and above. This overcomes the current problem of power grid monitoring being unable to simultaneously cover voltage levels of 10kV and above, making it difficult to support the coordinated operation of power sources, grids, loads, and storage in large power grids. To achieve comprehensive network-wide perception, the cloud system supports this capability by building a cloud-based data center and cloud-based smart grid operation monitoring.

[0068] a) The cloud data center is supported by cloud services such as RDS, OSS, and OTS provided by the cloud platform. By building a data access proxy application, it supports the centralization of data or files from different sources, formats, and qualities, such as network, provincial, and local edge clusters and other external systems, into a unified data resource pool, enabling multi-source data access to the cloud system and providing data service support for the comprehensive perception of the cloud system. In addition, compared with conventional data query methods, the data center adopts data resource indexing technology to achieve rapid access and call of various types of data across the entire network in the cloud system based on data resource indexing, thereby improving data access efficiency.

[0069] b) Cloud-based smart grid operation monitoring realizes operation monitoring of the source, grid, load and storage of the entire network at voltage levels of 10kV and above based on the operation data of the entire network and a panoramic model. It provides a centralized display of key information of the entire network in different dimensions, such as key information of the entire network operation, key indicators of the entire network operation, key alarms, key information charts, etc., and supports classified monitoring according to the type of monitored objects, including network-wide power generation and transmission operation monitoring, network-wide market entity operation monitoring, network-wide energy storage operation monitoring, network-wide new energy operation monitoring, and new power system key indicator monitoring.

[0070] 3) Cloud-based intelligent prediction uses artificial intelligence technology to train prediction models and provide short-term and ultra-short-term prediction services to the entire network. Leveraging the computing resources of the cloud platform's CPU and GPU, the cloud platform's DaaS big data platform components, and the cloud data center's massive network-wide operational data, the cloud system provides computing resources, a computing framework, and data resource services for training the cloud system's intelligent prediction models. This supports a range of intelligent prediction tasks, including data acquisition, feature engineering, prediction model construction, prediction model training, prediction model release, prediction model fusion, prediction model packaging, prediction result generation, and visualization. Cloud-based intelligent prediction supports network-wide prediction model training and forecasting, including network-wide 110kV and above busbar load forecasting, network-wide renewable energy power forecasting, and network-wide regional system load forecasting. The forecast results are used to support global optimization of the entire network. Furthermore, it supports model training for busbar load forecasting, renewable energy power forecasting, system load forecasting, and distributed energy forecasting within the cloud system's edge cluster jurisdiction based on cloud-based artificial intelligence technology. The trained models are then deployed to edge clusters through a two-level cloud-edge fusion approach, supporting localized predictions within edge clusters.

[0071] 4) The cloud system leverages the abundant computing resources of the cloud platform to build ultra-large-scale global optimization capabilities for the entire 10kV network and above. Based on a comprehensive grid model, grid-wide source, grid, load, and storage operational data, and grid-wide performance forecasts, this global optimization performs 5-minute or 15-minute rolling optimization of active power across the entire network. This generates regulation demands for both market and non-market scenarios, and transmits these demands to edge clusters for execution through a two-level cloud-edge integration approach. The cloud system's global optimization encompasses both planned rolling control and real-time control.

[0072] 5) Cloud-based intelligent operation and maintenance builds an efficient operation and maintenance system for the cloud system and network, provincial, and local edge clusters, including the entire network's main station system, to support comprehensive monitoring of the operating status of the entire network's main station system, remote operation and maintenance, and comprehensive analysis. The cloud system establishes an operation and maintenance data interface with the cloud platform base, virtual machine services, container management services, and various cloud application modules to obtain operation logs and alarm information, etc., to monitor the cloud platform resources, virtual machines, container groups, containers, and other different object information and operating status used by the cloud system; the cloud system collects different monitoring object information, operating status, etc., such as the edge cluster software and hardware ledgers, container groups, containers, and service components of the entire network through a two-level cloud-edge fusion method, to achieve centralized operation and maintenance monitoring of the entire network's edge clusters.

[0073] The edge cluster is independently built by grid dispatching organizations at the network, provincial, and local levels based on container technologies. It adopts a hierarchical design and consists of three layers from bottom to top: the public resource layer, the business platform layer, and the business application layer.

[0074] Specifically, edge clusters are real-time grid monitoring systems deployed at the network, provincial, and local dispatching agencies to adapt to the new power system. They enable flexible access, real-time monitoring, and collaborative control of controlled objects such as traditional power plants, centralized new energy sites, and a large number of emerging market players within the dispatching scope. The edge cluster consists of three parts: the public resource layer, the business platform layer, and the business application layer. The public resource layer provides independent basic software and hardware resources for the edge cluster, the business platform layer provides standardized and unified public services for business applications, and the business application layer includes advanced applications such as operation monitoring, operation prediction, operation control, and computational analysis. The edge cluster interacts with cloud systems upstream and edge gateways downstream for data exchange and application collaboration.

[0075] (1) In terms of technology, the edge cluster provides on-demand allocation capabilities of operating resources based on container technology, realizes the elastic expansion of edge cluster business applications, and realizes "plug and play" of application modules based on the standard interfaces provided by distributed service buses and message queues.

[0076] 1) Edge cluster container operating environment: Security Zone I and Security Zone I security access zone use a customized container operating environment that meets the security protection requirements of the power monitoring system control area; Security Zone II and Security Zone III build a container operating environment based on general container service technology, which is compatible with cloud system application download and deployment.

[0077] 2) Data storage of edge cluster application modules: Data is stored in the storage volume mounted on the container and managed by the application module itself. The format can be customized. Data is stored in the edge cluster database and storage service mounted on the container.

[0078] 3) Data exchange between edge cluster application modules: Messages between application modules supported by different container groups are transmitted via a message bus. Application modules supported by the same container group have strong functional dependencies and can exchange data through shared memory, databases, and storage.

[0079] 4) Service invocation of edge cluster application modules: A distributed service proxy is configured within the application module container group. The application module provides functional services externally through the distributed service proxy. Each application module invokes related services through the distributed service proxy on demand.

[0080] (2) In terms of business, the edge cluster is built with elastically scalable front-end collection based on technologies such as containers and microservices, supporting the dynamic elastic access of controlled objects such as traditional power plants, centralized new energy sites and massive emerging grid-connected entities; based on technologies such as containers and message buses, it builds a front-end data parallel processing mechanism to support large-scale access object monitoring and control; based on the cloud-edge fusion collaborative control system, it supports real-time collaborative control of controlled objects such as traditional power plants, centralized new energy sites and massive emerging grid-connected entities.

[0081] 1) The edge cluster utilizes technologies such as containers, microservices, and container orchestration to split and containerize front-end collection. Edge cluster front-end collection is split into data collection microservices, data aggregation microservices, and channel message storage microservices based on business logic, and runs in containers. Front-end collection supports dynamic access to a large number of collection channels through online elastic scaling of microservice container instances. Each microservice container instance for front-end collection supports customizing the number of plant and station channels to be taken over, with a single container instance processing messages from multiple plant and station channels.

[0082] 2) A data preprocessing method for a class of dynamic containers was designed. The edge cluster uses this method to process the data received in real time to achieve plant operation monitoring and operation control. According to the business logic, the data pre-processed by the edge cluster includes: remote control data S1, manual operation data S2, real-time statistical data S3 and related historical data S4. The edge cluster needs to process the above data in parallel. However, the data volume of different types of data is completely different, and the resources occupied are also completely different. Differentiated container capacity is dynamically allocated for different data to achieve a match between data and computing resources. This can minimize the data preprocessing delay while ensuring the minimum data processing resource overhead. The specific operations are as follows:

[0083] a) At the beginning of each time slot, data preprocessing is performed on remote control data S1, manual operation data S2, real-time statistical data S3, and related historical data S4 to identify abnormal values ​​and fill in related missing values, ultimately obtaining the corresponding data volume value.

[0084] b) The speed of data generation of the four types is expressed as: [V1, V2, V3, V4]. When the data generation speed is lower than the processing speed, At this time, the container capacity R i Only the minimum requirements need to be met. Conversely, the capacity of the container needs to be increased to avoid data loss.

[0085] c) We plan to jointly optimize the computing resources [C1, C2, C3, C4] allocated to the four types of data in parallel on the edge cluster, as well as the corresponding container capacities [R1, R2, R3, R4] allocated to the four types of data, to minimize the task preprocessing latency. The optimization problem is expressed as:

[0086]

[0087]

[0088]

[0089] In the above formula, δ represents the duration of the task processing time slot. Constraint (1) indicates that the container capacity allocated to each type of data must be greater than the rate at which the task is generated minus the data processing rate to ensure that the relevant data cannot be discarded. Constraint (2) represents the capacity size limit and the computing resource limit.

[0090] d) This optimization problem is a typical mixed integer programming problem, and a genetic algorithm is proposed to optimize and solve it.

[0091] e) After dynamic container calculations are complete, the results are automatically calculated based on real-time operational data and sent to a message queue for use by other businesses. In particular, for historical data, a combination of microservices and dynamic containers is used to improve the collection and storage of historical data and historical alarm information.

[0092] 3) The edge cluster supports two ways to carry out operation prediction. The first way is to remotely call the cloud-edge fusion cloud system intelligent prediction module to carry out load forecasting and new energy forecasting through the cloud-edge two-level fusion method and obtain the prediction results directly from the cloud system. The second way is to carry out model training of bus load forecasting, new energy power forecasting, system load forecasting and distributed energy forecasting within the jurisdiction of the edge cluster in the cloud system based on cloud artificial intelligence technology, and download the trained model to the edge cluster through the cloud-edge two-level fusion method. After the edge cluster is deployed, prediction is carried out based on the operation data of the edge cluster to obtain prediction results.

[0093] 4) Based on cloud-edge collaborative control, under a planned rolling control mode, the edge cluster, based on the total edge cluster regulation demand or market bidding results issued by the cloud system, further decomposes the rolling plan for the various control units of the source, grid, load, and storage under its jurisdiction based on the local grid model and operating data, and issues it to the regulated objects for execution. Under a real-time control mode, the edge cluster receives the real-time active power regulation demand issued by the cloud unified frequency regulation, decomposes it into real-time control instructions for the various control units of the source, grid, load, and storage under its jurisdiction, and issues them to the regulated objects for execution. In addition to conventional automatic control such as AGC and AVC, edge cluster control takes into account the participation of the aggregated control units of massive emerging grid-connected entities in grid operation control. By building an adjustable load control function, real-time regulation instructions are issued from the dispatching end to the adjustable load / massive emerging grid-connected entity aggregated control units, realizing automatic regulation of the active power of the adjustable load resources / massive emerging grid-connected entity aggregated control units within the regulation range. This is a closed-loop control process that meets the real-time security balance and frequency deviation regulation requirements of the grid.

[0094] In summary, it is proposed to use a dynamic container strategy to build an elastically scalable edge cluster as a real-time operation control system for power grid dispatching agencies at the network, provincial, and local levels, supporting the elastic access and real-time monitoring of controlled objects such as traditional power plants, centralized new energy, and massive emerging grid-connected entities by power grid dispatching agencies at all levels.

[0095] The edge gateway is a data interaction gateway device deployed at factories, stations, and market entities. It connects to the edge cluster upwards and the controlled objects downwards. The edge gateway adopts a unified technical architecture and provides local decision-making, local autonomy, and external services.

[0096] The cloud system and edge cluster conduct cloud-edge two-level business integration interaction through the cloud-edge interaction cloud service and the cloud-edge interaction edge cluster agent;

[0097] Specifically, edge gateways are deployed at grid connection points of emerging grid-connected entities such as substations, distribution stations / stations / towers, and distributed energy resources. As data exchange gateway devices for data export and command reception, they are responsible for collecting local grid operation data and marketing data, and sending them to the edge cluster to which they belong. This supports grid operation mode switching, frequency regulation / peak regulation, and other command operations, as well as grid dispatching and power market operations.

[0098] Build a standardized cloud-edge interaction channel to achieve rapid push of cloud data and services and high-concurrency upload of edge cluster data; build a cloud-edge fusion business model for new power system operation monitoring, prediction, collaborative control, intensive operation and maintenance, etc.

[0099] To support efficient and secure interaction between cloud systems and edge clusters at the network, provincial, and local levels, we build cloud-edge interactive cloud services in the cloud system, including cloud base services, cloud product warehouse services, cloud application interaction services, and cloud monitoring center services. We also build four types of interactive agents in the edge clusters using proxy technology: base interaction agents, product interaction agents, application interaction agents, and monitoring interaction agents.

[0100] The cloud-edge interaction cloud service and the edge cluster interaction agent jointly complete the configuration and establishment of four types of dedicated interaction channels, including base installation, product download and upload, cloud-edge application interaction, and cloud-edge operation and maintenance monitoring interaction. They serve as unified channel exits for various types of cloud-edge interaction services, realize the secure transmission of cloud-edge interaction services, and provide on-demand interaction services for cloud-edge interaction business applications.

[0101] Cloud-edge business integration and interaction include cloud-edge integration operation monitoring, cloud-edge integration prediction, cloud-edge integration collaborative control, cloud-edge blockchain services and cloud-edge integration intelligent operation and maintenance.

[0102] (1) Cloud-edge fusion operation monitoring: The cloud system collects the operation data of the entire network from the network, provincial and local edge clusters through the cloud-edge application interaction channel, and realizes real-time monitoring of 10kV and above traditional power plants, centralized new energy and emerging grid-connected entities in the entire network based on the panoramic model; the network, provincial and local edge clusters realize real-time monitoring of traditional power plants, centralized new energy and emerging grid-connected entities within the control range.

[0103] (2) Cloud-edge fusion prediction: The cloud system provides an intensive artificial intelligence training environment, and completes the prediction model training based on the panoramic model and massive operating data, and provides prediction results or prediction models to the edge cluster; the edge cluster obtains the required prediction results from the cloud system through the cloud-edge application interaction channel, or obtains the prediction model from the cloud system through the cloud-edge product interaction channel, deploys the prediction model locally in a containerized manner, and generates localized prediction results.

[0104] (3) Cloud-edge fusion collaborative control: The cloud system realizes rolling optimization of the source, grid, load and storage of the entire network and generates an adjustment plan based on the panoramic model of the power grid, the operation data of the source, grid, load and storage of the entire network and the prediction results of the entire network. The generated adjustment plan is sent to the network, provincial and local edge clusters through the cloud-edge application interaction channel; the network, provincial and local edge clusters decompose the adjustment plan sent by the cloud system and issue control instructions to various control units of the source, grid, load and storage, so as to realize the operation control of the source, grid, load and storage under cloud-edge fusion collaboration.

[0105] (4) Cloud-edge blockchain service: The cloud system provides standardized blockchain applications to monitor the operation of blockchain applications across the entire network. The edge cluster manages the blockchain and sends the blockchain operation status to the cloud. The edge gateway and the aggregated emerging grid-connected entities deploy standardized blockchain applications to automatically upload key data from the terminals of the aggregated emerging grid-connected entities to the blockchain. Through the cloud-edge blockchain service, the operation status and control response of the emerging grid-connected entities across the entire network can be fully monitored and recorded with credibility.

[0106] (5) Cloud-edge integrated intelligent operation and maintenance: The cloud system collects information such as the status and alarms of the edge clusters of the entire network through cloud-edge operation and maintenance monitoring interaction, and realizes comprehensive monitoring and analysis of the operation of the edge clusters of the entire network; the edge cluster realizes secondary operation monitoring of the cluster and edge gateway, and sends configurations through standardized remote operation and maintenance interfaces to realize intensive remote maintenance of the edge gateway.

[0107] To summarize, it is proposed to build a cloud-edge fusion interaction system using computer technologies such as service proxy, establish a flat, standardized, and highly secure cloud-edge interaction channel at the cloud and edge levels, and support the efficient and collaborative operation of cloud-edge business fusion interaction at the two levels.

[0108] The contents of the above method embodiments are all applicable to the present system embodiments. The functions specifically implemented by the present system embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0109] The above is a specific description of the preferred implementation of the present invention, but the invention is not limited to the embodiments. Those skilled in the art can make various equivalent modifications or substitutions without violating the spirit of the present invention. These equivalent modifications or substitutions are all included in the scope defined by the claims of this application.

Claims

1. A smart grid system based on cloud-edge fusion architecture, characterized by: include: The cloud system module is used to connect network, provincial, and local edge clusters, collect network-wide operation data, implement global optimization analysis, network-wide scheduling decisions, and panoramic information display, supporting network-wide sharing and service sharing; The edge system module is used for flexible access, real-time monitoring and collaborative control of traditional power plants, centralized new energy stations and a large number of emerging market players within the scope of regulation; The edge gateway module is responsible for collecting local power grid operation data and marketing data, and sending them to the edge cluster to support power grid operation mode switching, frequency and peak regulation, and support power grid dispatching and power market operations; The edge system module provides flexible access to traditional plants, centralized new energy sites, and a large number of emerging market players within the scope of management, specifically including: The edge cluster provides on-demand resource allocation capabilities based on container technology, enabling elastic expansion of edge cluster business applications. Based on the standard interfaces provided by the distributed service bus and message queue, it enables "plug and play" of application modules. (1) Edge cluster container operating environment; Security Zone I and the Security Zone I secure access zone use a customized container operating environment that meets the security protection requirements of the power monitoring system control area. Security Zone II and Security Zone III use general container service technologies to build a container operating environment that is compatible with cloud system application download and deployment. (2) Data storage for edge cluster application modules; Data is stored in the storage volume mounted on the container and managed by the application module itself. The format can be customized. Data is stored in the edge cluster database and storage service mounted on the container. (3) Data interaction between edge cluster application modules; Message transmission between application modules supported by different container groups is supported through a message bus. Application modules supported by the same container group have strong functional dependencies and can exchange data through shared memory, database, and storage. (4) Service calls of edge cluster application modules; A distributed service proxy is configured in the application module container group. The application module provides functional services to the outside world through the distributed service proxy. Each application module calls related services through the distributed service proxy on demand.

2. A smart grid scheduling method based on cloud-edge fusion architecture, characterized in that: include: Constructing a smart grid system based on a cloud-edge fusion architecture as described in claim 1; Collect demand data of various control units of source, grid, load and storage under the smart grid system; The various control units send computing task requests to the edge system module and to the cloud system module through the edge system module according to local computing capabilities; In the edge system module, under the planned rolling control mode, the edge cluster further decomposes the total edge cluster regulation demand or market bidding results issued by the cloud system into rolling plans for various control units of source, grid, load and storage under its jurisdiction based on the local grid model and operation data, and sends them to the regulated objects for execution; In the cloud system module, under the real-time control mode, the edge cluster receives the real-time active power regulation requirements issued by the cloud unified frequency modulation, decomposes them into real-time control instructions for various control units of source, grid, load and storage under its jurisdiction, and sends them to the control objects for execution.

3. The smart grid scheduling method based on cloud-edge fusion architecture according to claim 2 is characterized in that: The edge system module further includes: Container and microservices-based technologies enable flexible and scalable front-end data collection, supporting dynamic and flexible access to traditional power plants, centralized new energy stations, and a large number of emerging grid-connected entities. Build a front-end data parallel processing mechanism based on container and message bus technologies to support large-scale access object monitoring and control; Based on the cloud-edge integrated collaborative control system, it supports real-time collaborative control of traditional plants, centralized new energy stations and massive emerging grid-connected entities.

4. The smart grid scheduling method based on cloud-edge fusion architecture according to claim 3 is characterized in that: The edge system module's dynamic and flexible access to the controlled object also includes designing a data preprocessing method for a class of dynamic containers, specifically: Collect edge cluster pre-data, including remote control data S1, manual operation data S2, real-time statistical data S3 and related historical data S4; At the beginning of each time slot, data preprocessing is performed on remote control data S1, manual operation data S2, real-time statistical data S3 and related historical data S4, abnormal values ​​are proposed, and related missing values ​​are filled to obtain the corresponding data volume value. The speed of data generation is expressed as [V1, V2, V3, V4]. When the data generation speed is lower than the processing speed, that is, At this time, the container capacity R i No processing is required. When the data generation rate is higher than the processing rate, the capacity of the container is increased; We plan to jointly optimize the computing resources [C1, C2, C3, C4] allocated to the four types of data in parallel on the edge cluster, as well as the corresponding container capacity [R1, R2, R3, R4] allocated to the four types of data, to minimize the task preprocessing latency. The optimization problem is expressed as: In the above formula, δ represents the duration of the task processing time slot, constraint (1) indicates that the container capacity allocated to each type of data must be greater than the rate at which the task is generated minus the data processing rate to ensure that the relevant data cannot be discarded, and constraint (2) represents the capacity size limit and computing resource limit; The optimization problem is optimized and solved through genetic algorithms. After the dynamic container calculation is completed, the calculated results are automatically calculated based on the real-time operation data and sent to the message queue for use by other businesses.

5. The smart grid scheduling method based on cloud-edge fusion architecture according to claim 4 is characterized in that: It also includes edge clusters that support two ways to run predictions: The first method is to remotely call the intelligent prediction module of the cloud-edge fusion cloud system to perform load forecasting and new energy forecasting through a two-level cloud-edge fusion method and obtain the forecast results directly from the cloud system; The second method is to carry out model training for bus load forecasting, new energy power forecasting, system load forecasting and distributed energy forecasting within the jurisdiction of the edge cluster in the cloud system based on cloud artificial intelligence technology, and download the trained model to the edge cluster through the cloud-edge two-level fusion method. After the edge cluster is deployed, prediction is carried out based on the operating data of the edge cluster to obtain the prediction results.

6. The smart grid scheduling method based on cloud-edge fusion architecture according to claim 2 is characterized in that: The cloud system adopts a "cloud native" solution, based on a container operating environment, to support application "microservices" and elastic system expansion and upgrades. Based on the open and standard platform interfaces provided by the service bus and message queue, it builds "plug-and-play" application modules to support the joint construction of cloud applications across the entire network. Specifically: (1) Application module deployment in cloud systems; Deployment is based on the container runtime environment of the cloud platform, which also manages the container cluster, including internal communication, scheduling, start / stop, and horizontal scaling. (2) Communication between cloud system application modules; Communication is based on the virtual IP provided by the cloud platform VPC network. Application module services composed of multiple container groups can be mounted under a balancing IP to achieve load balancing of the application module's external service traffic. (3) Data storage of cloud system application modules; Data is stored in the storage volume mounted on the container and managed by the application module itself. The format can be customized. Data is stored in the cloud platform database and storage service mounted on the container. (4) Data interaction between cloud system application modules; Message transmission between application modules supported by different container groups is supported through a message bus. Application modules supported by the same container group have strong functional dependencies and can exchange data through shared memory, databases, and storage. (5) Service calls of cloud system application modules; It supports encapsulating all or a certain sub-functional logic of the application module and mounting it on the cloud service bus to provide functional services to the outside world. Each application module calls the services mounted on the cloud service bus through the cloud service bus on demand.

Images