New energy cloud side-end collaborative concentrated power prediction system, method, device and medium

Abstract

The invention provides a new energy cloud side-end collaborative concentrated power prediction system, method and device and a medium, and the method comprises a central cloud which is deployed in a production management region, the data processing module is used for executing macroscopic meteorological data analysis, historical data storage, global AI model training and iteration, global power prediction optimization, whole company range monitoring alarm and power transaction strategy simulation; the edge node is deployed in a large production control area of the area centralized control center and is used for receiving a model and an instruction of the central cloud, carrying out real-time aggregation and cleaning on data of a new energy station in an area under administration, executing area-level power prediction and carrying out safety data communication with the central cloud and the station end; and the station end is in communication connection with the edge node and is used for acquiring the equipment operation state and environmental data in real time and executing the issued control instruction, so that the problem that the network bandwidth pressure and the central cloud computing power bottleneck need to be reduced while the data real-time performance and the system response speed need to be ensured during new energy concentrated power prediction is solved.

Description

Technical Field

[0001] This invention relates to the fields of wind power generation and photovoltaic power generation technology, and in particular to a new energy cloud-edge-device collaborative centralized power prediction system, method, equipment and medium. Background Technology

[0002] With the large-scale development of my country's new energy industry, new energy power plants, represented by wind power and photovoltaic power, are characterized by cross-regional and distributed layouts. To achieve intensive and efficient management of numerous power plants, the industry generally adopts the solution of building provincial or group-level centralized power prediction systems. Such traditional solutions typically employ two technical architectures: one is a purely cloud-based centralized architecture, which transmits real-time data from all power plants to a central cloud platform for unified processing and prediction; the other is a completely decentralized power plant-level architecture, where prediction systems are deployed independently at each power plant, and the central platform performs simple data aggregation and display.

[0003] However, the aforementioned traditional architectures have revealed many inherent flaws in practical applications. For a purely cloud-based centralized architecture, the need to transmit high-frequency data (second-level and minute-level) from all sites to the central cloud via a dedicated network in real time leads to immense network bandwidth pressure, significant cross-regional data transmission latency, and difficulty in meeting the stringent real-time requirements of power forecasting, especially for ultra-short-term forecasting. Simultaneously, processing massive amounts of data in the central cloud places a huge computational load on the servers, resulting in high system expansion costs. While a completely decentralized site-level architecture alleviates the central pressure, it creates data silos, hindering cross-site collaborative forecasting and global optimization, and leads to fragmented system maintenance and low operational efficiency.

[0004] Therefore, there is an urgent need to propose a new energy cloud-edge-device collaborative centralized power prediction system to solve the technical problems of ensuring data real-time performance and system response speed while reducing network bandwidth pressure and central cloud computing power bottlenecks during centralized power prediction of new energy, and to achieve collaborative management and efficient operation and maintenance of cross-regional site data. Summary of the Invention

[0005] To overcome the problems existing in related technologies, this disclosure provides a new energy cloud-edge-device collaborative centralized power prediction system, device, equipment, and medium to solve the technical problems in related technologies that require ensuring data real-time performance and system response speed while reducing network bandwidth pressure and central cloud computing power bottlenecks, and achieving collaborative management and efficient operation and maintenance of cross-regional site data when performing centralized power prediction of new energy.

[0006] This specification provides one or more embodiments of a new energy cloud-edge-device collaborative centralized power prediction system, including: The central cloud is deployed in the production management area to perform macro-meteorological data analysis, historical data storage, global AI model training and iteration, global power prediction optimization, company-wide monitoring and alarms, and power trading strategy simulation. Edge nodes are deployed in the production control area of ​​the regional control center. They are used to receive models and instructions from the central cloud, perform real-time aggregation and cleaning of data from new energy power stations within their jurisdiction, execute regional power prediction, and conduct secure data communication with the central cloud and the power station terminals. The station terminal communicates with the edge nodes to collect real-time equipment operating status and environmental data, and executes the issued control commands.

[0007] Preferably, the central cloud comprises: The digital twin model building unit is used to build a digital twin model for each new energy power station. The digital twin model integrates the station's geographical topology, terrain features, equipment parameters, and historical operating data. The power prediction unit is used to fuse high-resolution numerical weather forecasts, radar images, and satellite cloud images, and to use graph neural networks to simulate micro-meteorological effects in order to generate station-level power prediction results. The equipment status adaptive correction unit is used to access the real-time equipment status data of the SCADA system and automatically adjust the power prediction value output by the power prediction unit according to the equipment performance degradation or shutdown maintenance status.

[0008] Preferably, the central cloud further includes: The reinforcement learning trading auxiliary decision-making unit, which embeds a reinforcement learning agent, is used to learn and output the optimal power application strategy and corresponding risk assessment report by simulating the power trading market environment. The predictive model self-evolution unit is used to establish a correlation model between prediction error and market clearing price. When the potential loss caused by prediction deviation exceeds a threshold, it automatically triggers the retraining and parameter optimization of the global AI model, forming a value closed loop of prediction-trading-correction.

[0009] Preferably, it also includes a federated learning service provider evaluation module, which is used to deploy a unified evaluation algorithm container on the local servers of each prediction service provider, so that each service provider can calculate the performance index of its prediction results locally, and upload the encrypted performance index to the central cloud for aggregation and ranking, so as to achieve the selection of the best service provider under the protection of data privacy.

[0010] Preferably, the central cloud further includes: The extreme weather resilience response unit is used to access meteorological disaster early warning information and simulate the quantitative impact of disasters on equipment safety and power output through the digital twin model, and output a risk assessment report. The intelligent emergency response plan triggering unit has a built-in hierarchical emergency response plan library, which is used to automatically push matching response plans and generate an operation and maintenance task list when an early warning is triggered.

[0011] Preferably, the central cloud further includes: The all-dimensional intelligent monitoring and root cause analysis unit is used to monitor the entire chain status of equipment operation, data transmission, prediction accuracy and transaction execution. It adopts alarm classification and knowledge graph association mechanism to automatically analyze the cause of alarms.

[0012] This specification provides one or more embodiments of a new energy cloud-edge-device collaborative centralized power prediction method, including the following steps: The central cloud deployed in the production management area performs macro-meteorological data analysis, historical data storage, global AI model training and iteration, global power prediction optimization, company-wide monitoring and alarms, and power trading strategy simulation. By receiving models and instructions from the central cloud through edge nodes deployed in the production control area of ​​the regional control center, the data of new energy power stations in the region are aggregated and cleaned in real time, regional power prediction is performed, and secure data communication is conducted with the central cloud and the power station. The station terminal, which is connected to the edge node, collects equipment operating status and environmental data in real time and executes the issued control commands. The central cloud, edge nodes, and field stations work together to form a power prediction and management system that enables global optimization and rapid regional response. Preferably, the central cloud performs global power prediction optimization, specifically including the following steps: A digital twin model is constructed for each new energy power station, and the digital twin model integrates the station's geographical topology, terrain features, equipment parameters and historical operating data; By integrating high-resolution numerical weather prediction, radar images, and satellite cloud images, and using graph neural networks to simulate micro-meteorological effects, station-level power prediction results are generated. The system accesses real-time equipment status data from the SCADA system and automatically adjusts the power prediction value output by the power prediction unit based on equipment performance degradation or shutdown for maintenance.

[0013] This specification provides one or more embodiments of a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-described new energy cloud-edge-device collaborative centralized power prediction method.

[0014] This specification provides one or more embodiments of a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the above-described new energy cloud-edge-device collaborative centralized power prediction method.

[0015] This disclosure provides a new energy cloud-edge-device collaborative centralized power prediction system, method, equipment, and medium. Its advantages lie in the central cloud, deployed in the production management area, used for macro-meteorological data analysis, historical data storage, global AI model training and iteration, global power prediction optimization, company-wide monitoring and alarms, and power trading strategy simulation. Through massive data storage and multi-source meteorological data fusion analysis, it supports global AI... Continuous model iteration significantly improves the global power prediction accuracy across regions and multiple energy types. The end-to-end monitoring and alarm function enables visualized management and rapid root cause localization, reducing the blind spots in cross-regional management. The power trading strategy simulation function empowers value creation and helps improve trading revenue. The containerized architecture supports rapid access for new power plants and service providers, reducing hardware and expansion costs and adapting to the needs of large-scale asset expansion, laying the foundation for energy storage collaboration, virtual power plant operation, and integrated development of power generation, grid, load, and storage. Edge nodes, deployed in the production control area of ​​the regional control center, receive models and instructions from the central cloud, perform real-time aggregation and cleaning of data from new energy power plants within their jurisdiction, execute regional-level power prediction, and conduct secure data communication with the central cloud and power plants. By offloading computing power, they undertake regional-level real-time business, completing data aggregation, cleaning, and regional-level power prediction locally, sharing the computing load of the central cloud, and improving local performance. The system improves response speed in sudden weather events, reduces cross-regional data transmission pressure and bandwidth consumption, relies on dedicated lines in the production control area for secure communication, meets secondary security requirements, connects the central cloud and the field station, provides regional prediction correction basis, and forms a global optimization-regional execution collaborative closed loop to improve system adaptability and reliability. At the field station, it communicates with the edge nodes to collect real-time equipment operating status and environmental data, and executes issued control commands. Accurate collection of equipment operating status and environmental data provides high-quality foundational support for edge node data processing, regional prediction, and central cloud model training, helping to bridge the gap between theoretical power generation capacity and actual power generation. Existing equipment can be reused without large-scale modifications, reducing project implementation costs and timelines. Efficient execution of control commands ensures smooth operation of the prediction-decision-control closed loop. Encrypted communication design meets data security compliance requirements, strengthening the foundation for architecture awareness and execution. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in one or more embodiments of this specification or in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this specification. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 A schematic diagram of a new energy cloud-edge-device collaborative centralized power prediction system provided for one or more embodiments of this specification; Figure 2 A schematic diagram of a new energy cloud-edge-device collaborative centralized power prediction method provided in one or more embodiments of this specification; Figure 3 This is a schematic diagram of the structure of a computer device provided for one or more embodiments of this specification. Detailed Implementation

[0018] To enable those skilled in the art to better understand the technical solutions in one or more embodiments of this specification, the technical solutions in one or more embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this specification, and not all of the embodiments. Based on one or more embodiments of this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of this invention.

[0019] The present invention will now be described in detail with reference to specific embodiments and accompanying drawings.

[0020] Method Implementation Examples According to embodiments of the present invention, a new energy cloud-edge-device collaborative centralized power prediction system is provided, such as... Figure 1 The diagram shown is a flowchart of the new energy cloud-edge-device collaborative centralized power prediction system provided in this embodiment. The new energy cloud-edge-device collaborative centralized power prediction system according to this embodiment includes: Central Cloud 11, deployed in the production management area, is used to perform macro-meteorological data analysis, historical data storage, global AI model training and iteration, global power prediction optimization, company-wide monitoring and alarms, and power trading strategy simulation. It links with the edge nodes of Production Zone II to achieve a collaborative closed loop of global optimization and regional execution. Central Cloud adopts a containerized + microservice architecture, supports elastic expansion, and can quickly adapt to new sites and new service providers, reducing hardware costs and later expansion costs.

[0021] Edge node 12, deployed in the production control area of ​​the regional control center, is used to receive models and instructions from the central cloud 11, perform real-time aggregation and cleaning of data from new energy power plants within its jurisdiction, conduct preliminary corrections, reduce the computing load of the central cloud, perform regional short-term and ultra-short-term power forecasts, improve the response speed in local weather change scenarios, and conduct secure data communication with the central cloud 11, the dedicated line of Production II area, and the power plant terminal 13 to ensure the compliance of the transmission of real-time business data such as equipment status and forecast results. It can push real-time computing services to the edge, reduce data transmission latency, and solve the problems of lag and security risks in cross-regional large-scale data transmission.

[0022] The station terminal 13 communicates with the edge node 12 through a standardized interface to meet the security requirements of secondary security for end-side equipment. It is used to collect equipment operating status and environmental data in real time. Specifically, it collects operating status data such as pitch angle, oil temperature, and inverter efficiency of wind turbines / inverters, as well as environmental perception data such as real-time power, wind speed, and light intensity. It also executes control commands issued by edge nodes and other devices, and performs secure data synchronization with edge node 12 and central cloud 11 to ensure a closed loop of data acquisition and command execution at the end side. The system provided in this embodiment, Central Cloud 11, is deployed in the production management area to perform macro-meteorological data analysis, historical data storage, global AI model training and iteration, global power prediction optimization, company-wide monitoring and alarms, and power trading strategy simulation. It supports global AI through massive data storage and multi-source meteorological data fusion analysis. Continuous model iteration significantly improves the global power prediction accuracy across regions and multiple energy types. The end-to-end monitoring and alarm function enables visualized management and rapid root cause localization, reducing the blind spots in cross-regional management. The power trading strategy simulation function empowers value creation and helps improve trading revenue. The containerized architecture supports rapid access for new power plants and service providers, reducing hardware and expansion costs and adapting to the needs of large-scale asset expansion, laying the foundation for energy storage collaboration, virtual power plant operation, and integrated development of power generation, grid, load, and storage. Edge node 12, deployed in the production control area of ​​the regional control center, receives models and instructions from the central cloud 11, performs real-time aggregation and cleaning of data from new energy power plants within its jurisdiction, executes regional-level power prediction, and maintains secure data communication with the central cloud 11 and the power plant terminal 13. Through computing power decentralization, it undertakes regional-level real-time business, completing data aggregation, cleaning, and regional-level power prediction locally, sharing the computing load of the central cloud 11, and improving local weather-related sudden power prediction capabilities. The system improves response speed across changing scenarios, reduces cross-regional data transmission pressure and bandwidth consumption, relies on dedicated lines in the production control area for secure communication, meets secondary security requirements, connects the central cloud 11 and the site terminal 13, provides regionalized prediction correction basis, forms a global optimization-regional execution collaborative closed loop, and improves system adaptability and reliability. The site terminal 13 communicates with the edge node 12 to collect real-time equipment operating status and environmental data, and executes issued control commands. By accurately collecting equipment operating status and environmental data, it provides high-quality basic support for edge node 12 data processing, regional prediction, and central cloud 11 model training, helping the prediction to leap from theoretical power generation capacity to actual power generation capacity. Existing equipment can be reused without large-scale modification, reducing project implementation costs and time. It efficiently executes control commands, ensuring the smooth operation of the prediction-decision-control closed loop. The encrypted communication design meets data security compliance requirements, and strengthens the foundation for architecture perception and execution.

[0023] In one embodiment, the central cloud 11 includes: The digital twin model building unit is used to build a digital twin model for each new energy power station. The digital twin model integrates the geographical topology of the wind turbine / photovoltaic panel distribution, terrain features such as altitude and slope, equipment parameters such as rated power and efficiency curves, and historical operating data of the power station to achieve a precise mapping between the physical environment and operating status of the power station.

[0024] The power prediction unit is used to integrate multi-source meteorological data such as high-resolution numerical weather forecasts, radar images, and satellite cloud images, and uses graph neural networks to simulate micro-meteorological effects such as terrain shading, wind turbine wakes, and photovoltaic module shadows. This transforms regional weather forecasts into refined environmental simulations within the site, generating site-level power prediction results and improving the spatial accuracy of predictions.

[0025] The equipment status adaptive correction unit is used to access real-time equipment status data such as pitch angle, gearbox oil temperature, and inverter efficiency from the SCADA system. Based on the equipment performance degradation or shutdown maintenance status, it automatically adjusts the power prediction value output by the power prediction unit, realizing the leap from theoretical power generation capacity prediction to actual power generation prediction and reducing prediction deviation.

[0026] The system provided in this embodiment integrates the geographical topology, terrain features, equipment parameters, and historical operating data of the power station through a digital twin model building unit. This creates a precise digital mapping model for each power station. Combined with a power prediction unit, it integrates multi-source meteorological data and uses graph neural networks to simulate micro-meteorological effects, achieving refined environmental analysis and power station-level power prediction. Then, through an equipment status adaptive correction unit, it connects to the real-time data of the SCADA system and dynamically adjusts the predicted values ​​according to changes in equipment operation. This completes the leap from theoretical to actual power generation prediction, comprehensively improving the scenario adaptability, spatial accuracy, and practical value of power prediction.

[0027] In one embodiment, the central cloud 11 further includes: The reinforcement learning-based transaction auxiliary decision-making unit, which embeds a reinforcement learning agent, is used to learn application strategies under different forecast accuracies and market scenarios by simulating electricity price fluctuations, dispatch rules, and transaction risks in the electricity trading market environment. It then outputs the optimal electricity application strategy, recommended application curve, and corresponding risk assessment report.

[0028] The predictive model self-evolution unit is used to establish a correlation model between prediction error and market clearing price. When the potential penalty or loss of profit caused by prediction deviation exceeds the threshold, it automatically triggers the retraining and parameter optimization of the global AI model, adjusts the model parameters and feature weights, and forms a value closed loop of prediction-trading-correction, so that the prediction accuracy can be continuously improved iteratively.

[0029] The system provided in this embodiment features a central cloud-based reinforcement learning transaction auxiliary decision-making unit that, through an embedded reinforcement learning agent, simulates the power trading market environment and outputs the optimal application strategy and risk assessment report, providing intelligent support for transaction decisions. The prediction model self-evolution unit establishes a correlation between prediction error and market clearing price, automatically triggering model retraining and parameter optimization when potential losses exceed a threshold, forming a value closed loop of prediction-transaction-correction. This not only allows the prediction model to continuously adapt to market demands but also significantly improves the scientific nature and profitability of power trading.

[0030] In one embodiment, a federated learning service provider evaluation module is also included. This module uses federated learning technology to deploy a unified evaluation algorithm container on the local servers of each prediction service provider. This allows each service provider to calculate the performance metrics of its prediction results locally and upload the encrypted performance metrics to the central cloud for aggregation and ranking. This enables the selection of the best service provider while protecting data privacy. Based on the aggregated metric data, a service provider scenario suitability score is generated to help the company select high-quality service providers economically. The module also supports the encrypted distribution of the best prediction data, provides reference correction for edge nodes and site terminals, improves the efficiency of service provider selection, and strengthens privacy protection and compliance.

[0031] The system provided in this embodiment, through its federated learning service provider evaluation module, deploys a unified evaluation algorithm container locally on each prediction service provider's premises. This allows service providers to calculate prediction performance indicators locally and upload only the encrypted indicators to the central cloud 11 for aggregation and ranking. This achieves both data privacy protection and protection of business secrets, while ensuring the fairness and transparency of service provider evaluation, efficiently selecting high-quality service providers, and enhancing the attractiveness of cooperation with service providers. It also avoids the privacy leakage risks associated with traditional centralized data comparison.

[0032] In one embodiment, the central cloud 11 further includes: The extreme weather resilience response unit is used to receive early warning information on meteorological disasters such as sandstorms, blizzards, freezing, and extreme winds from meteorological observatories. It also uses a digital twin model to quickly simulate the quantitative impact of disasters on equipment safety and power output, and outputs a risk assessment report, which includes the risk level and the expected power loss.

[0033] The intelligent emergency response triggering unit has a built-in hierarchical emergency response plan library, divided into three levels: general, important, and urgent. When an early warning is triggered, it automatically pushes matching response plans such as adjusting the operating mode in advance, requesting a shutdown, and starting the backup power supply, and generates an operation and maintenance task list to guide operation and maintenance personnel to respond quickly, reduce power generation losses under extreme weather conditions, and reduce the incidence of equipment safety accidents.

[0034] The system provided in this embodiment, with its extreme weather resilience response unit in the central cloud 11, accesses meteorological disaster early warning information and relies on a digital twin model to accurately simulate the quantitative impact of disasters on equipment safety and power output, and outputs a risk assessment report. This upgrades the system from receiving early warning information to accurately predicting risks. The intelligent emergency plan triggering unit, with its built-in hierarchical emergency plan library, automatically pushes appropriate response plans and generates an operation and maintenance task list when an early warning is triggered, transforming passive response into proactive and precise handling. The two work together to build a closed loop of risk prevention and emergency response under extreme weather conditions, effectively improving the disaster resilience of the station, ensuring the safe operation of equipment, and reducing power generation losses caused by disasters.

[0035] In one embodiment, the central cloud 11 further includes: The all-dimensional intelligent monitoring and root cause analysis unit is used to monitor the entire chain of equipment operation, data transmission, prediction accuracy, and transaction execution. It enables monitoring of the entire chain from equipment operation to data transmission, prediction accuracy, and transaction execution, and supports multi-dimensional filtering and visualization. It employs an alarm hierarchy and knowledge graph association mechanism to automatically analyze the causes of alarms.

[0036] The system provided in this embodiment features a central cloud-based all-dimensional intelligent monitoring and root cause analysis unit that achieves full-link status monitoring of equipment operation, data transmission, prediction accuracy, and transaction execution. Through an alarm classification mechanism, it focuses on key issues and automatically traces the root cause of alarms using knowledge graph association technology. This breaks through the limitations of information fragmentation in traditional decentralized monitoring and the reliance on manual root cause investigation, significantly improving the efficiency and accuracy of anomaly detection and location, reducing the blindness of cross-regional management, and providing strong support for stable system operation and efficient maintenance.

[0037] Method Implementation Examples According to embodiments of the present invention, a new energy cloud-edge-device collaborative centralized power prediction method is provided, such as... Figure 2 The diagram shown is a flowchart illustrating the new energy cloud-edge-device collaborative centralized power prediction method provided in this embodiment. The new energy cloud-edge-device collaborative centralized power prediction method according to this embodiment includes: S210 performs macro-meteorological data analysis, historical data storage, global AI model training and iteration, global power prediction optimization, company-wide monitoring and alarms, and power trading strategy simulation through a central cloud deployed in the production management area.

[0038] S220 receives models and instructions from the central cloud through edge nodes deployed in the production control area of ​​the regional control center, performs real-time aggregation and cleaning of data from new energy power plants within its jurisdiction, executes regional power prediction, and conducts secure data communication with the central cloud and the power plants.

[0039] S230: The station terminal, which is connected to the edge node, collects equipment operating status and environmental data in real time and executes the issued control commands.

[0040] The central cloud, edge nodes, and field stations work together to form a power prediction and management system that enables global optimization and rapid regional response. The method provided in this embodiment performs macro-meteorological data analysis, historical data storage, global AI model training and iteration, global power prediction optimization, company-wide monitoring and alarming, and power trading strategy simulation through a central cloud deployed in the production management area. Through massive data storage and multi-source meteorological data fusion analysis, it supports continuous iteration of the global AI model, significantly improving the accuracy of global power prediction in cross-regional and multi-energy-type scenarios. The end-to-end monitoring and alarm function enables visualized control and rapid root cause localization, reducing the blind spots in cross-regional management. The power trading strategy simulation function empowers value creation and helps improve trading revenue. The containerized architecture supports rapid access for new power plants and service providers, reducing hardware and expansion costs, adapting to the needs of large-scale asset expansion, and laying the foundation for energy storage collaboration, virtual power plant operation, and integrated development of power generation, grid, load, and storage. Edge nodes deployed in the production control area of ​​the regional control center receive models and instructions from the central cloud, perform real-time aggregation and cleaning of data from new energy power plants within the jurisdiction, execute regional-level power prediction, and conduct secure data exchange with the central cloud and power plant terminals. The communication system, through the decentralization of computing power, undertakes regional-level real-time services, locally completing data aggregation, cleaning, and regional-level power prediction. This alleviates the computing load on the central cloud, improves response speed in scenarios of sudden local weather changes, reduces cross-regional data transmission pressure and bandwidth consumption, and achieves secure communication through dedicated lines in the production control area, meeting secondary security requirements. It connects the central cloud and the field stations, providing regionalized prediction correction basis, forming a global optimization-regional execution collaborative closed loop, and improving system adaptability and reliability. The field stations, through communication connections with edge nodes, collect real-time equipment operating status and environmental data, and execute issued control commands. By accurately collecting equipment operating status and environmental data, it provides high-quality foundational support for edge node data processing, regional-level prediction, and central cloud model training, helping predictions transition from theoretical power generation capacity to actual power generation. It reuses existing equipment without large-scale modifications, reducing project implementation costs and timelines, efficiently executes control commands, and ensures the smooth operation of the prediction-decision-control closed loop. The encrypted communication design meets data security compliance requirements, solidifying the foundation for architecture awareness and execution.

[0041] The embodiments of the present invention are device embodiments corresponding to the above method embodiments. The specific operations of each module processing step can be understood with reference to the description of the method embodiments, and will not be repeated here.

[0042] like Figure 3 As shown, the present invention also provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, it implements the new energy cloud-edge-device collaborative centralized power prediction system in the above embodiments, or when the computer program is executed by a processor, it implements the new energy cloud-edge-device collaborative centralized power prediction system in the above embodiments.

[0043] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0044] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, for apparatus or system embodiments, since they are basically similar to method embodiments, the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments. The apparatus and system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0045] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features. These modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and the contents not described in detail in the specification of the present invention are known to those skilled in the art.

Claims

1. A new energy cloud-edge-device collaborative centralized power prediction system, characterized in that, include: The central cloud is deployed in the production management area to perform macro-meteorological data analysis, historical data storage, global AI model training and iteration, global power prediction optimization, company-wide monitoring and alarms, and power trading strategy simulation. Edge nodes are deployed in the production control area of ​​the regional control center. They are used to receive models and instructions from the central cloud, perform real-time aggregation and cleaning of data from new energy power stations within their jurisdiction, execute regional power prediction, and conduct secure data communication with the central cloud and the power station terminals. The station terminal communicates with the edge nodes to collect real-time equipment operating status and environmental data, and executes the issued control commands.

2. The new energy cloud-edge-device collaborative centralized power prediction system as described in claim 1, characterized in that, The central cloud includes: The digital twin model building unit is used to build a digital twin model for each new energy power station. The digital twin model integrates the station's geographical topology, terrain features, equipment parameters, and historical operating data. The power prediction unit is used to fuse high-resolution numerical weather forecasts, radar images, and satellite cloud images, and to use graph neural networks to simulate micro-meteorological effects in order to generate station-level power prediction results. The equipment status adaptive correction unit is used to access the real-time equipment status data of the SCADA system and automatically adjust the power prediction value output by the power prediction unit according to the equipment performance degradation or shutdown maintenance status.

3. The new energy cloud-edge-device collaborative centralized power prediction system as described in claim 1, characterized in that, The central cloud also includes: The reinforcement learning trading auxiliary decision-making unit, which embeds a reinforcement learning agent, is used to learn and output the optimal power application strategy and corresponding risk assessment report by simulating the power trading market environment. The predictive model self-evolution unit is used to establish a correlation model between prediction error and market clearing price. When the potential loss caused by prediction deviation exceeds a threshold, it automatically triggers the retraining and parameter optimization of the global AI model, forming a value closed loop of prediction-trading-correction.

4. The new energy cloud-edge-device collaborative centralized power prediction system as described in claim 1, characterized in that, It also includes a federated learning service provider evaluation module, which is used to deploy a unified evaluation algorithm container on the local servers of each prediction service provider, so that each service provider can calculate the performance index of its prediction results locally, and upload the encrypted performance index to the central cloud for aggregation and ranking, so as to achieve the selection of the best service provider under the protection of data privacy.

5. The new energy cloud-edge-device collaborative centralized power prediction system as described in claim 2, characterized in that, The central cloud also includes: The extreme weather resilience response unit is used to access meteorological disaster early warning information and simulate the quantitative impact of disasters on equipment safety and power output through the digital twin model, and output a risk assessment report. The intelligent emergency response plan triggering unit has a built-in hierarchical emergency response plan library, which is used to automatically push matching response plans and generate an operation and maintenance task list when an early warning is triggered.

6. The new energy cloud-edge-device collaborative centralized power prediction system as described in claim 2, characterized in that, The central cloud also includes: The all-dimensional intelligent monitoring and root cause analysis unit is used to monitor the entire chain status of equipment operation, data transmission, prediction accuracy and transaction execution. It adopts alarm classification and knowledge graph association mechanism to automatically analyze the cause of alarms.

7. A centralized power prediction method for new energy cloud-edge-device collaboration, characterized in that, Includes the following steps: The central cloud deployed in the production management area performs macro-meteorological data analysis, historical data storage, global AI model training and iteration, global power prediction optimization, company-wide monitoring and alarms, and power trading strategy simulation. By receiving models and instructions from the central cloud through edge nodes deployed in the production control area of ​​the regional control center, the data of new energy power stations in the region are aggregated and cleaned in real time, regional power prediction is performed, and secure data communication is conducted with the central cloud and the power station. The station terminal, which is connected to the edge node, collects equipment operating status and environmental data in real time and executes the issued control commands. The central cloud, edge nodes, and field stations work together to form a power prediction and management system that enables global optimization and rapid regional response.

8. The new energy cloud-edge-device collaborative centralized power prediction method as described in claim 7, characterized in that, The central cloud performs global power prediction optimization, specifically including the following steps: A digital twin model is constructed for each new energy power station, and the digital twin model integrates the station's geographical topology, terrain features, equipment parameters and historical operating data; By integrating high-resolution numerical weather prediction, radar images, and satellite cloud images, and using graph neural networks to simulate micro-meteorological effects, station-level power prediction results are generated. The system accesses real-time equipment status data from the SCADA system and automatically adjusts the power prediction value output by the power prediction unit based on equipment performance degradation or shutdown for maintenance.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the new energy cloud-edge-device collaborative centralized power prediction method as described in any one of claims 7 to 8.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the new energy cloud-edge-device collaborative centralized power prediction method as described in any one of claims 7 to 8.

Images