Event recording method of intelligent flexible adjusting unit

Abstract

The invention discloses an event recording method of an intelligent flexible adjusting unit. According to the method, through precise deployment of the sensor network, monitoring of various parameters is comprehensively covered, and the multi-element load monitoring accuracy is improved; through deep learning and edge calculation, voltage fluctuation and potential faults can be rapidly identified, and early warning accuracy and response speed are improved. Through automatic recording and anomaly classification, automatic recording of threshold data is realized, and event processing sorting is realized based on random forest algorithm classification. On the basis of a hot and cold data layering strategy and an encryption technology, instant response and long-term safe storage are ensured; through authority management, based on role access control and dynamic adjustment, flexibility and security are enhanced; efficient data interaction is realized based on RESTful API and WebSocket through a remote query and analysis tool, and an advanced analysis model is integrated; through side-to-side communication and adaptive learning, communication stability is ensured based on MQTTCP / IP or Zigbee, an adaptive learning mechanism continuously updates a model, and the self-optimization capability of the system is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power system management, and in particular to an event recording method of an intelligent flexible regulating unit. Background Art

[0002] In the development and management framework of the new power system, the intelligent regulation unit, as a key component, undertakes the important task of ensuring accurate monitoring and interactive response of power loads. However, the current intelligent regulation unit has exposed obvious shortcomings in abnormal event monitoring and historical operation tracking, which is manifested in incomplete data recording and difficult tracing, which poses a serious challenge to system fault diagnosis and operation and maintenance efficiency. In particular, when dealing with key events such as overload voltage anomalies and sudden changes in current direction, there is a lack of real-time recording and analysis methods, which affects the stable operation of the power system, limits the response speed, and cannot meet the growing demand for intelligence and high efficiency.

[0003] In view of the above technical problems, the limitations of traditional methods are highlighted in incomplete real-time data capture, inaccurate processing, slow abnormal response, and weak event tracing capabilities. The lack of records of key events, such as overload and overcurrent abnormal processing, makes it difficult for the system to analyze in real time, which directly affects the stability and efficient operation and maintenance efficiency of the power system, and cannot effectively respond to the transformation trend of the power system from "source follows load" to "source and load interaction".

[0004] Therefore, there is an urgent need for an intelligent and efficient event recording and monitoring mechanism to solve the problem of difficulty in accurately monitoring multiple loads, improve the reliability and response speed of the power system, and provide technical support for the construction of a modern power supply service system and a new power system. Summary of the invention

[0005] The present invention provides an event recording method for an intelligent flexible regulation unit, which can provide an intelligent and efficient event recording and monitoring mechanism to solve the problem that multiple loads are difficult to accurately monitor, improve the reliability and response speed of the power system, and provide technical support for the construction of a modern power supply service system and a new power system.

[0006] In order to achieve the above object, the present invention provides an event recording method of an intelligent flexible regulating unit, the method comprising: Plan and deploy sensor networks for the power system to fully cover the voltage, current, and environmental parameter monitoring needs and ensure comprehensive monitoring of the power system status; Through edge computing models, combined with cloud computing and IoT technologies, real-time processing and analysis of sensor data can be achieved, and rule engines or machine learning algorithms can be used to quickly identify threshold-exceeding anomalies; Based on the event capture mechanism, abnormal data that exceeds the threshold or is identified by the model is automatically recorded, including the time period before and after the abnormality, timestamp, location, and device identification; Through machine learning classification algorithms, captured events are intelligently classified, based on historical data training, and based on abnormal characteristics, the urgency is quickly determined and sorted; Build a storage architecture through SSD cache and hard disk array or cloud storage to ensure immediate response and long-term storage of data based on data lifecycle management strategy; The AES-256-bit encryption algorithm is used to ensure the security of transmission and static data, and role-based access control is used to achieve detailed permission management.

[0007] Preferably, the planning and deployment of the sensor network for the power system is specifically as follows: The sensor network includes voltage, current, power and environmental parameter sensors, and each of the sensors is deployed in redundant quantities at key nodes, lines and inside equipment to ensure that monitoring continuity is not affected by single point failures.

[0008] Preferably, the edge computing model is as follows: Through deep learning algorithm models, complex voltage fluctuation patterns are deeply analyzed to identify potential circuit failure problems in advance.

[0009] Preferably, abnormal data exceeding the threshold or identified by the model is automatically recorded, specifically: Abnormal data is instantly stored locally and uploaded to the cloud, achieving local and cloud dual backup to ensure data security.

[0010] Preferably, the capture events are intelligently classified, specifically: Through the random forest algorithm, abnormal data is classified according to multiple dimensions such as frequency, intensity and duration to improve the accuracy of event classification.

[0011] Preferably, the storage architecture is constructed as follows: The storage architecture adopts an automatic tiering strategy for hot and cold data, automatically adjusts storage media based on data access frequency, and optimizes storage costs and reading speeds.

[0012] Preferably, permission management supports dynamic permission adjustment, that is, instantly adjusting role permissions according to changes in operation and maintenance scenarios to enhance system flexibility.

[0013] Preferably, it also includes remote query and analysis tool integration, specifically: Through the remote query interface, secure and efficient data transmission is achieved, and paging query and filter condition optimization are provided; Anomaly detection, root cause analysis, and time series forecasting are all available through analytical tools that integrate advanced predictive models.

[0014] Preferably, the remote query interface is specifically: The remote query interface complies with the RESTful API standard, supports HTTPS and WebSocket protocols, realizes two-way real-time data transmission, and improves interactivity and response speed.

[0015] Preferably, it also includes edge collaboration and adaptive learning mechanisms, specifically: Implement edge-to-edge communication in mesh networks based on MQTTCP / IP or Zigbee protocols to ensure low latency and stability of communication between devices; An adaptive learning mechanism is established based on continuous data feedback to continuously update the model and ensure that the model effectively adapts to load changes and abnormal types.

[0016] The embodiment of the present application discloses an event recording method for an intelligent flexible regulation unit. The method realizes all-round monitoring of voltage, current, and environmental parameters in the power system by carefully planning and deploying a sensor network, ensuring real-time and comprehensive grasp of the power system status, solving the problem that multiple loads are difficult to accurately monitor, and improving the integrity of data collection and system perception capabilities; a deep learning model combined with edge computing is used to quickly identify voltage fluctuation patterns and potential faults, improve the accuracy and response speed of fault warnings, and enable the system to more effectively handle complex power fluctuations and reduce losses caused by faults; by automatically recording over-threshold or abnormal data and using random forest algorithm classification, fast and accurate classification and sorting of events are achieved, helping dispatchers to quickly locate and prioritize emergency issues, and optimizing operation and maintenance efficiency; an efficient storage architecture is constructed, and a cold and hot data stratification strategy and encryption technology are used to ensure instant response and long-term secure storage of data, while reducing operation and maintenance costs and improving the economy and security of data management; role-based access control and dynamic permission adjustment are implemented, enhancing system flexibility and security, facilitating operation and maintenance operations, and adapting to different scenario requirements; remote query and analysis tools are integrated, and through RESTful API and WebSocket protocols improve the efficiency of data interaction, and combined with advanced analysis models, they realize accurate anomaly detection, root cause analysis and prediction, and enhance decision-making support capabilities. The use of MQTTCP / IP or Zigbee protocols to achieve edge-to-edge communication ensures low latency and stability between devices, and the adaptive learning mechanism enables the model to be continuously updated to maintain adaptability to load changes and anomaly types, thereby enhancing the system's intelligence level and self-optimization capabilities. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present invention. Moreover, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings: Figure 1 The present invention is a flowchart of an event recording method of an intelligent flexible regulating unit. DETAILED DESCRIPTION

[0018] As mentioned in the background technology, the limitations of traditional methods are incomplete real-time data capture, inaccurate processing, slow abnormal response, and weak event tracing capabilities. The lack of records of key events, such as overload and overcurrent abnormal processing, makes it difficult for the system to analyze in real time, which directly affects the stability and efficient operation and maintenance efficiency of the power system.

[0019] In order to solve the above problems, the embodiment of the present application proposes an event recording method of an intelligent flexible regulation unit. The solution provides an intelligent and efficient event recording and monitoring mechanism to solve the problem that multiple loads are difficult to accurately monitor, improve the reliability and response speed of the power system, and provide technical support for the construction of a modern power supply service system and a new power system.

[0020] like Figure 1 As shown, it is a flow chart of the event recording method of the above-mentioned intelligent flexible adjustment unit. Before introducing the specific scheme, the concepts involved below are introduced: SSD (Solid State Drive): is a storage device that uses solid-state memory as a permanent storage medium. Compared with traditional hard disk drives, SSDs have faster read and write speeds and higher durability, and are suitable for storing instant data response and long-term data retention.

[0021] AES (Advanced Encryption Standard): is a widely used symmetric key encryption algorithm. "25-bit" refers to the AES encryption method with a key length of 25 bits. It is the strongest security level in the AES family and is widely used for data encryption to ensure the security of data during transmission and storage.

[0022] RESTful API (Representational State Transfer) is a network application program interface design style that emphasizes simplicity, readability, statelessness, and HTTP-based interaction, which is conducive to cross-platform data exchange. RESTful API is often used in Web services to simplify service calls and enable data acquisition and manipulation.

[0023] HTTPS (HyperText Transfer Protocol Secure): It is an encrypted version of HTTP. It combines the SSL / TLS protocol to provide a secure communication environment, ensuring data encryption and authentication during transmission to prevent eavesdropping and tampering. It is often used for security-sensitive data interaction.

[0024] WebSocket protocol: WebSocket is a persistent two-way communication protocol that allows long connections to be established between the client and the server, enables real data exchange, and provides full-duplex two-way real-time data transmission. WebSocket overcomes the limitations of HTTP and improves the interactivity of real-time applications such as chat and real-time monitoring.

[0025] TT / IP protocol (Transport Control Protocol): It is a network control protocol based on TCP / IP protocol and is used for communication between devices. TT / IP is often used to implement the transmission of instructions, such as device control, to ensure the reliability and responsiveness of communication, and is suitable for industrial control and automation fields.

[0026] Zigbee protocol: Zigbee is a low-power wireless network protocol stack based on the IEEE 802.15.4 standard. It is designed for communication between devices and provides low-power, low-cost, and long-distance communication. It is commonly used in sensor networks, home automation, and industrial monitoring. The igbee protocol supports flexible self-organizing mesh networks to improve communication stability and efficiency.

[0027] Specifically, the method comprises the following steps: Step S101: Plan and deploy a sensor network for the power system to fully cover the voltage, current, and environmental parameter monitoring requirements and ensure comprehensive monitoring of the power system status.

[0028] As mentioned above, first of all, based on the characteristics and monitoring objectives of the power system, considering factors such as system scale, load type, and operating environment, clarify the range of parameters to be monitored, such as voltage, current, power, frequency, and environmental parameters (such as temperature and humidity). Then, according to the determined monitoring needs, select the appropriate sensor model, such as high-precision voltage sensors for voltage monitoring, Hall effect sensors for current measurement, and temperature and humidity sensors for monitoring environmental conditions. Then design the sensor network layout, considering the distribution density, location, and connection method of the sensors, to ensure full coverage of key nodes and lines of the power system, including substations, transmission lines, distribution boxes, important load points, etc., and consider redundant design to deal with single point failures. Finally, perform physical installation according to the plan, including sensor fixation, wiring, and debugging, and ensure that the sensors are correctly connected to the data acquisition equipment or edge computing unit to prepare for subsequent data transmission and processing.

[0029] For example, in order to fully monitor the operating status of wind turbines in a wind farm, a sensor network needs to be deployed. First, analyze the monitoring needs of wind turbines to determine the voltage, current, frequency, power output, generator temperature, and humidity in the tower that need to be monitored. Weather-resistant voltage and current sensors are selected and installed at the generator outlet, high-precision temperature sensors are placed inside the generator, and humidity sensors are installed in the tower. When planning the network, ensure that each wind turbine has at least three sets of sensor groups, covering key parts, and connected to the edge computing unit in the central monitoring room through the LoRa wireless communication module. During deployment, install and test them one by one to ensure that the data is accurately transmitted to the central monitoring system to complete the entire deployment.

[0030] It should be noted that an intelligent management mechanism for sensor networks is adopted, that is, the Internet of Things technology is introduced. The sensor network is not only responsible for data collection, but also can self-monitor its own status, such as health status and battery power, and transmit it back to the management platform wirelessly for easy maintenance; an edge computing and sensor fusion mechanism is adopted, that is, the sensor and edge computing units are more tightly integrated, and the sensor data is directly processed and abnormalities are judged on the edge side, reducing the data transmission burden and speeding up the response speed; a dynamic optimization layout mechanism is adopted, that is, the sensor network can be dynamically adjusted according to actual operation data feedback. For example, if blind spots or excessive monitoring are found in a certain area, the network layout can be optimized by adjusting the sensor position or increasing or decreasing the amount to improve efficiency and accuracy. All of these fall within the scope of protection of this application.

[0031] Step S102: Through the edge computing model, combined with cloud computing and Internet of Things technology, real-time processing and analysis of sensor data is achieved, and a rule engine or machine learning algorithm is used to quickly identify threshold exceeding anomalies.

[0032] As mentioned above, first, the edge computing device receives the data stream from the sensor network, performs preliminary cleaning and format conversion, removes noise, ensures data quality, and prepares for subsequent analysis; deploys models on edge devices based on rule engines (setting threshold comparisons) or more complex machine learning algorithms (such as support vector machines, neural networks), aiming to quickly analyze data and identify specific patterns or anomalies; the model analyzes the data stream in real time, and when it is found that the data point exceeds the preset threshold or conforms to an abnormal pattern, it immediately triggers an alarm or takes action, such as adjusting the control strategy, recording abnormal events, etc.

[0033] For example, in the environmental monitoring system of a smart factory, multiple sensors are deployed to monitor the temperature, humidity, and concentration of harmful gases on the production line. The edge computing model is integrated into the control cabinet of the production line and receives sensor data in real time. The model includes a rule engine-based part, such as setting the upper temperature limit to 35 degrees Celsius. Once the temperature exceeds this value, the model immediately sends instructions to the control system to reduce the operating frequency of the cooling system; at the same time, the model also uses machine learning algorithms and can identify certain complex patterns through historical data training. For example, changes in the combination of temperature and humidity may indicate a precursor to overheating of a specific machine, thereby providing early warning and reducing the risk of downtime.

[0034] It should be noted that the adoption of a model adaptive learning mechanism, that is, the edge computing model supports online learning, continuously optimizes itself from new data, adapts to changes in data patterns caused by environmental changes or equipment aging, and improves recognition accuracy; the adoption of an edge collaborative processing mechanism, that is, multiple edge computing nodes achieve collaboration, share processing power, and optimize resource allocation. For example, when a node is overloaded, some processing tasks are dynamically transferred to adjacent idle nodes; the adoption of a low-latency response mechanism, that is, for certain control scenarios that require extremely low latency, such as emergency stop operations, edge computing directly controls to reduce the time it takes for data to travel to and from the cloud and achieve millisecond-level response, all of which fall within the scope of protection of this application.

[0035] Step S103: Based on the event capture mechanism, abnormal data that exceeds the threshold or is identified by the model is automatically recorded, including the time period before and after the abnormality, timestamp, location and device identification.

[0036] As mentioned above, the normal working range thresholds of various parameters, such as current, voltage, temperature, etc., are pre-set in the system. Anything beyond this range is considered an abnormality. The system continuously monitors the sensor data stream and compares it with the threshold. At the same time, the edge computing model analyzes the data pattern in real time to identify non-threshold anomalies. Once an anomaly is detected, the event capture mechanism is immediately triggered to automatically record the abnormal data and its related metadata. The record content includes the start and end time of the abnormal data, the precise timestamp, the specific location where the abnormality occurred (such as equipment number, floor, room number), and the device identifier involved.

[0037] For example, the air conditioning system on the floor is equipped with temperature and humidity sensors. At 3:27 pm one day, the temperature sensor detected that the temperature in a room suddenly rose to 35°C, exceeding the preset comfort zone (20°C-26°C). The edge computing model analyzed historical data and current environmental factors and confirmed that this was an anomaly rather than a seasonal temperature fluctuation. At this time, the event capture mechanism was activated and the abnormal data was automatically recorded: the abnormal time period was from 3:07 to 3:15, the timestamp was 2022-03-15:07:15, the location information was Room 5, 3rd Floor, Building B, and the equipment identification was AC-001. This record was immediately used to notify the maintenance team to quickly check whether the air conditioning system was faulty or whether there was an abnormal temperature caused by external reasons.

[0038] It should be noted that the use of intelligent prediction and early warning mechanisms, that is, based on historical event data, model learning predicts possible abnormal trends, issues early warnings, and reduces the impact of sudden failures; the use of multi-source data fusion mechanisms, that is, integrating multi-dimensional data such as video surveillance and sound to provide more comprehensive contextual information for abnormal events and assist in decision-making; the use of data mining and optimization mechanisms, that is, the use of long-term accumulated abnormal data for in-depth analysis, identification of system weaknesses, optimization of configuration, improvement of overall operational efficiency, or adjustment of threshold settings, all fall within the scope of protection of this application.

[0039] Step S104: Intelligently classify the captured events through machine learning classification algorithms, quickly determine the urgency and sort them based on historical data training and abnormal characteristics.

[0040] As mentioned above, key features are extracted from the captured abnormal data. These features may include the duration, frequency, amplitude, rate of change, associated environmental parameters, etc. of the abnormality; historical data sets containing marked abnormal events and their urgency are used to train machine learning models (such as decision trees, random forests, support vector machines, or deep learning models); new abnormal event data is input into the trained model, and the model automatically classifies abnormal events and determines their types, such as equipment failure, environmental problems, or human operational errors; the model assesses the urgency of the event based on the feature weights, such as abnormally high equipment temperature may be more urgent than a small humidity level; events are sorted according to the urgency, and the most urgent anomalies are displayed to the operator first to facilitate rapid response and processing.

[0041] It should be noted that the use of adaptive learning and model optimization mechanisms, that is, the system continuously receives new data, the model automatically updates and self-optimizes, and improves the accuracy of classification and the accuracy of urgency judgment; the use of a multi-dimensional analysis mechanism, that is, combining external data (such as weather forecasts, holidays) to enhance the model's prediction of the impact of abnormal events and the judgment of the degree of urgency; the use of a predictive maintenance mechanism, that is, through the classification and urgency of abnormal events, the model predicts possible future failure trends, promotes proactive maintenance, and reduces unexpected downtime, all of which fall within the scope of protection of this application.

[0042] Step S105: Build a storage architecture through SSD cache and hard disk array or cloud storage to ensure instant response and long-term storage of data based on data lifecycle management strategy.

[0043] As mentioned above, by combining the fast response capability of SSD cache with the long-term preservation advantages of hard disk array or cloud storage, an efficient and reliable storage architecture is built to meet the needs of immediate data processing and long-term archiving. Based on the data lifecycle management strategy, the system automatically determines the storage location and period of the data to ensure the best use of resources and the long-term availability of data.

[0044] For example, using solid-state drives (SSDs) as caches, due to their high-speed read and write performance, are suitable for storing frequently accessed real-time data to ensure fast system response; hard disk array technology (such as RAID) increases storage capacity and fault tolerance by having multiple hard disks work in parallel, and is used to store data that is not frequently accessed but needs to be retained for a long time; combined with cloud storage services, it provides elasticity and scalability for data, especially for remote storage of massive or backup data, to achieve off-site disaster recovery; define full-cycle strategies for data from generation, active use, archiving to final abandonment, and automatically or manually migrate data to the appropriate storage layer; automatically move data between SSDs, hard disk arrays, and cloud storage based on data access frequency and importance to balance performance and cost.

[0045] It should be noted that the use of an intelligent tiered storage mechanism, that is, the introduction of a more intelligent tiered storage solution, such as an automatic learning algorithm to predict data access patterns, dynamically adjust storage strategies, and optimize data distribution; the use of a data compression and deduplication mechanism, that is, compressing and deduplicating data before storage to reduce storage occupancy, especially in cloud storage and hard disk arrays, to improve efficiency; the use of a disaster recovery mechanism, that is, combining the characteristics of cloud storage to design a disaster recovery plan to achieve rapid data recovery, ensure business continuity, and reduce the impact of force majeure events, all fall within the scope of protection of this application.

[0046] Step S106: Use the AES-256 bit encryption algorithm to ensure the security of transmission and static data, and implement detailed permission management through role-based access control.

[0047] As mentioned above, the AES-256 bit encryption algorithm is used to ensure the security of data during transmission and storage, and the role-based access control (RBAC) mechanism is used to achieve refined management of different user permissions. This measure is designed to protect sensitive information from illegal access and leakage, while ensuring that legitimate users can operate efficiently with appropriate permissions.

[0048] For example, the 256-bit key length version of the Advanced Encryption Standard (AES) is used to perform high-strength encryption on data in transit to ensure the security of information transmission. This encryption technology is also used when storing static data. Even if the data is stolen, it cannot be decrypted without the key. Different roles are defined in the system, such as administrators, operation and maintenance personnel, analysts, ordinary users, etc. Each role represents a user group and their appropriate operating permissions. According to the role definition, permissions are automatically or manually assigned, such as read-only, read-write, modify, and manage permissions, to ensure that users can only access and operate resources permitted by their roles. When a user tries to access a resource, the system verifies it based on its role permissions, allowing access if it meets the permissions, otherwise denying access, thereby achieving fine-grained access control. User access logs are recorded and permission usage is audited to help monitor and trace potential security issues and ensure compliance with permission management.

[0049] It should be noted that the adoption of a dynamic permission management mechanism, that is, dynamically adjusting user permissions according to time, location or specific conditions, such as restricting sensitive operation permissions outside working hours to improve security; the adoption of a two-factor authentication mechanism, that is, combining a two-factor authentication mechanism, such as a password plus an SMS verification code or biometrics, to further enhance account security, and even if the password is leaked, it is not easy to be illegally logged in; the adoption of a data desensitization processing mechanism, that is, desensitizing sensitive data, such as hiding some fields and providing them to users who do not need complete sensitive information to protect privacy, all fall within the scope of protection of this application.

[0050] Preferably, a sensor network is planned and deployed for the power system, specifically the sensor network includes voltage, current, power and environmental parameter sensors, each of which is deployed in redundant quantity at key nodes, lines and inside equipment to ensure monitoring continuity is not affected by single point failures.

[0051] As mentioned above, in order to achieve comprehensive power system monitoring, the sensor network design should include high-precision voltage and current sensors to monitor the voltage fluctuations and current load conditions of the power grid; power sensors monitor energy conversion efficiency and consumption. Environmental parameter sensors, such as temperature and humidity sensors, are placed near key equipment to prevent overheating. When deployed, sensors are arranged in a redundant manner, that is, more than two or more sensors are installed at each key location to ensure that even if one of them fails, the monitoring task can still be carried out seamlessly, ensuring the stability of continuous monitoring of the system.

[0052] Preferably, an edge computing model, specifically a deep learning algorithm model, is used to deeply analyze complex voltage fluctuation patterns and identify potential circuit failure problems in advance.

[0053] As mentioned above, the edge computing model uses deep neural networks (such as convolutional neural networks and recurrent neural networks) to analyze the voltage data collected in real time. By learning the patterns in historical fault data, it can identify subtle changes in voltage fluctuations, such as periodicity and harmonic waveform distortion, as precursors to faults, issue early warnings, and reduce power outages caused by circuit problems.

[0054] Preferably, abnormal data exceeding the threshold or identified by the model is automatically recorded, specifically, the abnormal data is instantly stored locally and uploaded to the cloud, so as to realize local and cloud dual backup and ensure data security.

[0055] As mentioned above, when sensor data exceeds the preset threshold or the model recognizes an abnormality, the system immediately activates the local cache mechanism, records the data to the SSD cache, and synchronizes it to the cloud in real time through an encrypted channel to achieve dual backup. In this way, even if the local system encounters an unforeseen failure, the cloud still has a complete record to ensure data continuity and security.

[0056] Preferably, the captured events are intelligently classified, specifically by using a random forest algorithm to classify the abnormal data according to multiple dimensions such as frequency, intensity and duration, so as to improve the accuracy of event classification.

[0057] As mentioned above, the random forest classification algorithm is used to perform multi-dimensional analysis on recorded abnormal events, such as abnormal frequency, intensity, duration, fluctuation amplitude, etc., and classify the events into different categories, such as equipment failure, environmental impact or temporary interference. The algorithm improves the accuracy of classification through the voting mechanism of a large number of decision trees, helping to quickly identify and prioritize high-risk and urgent events.

[0058] Preferably, a storage architecture is constructed, specifically, the storage architecture adopts an automatic tiering strategy for hot and cold data, automatically adjusts the storage medium based on the access frequency of the data, and optimizes the storage cost and reading speed.

[0059] As mentioned above, in the storage architecture design, hot data that is frequently accessed recently is stored in SSD or fast-access disk to reduce read latency, while cold data that has not been accessed for a long time is automatically migrated to low-cost hard disk arrays or cloud storage. Intelligent algorithms monitor data access patterns and automatically adjust data storage locations, which optimizes costs and ensures access efficiency.

[0060] Preferably, permission management supports dynamic permission adjustment, that is, instantly adjusting role permissions according to changes in operation and maintenance scenarios to enhance system flexibility.

[0061] As mentioned above, the role-based access control (RBAC) system is used to dynamically adjust role permissions according to operation and maintenance tasks or emergency situations. For example, during emergency troubleshooting, more operation permissions are temporarily granted to front-line engineers to quickly respond to problems, and the original permission configuration is restored on a daily basis to ensure the principle of minimizing permissions and enhance system security and flexibility.

[0062] Preferably, it also includes the integration of remote query and analysis tools, specifically, through the remote query interface, to achieve safe and efficient data transmission, provide paging query and filter condition optimization; through the integration of advanced prediction model analysis tools, to achieve anomaly detection, root cause analysis and time series prediction functions.

[0063] As mentioned above, develop a secure remote API interface, follow the RESTful specification, support HTTPS encrypted transmission, and combine WebSocket to achieve real-time two-way data push, improve query response speed and interactivity. The analysis tool integrates machine learning models, such as time series prediction models (ARIMA, LSTM), supports anomaly detection and root cause analysis, and helps operation and maintenance personnel quickly browse and filter data through a graphical interface, identify problem trends, and guide operation and maintenance decisions.

[0064] Preferably, the remote query interface, specifically the remote query interface follows the RESTful API standard, supports HTTPS and WebSocket protocols, realizes two-way real-time data transmission, and improves interactivity and response speed.

[0065] As mentioned above, the remote query interface design follows the RESTful API standard to ensure that the interface is concise, easy to understand and call, and supports HTTPS encrypted transmission to ensure data security. At the same time, the WebSocket protocol is integrated to achieve two-way real-time communication. The data exchange between the client and the server is no longer limited to the request-response mode, but is pushed and updated in real time, greatly improving the interactive experience and data synchronization speed.

[0066] Preferably, it also includes edge collaboration and adaptive learning mechanisms, specifically, implementing edge-to-edge communication of the mesh network based on MQTTCP / IP or Zigbee protocol to ensure low latency and stability of communication between devices; establishing an adaptive learning mechanism based on continuous data feedback to achieve continuous model updating and ensure that the model effectively adapts to load changes and abnormal types.

[0067] As mentioned above, the low power consumption and high reliability characteristics of MQTTCP / IP or Zigbee protocols are used to build a grid communication network to ensure direct, fast and stable data exchange between smart terminals, which can be transmitted through multiple paths even in the event of a single point failure. The adaptive learning mechanism continuously collects the operating data of edge devices and feeds it back to the cloud model training. It uses online learning or incremental learning strategies to continuously optimize model parameters to adapt to dynamic changes in power system loads and emerging abnormal patterns, thereby improving prediction and response capabilities.

[0068] Compared with the existing technology, this method realizes all-round monitoring of voltage, current and environmental parameters in the power system by carefully planning and deploying sensor networks, ensuring real-time and comprehensive grasp of the power system status, solving the problem of difficulty in accurately monitoring multiple loads, and improving the integrity of data collection and system perception capabilities; it uses deep learning models combined with edge computing to quickly identify voltage fluctuation patterns and potential faults, improves the accuracy and response speed of fault warnings, and enables the system to more effectively handle complex power fluctuations and reduce losses caused by faults; it automatically records over-threshold or abnormal data and uses random forest algorithm classification to achieve fast and accurate classification and sorting of events, helping dispatchers to quickly locate and prioritize emergency issues and optimize operation and maintenance efficiency; it builds an efficient storage architecture, adopts hot and cold data stratification strategies and encryption technology, ensures instant response and long-term secure storage of data, reduces operation and maintenance costs, and improves the economy and security of data management; it implements role-based access control and dynamic permission adjustment, enhances system flexibility and security, facilitates operation and maintenance operations, and adapts to the needs of different scenarios; it integrates remote query and analysis tools, and uses RESTful API and WebSocket protocols improve the efficiency of data interaction, and combined with advanced analysis models, they realize accurate anomaly detection, root cause analysis and prediction, and enhance decision-making support capabilities. The use of MQTTCP / IP or Zigbee protocols to achieve edge-to-edge communication ensures low latency and stability between devices, and the adaptive learning mechanism enables the model to be continuously updated to maintain adaptability to load changes and anomaly types, thereby enhancing the system's intelligence level and self-optimization capabilities.

[0069] Through the description of the above implementation methods, those skilled in the art can clearly understand that the present invention can be implemented by hardware, or by software plus a necessary general hardware platform. Based on such an understanding, the technical solution of the present invention can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.), including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each implementation scenario of the present invention.

[0070] Those skilled in the art will appreciate that the accompanying drawings are merely schematic diagrams of a preferred implementation scenario, and the modules or processes in the accompanying drawings are not necessarily required for implementing the present invention.

[0071] Those skilled in the art will appreciate that the modules in the device in the implementation scenario can be distributed in the device in the implementation scenario according to the implementation scenario description, or can be changed accordingly and located in one or more devices different from the implementation scenario. The modules in the above implementation scenario can be combined into one module, or can be further split into multiple sub-modules.

[0072] The above serial numbers of the present invention are only for description and do not represent the advantages or disadvantages of the implementation scenarios.

[0073] The above disclosure is only a few specific implementation scenarios of the present invention, but the present invention is not limited thereto, and any changes that can be conceived by those skilled in the art should fall within the protection scope of the present invention.

Claims

1. An event recording method for an intelligent flexible adjustment unit, It is characterized in that include: Plan and deploy sensor networks for the power system to fully cover the voltage, current, and environmental parameter monitoring needs and ensure comprehensive monitoring of the power system status; Through edge computing models, cloud computing and IoT technologies are combined to achieve real-time processing and analysis of sensor data, and rule engines or machine learning algorithms are used to quickly identify threshold-exceeding anomalies; Based on the event capture mechanism, abnormal data that exceeds the threshold or is identified by the model is automatically recorded, including the time period before and after the abnormality, timestamp, location, and device identification; Through machine learning classification algorithms, captured events are intelligently classified, based on historical data training, and based on abnormal characteristics, the urgency is quickly determined and sorted; Build a storage architecture through SSD cache and hard disk array or cloud storage to ensure immediate response and long-term storage of data based on data lifecycle management strategy; The AES-256-bit encryption algorithm is used to ensure the security of transmission and static data, and role-based access control is used to achieve detailed permission management.

2. According to the method of claim 1, the planning and deployment of the sensor network for the power system is specifically: The sensor network includes voltage, current, power and environmental parameter sensors, and each of the sensors is deployed in redundant quantities at key nodes, lines and inside equipment to ensure that monitoring continuity is not affected by single point failures.

3. According to the method of claim 1, the edge computing model is specifically: Through deep learning algorithm models, complex voltage fluctuation patterns are deeply analyzed to identify potential circuit failure problems in advance.

4. According to the method of claim 1, the automatic recording of abnormal data exceeding the threshold or identified by the model is specifically: Abnormal data is instantly stored locally and uploaded to the cloud, achieving dual backup of local and cloud data to ensure data security.

5. According to the method of claim 1, the intelligent classification of the captured events is specifically: Through the random forest algorithm, abnormal data is classified according to multiple dimensions such as frequency, intensity and duration to improve the accuracy of event classification.

6. According to the method of claim 1, the building of the storage architecture comprises: The storage architecture adopts an automatic tiering strategy for hot and cold data, automatically adjusts the storage medium based on the access frequency of the data, and optimizes storage costs and reading speeds.

7. According to the method described in claim 1, the permission management supports dynamic permission adjustment, that is, the role permissions are adjusted in real time according to changes in operation and maintenance scenarios to enhance system flexibility.

8. The method according to claim 1 further comprises integrating remote query and analysis tools, specifically: Through the remote query interface, secure and efficient data transmission is achieved, and paging query and filter condition optimization are provided; Anomaly detection, root cause analysis, and time series forecasting are all available through analytical tools that integrate advanced predictive models.

9. According to the method of claim 8, the remote query interface is specifically: The remote query interface complies with the RESTful API standard, supports HTTPS and WebSocket protocols, realizes two-way real-time data transmission, and improves interactivity and response speed.

10. The method according to claim 1 further comprises an edge collaboration and adaptive learning mechanism, specifically: Implement edge-to-edge communication in mesh networks based on MQTTCP / IP or Zigbee protocols to ensure low latency and stability of communication between devices; An adaptive learning mechanism is established based on continuous data feedback to continuously update the model and ensure that the model effectively adapts to load changes and abnormal types.

Images