Power edge optimization control method and system based on IoT cloud-edge collaboration

Abstract

The present invention discloses an electric power edge optimization control method and system based on the cloud-edge collaboration of the Internet of Things, which relates to the relevant fields of the electric power Internet of Things. The method includes: constructing a cloud-based Internet of Things management platform and an edge-side general platform; calling the flow log, performing log analysis, and establishing a time-series cloud-based policy meta-feature; performing edge node self-checking, constructing independent monitoring abnormal indicators, performing impact clustering, and generating similar clustering results; configuring an abnormal monitoring network, performing abnormal analysis, and obtaining abnormal values; calling the cloud-based Internet of Things management platform, performing real-time edge node data aggregation and reading, and constructing a central data set; performing cloud-edge collaborative analysis, generating control warning results, and performing electric power edge optimization control management. The method solves the technical problems of limited data processing capability, insufficient real-time abnormal response capability, and poor system stability in existing electric power edge control, and achieves the technical effects of improving data processing capability, enhancing abnormal real-time response capability, and enhancing system stability.

Description

Technical Field

[0001] The present application relates to fields related to the electric power Internet of Things, and in particular to an electric power edge-side optimization control method and system based on the Internet of Things cloud-edge collaboration. Background Art

[0002] With the rapid development of the power Internet of Things, the power system is facing increasingly complex monitoring and management needs. Traditional power monitoring methods mainly rely on centralized data processing systems. This model is competent when the amount of data is relatively small, but when faced with massive amounts of power data, its processing capacity is limited and it is difficult to complete the analysis and processing of large amounts of data in a short period of time. In addition, centralized data processing systems usually adopt a centralized architecture, and the transmission and processing of data are often accompanied by significant delays, resulting in untimely responses to abnormal conditions of power-side equipment and even missing the best time for intervention.

[0003] Among the current related technologies, power side control has technical problems such as limited data processing capabilities, insufficient real-time abnormal response capabilities, and poor system stability. Summary of the invention

[0004] This application provides an electric power edge optimization control method and system based on IoT cloud-edge collaboration, adopts technical means such as building a cloud-based IoT management platform and an edge-side general platform, establishing time-series cloud-based policy meta-features, building independent monitoring abnormality indicators for edge nodes, performing abnormality analysis, building a central data set, and conducting cloud-edge collaborative analysis, thereby achieving cloud-edge collaborative work and achieving the technical effect of improving data processing capabilities, enhancing real-time response capabilities to abnormalities, and enhancing system stability.

[0005] This application provides a power edge optimization control method based on IoT cloud-edge collaboration, including:

[0006] Construct a cloud IoT management platform and an edge general platform, and configure a standard API interface, wherein the cloud IoT management platform and the edge general platform communicate digitally based on the standard API interface; configure a log verification cycle, call the flow log of the cloud IoT management platform during the log verification cycle, perform log analysis, and establish a time-series cloud-based policy meta-feature; perform edge node self-check based on the edge general platform, construct an independent monitoring abnormality index of the edge node, cluster the impact of the edge node through the independent monitoring abnormality index and the time-series cloud-based policy meta-feature, and generate similar clustering results; configure an abnormal monitoring network, input the similar clustering results into the abnormal monitoring network, perform abnormal analysis, and obtain abnormal values, wherein the abnormal values ​​carry an abnormal type identifier; determine whether the abnormal value meets a preset abnormal threshold, and if the abnormal value meets the preset abnormal threshold, call the cloud IoT management platform, perform real-time edge node data aggregation and reading, and construct a central data set; perform cloud-edge collaborative analysis based on the central data set, the independent monitoring abnormality index and the abnormal value, generate a control warning result, and perform power edge optimization control management based on the control warning result.

[0007] In a possible implementation, the configuration abnormality monitoring network performs the following processing:

[0008] The anomaly monitoring network is constructed based on the anomaly evaluation function, and the anomaly evaluation function formula is as follows: Among them, A is the abnormal evaluation function, n is the total number of edge nodes, i represents any edge node, C i represents the independent monitoring abnormality index of the i-th edge node, C cloud is the global anomaly index constructed based on log analysis, ζ is the time series decay factor, k is the number of nearest neighbor edge nodes considered in the adjacent analysis, j represents the jth edge node, S ij is the similarity score between the ith edge node and the jth edge node constructed based on similar clustering results, v1, v2, ω Sim are the weights of independent monitoring anomaly indicators, global anomaly indicators and similarity scores respectively.

[0009] In a possible implementation, the impact clustering of edge nodes is performed by using the independent monitoring abnormal indicators and the time series cloud-based strategy meta-features to generate similar clustering results, and the following processing is performed:

[0010] Construct the time series decay function as follows: σ(t) = e -ρ(T-t); Wherein, σ(t) is the time series decay function, ρ is the decay rate, T is the current time, and t is the feature timestamp of the time series cloud-based strategy meta-feature; the time series cloud-based strategy meta-feature is used as the global feature, and the independent monitoring abnormality indicator is used as the original feature. Feature fusion is performed based on the original feature and the global feature to construct a fusion feature; the distance between edge nodes is calculated through the fusion feature, and the initial similarity score is constructed. The time series time impact analysis of the initial similarity score is performed through the time series decay function to generate a similar clustering result.

[0011] In a possible implementation, the distance between edge nodes is calculated by the fusion feature, and an initial similarity score is constructed. The time series time impact analysis of the initial similarity score is performed by a time series decay function to generate a similar clustering result, and the following processing is performed:

[0012] Construct the distance calculation formula as follows: Among them, D ij represents the distance between any two edge nodes ij, m is the total number of features in the fusion feature, h is the index of traversing each fusion feature, z ih is the fused feature value of edge node i on the hth feature, z jh is the fused feature value of edge node j on the hth feature; construct the similarity score calculation formula as follows: Among them, β is a constant, which is used to adjust the rate at which the similarity decreases with increasing distance; the distance between edge nodes is calculated by the distance calculation formula and the similarity score calculation formula, and the initial similarity score is constructed.

[0013] In a possible implementation, the determining whether the abnormal value meets a preset abnormal threshold value is to perform the following processing:

[0014] If the abnormal value fails to meet the preset abnormal threshold, abnormal attention data is reported; an abnormal identification window is configured through the abnormal attention data, periodic sensitive monitoring is performed through the abnormal identification window, and a sensitive monitoring data set is constructed; cloud-edge fusion analysis is performed on the sensitive monitoring data set to generate additional warning information; and power edge optimization control management is performed through the additional warning information.

[0015] In a possible implementation, the following processing is performed:

[0016] If no additional warning information is created, a time series attenuation sensitivity factor is generated; and continuous abnormality monitoring is performed through the time series attenuation sensitivity factor.

[0017] In a possible implementation, the cloud-edge collaborative analysis is performed based on the central data set, the independent monitoring abnormal indicators and abnormal values, a control warning result is generated, and the following processing is performed:

[0018] The abnormal collaborative database is called through the abnormal value; based on the abnormal collaborative database, the independent monitoring abnormal index and the central data set are subjected to abnormal collaborative identification to generate a control warning result.

[0019] This application also provides a power edge optimization control system based on IoT cloud-edge collaboration, including:

[0020] A platform construction module, the platform construction module is used to build a cloud-based IoT management platform and an edge-side general platform, and configure a standard API interface. The cloud-based IoT management platform and the edge-side general platform communicate digitally based on the standard API interface; a timing cloud-based policy meta-feature establishment module, the timing cloud-based policy meta-feature establishment module is used to configure a log verification cycle, call the flow log of the cloud-based IoT management platform during the log verification cycle, perform log analysis, and establish a timing cloud-based policy meta-feature; an impact clustering module, the impact clustering module is used to perform edge node self-check based on the edge-side general platform, build independent monitoring anomaly indicators of edge nodes, cluster the impact of edge nodes through the independent monitoring anomaly indicators and the timing cloud-based policy meta-features, and generate similar clustering results; Anomaly analysis module, the anomaly analysis module is used to configure an anomaly monitoring network, input the similarity clustering results into the anomaly monitoring network, perform anomaly analysis, and obtain anomaly values, wherein the anomaly values ​​carry an anomaly type identifier; a central data set construction module, the central data set construction module is used to determine whether the anomaly value meets a preset anomaly threshold. If the anomaly value meets the preset anomaly threshold, the cloud-based IoT management platform is called to perform real-time edge node data aggregation and reading to construct a central data set; an electric power edge-side optimization control management module, the electric power edge-side optimization control management module is used to perform cloud-edge collaborative analysis based on the central data set, the independent monitoring anomaly indicators and anomaly values, generate control warning results, and perform electric power edge-side optimization control management based on the control warning results.

[0021] The power edge optimization control method and system based on the Internet of Things cloud-edge collaboration proposed in this application is to first build a cloud-based IoT management platform and an edge-side general platform, and configure a standard API interface. The cloud-based IoT management platform and the edge-side general platform communicate digitally based on the standard API interface, and then configure the log verification cycle. During the log verification cycle, the flow log of the cloud-based IoT management platform is called, and log analysis is performed to establish a time-series cloud-based policy meta-feature. Then, based on the edge-side general platform, edge node self-inspection is performed, and independent monitoring abnormality indicators of edge nodes are constructed. The impact of edge nodes is clustered through independent monitoring abnormality indicators and time-series cloud-based policy meta-features to generate similar clusters. Classify the results, then configure the anomaly monitoring network, input the similar clustering results into the anomaly monitoring network, perform anomaly analysis, obtain the anomaly value, the anomaly value has the anomaly type identifier, and then judge whether the anomaly value meets the preset anomaly threshold. If the anomaly value meets the preset anomaly threshold, call the cloud IoT management platform, perform real-time edge node data aggregation and reading, build a central data set, and finally perform cloud-edge collaborative analysis based on the central data set, independent monitoring anomaly indicators and anomaly values ​​to generate control warning results. Based on the control warning results, optimize the control management of the power side, which achieves the technical effect of improving data processing capabilities, enhancing real-time response capabilities to anomalies, and enhancing system stability. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solution of the embodiment of the present invention, the accompanying drawings of the embodiment of the present invention will be briefly introduced below. A flow chart is used in the present application to illustrate the operations performed by the system according to the embodiment of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps can be processed in reverse order or simultaneously as needed. At the same time, other operations can also be added to these processes, or one or more operations can be removed from these processes.

[0023] Figure 1 A flow chart of a power edge-side optimization control method based on IoT cloud-edge collaboration provided in an embodiment of the present application.

[0024] Figure 2 A schematic diagram of the structure of a power edge-side optimization control system based on IoT cloud-edge collaboration provided in an embodiment of the present application.

[0025] Explanation of the accompanying drawings: platform construction module 10, time-series cloud-based strategy meta-feature establishment module 20, impact clustering module 30, anomaly analysis module 40, central data set construction module 50, power edge optimization control management module 60. DETAILED DESCRIPTION

[0026] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below.

[0027] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings. The described embodiments should not be regarded as limiting the present application. All other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of this application.

[0028] In the following description, reference is made to "some embodiments", which describe a subset of all possible embodiments, but it is understood that "some embodiments" may be the same subset or different subsets of all possible embodiments, and may be combined with each other without conflict. The terms "including" and "having" and any variations are intended to cover non-exclusive inclusions. For example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or modules that are not clearly listed or inherent to these processes, methods, products, or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by technicians in the technical field of this application. The terms used herein are for the purpose of describing the embodiments of the present application only.

[0029] The present application embodiment provides a power edge optimization control method based on IoT cloud-edge collaboration, such as Figure 1 As shown, the method includes:

[0030] Step S100, construct a cloud IoT management platform and a side general platform, and configure a standard API interface, the cloud IoT management platform and the side general platform are based on the standard API interface digital communication. Specifically, design and develop a cloud IoT management platform, the cloud IoT management platform is located in the cloud, and is an integrated platform that integrates device connection, message routing, and data processing and analysis capabilities. It supports connecting massive IoT devices downward, collecting device data on the cloud, and provides standard API interface (application programming interface, which defines the interaction method and data format between software programs) services upward. It can send instructions to the IoT device end by calling the platform API interface to achieve remote control. The cloud IoT management platform provides data collection and aggregation, data caching, data service capabilities, as well as IoT application development, platform operation and maintenance, security management and other capabilities. Construct an edge-side general platform, which is located at the edge side of the IoT device and provides edge computing capabilities extended from the cloud IoT management platform. Based on edge intelligent gateways or devices, it can connect various IoT sensor terminals, realize access and conversion of various IoT protocols, and realize data collection and aggregation; on the other hand, it can collaborate with the cloud IoT management platform to provide data services, intelligent algorithm analysis and other extended application capabilities. The cloud IoT management platform and the edge-side general platform perform efficient and stable data exchange and communication through the defined standard API interface.

[0031] Step S200, configure the log verification cycle, call the flow log of the cloud IoT management platform during the log verification cycle, perform log analysis, and establish the time-series cloud-based policy meta-feature. Specifically, according to business needs and the frequency of system log generation, determine the log verification cycle, that is, the time interval for verifying the logs generated by the system or application, such as daily, hourly, every minute, etc., set a scheduled task in the system or application so that the log verification process can be automatically triggered when the set log verification cycle arrives. In each log verification cycle, send a request to the cloud IoT management platform to request to obtain the flow log within the specified time range, receive the response of the cloud IoT management platform, and obtain the flow log data, wherein the flow log refers to the log generated during the use of the IoT device, which records the device status changes, operation records and other information. Perform pre-processing operations such as cleaning, sorting, and formatting on the received flow log to ensure data quality and consistency, process and analyze the flow log data according to the defined rules and logic (such as identifying abnormal patterns, calculating statistical indicators, etc.), extract valuable information, identify abnormalities or potential problems, and evaluate system performance. Extract business-related features with time series characteristics from the results of log analysis, encode and standardize the extracted features to meet the requirements of time series cloud-based strategies. The time series cloud-based strategies refer to management and control strategies based on time series data (such as log data generated by IoT devices); meta-features refer to features with specific meanings and uses used when building models or formulating strategies. In time series cloud-based strategies, meta-features are used to describe the characteristics or patterns of time series data. Therefore, the time series cloud-based strategy meta-features are key features extracted after time series analysis of a large amount of historical data based on the cloud-based IoT management platform. These features reflect trends and laws at the global or macro level.

[0032] Step S300, based on the edge general platform, perform edge node self-check, construct independent monitoring abnormality indicators of edge nodes, cluster the impact of edge nodes through the independent monitoring abnormality indicators and the time-series cloud-based policy meta-features, and generate similar clustering results. Specifically, on the edge general platform, start the self-check program for edge nodes (terminal devices in the Internet of Things system, such as sensors, actuators, etc.), the self-check program collects various operating data of edge nodes, such as CPU usage, memory usage, disk space, network status, etc., analyzes the collected data, and determines whether the edge nodes are operating normally or have potential problems. According to the characteristics and business requirements of edge nodes, a series of independent monitoring abnormality indicators are defined. The independent monitoring abnormality indicators are standards or parameters defined separately for each edge node to determine whether it is abnormal, such as high CPU usage, abnormal memory usage, network delay, etc., set a reasonable threshold for each independent monitoring abnormality indicator, when the data exceeds the threshold, it is considered that the independent monitoring abnormality indicator is abnormal, and the various data of the edge node are monitored in real time through the self-check program, and compared with the abnormality indicator threshold. The independent monitoring abnormal indicator data and the time-series cloud-based strategy meta-feature data are integrated together, and the features that have an important impact on the clustering results are selected from the integrated data. According to the characteristics of the data and clustering requirements, appropriate clustering algorithms such as K-means and DBSCAN are used to perform impact clustering on edge nodes. That is, according to the abnormal conditions and global trends of edge nodes, edge nodes with similar impacts or similar anomalies are clustered into one category, and the clustering results are output in a visual way or in the form of data reports to obtain similar clustering results.

[0033] In a possible implementation, the impact clustering of edge nodes is performed by the independent monitoring abnormal indicators and the time series cloud-based strategy meta-features to generate similar clustering results. Step S300 further includes step S310, constructing a time series attenuation function, as follows:

[0034] σ(t)=e -ρ(T-t) ;

[0035] Among them, σ(t) is the time-series decay function, ρ is the decay rate, T is the current time, and t is the feature timestamp of the time-series cloud-based strategy meta-feature. Specifically, the time-series decay function is used to calculate the influence of time difference on feature weight, the decay rate represents the speed at which the influence of time on the feature decays, the current time is the time point when cluster analysis is performed, and the feature timestamp is the time point when the time-series cloud-based strategy meta-feature is recorded or generated. Step S320, taking the time-series cloud-based strategy meta-feature as a global feature, taking the independent monitoring abnormality index as an original feature, and performing feature fusion based on the original feature and the global feature to construct a fused feature. Specifically, the time-series cloud-based strategy meta-feature is taken as a global feature because the time-series cloud-based strategy meta-feature reflects the state of the entire system or network; the independent monitoring abnormality index is taken as an original feature because the independent monitoring abnormality index directly reflects the state of the edge node. Use feature fusion methods (such as concatenation, addition, weighting, etc.) to combine global features and original features to construct fused features, and the fused features contain both global information and specific information of edge nodes. Step S330, calculate the distance between edge nodes through the fusion feature, and construct an initial similarity score, perform temporal time impact analysis of the initial similarity score through the time series decay function, and generate similar clustering results. Specifically, use a distance measurement method (such as Euclidean distance, cosine similarity, etc.) to calculate the distance between edge nodes based on the fusion feature, and construct an initial similarity score for each edge node pair based on the calculated distance. The closer the distance, the higher the similarity score. Combine the initial similarity score with the time series decay function to perform a temporal time impact analysis. The initial similarity score can be multiplied by the time series decay function to reflect the impact of time on the initial similarity. Based on the adjusted initial similarity score, a clustering algorithm is used to cluster the edge nodes to generate similar clustering results. This implementation method generates more accurate similar clustering results by considering time factors and fusion features, thereby achieving the technical effect of improving the accuracy of clustering.

[0036] In a possible implementation, step S330 further includes step S331, constructing a distance calculation formula as follows:

[0037]

[0038] Among them, D ij represents the distance between any two edge nodes ij, m is the total number of features in the fusion feature, h is the index of traversing each fusion feature, z ih is the fused feature value of edge node i on the hth feature, z jh is the fused feature value of edge node j on the hth feature; Step S332, construct a similarity score calculation formula as follows:

[0039]

[0040] Among them, β is a constant used to adjust the rate at which the similarity decreases with increasing distance; step S333, calculate the distance between edge nodes through the distance calculation formula and the similarity score calculation formula, and construct an initial similarity score. Specifically, use the distance calculation formula constructed in step S331 to calculate the distance between any two edge nodes, substitute the calculated distance into the similarity score calculation formula constructed in step S332, and obtain the initial similarity score between edge nodes. This implementation method accurately quantifies the difference in fusion features between any two edge nodes by constructing a distance calculation formula, and by adjusting the constant β, it can flexibly control the rate at which the similarity changes with distance, thereby achieving the technical effect of reasonably measuring the similarity between edge nodes.

[0041] Step S400, configure an anomaly monitoring network, input the similar clustering results into the anomaly monitoring network, perform anomaly analysis, and obtain anomaly values, which carry an anomaly type identifier. Specifically, according to business needs and data analysis results, the structure and parameters of the anomaly monitoring network are designed, including selecting a suitable machine learning or deep learning model, such as a neural network, decision tree, random forest, etc., and determining the input and output of the model, using historical data or simulated data to train the model so that the model learns the patterns of normal and abnormal behaviors so that new abnormal events can be accurately identified, and the trained model is deployed in the anomaly monitoring network, which is used to monitor and analyze abnormal events in data streams in real time. The generated similar clustering results are used as input data, and after preprocessing (such as standardization, normalization, encoding, etc.), they are input into the anomaly monitoring network to perform anomaly analysis. The anomaly monitoring network evaluates the input data according to the learned normal and abnormal modes, and outputs the corresponding anomaly value, which is a quantitative indicator of the severity of the data point or event detected in the anomaly monitoring process that is significantly different from the normal mode, indicating that the system or equipment has a fault, error or abnormal situation, including the identifier of the abnormal event, the degree of abnormality, the type of abnormality, and other information. Among them, each anomaly value carries an abnormality type identifier, that is, a label or identifier that identifies the type of abnormal event, such as voltage anomaly, current anomaly, etc., which is used to help understand the nature and impact of the abnormal event.

[0042] In a possible implementation, the step S400 of configuring the abnormality monitoring network further includes a step S410 of constructing the abnormality monitoring network based on an abnormality evaluation function, wherein the abnormality evaluation function formula is as follows:

[0043]

[0044] Among them, A is the abnormal evaluation function, n is the total number of edge nodes, i represents any edge node, C irepresents the independent monitoring abnormality index of the i-th edge node, C cloud is the global anomaly index constructed based on log analysis, ζ is the time series decay factor, k is the number of nearest neighbor edge nodes considered in the adjacent analysis, j represents the jth edge node, S ij is the similarity score between the ith edge node and the jth edge node constructed based on similar clustering results, ω1, ω2, ω sim They are the weights of the independent monitoring anomaly index, the global anomaly index and the similarity score. Specifically, the anomaly evaluation function is a function used to quantify the degree of abnormality of edge nodes, the independent monitoring anomaly index is a standard or parameter defined separately for each edge node and used to determine whether it is abnormal, the global anomaly index is an indicator based on the data of the entire system or network that reflects the overall abnormal state, the similarity score is a numerical value indicating the degree of similarity between two edge nodes, the weight parameter is a parameter used to adjust the relative importance of different indicators in the anomaly evaluation, and the time series attenuation factor is a factor used to consider the impact of historical data on the current anomaly evaluation. This implementation method comprehensively evaluates the abnormal state of edge nodes by comprehensively considering the independent monitoring anomaly index, the global anomaly index and the similarity score, achieving the technical effect of improving the accuracy and reliability of anomaly analysis.

[0045] Step S500, determine whether the abnormal value meets the preset abnormal threshold. If the abnormal value meets the preset abnormal threshold, call the cloud IoT management platform, perform real-time edge node data aggregation and reading, and build a central data set. Specifically, according to business needs or system characteristics, a threshold for judging whether the abnormal value has reached a certain severity or needs to take further measures is pre-set, and the acquired abnormal value is compared with the preset abnormal threshold. When the abnormal value meets the preset abnormal threshold, the operation of calling the cloud IoT management platform is triggered by sending an API request, triggering a message in a message queue, or using other communication mechanisms, and a communication connection is established with the cloud IoT management platform. Through the cloud IoT management platform, the data of the relevant edge nodes, including the operating status of the equipment, sensor data, logs, etc., are read in real time, and the read data is aggregated and sorted, including data cleaning, deduplication, format conversion and other operations. After aggregation, a central data set is obtained, that is, a data set aggregated from multiple edge nodes for further analysis.

[0046] In a possible implementation, the step S500 further includes step S510, if the abnormal value fails to meet the preset abnormal threshold, abnormal attention data is reported. Specifically, if the abnormal value fails to meet the preset abnormal threshold (for example, the abnormal value is less than the preset abnormal threshold, or the rate of increase of the abnormal value is lower than a certain set value), it is considered that the abnormal value exists, but it is not enough to trigger the control warning. In this case, abnormal attention data is reported, including the specific value of the abnormal value, the time of occurrence, the relevant equipment or system information, etc. Step S520, configure the abnormal recognition window through the abnormal attention data, perform periodic sensitive monitoring through the abnormal recognition window, and construct a sensitive monitoring data set. Specifically, after reporting the abnormal attention data, an abnormal recognition window is configured according to the abnormal attention data, and the abnormal recognition window can be a time range or a monitoring range of a specific device or system. Periodic sensitive monitoring is performed in the abnormal recognition window, that is, the data changes in the abnormal recognition window are continuously monitored according to a certain time interval or condition. Through the monitoring process, a sensitive monitoring data set is collected and constructed, and the sensitive monitoring data set contains a series of abnormal values ​​or abnormal events generated in the abnormal recognition window. Step S530, perform cloud-edge fusion analysis on the sensitive monitoring data set to generate additional warning information. Specifically, after collecting the sensitive monitoring data set, various data analysis algorithms, machine learning models, etc. in the cloud-edge fusion analysis technology are used to process the sensitive monitoring data set, combining the powerful data processing capabilities of cloud computing with the real-time and low latency of edge computing, and deeply analyzing and mining the sensitive data to generate additional warning information, that is, information used to guide abnormal processing or optimization. Step S540, optimize the control and management of the power side through the additional warning information. Specifically, the additional warning information is used for the optimization control and management of the power side (i.e., the distribution network or the power consumption side), including adjusting the operating parameters of the power equipment, optimizing the energy allocation strategy, and early warning of potential faults. This implementation method avoids underreporting by reporting abnormal attention data, more accurately identifies abnormal events that need attention, and achieves the technical effect of improving the comprehensiveness and accuracy of abnormal detection.

[0047] In a possible implementation, step S530 further includes step S531, if no additional warning information is created, a time-series decay sensitivity factor is generated. Specifically, when the cloud-edge fusion analysis fails to directly generate additional warning information, a time-series decay sensitivity factor is generated based on the analysis results of the abnormal data, and the time-series decay sensitivity factor is used to reflect the characteristics of abnormal events decaying over time, so as to more accurately evaluate the current attention of abnormal events, that is, abnormal events may have a high degree of attention for a period of time after they occur, but as time goes by, their attention will gradually decrease. Step S532, continuous abnormal monitoring is performed through the time-series decay sensitivity factor. Specifically, the attention to abnormal events is dynamically adjusted according to the time-series decay sensitivity factor, that is, the data in the sensitive monitoring data set is continued to be monitored, and the current attention of each abnormal event is evaluated according to the time-series decay sensitivity factor. If the attention of a certain abnormal event exceeds the preset threshold, the corresponding additional warning information is generated. This implementation method more accurately assesses the current attention of abnormal events by generating time-series attenuation sensitivity factors and performing continuous anomaly monitoring, ensures timely response and effective handling of abnormal situations, reduces unnecessary waste of resources, and achieves the technical effect of efficient use of resources.

[0048] Step S600, based on the central data set, the independent monitoring abnormal indicators and abnormal values, cloud-edge collaborative analysis is performed to generate control warning results, and power edge optimization control management is performed based on the control warning results. Specifically, the central data set, the independent monitoring abnormal indicators and abnormal values ​​are integrated to maintain the consistency and complementarity of these data sources in logic and structure, and the powerful computing power and data resources of the cloud are used to perform in-depth analysis on the integrated data, including but not limited to algorithms and technologies such as data mining and pattern recognition, and to make collaborative decisions in combination with the real-time nature of edge computing and the global nature of cloud computing, wherein edge computing is used to provide real-time local data support, and cloud computing is used to provide global data analysis and strategy recommendations. Based on the results of cloud-edge collaborative analysis, the existing abnormalities or potential problems are identified, and the corresponding control warning results are generated according to the type, degree and potential impact of the abnormalities, that is, the warning information used to guide the optimization control management of the power edge, including the specific description of the abnormality, the time of occurrence, the scope of impact and other information. According to the control warning results, the corresponding optimization control strategy is generated, including adjusting power distribution, optimizing equipment operating parameters, switching operating modes, etc., and the effect of the optimization control strategy after execution is continuously monitored and evaluated to verify the effectiveness and sustainability of the strategy. The embodiments of the present application adopt technical means such as building a cloud-based IoT management platform and an edge-side general platform, establishing time-series cloud-based policy meta-features, building independent monitoring anomaly indicators for edge nodes, performing anomaly analysis, building a central data set, and conducting cloud-edge collaborative analysis, thereby realizing cloud-edge collaborative work and achieving the technical effects of improving data processing capabilities, enhancing real-time response capabilities to anomalies, and enhancing system stability.

[0049] In a possible implementation, the cloud-edge collaborative analysis is performed based on the central data set, the independent monitoring abnormal indicators and the abnormal values ​​to generate a control warning result. Step S600 further includes step S610, calling the abnormal collaborative database through the abnormal value. Specifically, according to the type, source and context information of the abnormal value, the abnormal collaborative database related to it is selected. The abnormal collaborative database is a pre-established database that stores various abnormal modes, causes, effects and processing strategies, which is used to support abnormal identification, analysis and processing. Through a standard API interface or other means, call the data and information related to the current abnormal value from the abnormal collaborative database. Step S620, based on the abnormal collaborative database, perform abnormal collaborative identification on the independent monitoring abnormal indicators and the central data set to generate a control warning result. Specifically, the independent monitoring abnormal indicators, the central data set and the data and information called from the abnormal collaborative database are merged to form a comprehensive data set, and the algorithms and models in the abnormal collaborative database are used to perform abnormal collaborative identification on the comprehensive data set. According to the results of the abnormal collaborative identification, a control warning result is generated, including the type, level, possible cause, and recommended treatment measures of the abnormality. This implementation method calls the anomaly collaborative database, utilizes historical data and expert knowledge to assist current anomaly identification and analysis, and achieves the technical effect of improving the accuracy and reliability of anomaly identification.

[0050] In the above, refer to Figure 1 The power edge optimization control method based on the Internet of Things cloud-edge collaboration according to an embodiment of the present invention is described in detail. Figure 2 Describe the power edge-side optimization control system based on the Internet of Things cloud-edge collaboration according to an embodiment of the present invention.

[0051] The power edge optimization control system based on the cloud-edge collaboration of the Internet of Things according to the embodiment of the present invention is used to solve the technical problems of limited data processing capability, insufficient real-time abnormal response capability, and poor system stability in the existing power edge control, so as to achieve the technical effects of improving data processing capability, enhancing abnormal real-time response capability, and enhancing system stability. The power edge optimization control system based on the cloud-edge collaboration of the Internet of Things includes: a platform construction module 10, a time-series cloud-based strategy meta-feature establishment module 20, an impact clustering module 30, an abnormality analysis module 40, a central data set construction module 50, and a power edge optimization control management module 60.

[0052] The platform construction module 10 is used to build a cloud-based IoT management platform and an edge-side general platform, and configure a standard API interface. The cloud-based IoT management platform and the edge-side general platform communicate digitally based on the standard API interface; the timing cloud-based policy meta-feature establishment module 20 is used to configure a log verification cycle, call the flow log of the cloud-based IoT management platform during the log verification cycle, and perform log analysis to establish a timing cloud-based policy meta-feature; the impact clustering module 30 is used to perform edge node self-inspection based on the edge-side general platform, build independent monitoring anomaly indicators of edge nodes, cluster the impact of edge nodes through the independent monitoring anomaly indicators and the timing cloud-based policy meta-features, and generate similar clustering results. ; The anomaly analysis module 40 is used to configure the anomaly monitoring network, input the similar clustering results into the anomaly monitoring network, perform anomaly analysis, and obtain anomaly values, wherein the anomaly values ​​carry an anomaly type identifier; the central data set construction module 50 is used to determine whether the anomaly value meets the preset anomaly threshold. If the anomaly value meets the preset anomaly threshold, the cloud-based IoT management platform is called to perform real-time edge node data aggregation and reading to construct a central data set; the power edge optimization control management module 60 is used to perform cloud-edge collaborative analysis based on the central data set, the independent monitoring anomaly indicators and anomaly values, generate control warning results, and perform power edge optimization control management based on the control warning results.

[0053] The specific configuration of the abnormality analysis module 40 will be described in detail below. As described above, the abnormality monitoring network is configured, and the abnormality analysis module 40 may further include: an abnormality monitoring network construction unit is used to construct the abnormality monitoring network based on the abnormality evaluation function, and the abnormality evaluation function formula is as follows: Among them, A is the abnormal evaluation function, n is the total number of edge nodes, i represents any edge node, C i represents the independent monitoring abnormality index of the i-th edge node, C cloud is the global anomaly index constructed based on log analysis, ζ is the time series decay factor, k is the number of nearest neighbor edge nodes considered in the adjacent analysis, j represents the jth edge node, S ij is the similarity score between the ith edge node and the jth edge node constructed based on similar clustering results, ω1, ω2, ω sim are the weights of independent monitoring anomaly indicators, global anomaly indicators and similarity scores respectively.

[0054] The specific configuration of the impact clustering module 30 will be described in detail below. As described above, the impact clustering of edge nodes is performed through the independent monitoring abnormal indicators and the time series cloud-based strategy meta-features to generate similar clustering results. The impact clustering module 30 may further include: a time series attenuation function construction unit for constructing a time series attenuation function, as follows: σ(t) = e -ρ(T-t), where σ(t) is the time series decay function, ρ is the decay rate, T is the current time, and t is the feature timestamp of the time series cloud-based strategy meta-feature; the feature fusion unit is used to take the time series cloud-based strategy meta-feature as the global feature, take the independent monitoring abnormality indicator as the original feature, perform feature fusion based on the original feature and the global feature, and construct a fusion feature; the time series time impact analysis unit is used to calculate the distance between edge nodes through the fusion feature, and construct an initial similarity score, perform time series time impact analysis on the initial similarity score through the time series decay function, and generate a similar clustering result.

[0055] Among them, the distance between edge nodes is calculated by the fusion feature, and an initial similarity score is constructed, and the time series time impact analysis of the initial similarity score is performed through the time series decay function to generate a similar clustering result. The time series time impact analysis unit can further include: a distance calculation formula construction subunit is used to construct a distance calculation formula, as follows: Among them, D ij represents the distance between any two edge nodes ij, m is the total number of features in the fusion feature, h is the index of traversing each fusion feature, z ih is the fused feature value of edge node i on the hth feature, z jh is the fused feature value of edge node j on the hth feature; the similarity score calculation formula construction subunit is used to construct the similarity score calculation formula, as follows: Among them, β is a constant used to adjust the rate at which the similarity decreases with increasing distance; the initial similarity score construction subunit is used to calculate the distance between edge nodes through the distance calculation formula and the similarity score calculation formula, and construct the initial similarity score.

[0056] The specific configuration of the central data set construction module 50 will be described in detail below. As described above, to determine whether the abnormal value meets the preset abnormal threshold, the central data set construction module 50 may further include: an abnormal concern data reporting unit for reporting abnormal concern data if the abnormal value fails to meet the preset abnormal threshold; a periodic sensitive monitoring unit for configuring an abnormal recognition window through the abnormal concern data, performing periodic sensitive monitoring through the abnormal recognition window, and constructing a sensitive monitoring data set; a cloud-edge fusion analysis unit for performing cloud-edge fusion analysis on the sensitive monitoring data set to generate additional warning information; and a power side optimization control management unit for performing power side optimization control management through the additional warning information.

[0057] Among them, the cloud-edge fusion analysis unit may further include: a timing attenuation sensitivity factor generation subunit for generating a timing attenuation sensitivity factor if additional warning information is not created; and a continuous abnormality monitoring subunit for performing continuous abnormality monitoring through the timing attenuation sensitivity factor.

[0058] The specific configuration of the power edge optimization control management module 60 will be described in detail below. As described above, based on the central data set, the independent monitoring abnormal indicators and abnormal values, cloud-edge collaborative analysis is performed to generate a control warning result. The power edge optimization control management module 60 may further include: an abnormal collaborative database calling unit is used to call the abnormal collaborative database through the abnormal value; an abnormal collaborative identification unit is used to perform abnormal collaborative identification on the independent monitoring abnormal indicators and the central data set based on the abnormal collaborative database to generate a control warning result.

[0059] The power edge-side optimization control system based on the Internet of Things cloud-edge collaboration provided by the embodiment of the present invention can execute the power edge-side optimization control method based on the Internet of Things cloud-edge collaboration provided by any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0060] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, any number of different modules may be used and run on the user terminal and / or server, and the various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of the present invention.

[0061] The above specific implementation manner does not constitute a limitation to the protection scope of the present application. It should be understood by those skilled in the art that various modifications, combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present application should be included in the protection scope of the present application. In some cases, the actions or steps recorded in the present application can be performed in an order different from that in the embodiment and can still achieve the desired results. In addition, the process depicted in the accompanying drawings does not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

Claims

1. The power edge optimization control method based on the Internet of Things cloud-edge collaboration is characterized by: The method comprises: Construct a cloud IoT management platform and a side universal platform, and configure a standard API interface, wherein the cloud IoT management platform and the side universal platform communicate digitally based on the standard API interface; Configure a log verification cycle, call the flow log of the cloud IoT management platform during the log verification cycle, perform log analysis, and establish a time-series cloud-based policy meta-feature; Based on the edge-side universal platform, edge node self-check is performed to build independent monitoring abnormality indicators of edge nodes, and the impact of edge nodes is clustered through the independent monitoring abnormality indicators and the time-series cloud-based strategy meta-features to generate similar clustering results; configuring an anomaly monitoring network, inputting the similarity clustering result into the anomaly monitoring network, performing anomaly analysis, and obtaining anomaly values, wherein the anomaly values ​​are provided with an anomaly type identifier; Determine whether the abnormal value meets the preset abnormal threshold. If the abnormal value meets the preset abnormal threshold, call the cloud IoT management platform to perform real-time edge node data aggregation and reading to build a central data set; Perform cloud-edge collaborative analysis based on the central data set, the independent monitoring abnormal indicators and abnormal values, generate control warning results, and perform power edge optimization control management based on the control warning results; The step of clustering the impact of edge nodes by using the independent monitoring abnormal indicators and the time series cloud-based strategy meta-features to generate similar clustering results includes: Construct a time-series decay function as follows: σ(t)=e -ρ(T-t) 4 Among them, σ(t) is the time series decay function, ρ is the decay rate, T is the current time, and t is the feature timestamp of the time series cloud-based strategy meta-feature; Taking the time series cloud-based strategy meta-feature as a global feature, taking the independent monitoring abnormality indicator as an original feature, performing feature fusion based on the original feature and the global feature, and constructing a fusion feature; The distance between edge nodes is calculated by the fusion feature, and an initial similarity score is constructed. The time series time impact analysis of the initial similarity score is performed by the time series decay function to generate similar clustering results, including: Construct the distance calculation formula as follows: Among them, D ij represents the distance between any two edge nodes ij, m is the total number of features in the fusion feature, h is the index of traversing each fusion feature, z ih is the fused feature value of edge node i on the hth feature, z jh is the fused feature value of edge node j on the hth feature; Construct the similarity score calculation formula as follows: Among them, β is a constant used to adjust the rate at which similarity decreases with increasing distance; The distance between edge nodes is calculated using the distance calculation formula and the similarity score calculation formula, and the initial similarity score is constructed.

2. The power edge optimization control method based on IoT cloud-edge collaboration as claimed in claim 1 is characterized in that: The configuration anomaly monitoring network further includes: The anomaly monitoring network is constructed based on the anomaly evaluation function, and the anomaly evaluation function formula is as follows: Among them, A is the abnormal evaluation function, n is the total number of edge nodes, u represents any edge node, C i represents the independent monitoring abnormality index of the i-th edge node, C cloud is the global anomaly index constructed based on log analysis, ζ is the time series decay factor, k is the number of nearest neighbor edge nodes considered in the adjacent analysis, j represents the jth edge node, S ij is the similarity score between the ith edge node and the jth edge node constructed based on similar clustering results, ω1, ω2, ω sim are the weights of independent monitoring anomaly indicators, global anomaly indicators and similarity scores respectively.

3. The power edge optimization control method based on IoT cloud-edge collaboration as claimed in claim 1 is characterized in that: The determining whether the abnormal value meets a preset abnormal threshold value further includes: If the abnormal value fails to meet the preset abnormal threshold, abnormal concern data is reported; An abnormality identification window is configured through the abnormal attention data, and periodic sensitive monitoring is performed through the abnormality identification window to construct a sensitive monitoring data set; Performing cloud-edge fusion analysis on the sensitive monitoring data set to generate additional warning information; The additional warning information is used to perform power-side optimization control management.

4. The power edge optimization control method based on IoT cloud-edge collaboration as claimed in claim 3 is characterized in that: The method further comprises: If no additional warning information is created, a time series attenuation sensitivity factor is generated; Continuous abnormality monitoring is performed through the time series attenuation sensitivity factor.

5. The power edge optimization control method based on IoT cloud-edge collaboration as claimed in claim 1 is characterized in that: The performing cloud-edge collaborative analysis based on the central data set, the independently monitored abnormal indicators and abnormal values ​​to generate a control warning result also includes: calling an abnormal collaborative database through the abnormal value; Based on the abnormal collaborative database, the independent monitoring abnormal indicators and the central data set are subjected to abnormal collaborative identification to generate a control warning result.

6. The power edge optimization control system based on IoT cloud-edge collaboration is characterized by: The system is used to implement the power edge optimization control method based on the Internet of Things cloud-edge collaboration according to any one of claims 1 to 5, and the system includes: A platform construction module, which is used to construct a cloud IoT management platform and a side universal platform, and configure a standard API interface. The cloud IoT management platform and the side universal platform communicate digitally based on the standard API interface; A time-series cloud-based policy meta-feature establishment module, wherein the time-series cloud-based policy meta-feature establishment module is used to configure a log verification cycle, call the flow log of the cloud-based IoT management platform during the log verification cycle, perform log analysis, and establish a time-series cloud-based policy meta-feature; An impact clustering module, the impact clustering module is used to perform edge node self-check based on the edge-side universal platform, build independent monitoring abnormality indicators of edge nodes, perform impact clustering of edge nodes through the independent monitoring abnormality indicators and the time-series cloud-based strategy meta-features, and generate similar clustering results; An anomaly analysis module, the anomaly analysis module is used to configure an anomaly monitoring network, input the similarity clustering result into the anomaly monitoring network, perform an anomaly analysis, and obtain an anomaly value, wherein the anomaly value has an anomaly type identifier; A central data set construction module, wherein the central data set construction module is used to determine whether the abnormal value meets a preset abnormal threshold. If the abnormal value meets the preset abnormal threshold, the cloud IoT management platform is called to perform real-time edge node data aggregation and reading to construct a central data set; The power edge optimization control management module is used to perform cloud-edge collaborative analysis based on the central data set, the independently monitored abnormal indicators and abnormal values, generate control warning results, and perform power edge optimization control management based on the control warning results.

Images