Multi-dimensional Intelligent Early Warning Method for Power Transmission and Transformation Safety

Abstract

The present application provides a multi-dimensional intelligent early warning method for power transmission and transformation safety, which relates to the technical field of power transmission and transformation safety, and includes: establishing an information database of power transmission and transformation lines; establishing line collaborative associations of power transmission and transformation lines; establishing a basic linkage unit for sensors, and establishing a set of weak classifiers mapped thereto through the basic linkage unit; establishing an abnormal recognition result; establishing a trust penalty weight for the abnormal recognition result; performing collaborative authentication on the abnormal recognition result through real-time line collaborative associations, establishing a collaborative authentication coefficient, inputting the collaborative authentication coefficient, the trust penalty weight, and the abnormal recognition result into an abnormal analysis network, and establishing an abnormal early warning. Through the present application, the technical problem in the prior art that false alarms and missed alarms are easily generated due to the simple early warning mechanism, further affecting the accuracy of early warning can be solved, the comprehensive monitoring and intelligent early warning of the operation state of power transmission and transformation lines can be realized, and the technical effects of reducing false alarms and missed alarms and improving the accuracy of early warning can be achieved.

Description

Technical Field

[0001] This application relates to the technical field of power transmission and transformation safety, and particularly to a multi-dimensional intelligent early warning method for power transmission and transformation safety. Background Art

[0002] Power transmission and transformation lines are an important part of the power system, and their safe and stable operation is crucial for ensuring power supply. However, during operation, power transmission and transformation lines may be affected by various factors, such as weather conditions, equipment aging, and human damage. These factors may cause line failures and affect the normal operation of the power system.

[0003] Currently, for the safety monitoring of power transmission and transformation lines, a variety of monitoring technologies and early warning methods have been developed. These technologies usually involve deploying various monitoring sensors along the power transmission and transformation lines, and analyzing the line status by collecting the operation data of the lines. However, the power transmission line is a complex structure with cross-connections between different parts. Depending on simple threshold judgments, false alarms and missed alarms are likely to occur.

[0004] In summary, there is a technical problem in the prior art that due to the simple early warning mechanism, false alarms and missed alarms are likely to occur, further affecting the accuracy of early warning. Summary of the Invention

[0005] The purpose of this application is to provide a multi-dimensional intelligent early warning method for power transmission and transformation safety, so as to solve the technical problem in the prior art that due to the simple early warning mechanism, false alarms and missed alarms are likely to occur, further affecting the accuracy of early warning.

[0006] In view of the above problems, this application provides a multi-dimensional intelligent early warning method for power transmission and transformation safety, including: establishing an information database of the power transmission and transformation line, where the information database stores the line layout information and multi-dimensional monitoring sensor layout information of the power transmission and transformation line; calling the line layout information in the information database, performing line layout information parsing, and establishing the line collaborative association of the power transmission and transformation line; establishing a basic linkage unit of the sensor based on the multi-dimensional monitoring sensor layout information, and establishing a set of weak classifiers mapped thereto through the basic linkage unit; obtaining the line tasks of the power transmission and transformation line, performing line collaborative association calls based on the line tasks, establishing real-time line collaborative association, calling the time-series monitoring data of the multi-dimensional monitoring sensors, inputting the time-series monitoring data into the corresponding weak classifier, performing abnormal identification of the weak classifier, and establishing an abnormal identification result; calling the time-series verification window of the set of weak classifiers according to the abnormal identification result, establishing the trust penalty weight of the abnormal identification result through the time-series verification window; performing collaborative authentication of the abnormal identification result through real-time line collaborative association, establishing a collaborative authentication coefficient, and inputting the collaborative authentication coefficient, trust penalty weight, and abnormal identification result into the abnormal analysis network to establish an abnormal early warning.

[0007] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0008] An information database of the power transmission and transformation line is established, and the information database stores the line layout information of the power transmission and transformation line and the layout information of multi-dimensional monitoring sensors; the line layout information in the information database is called, the line layout information parsing is executed, and the line collaborative association of the power transmission and transformation line is established; a basic linkage unit of the sensor is established based on the layout information of the multi-dimensional monitoring sensors, and a set of weak classifiers mapped to it is established through the basic linkage unit; the line task of the power transmission and transformation line is obtained, the line collaborative association call is performed based on the line task, a real-time line collaborative association is established, the time-series monitoring data of the multi-dimensional monitoring sensors is called, and the time-series monitoring data is input into the corresponding weak classifier to execute the anomaly recognition of the weak classifier and establish an anomaly recognition result; according to the anomaly recognition result, the time-series verification window of the set of weak classifiers is called, and the trust penalty weight of the anomaly recognition result is established through the time-series verification window; the collaborative authentication of the anomaly recognition result is performed through the real-time line collaborative association, a collaborative authentication coefficient is established, and the collaborative authentication coefficient, the trust penalty weight, and the anomaly recognition result are input into the anomaly analysis network to establish an anomaly warning. By calling the information database to establish the line collaborative association of the power transmission and transformation line and the establishment of the basic linkage unit, a corresponding set of weak classifiers is established based on the basic linkage unit, the collaborative association and weak classifier are called according to the line task for anomaly recognition, and anomaly warning is performed by combining time-series verification and collaborative authentication, so as to realize the comprehensive monitoring and intelligent warning of the operation state of the power transmission and transformation line, and achieve the technical effect of reducing false alarms and missed alarms and improving the accuracy of early warning.

[0009] The above description is only an overview of the technical solution of this application. In order to be able to understand the technical means of this application more clearly, it can be implemented according to the content of the specification. And in order to make the above and other purposes, features and advantages of this application more obvious and understandable, the specific embodiments of this application are given below. It should be understood that the content described in this part is not intended to identify the key or important features of the embodiments of this application, nor is it used to limit the scope of this application. Other features of this application will become easily understood through the following description. Description of the Drawings

[0010] In order to more clearly illustrate the technical solutions in this application or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only exemplary, and for those of ordinary skill in the art, other drawings can be obtained according to the provided drawings without creative efforts.

[0011] Figure 1 It is a schematic flowchart of a multi-dimensional intelligent early warning method for power transmission and transformation safety in this application;

[0012] Figure 2 This is a schematic flow chart for establishing line collaboration association in the multi-dimensional intelligent early warning method for power transmission and transformation safety of this application. Specific implementation manner

[0013] By providing a multi-dimensional intelligent early warning method for power transmission and transformation safety, this application solves the technical problem in the prior art that due to the simple early warning mechanism, false alarms and missed alarms are likely to occur, further affecting the accuracy of early warning. By calling the information database to establish the line collaboration association of the power transmission and transformation line and the establishment of the basic linkage unit, based on the basic linkage unit, a corresponding weak classifier set is established, and the collaboration association and weak classifier are called according to the line task for anomaly identification, and anomaly early warning is carried out by combining time series verification and collaboration authentication, realizing the comprehensive monitoring and intelligent early warning of the operation state of the power transmission and transformation line, achieving the technical effect of reducing false alarms and missed alarms and improving the accuracy of early warning.

[0014] Next, the technical solutions in this application will be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all the embodiments of this application. It should be understood that this application is not limited by the example embodiments described here. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the scope of protection of this application. Additionally, it should be noted that for the sake of description, only the parts related to this application are shown in the accompanying drawings rather than all of them.

[0015] Please refer to the attached Figure 1 , this application provides a multi-dimensional intelligent early warning method for power transmission and transformation safety, specifically including the following steps:

[0016] Step 1: Establish an information database for the power transmission and transformation line, and the information database stores the line layout information of the power transmission and transformation line and the layout information of multi-dimensional monitoring sensors.

[0017] Specifically, the line layout information of the power transmission and transformation line includes basic information such as the line's direction, length, pole tower location, conductor type, insulator configuration, etc. This information can be obtained through channels such as the design drawings, construction records, and operation and maintenance files of the power transmission and transformation line. The multi-dimensional monitoring sensor layout information includes the detailed information of various sensors installed at key parts of the power transmission and transformation line, such as location, model, function, communication method, etc. The establishment of the information database can use a database management system, such as MySQL, Oracle, etc., to facilitate the storage, query, update, and maintenance of information. In the database, multiple tables can be designed to store different types of information. For example, one table is specifically used to store the line layout information, and another table is used to store the sensor layout information. In each table, corresponding fields can be set to correspond to different information attributes. For example, in the line layout information table, there can be fields such as line number, starting point, ending point, and number of pole towers.

[0018] In addition, in order to achieve the effective management and utilization of information, a user interface can also be provided for information input, query, statistics, and analysis. For example, users can query the detailed information of a certain section of the line or count the layout of sensors in a certain area through the user interface. Thus, the establishment of the information database of the power transmission and transformation line realizes the systematic and digital management of the information related to the power transmission and transformation line, providing a basis for the abnormal warning of the power transmission and transformation line.

[0019] Step 2: Call the line layout information in the information database, execute the parsing of the line layout information, and establish the line collaborative association of the power transmission and transformation line.

[0020] Specifically, extract the line layout information from the information database. The parsing of the line layout information refers to understanding and processing the line layout information to facilitate the establishment of the line collaborative association. The parsing process can include: analyzing the topological structure of the line to determine the connection relationship and branch situation of the line; identifying the length, type, etc. of the line; and counting the operating environment of the line. Establishing the line collaborative association of the power transmission and transformation line is to construct a collaborative relationship model between the lines based on the parsed information. This model can reflect the mutual influence and dependence relationship between the lines. For example, the failure of a certain line may affect the operation of other lines connected to it; the overload of a certain line may require support from other lines that cooperate with it. The establishment of the collaborative association helps to improve the reliability and economy of the line operation and provides a basis for the optimal scheduling and fault handling of the line. Specifically, existing mathematical tools and methods such as graph theory and network analysis can be used, combined with professional knowledge, to establish the line collaborative association to assist in the abnormal analysis and warning of the power transmission and transformation line.

[0021] Step 3: Establish the basic linkage unit of the sensor with the multi-dimensional monitoring sensor layout information, and establish a set of weak classifiers mapped to it through the basic linkage unit.

[0022] Specifically, the multi-dimensional monitoring sensor layout information includes detailed information such as the specific location, type, function, and communication method of the sensors on the power transmission and transformation lines. A basic linkage unit for the sensors is established, and the basic linkage unit includes a combination of sensors at the same location. For example, the temperature sensor and voltage sensor at the same location can be used as a basic linkage unit.

[0023] The sensors within each basic linkage unit can share some basic information and processing logics. For example, the same communication protocol can be used for data transmission, or they can have similar data processing algorithms. By establishing the basic linkage unit, the management and maintenance of the sensors can be simplified, and the efficiency of data processing can be improved.

[0024] Next, a set of weak classifiers mapped to it is established through the basic linkage unit. A weak classifier is a simple classifier, and its accuracy in classification problems may be lower than that of a strong classifier, but it is easier to train and has a lower computational cost. In machine learning, weak classifiers are usually used to construct ensemble learning algorithms such as random forest or AdaBoost. Each basic linkage unit can correspond to one or more weak classifiers. These weak classifiers make a preliminary judgment on the operating state of the power transmission and transformation lines based on the data characteristics collected by the sensors. For example, the weak classifier judges whether there are overheating phenomena, voltage abnormalities, etc. in the line based on the sensing data collected by the basic linkage unit. In this way, the complex monitoring task can be decomposed into multiple simple classification tasks, and each task is processed by a weak classifier. Finally, the outputs of all weak classifiers are integrated to obtain the final judgment on the operating state of the power transmission and transformation lines, improving the monitoring accuracy and robustness while reducing the computational complexity.

[0025] Step 4: Obtain the line tasks of the power transmission and transformation lines, perform line collaborative association calls based on the line tasks, establish real-time line collaborative associations, call the time-series monitoring data of the multi-dimensional monitoring sensors, input the time-series monitoring data into the corresponding weak classifiers, execute the anomaly recognition of the weak classifiers, and establish anomaly recognition results.

[0026] Specifically, the line task of the power transmission and transformation line refers to the power transmission task, which is usually formulated by the power grid dispatching center or the operation and maintenance department according to the operation requirements of the power grid and the actual situation of the line. Specifically, it may include line branches that need to perform power transmission, etc. Conducting line collaborative association calls based on the line task means extracting the collaborative association corresponding to the real-time line of the currently executed line task from the previously established line collaborative associations according to the requirements of the line task, that is, the collaborative relationship of the real-time line of the currently executed line task, and using this as the real-time line collaborative association. Then, according to the real-time line collaborative association, the basic linkage unit at the corresponding position is called, and the line operation parameters are monitored through the multi-dimensional monitoring sensors in the basic linkage unit, and the time-series monitoring data is generated based on the monitoring results. According to the foregoing records, one basic linkage unit corresponds to a set of weak classifiers. The time-series monitoring data collected by the basic linkage unit is input into the corresponding weak classifier to perform the anomaly recognition of the weak classifier. The weak classifier will judge whether the data is abnormal according to the characteristics of the data, such as the change trend, amplitude, etc. The result of the anomaly recognition can be binary classification, such as normal or abnormal; it can also be multi-classification, such as different anomaly types and degrees of anomaly, etc. Specifically, the type of the recognition result can be determined by professionals in this field to train the weak classifier.

[0027] Finally, an anomaly recognition result is established, that is, the outputs of all weak classifiers are summarized and integrated to form a final judgment on the line operation state, including the anomaly type and degree of anomaly, realizing the intelligent monitoring and health management of the power transmission and transformation line, and improving the operation efficiency and reliability of the power grid.

[0028] Step Five: Call the time-series verification window of the set of weak classifiers according to the anomaly recognition result, and establish the trust penalty weight of the anomaly recognition result through the time-series verification window.

[0029] Specifically, starting from the recognition node of the anomaly recognition result, according to the magnitude of the anomaly value of the anomaly recognition result, the window length is set. The larger the anomaly value, the larger the window length, and thus the time-series verification window is established. Establishing the trust penalty weight of the anomaly recognition result is a process of quantifying the reliability of the anomaly recognition result. This weight can reflect the trust degree of the anomaly recognition result. The higher the weight, the higher the trust degree of the anomaly recognition result; the lower the weight, the lower the trust degree of the anomaly recognition result. According to the data within the time-series verification window and the anomaly recognition result, the recognition analysis of the weak classifier is carried out to obtain the recognition accuracy and establish the trust penalty weight, improving the accuracy and robustness of the anomaly monitoring, reducing false alarms and missed alarms, and providing stronger support for the safe operation of the power transmission and transformation line.

[0030] Step 6: Conduct collaborative authentication of the anomaly recognition results through real-time line collaboration association, establish a collaborative authentication coefficient, input the collaborative authentication coefficient, trust penalty weight, and anomaly recognition results into the anomaly analysis network, and establish an anomaly warning.

[0031] Specifically, collaborative authentication takes into account the mutual influence and dependence relationships among various parts of the line, and can evaluate abnormal situations more comprehensively. Analyze the anomaly recognition results using the previously established real-time line collaboration association, that is, analyze whether there is a collaborative relationship in the real-time line collaboration association for the abnormal line positions in the anomaly recognition results. The collaborative authentication coefficient is a quantitative indicator that reflects the matching degree between the anomaly recognition results and the real-time line collaboration association. The higher the coefficient, the more reliable the anomaly recognition results; the lower the coefficient, the more likely there are problems with the anomaly recognition results. Specifically, it can be analyzed whether the abnormal line positions in the anomaly recognition results are consistent with the real-time line collaboration association, that is, whether the line nodes with anomalies have a connection relationship in the real-time line collaboration association. The higher the degree of consistency with the real-time line collaboration association, the larger the corresponding collaborative authentication coefficient. The proportion of abnormal nodes consistent with the real-time line collaboration association in the total number of nodes in the anomaly recognition results can be statistically calculated as the collaborative authentication coefficient.

[0032] Input the collaborative authentication coefficient, trust penalty weight, and anomaly recognition results into the anomaly analysis network. The anomaly analysis network is a data processing model, which can be a rule-based logic model or a data-driven machine learning model. Exemplarily, if it is a rule-based logic model, the collaborative authentication coefficient and trust penalty weight can be used as weight factors to compensate and correct the anomaly recognition results, and conduct anomaly warning based on the corrected anomaly recognition results. Specifically, warning information can be sent to the management personnel to achieve intelligent monitoring and warning of the anomalies in the power transmission and transformation lines, and improve the operation efficiency and reliability of the power grid.

[0033] Further, as shown in the appendix Figure 2 it shows that Step 2 of this application includes:

[0034] Obtain the line topology structure according to the parsing result of the line layout information, establish a line adjacency matrix based on the line topology structure; call the layout environment information with the line layout information, and extract the layout environment characteristics; read the line information in the information library, and extract the line characteristics according to the line information, where the line characteristics include line length characteristics, line type characteristics, and equipment type characteristics; establish a line collaboration association according to the line characteristics, layout environment characteristics, and line adjacency matrix.

[0035] Specifically, first identify the key elements in the line layout information, such as the positions of poles and towers, the connection methods of conductors, branch points, etc., so as to construct a topological graph of the line. This topological graph reflects the connection relationships between various nodes (such as poles and towers, substations, etc.) in the line. Establishing a line adjacency matrix based on the line topological structure is to transform the topological graph into a matrix form, where the rows and columns of the matrix represent the nodes in the line, and the elements in the matrix represent the connection states between the nodes. If two nodes are directly connected, the corresponding matrix element is 1, otherwise it is 0. The adjacency matrix is a commonly used representation method in graph theory, which is convenient for mathematical calculations and analyses. Invoking the layout environment information with the line layout information requires obtaining the environmental parameters during line layout from the information library, such as terrain, climate, and vegetation as layout environment characteristics, reading the line information in the information library, and extracting line characteristics according to the line information, including line length characteristics, line type characteristics, equipment type characteristics, etc. The line length characteristics may affect the resistance and inductance of the line; the line type characteristics (such as AC or DC) affect the operation mode and maintenance requirements of the line; the equipment type characteristics (such as insulator, conductor model) affect the electrical performance and safety factor of the line.

[0036] Finally, establish a line collaborative association based on the line characteristics, layout environment characteristics, and line adjacency matrix, that is, integrate all the above information, mark the line characteristics and layout environment characteristics to the line adjacency matrix, obtain the connection relationships between different nodes, as well as the corresponding line types and layout environments, and use this as the line collaborative association to facilitate subsequent line anomaly analysis and improve the accuracy of anomaly analysis.

[0037] Furthermore, step three of this application includes:

[0038] Establish a position tolerance distance at the same location, perform authentication of monitoring sensors at the same location through the position tolerance distance and multi-dimensional monitoring sensor layout information, and generate a same-location authentication result; establish a historical anomaly data set for each location; establish associated locations according to the line topological structure, call the historical anomaly authentication data set through the associated locations, and establish position anomalies and associated anomalies; perform sensor identification combination with the position anomalies and associated anomalies, the sensor identification combination is a combination under the same-location authentication result and is the combination with the least number that can be identified, and establish a basic linkage unit based on the sensor identification combination.

[0039] Specifically, the position tolerance distance is a threshold used to define the maximum distance difference at which two sensor positions are considered to be at the same location. For example, if the set position tolerance distance is 5 meters, then sensors with a distance less than or equal to 5 meters will be considered to be deployed at the same location, which can be specifically determined by those skilled in the art in combination with practical experience. Compare the actual position of each sensor with the positions of other sensors to determine whether it is within the position tolerance distance. If so, these sensors are certified as sensors at the same location. Generate a same-location certification result, that is, list all the sensors certified as being at the same location. Collect and organize the abnormal data that occurred at each location over a period of time in the past to form a historical abnormal data set for each location. These data may include fault records, maintenance logs, monitoring data, etc. The establishment of the historical abnormal data set helps to analyze the abnormal patterns and trends of each location and provides data support for subsequent abnormal identification and prediction.

[0040] Establishing associated locations according to the line topology structure means analyzing the electrical connection relationship and physical layout of the line to determine which locations are directly or indirectly associated in power transmission or fault propagation. For example, adjacent poles and towers, nodes within the same line loop network, etc. may be defined as associated locations. Associated locations are the locations where the line topology structure has a connection relationship.

[0041] By invoking the historical abnormal certification data set through associated locations, the abnormal propagation and mutual influence between these locations can be analyzed, and location anomalies and associated anomalies can be established. Location anomalies are the line nodes where anomalies occur, and associated anomalies refer to the line nodes with abnormal associations, that is, to analyze whether the anomaly at a certain location is related to the anomalies at other locations. The association relationship between the anomalies at different locations can be specifically analyzed through existing management analysis methods. Combining sensor identification with location anomalies and associated anomalies means determining the number of sensors for anomaly identification at the same location according to the historical abnormal certification data set. The sensor identification combination should be a combination under the same-location certification result and the combination with the minimum number that can be identified to reduce the complexity and cost of data analysis.

[0042] Finally, based on the sensor identification combination, a basic linkage unit is established, that is, the sensors that need to work together are taken as a basic linkage unit. Such a basic linkage unit can more effectively conduct anomaly monitoring and fault diagnosis, improve the accuracy and efficiency of the monitoring system, and provide guarantee for the safe operation of the power transmission and transformation line.

[0043] Furthermore, the present application further includes the following steps:

[0044] Invoke the historical monitoring dataset of the basic linkage unit; use the historical anomaly dataset to perform supervised labeling on the historical monitoring dataset, and establish a training dataset and a validation dataset based on the supervised labeling results; perform basic model adaptation decision-making for the basic linkage unit through the decision-making channel, and complete the construction of the weak classifier set based on the adaptation decision-making results, the training dataset, and the validation dataset.

[0045] Specifically, after establishing the basic linkage unit, it is also necessary to establish a corresponding weak classifier set. The specific steps are as follows: Invoking the historical monitoring dataset of the basic linkage unit means obtaining the historical monitoring data corresponding to the historical anomaly dataset, which may include time-series data such as temperature, voltage, and current collected by sensors. The historical anomaly dataset includes information such as anomaly types and degrees of anomaly. Use the information in the historical anomaly dataset to label the data in the historical monitoring dataset. The results of supervised labeling can be different types of anomalies and degrees of anomaly. Establishing a training dataset and a validation dataset based on the supervised labeling results is for the training and validation of the machine learning model. The training dataset is used to train the model, and the validation dataset is used to evaluate the performance of the model. Usually, the training dataset and the validation dataset are divided from the original dataset in a certain proportion, such as 70% of the data for training and 30% of the data for validation.

[0046] Performing basic model adaptation decision-making for the basic linkage unit through the decision-making channel means selecting a suitable machine learning model or algorithm according to the characteristics of the basic linkage unit. Factors that may be considered in the adaptation decision include the characteristics of the data and the complexity of the model. Specifically, professionals in the field can select multiple machine learning models and configure corresponding data characteristics for each machine learning model, so as to match the basic model according to the data characteristics of the sensors corresponding to the basic linkage unit, and use the training dataset and the validation dataset to train the model to obtain a converged model to form a weak classifier set. Construct a weak classifier set suitable for the characteristics of the basic linkage unit to improve the accuracy and efficiency of anomaly monitoring and fault diagnosis.

[0047] Furthermore, step five of this application includes:

[0048] Obtain the recognition node of the anomaly recognition result, and use the recognition node as the time series zero point; read the anomaly value of the anomaly recognition result, determine the window length based on the anomaly value, and construct a time series verification window based on the window length and the time series zero point; determine the anomaly type according to the anomaly recognition result, and configure the influence weight of the anomaly type; perform recognition analysis of the weak classifier within the time series verification window through the anomaly type and the influence weight, and establish a trust penalty weight.

[0049] Specifically, the identified nodes in the anomaly recognition result are used as the time series zero point, that is, the starting point of the window. The anomaly values in the anomaly recognition result are read, that is, the degrees of anomaly in the anomaly recognition result are extracted, such as the amplitude and duration of the anomaly. Determining the window length based on the anomaly value is to construct a time window for more in-depth analysis. The window length can be determined according to the size of the anomaly value. The window length is proportional to the anomaly value. Specifically, those skilled in the art can configure the corresponding relationship between the anomaly value and the window length based on experience. A time series verification window is constructed based on the window length and the time series zero point, that is, the time series zero point is used as the starting point of the window, and the corresponding time length is extended backward to form a time period, which is the time series verification window.

[0050] Determine the anomaly type according to the anomaly recognition result and configure the influence weight of the anomaly type to perform more targeted analysis within the time series verification window. Different anomaly types may have different degrees of influence on the power grid. Therefore, an influence weight needs to be assigned to each anomaly type to reflect its importance and urgency. Specifically, historical anomaly accidents corresponding to different anomaly types can be collected, the affected range and losses caused by the historical anomaly accidents can be analyzed, and different influence weights can be configured for different anomaly types. Specifically, this can be configured by experts in the field based on experience.

[0051] Perform the identification and analysis of weak classifiers within the time series verification window through the anomaly type and the influence weight, that is, use the weak classifier to process and analyze the data within the window, and output the identification results corresponding to consecutive time nodes. Under normal circumstances, the output results are basically the same. Therefore, the difference degree of the identification results is obtained through difference identification, and it is weighted with the influence weight to generate the final trust penalty weight. This weight reflects the reliability of the anomaly recognition result. The higher the weight, the higher the trust in the anomaly recognition result; the lower the weight, the lower the trust in the anomaly recognition result. Realize in-depth analysis and verification of the anomaly recognition result, and improve the accuracy and robustness of anomaly monitoring.

[0052] Furthermore, the present application further includes the following steps:

[0053] Establish a basic influence weight, where the basic influence weight is the influence weight of non-anomaly types; perform the identification and analysis of weak classifiers within the time series verification window according to the basic influence weight, and establish a trust penalty weight according to the identification and analysis results of anomaly types and non-anomaly types.

[0054] Specifically, establishing the basic influence weight, that is, determining the influence weight of non-abnormal types, is for the purpose of identifying and analyzing non-abnormal types within the time-series verification window to compare and verify the identification results of abnormal types. The non-abnormal type refers to the normal operation condition of the line. The basic influence weight reflects the degree of influence of the non-abnormal state on the power grid operation, and the corresponding basic influence weight can be generated by collecting historical line operation records in the non-abnormal state for correlation influence analysis. According to the basic influence weight, the weak classifier identification analysis within the time-series verification window is carried out, that is, the weak classifier is used to process and analyze the non-abnormal type data within the window. The purpose of this step is to evaluate the performance of the system in the non-abnormal state and the monitoring ability of the system for the normal operation state. According to the identification analysis results of abnormal types and non-abnormal types, a trust penalty weight is established, that is, considering the identification results of abnormal types and non-abnormal types comprehensively, a trust weight is assigned to the abnormal identification result. Exemplarily, the identification error of the weak classifier for non-abnormal types is statistically analyzed, weighted in combination with the basic influence weight, and averaged and weighted with the trust penalty weight established by the above-mentioned weak classifier identification analysis through the abnormal type and influence weight within the time-series verification window to obtain the final trust penalty weight, improving the reliability of abnormal early warning.

[0055] Furthermore, the present application further includes the following steps:

[0056] Set an abnormal early warning threshold, which is constructed by the historical abnormal data set of the power transmission and transformation line; trigger and discriminate the abnormal early warning through the abnormal early warning threshold; establish a response strategy according to the trigger discrimination result, and conduct abnormal early warning management based on the response strategy.

[0057] Specifically, setting the abnormal warning threshold is to accurately determine when to issue a warning when an abnormal situation is detected. The abnormal warning threshold is constructed by analyzing the historical abnormal data set of the power transmission and transformation line, including statistical analysis, trend analysis, etc. of the historical abnormal data, determining the minimum abnormal value that will cause a line fault, and setting it as the abnormal warning threshold. The abnormal warning is triggered and judged through the abnormal warning threshold, that is, when the abnormal value of the abnormal warning exceeds or reaches the abnormal warning threshold, the abnormal warning is triggered. Establishing a response strategy according to the trigger judgment result means formulating corresponding countermeasures according to the type and severity of the abnormality. The response strategy may include notifying the operation and maintenance personnel, starting standby equipment, adjusting operation parameters, etc. Specifically, it needs to be determined in combination with the actual warning response strategy. Professionals in this field can configure different response strategies for different abnormal types and abnormal degrees according to actual experience. When the abnormal warning is triggered, the response strategy is generated according to the abnormal type and abnormal degree. Managing the abnormal warning based on the response strategy means that after detecting an abnormality and triggering a warning, notifying the relevant staff to process it according to the predetermined response strategy, realizing the timely warning and effective management of the abnormalities of the power transmission and transformation line, and improving the operation efficiency and safety of the power grid.

[0058] Furthermore, the present application further includes the following steps:

[0059] Making a warning record for the abnormal warning and establishing a continuous abnormality identification; reporting the line fault of the power transmission and transformation line through the continuous abnormality identification, and managing the power transmission and transformation line according to the line fault.

[0060] Specifically, making a warning record for the abnormal warning means that after the abnormal warning, recording the relevant warning information, including the time, location, abnormal type, abnormal degree, etc. of the warning occurrence. The warning record helps subsequent fault analysis and processing, and also provides a basis for the management and analysis of historical data. Arranging the warning records continuously in chronological order to generate the continuous abnormality identification result. Reporting the line fault of the power transmission and transformation line through the continuous abnormality identification, that is, when an abnormal situation is continuously detected, a fault alarm is issued, that is, the line fault is reported. The method of reporting the line fault can include any form of sound, light, and electricity to remind relevant personnel of the timely response and effective management of the power transmission and transformation line fault, ensuring the stable operation and safety of the power grid.

[0061] In summary, the multi-dimensional intelligent warning method for power transmission and transformation safety provided by the present application has the following technical effects:

[0062] Establish an information database for power transmission and transformation lines. The information database stores the line layout information and multi-dimensional monitoring sensor layout information of the power transmission and transformation lines. Call the line layout information in the information database, perform line layout information parsing, and establish the line collaborative association of the power transmission and transformation lines. Establish the basic linkage unit of the sensor based on the multi-dimensional monitoring sensor layout information, and establish a set of weak classifiers mapped to it through the basic linkage unit. Obtain the line tasks of the power transmission and transformation lines, perform line collaborative association calls based on the line tasks, establish real-time line collaborative association, call the time-series monitoring data of the multi-dimensional monitoring sensors, input the time-series monitoring data into the corresponding weak classifiers, perform anomaly recognition of the weak classifiers, and establish anomaly recognition results. Call the time-series verification window of the set of weak classifiers according to the anomaly recognition results, and establish the trust penalty weight of the anomaly recognition results through the time-series verification window. Perform collaborative authentication of the anomaly recognition results through real-time line collaborative association, establish a collaborative authentication coefficient, and input the collaborative authentication coefficient, trust penalty weight, and anomaly recognition results into the anomaly analysis network to establish an anomaly warning. By calling the information database to establish the line collaborative association and the establishment of the basic linkage unit of the power transmission and transformation lines, establishing a corresponding set of weak classifiers based on the basic linkage unit, calling the collaborative association and weak classifiers for anomaly recognition according to the line tasks, and combining time-series verification and collaborative authentication for anomaly warning, the overall monitoring and intelligent warning of the operating state of the power transmission and transformation lines are realized, achieving the technical effects of reducing false alarms and missed alarms and improving the accuracy of early warning.

[0063] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to these embodiments shown herein, but will be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0064] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalent technologies, the present application is also intended to include these changes and modifications.

Claims

1. A multi-dimensional intelligent early warning method for power transmission and transformation safety, characterized in that: The multi-dimensional intelligent early warning method for power transmission and transformation safety includes: Establishing an information base of power transmission and transformation lines, wherein the information base stores line layout information of the power transmission and transformation lines and layout information of multi-dimensional monitoring sensors; Calling the line layout information in the information database, performing line layout information parsing, and establishing line collaborative association of the transmission and transformation lines, including: Acquire a line topology structure according to the analysis result of the line layout information, and establish a line adjacency matrix based on the line topology structure; Using the line layout information to call layout environment information and extract layout environment features; Reading line information in the information database, and extracting line features according to the line information, wherein the line features include line length features, line type features, and equipment type features; Establishing line collaborative association according to the line characteristics, layout environment characteristics and line adjacency matrix; Establishing a basic linkage unit of sensors with the multi-dimensional monitoring sensor deployment information, and establishing a set of weak classifiers mapped thereto through the basic linkage unit; The basic linkage unit includes a combination of sensors in the same position, and the sensors in each basic linkage unit share some basic information and processing logic; Acquire the line task of the power transmission and transformation line, perform line collaborative association call based on the line task, establish real-time line collaborative association, call the time series monitoring data of the multi-dimensional monitoring sensor, input the time series monitoring data into the corresponding weak classifier, perform abnormality recognition of the weak classifier, and establish abnormality recognition results; Calling the time series verification window of the weak classifier set according to the abnormal recognition result, and establishing the trust penalty weight of the abnormal recognition result through the time series verification window; Through real-time line collaborative association, collaborative authentication of anomaly identification results is performed, a collaborative authentication coefficient is established, and the collaborative authentication coefficient, trust penalty weight and anomaly identification results are input into the anomaly analysis network to establish an anomaly warning.

2. The multi-dimensional intelligent early warning method for power transmission and transformation safety according to claim 1, characterized in that: The basic linkage unit for establishing sensors based on the multi-dimensional monitoring sensor layout information also includes: Establishing a position tolerance distance for the same position, performing monitoring sensor authentication at the same position through the position tolerance distance and multi-dimensional monitoring sensor deployment information, and generating a same position authentication result; The position tolerance distance is a threshold value used to define the maximum distance difference between two sensor positions for being considered as the same position; Build historical anomaly datasets for each location; Establishing an associated position according to the line topology structure, calling a historical abnormality authentication data set through the associated position, and establishing a position abnormality and an associated abnormality; A sensor identification combination is performed based on the position anomaly and the associated anomaly. The sensor identification combination is a combination under the same position authentication result and is the minimum number of identifiable combinations. A basic linkage unit is established based on the sensor identification combination.

3. The multi-dimensional intelligent early warning method for power transmission and transformation safety according to claim 2, characterized in that: The step of establishing a weak classifier set mapped thereto through the basic linkage unit also includes: Call the historical monitoring data set of the basic linkage unit; Using the historical abnormal data set to perform supervisory identification on the historical monitoring data set, and establishing a training data set and a verification data set according to the supervisory identification results; The basic model adaptation decision of the basic linkage unit is made through the decision channel, and the weak classifier set is constructed based on the adaptation decision results and the training data set and the verification data set.

4. The multi-dimensional intelligent early warning method for power transmission and transformation safety according to claim 1, characterized in that: The step of calling the time series verification window of the weak classifier set according to the abnormal recognition result and establishing the trust penalty weight of the abnormal recognition result through the time series verification window further includes: Obtaining an identification node of an abnormal identification result, and taking the identification node as a timing zero point; Reading an abnormal value of an abnormal recognition result, determining a window length based on the abnormal value, and constructing a timing verification window based on the window length and the timing zero point; Determine the abnormality type according to the abnormality identification result, and configure the impact weight of the abnormality type; The identification and analysis of the weak classifier is performed within the time series verification window through the abnormal type and the impact weight, and the trust penalty weight is established.

5. The multi-dimensional intelligent early warning method for power transmission and transformation safety according to claim 4, characterized in that: The multi-dimensional intelligent early warning method for power transmission and transformation safety also includes: Establishing a basic influence weight, wherein the basic influence weight is an influence weight of a non-abnormal type; The weak classifier identification analysis within the time series verification window is performed according to the basic impact weight, and the trust penalty weight is established according to the identification analysis results of the abnormal type and the identification analysis results of the non-abnormal type.

6. The multi-dimensional intelligent early warning method for power transmission and transformation safety according to claim 1, characterized in that: The multi-dimensional intelligent early warning method for power transmission and transformation safety also includes: Setting an abnormal warning threshold, wherein the abnormal warning threshold is constructed by using a historical abnormal data set of the power transmission and transformation line; Triggering and judging the abnormal warning by using the abnormal warning threshold; Establish a response strategy based on the trigger identification results, and perform abnormal warning management based on the response strategy.

7. The multi-dimensional intelligent early warning method for power transmission and transformation safety according to claim 1, characterized in that: The multi-dimensional intelligent early warning method for power transmission and transformation safety also includes: Record the abnormal warning and establish continuous abnormality recognition; The line fault of the power transmission and transformation line is reported through the continuous abnormal identification, and the power transmission and transformation line is managed according to the line fault.

Images