Lightning arrester monitoring method and system

Abstract

The embodiment of the present invention provides a lightning arrester monitoring method and system, which can reflect the cyclic characteristic changes of the working mode of the lightning arrester by obtaining the electrical state snapshots of the lightning arrester in multiple monitoring sections and extracting the state mode vectors of these snapshots. Furthermore, these electrical state snapshots and the corresponding operation mode vectors are loaded into the lightning arrester operation trend estimation network, and an estimated electrical state snapshot of the lightning arrester in the monitoring section to be evaluated can be generated. Finally, based on the electrical state snapshots of the lightning arrester in multiple monitoring sections and the estimated electrical state snapshots in the monitoring section to be evaluated, the performance change trend of the lightning arrester is generated, which effectively improves the accuracy of the state monitoring of the lightning arrester and the ability to predict the performance change trend, helps to discover possible problems in advance and take corresponding maintenance measures, thereby improving the operational safety and stability of the lightning arrester.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of artificial intelligence technology, and more specifically, to a lightning arrester monitoring method and system. Background Art

[0002] In the power system, the arrester is an important protective device used to protect electrical equipment from damage caused by lightning overvoltage. In order to ensure the normal operation and effective protection of the arrester, it needs to be monitored and evaluated regularly. However, traditional arrester monitoring methods are usually based on regular manual inspections or simple electrical parameter measurements. These methods are not only inefficient, but also difficult to detect potential faults and performance change trends of the arrester in a timely and accurate manner. For example, some methods fail to effectively extract and utilize the characteristic changes in this information, resulting in the accuracy and reliability of the monitoring results to be improved; in addition, some methods fail to make full use of the information of the operating mode vector when predicting the future working state of the arrester, and cannot accurately reflect the cyclic characteristic changes of the arrester working mode. Therefore, the prediction of the working state and performance of the arrester is not accurate, and it cannot provide sufficient information to predict the performance change trend of the arrester.

[0003] In general, the existing technical methods have certain limitations in prediction accuracy and information utilization efficiency. Therefore, it is a current technical challenge to develop a new lightning arrester monitoring method that can accurately reflect the working status and performance change trend of the lightning arrester and improve monitoring accuracy and information utilization efficiency. Summary of the invention

[0004] In order to at least overcome the above-mentioned deficiencies in the prior art, an object of the embodiments of the present application is to provide a lightning arrester monitoring method and system.

[0005] In a first aspect, an embodiment of the present application provides a lightning arrester monitoring method, the method comprising:

[0006] Obtaining electrical status snapshots of the arrester in multiple monitoring sections; the multiple monitoring sections are monitoring sections located forward of an Xth monitoring section, the Xth monitoring section is a monitoring section to be evaluated, and X is greater than 1;

[0007] Extracting state mode vectors from the electrical state snapshots of the plurality of monitoring segments, and generating operation mode vectors corresponding to the electrical state snapshots of the plurality of monitoring segments; the operation mode vectors reflect cyclic characteristic changes of the arrester working mode of the arrester;

[0008] Loading the electrical state snapshots of the plurality of monitoring segments and the respective corresponding operation mode vectors into the arrester operation trend estimation network to generate an estimated electrical state snapshot of the arrester in the Xth monitoring segment;

[0009] A performance change trend of the arrester is generated according to the electrical state snapshots of the arrester in the plurality of monitoring segments and the estimated electrical state snapshot in the Xth monitoring segment.

[0010] In a possible implementation manner of the first aspect, the operation mode vector includes a fluctuation rate vector;

[0011] The extracting state mode vectors of the electrical state snapshots of the plurality of monitoring segments to generate operation mode vectors corresponding to the electrical state snapshots of the plurality of monitoring segments respectively includes:

[0012] Performing order-preserving discrete cosine transformation on the electrical state snapshots of the multiple monitoring segments to generate the fluctuation rate vectors corresponding to the electrical state snapshots of the multiple monitoring segments;

[0013] The order-preserving discrete cosine transform is a discrete cosine transform with an order-preserving property, and the order-preserving property is used to indicate that the generated feature vector has a property of having a consistent change effect with the feature vector loaded thereto.

[0014] In a possible implementation of the first aspect, the multiple monitoring segments include an x-th monitoring segment, where x is less than X; the lightning arrester includes Y lightning protection working events, and the lightning arrester includes the Y lightning protection working events in the electrical state snapshots of the multiple monitoring segments at the electrical state nodes corresponding to the Y lightning protection working events in the x-th monitoring segment, respectively, and Y is not less than 1; the fluctuation rate vector includes a fluctuation rate linkage parameter between the Y lightning protection working events;

[0015] The performing order-preserving discrete cosine transformation on the electrical state snapshots of the plurality of monitoring segments to generate the fluctuation rate vectors corresponding to the electrical state snapshots of the plurality of monitoring segments respectively includes:

[0016] Generate a weight vector of the zth lightning protection working event at the i-th fluctuation rate according to the electrical state node of the zth lightning protection working event among the Y lightning protection working events at the x-th monitoring segment and the attention electrical state node of the Y lightning protection working events at the x-th monitoring segment;

[0017] According to the weight vector of the zth lightning protection working event on the i-th fluctuation rate, and the weight vector of the yth lightning protection working event among the Y lightning protection working events on the i-th fluctuation rate, the fluctuation rate linkage parameter between the zth lightning protection working event and the yth lightning protection working event is generated, the i-th fluctuation rate is a fluctuation rate on the fluctuation rate domain, the i is a positive integer between 0 and X-1, and the fluctuation rate domain is represented by 0 to X-1 fluctuation rates; the z and the y are not greater than the Y, and the z is not equal to the y.

[0018] In a possible implementation manner of the first aspect, generating a weight vector of the zth lightning protection working event at the i-th fluctuation rate based on the zth lightning protection working event among the Y lightning protection working events at the electrical state node of the x-th monitoring segment and the attention electrical state node of the Y lightning protection working events at the x-th monitoring segment includes:

[0019] Determine the deviation parameter between the electrical state node of the zth lightning protection working event among the Y lightning protection working events in the xth monitoring segment and the attention electrical state node of the Y lightning protection working events in the xth monitoring segment, and generate the deviation parameter of the zth lightning protection working event in the xth monitoring segment; the deviation parameter reflects the migration order-preserving property of the electrical state node;

[0020] The deviation parameter of the zth lightning protection working event in the xth monitoring section is mapped to the ith fluctuation rate in the fluctuation rate domain, and a weight vector of the zth lightning protection working event on the ith fluctuation rate is generated.

[0021] In a possible implementation manner of the first aspect, generating the fluctuation rate linkage parameter between the zth lightning protection working event and the yth lightning protection working event based on the weight vector of the zth lightning protection working event at the ith fluctuation rate and the weight vector of the yth lightning protection working event among the Y lightning protection working events at the ith fluctuation rate includes:

[0022] Performing a dot product operation on the weight vector of the zth lightning protection working event at the ith fluctuation rate and the weight vector of the yth lightning protection working event among the Y lightning protection working events at the ith fluctuation rate to generate a dot product result;

[0023] The vector amplitude corresponding to the dot product result is output as the fluctuation rate linkage parameter between the z-th lightning protection working event and the y-th lightning protection working event.

[0024] In a possible implementation manner of the first aspect, the fluctuation rate vector further includes a fluctuation rate change value of the zth lightning protection working event;

[0025] The method further comprises:

[0026] The square value of the Euclidean length corresponding to the weight vector of the z-th lightning protection working event on the i-th fluctuation rate is calculated to generate the fluctuation rate change value of the z-th lightning protection working event.

[0027] In a possible implementation manner of the first aspect, loading the electrical state snapshots of the multiple monitoring segments and the respective corresponding operation mode vectors into the arrester operation trend estimation network to generate the estimated electrical state snapshot of the arrester in the Xth monitoring segment includes:

[0028] Loading the electrical state snapshots of the plurality of monitoring segments and the corresponding operation mode vectors to the arrester operation trend estimation network, so that the operation mode vector is transmitted to each layer of the arrester operation trend estimation network in the forward propagation, and generating an estimated electrical state node of the arrester in the Xth monitoring segment;

[0029] According to the estimated electrical state node of the lightning arrester in the Xth monitoring section, a snapshot of the estimated electrical state of the lightning arrester in the Xth monitoring section is generated.

[0030] In a possible implementation manner of the first aspect, the operation mode vector includes at least one of a fluctuation rate linkage parameter between a zth lightning protection working event and a yth lightning protection working event among the Y lightning protection working events and a fluctuation rate change value of the zth lightning protection working event;

[0031] The method further comprises:

[0032] Adding the fluctuation rate linkage parameter between the z-th lightning protection working event and the y-th lightning protection working event in the Y lightning protection working events as a state link vector in the electrical state snapshot;

[0033] And, the fluctuation rate change value of the zth lightning protection working event is added as the state member vector in the electrical state snapshot, so that the operation mode vector is transmitted to each layer of the lightning arrester operation trend estimation network in the forward propagation.

[0034] In the second aspect, an embodiment of the present application also provides a lightning arrester monitoring system, which includes a processor and a machine-readable storage medium, in which machine-readable storage medium is stored machine-executable instructions, which are loaded and executed by the processor to implement the lightning arrester monitoring method in any possible implementation manner in the first aspect.

[0035] In any of the above aspects, the present application can reflect the cyclical characteristic changes of the arrester's operating mode by obtaining snapshots of the arrester's electrical state in multiple monitoring segments and extracting state mode vectors from these snapshots. Furthermore, these electrical state snapshots and the corresponding operating mode vectors are loaded into the arrester operation trend estimation network to generate an estimated electrical state snapshot of the arrester in the monitoring segment to be evaluated. Finally, based on the electrical state snapshots of the arrester in multiple monitoring segments and the estimated electrical state snapshots in the monitoring segment to be evaluated, the performance change trend of the arrester is generated, which effectively improves the accuracy of the arrester state monitoring and the ability to predict performance change trends.

[0036] That is to say, the present application can obtain the electrical state snapshot of the arrester in multiple monitoring sections, and by extracting the state mode vector of the electrical state snapshot, the cyclic characteristic changes of the arrester working mode can be reflected, which is conducive to more accurately understanding and predicting the working state of the arrester. Then, the electrical state snapshot and the corresponding operation mode vector are loaded into the arrester operation trend estimation network, and the estimated electrical state snapshot of the arrester in the monitoring section to be evaluated can be generated, so as to effectively predict the possible working state of the arrester in the future. Finally, according to the electrical state snapshot of the arrester in multiple monitoring sections and the estimated electrical state snapshot of the monitoring section to be evaluated, the performance change trend of the arrester can be generated, which can assist in judging the historical working performance of the arrester and provide a reference for predicting the future performance of the arrester. Thus, the working state and performance change trend of the arrester can be accurately reflected, which helps to discover possible problems in advance and take corresponding maintenance measures, thereby improving the operational safety and stability of the arrester. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required to be used in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be extracted in combination with these drawings without creative work.

[0038] Figure 1 A schematic diagram of a flow chart of a lightning arrester monitoring method provided by an embodiment of the present invention;

[0039] Figure 2 A schematic block diagram of the structure of a lightning arrester monitoring system for implementing the above-mentioned lightning arrester monitoring method provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0040] The following description is intended to enable one of ordinary skill in the art to implement and incorporate the present invention, and the description is provided in the context of a specific application scenario and its requirements. It will be apparent to one of ordinary skill in the art that various changes may be made to the disclosed embodiments, and that the general principles defined in the present invention may be applied to other embodiments and application scenarios without departing from the principles and scope of the present invention. Therefore, the present invention is not limited to the described embodiments, but should be given the broadest scope consistent with the claims.

[0041] Figure 1 FIG. 1 is a flow chart of a lightning arrester monitoring method provided by an embodiment of the present invention. The lightning arrester monitoring method is introduced in detail below.

[0042] Step S110, obtaining snapshots of electrical states of the arrester in multiple monitoring sections, wherein the multiple monitoring sections are monitoring sections located forward of an Xth monitoring section, the Xth monitoring section is a monitoring section to be evaluated, and X is greater than 1.

[0043] In this embodiment, in an electric power system, the arrester monitoring system is responsible for managing and maintaining the operating data of various electrical equipment. These electrical equipment include multiple arresters, each of which is provided with different monitoring segments to monitor its working status in real time. Now, the arrester monitoring system needs to evaluate the performance change trend of a specific arrester (referred to as the "target arrester") in the Xth monitoring segment (the monitoring segment to be evaluated). In order to achieve this goal, the arrester monitoring system first needs to obtain a snapshot of the electrical status of the arrester in multiple monitoring segments before the Xth monitoring segment.

[0044] For example, the arrester monitoring system communicates through the sensor network connected to it, sending data requests to the sensors installed in each monitoring section of the arrester. These sensors can measure and record the working status of the arrester under various electrical parameters (such as voltage, current, impedance, etc.) in real time. After receiving the request from the arrester monitoring system, the sensor sends the electrical status data (i.e., "electrical status snapshot") within the recent period to the arrester monitoring system. The arrester monitoring system stores this data in memory for subsequent analysis and processing.

[0045] That is, this process involves using sensors or other monitoring devices to capture and record the electrical status data of the lightning arrester at different points in time. These electrical status data are usually instantaneous values ​​of key parameters such as voltage, current, temperature, impedance, etc., which together constitute a snapshot of the electrical status of the lightning arrester. For example, assuming that the lightning arrester has five monitoring sections (A, B, C, D, E), the lightning arrester monitoring system needs to evaluate the performance of the E monitoring section (the Xth monitoring section). Prior to this, electrical status snapshots of the A, B, C, and D monitoring sections (multiple monitoring sections before E) can be collected. For example, in the B monitoring section, the lightning arrester monitoring system may record the voltage fluctuations and current peaks of the lightning arrester under specific thunderstorm weather.

[0046] Forward refers to the monitoring segments that are located before the monitoring segment to be evaluated in the time series or physical layout. These monitoring segments provide information about the working status of the lightning arrester at previous time points or locations, which is essential for predicting future status or evaluating performance trends. For example, if the Xth monitoring segment is E, then A, B, C, and D are the monitoring segments located in the forward direction of E. The lightning arrester monitoring system will pay attention to the data of these monitoring segments because they contain the working history of the lightning arrester before reaching the E monitoring segment.

[0047] The monitoring segment to be evaluated refers to the operating segment of the arrester that currently needs to be analyzed or predicted. In this context, monitoring segment X specifically refers to a specific area or time point on the arrester, and the arrester monitoring system will evaluate its performance. Continuing with the above example, if the arrester monitoring system wants to predict the performance of the arrester in monitoring segment E, then E is the monitoring segment to be evaluated. All data collection and analysis work will be carried out around this goal.

[0048] Step S120, extracting state mode vectors from the electrical state snapshots of the plurality of monitoring segments, and generating operation mode vectors corresponding to the electrical state snapshots of the plurality of monitoring segments, wherein the operation mode vectors reflect the cyclic characteristic changes of the arrester working mode of the arrester.

[0049] In this embodiment, after obtaining the electrical state snapshots of the target arrester in multiple monitoring sections, the arrester monitoring system needs to further analyze the data of these electrical state snapshots to extract useful information. The purpose of this step is to extract the operating mode vector that can reflect the cyclical characteristic changes of the arrester working mode from the original electrical state data.

[0050] For example, the arrester monitoring system runs a pre-set algorithm to perform data processing and pattern recognition on the collected electrical state snapshots. These algorithms can include Fourier transform, wavelet analysis, machine learning classifiers, etc., which are used to identify specific patterns or features from the complex electrical signals of the electrical state snapshots. Through these processes, the arrester monitoring system is able to extract the operating mode vectors corresponding to each monitoring segment electrical state snapshot, and these operating mode vectors contain characteristic information of the cyclic changes of the arrester under different operating modes. For example, the arrester monitoring system may use Fourier analysis to identify periodic fluctuation patterns in the arrester electrical state snapshots and encode these periodic fluctuation patterns into vectors. For example, a specific vector may represent the voltage and current fluctuation patterns of the arrester under normal working conditions.

[0051] The electrical state snapshot of each monitoring segment is converted into an operating mode vector. This operating mode vector is a mathematical description of the operating state of the lightning arrester in the monitoring segment, which can capture the cyclical characteristic changes of the lightning arrester in different operating modes. For example, if the lightning arrester exhibits a specific voltage fluctuation pattern in the B monitoring segment, the lightning arrester monitoring system will generate a corresponding operating mode vector. This vector may contain information about the fluctuation frequency, amplitude and phase, which together define the operating mode of the lightning arrester in the B monitoring segment. In addition, the operating mode vector not only describes the state of the lightning arrester at a certain moment, but also captures the cyclical characteristics of its operating mode. This means that if the operating mode of the lightning arrester shows repetitive or periodic changes over time, these changes will be encoded in the operating mode vector. For example, suppose that the lightning arrester will experience a short recovery phase after each thunderstorm, during which its electrical parameters will show a specific fluctuation pattern. The lightning arrester monitoring system can identify and track this pattern through the generated operating mode vector, thereby predicting the future behavior of the lightning arrester.

[0052] Step S130, loading the electrical state snapshots of the plurality of monitoring segments and the respective corresponding operation mode vectors into the arrester operation trend estimation network, and generating an estimated electrical state snapshot of the arrester in the Xth monitoring segment.

[0053] After extracting the operation mode vector, the arrester monitoring system needs to use the electrical state snapshots of the multiple monitoring segments and the corresponding operation mode vectors to predict the working state of the arrester in the future (i.e., the Xth monitoring segment). To this end, the arrester monitoring system uses a pre-trained arrester operation trend estimation network.

[0054] In detail, the arrester monitoring system takes the electrical state snapshot and the operation mode vector extracted in the previous step as input and loads them into the arrester operation trend estimation network. This arrester operation trend estimation network may be a deep learning model, such as a recurrent neural network (RNN) or a long short-term memory network (LSTM), which is suitable for processing sequence data and predicting the future state of time series. After processing, the arrester operation trend estimation network outputs an estimated electrical state snapshot, which predicts the working state of the arrester in the Xth monitoring section.

[0055] For example, suppose there is a deep learning model that has been trained to recognize patterns in the electrical state snapshots and operating mode vectors of a lightning arrester. Now, the data (electrical state snapshots and corresponding operating mode vectors) collected from monitoring segments A, B, C, and D are input into this deep learning model. The deep learning model analyzes this data and predicts the possible electrical state of the lightning arrester in monitoring segment E (Monitoring segment X) based on the rules learned internally. Then, the electrical state of the lightning arrester at a certain point in the future (Monitoring segment X) can be predicted based on the input data. This prediction result is an estimate based on the analysis and understanding of historical data by the deep learning model. Continuing with the above example, the deep learning model will output an estimated electrical state snapshot, which describes the state of the possible electrical parameters such as voltage and current of the lightning arrester in monitoring segment E. This estimate is based on the model's analysis of the data from monitoring segments A, B, C, and D.

[0056] Step S140: generating a performance change trend of the arrester according to the electrical state snapshots of the arrester in the plurality of monitoring segments and the estimated electrical state snapshot in the Xth monitoring segment.

[0057] Finally, the arrester monitoring system needs to integrate all the information collected and analyzed in the previous steps to generate a comprehensive report on the performance trends of the target arrester.

[0058] In detail, the arrester monitoring system compares and analyzes the actual detected electrical state snapshots (from multiple monitoring segments) with the estimated electrical state snapshots (for the Xth monitoring segment) generated by the operation trend estimation network. By comparing the electrical parameters in these electrical state snapshots (such as the changing trends of voltage and current, the fluctuation range of impedance, etc.), the arrester monitoring system can evaluate the changing trends of the arrester performance, such as whether there are problems such as aging and insulation performance degradation. The arrester monitoring system outputs these analysis results in charts, reports or other visual forms for reference by the operation and maintenance personnel of the power system so that they can take necessary maintenance measures in a timely manner.

[0059] For example, the arrester monitoring system will compare the generated snapshot of the estimated electrical state of the E monitoring section with the actual electrical state snapshot measured in the E monitoring section. By comparing and analyzing these data, the arrester monitoring system can detect the changing trend of the arrester performance, such as whether the voltage is stable, whether the current is abnormal, etc. If an abnormality is found, the arrester monitoring system may issue an alarm to notify the operation and maintenance personnel to conduct further inspection and maintenance. Through these steps, the arrester monitoring system can realize real-time monitoring and prediction of the arrester performance, thereby ensuring the safe and stable operation of the power system.

[0060] Based on the above steps, the present application can reflect the cyclic characteristic changes of the arrester's operating mode by obtaining snapshots of the arrester's electrical state in multiple monitoring segments and extracting state mode vectors from these snapshots. Furthermore, these electrical state snapshots and the corresponding operating mode vectors are loaded into the arrester operating trend estimation network to generate an estimated electrical state snapshot of the arrester in the monitoring segment to be evaluated. Finally, based on the electrical state snapshots of the arrester in multiple monitoring segments and the estimated electrical state snapshots in the monitoring segment to be evaluated, the performance change trend of the arrester is generated, which effectively improves the accuracy of the arrester state monitoring and the ability to predict performance change trends.

[0061] That is to say, the present application can obtain the electrical state snapshot of the arrester in multiple monitoring sections, and by extracting the state mode vector of the electrical state snapshot, the cyclic characteristic changes of the arrester working mode can be reflected, which is conducive to more accurately understanding and predicting the working state of the arrester. Then, the electrical state snapshot and the corresponding operation mode vector are loaded into the arrester operation trend estimation network, and the estimated electrical state snapshot of the arrester in the monitoring section to be evaluated can be generated, so as to effectively predict the possible working state of the arrester in the future. Finally, according to the electrical state snapshot of the arrester in multiple monitoring sections and the estimated electrical state snapshot of the monitoring section to be evaluated, the performance change trend of the arrester can be generated, which can assist in judging the historical working performance of the arrester and provide a reference for predicting the future performance of the arrester. Thus, the working state and performance change trend of the arrester can be accurately reflected, which helps to discover possible problems in advance and take corresponding maintenance measures, thereby improving the operational safety and stability of the arrester.

[0062] In a possible implementation manner, the operating mode vector includes a fluctuation rate vector.

[0063] Step S120 may include: performing order-preserving discrete cosine transformation on the electrical state snapshots of the plurality of monitoring segments to generate the fluctuation rate vectors corresponding to the electrical state snapshots of the plurality of monitoring segments, wherein the order-preserving discrete cosine transformation is a discrete cosine transformation with an order-preserving property, and the order-preserving property is used to indicate that the generated feature vector has a property of consistent change effect with the loaded feature vector.

[0064] In this scenario, the operating mode vector processed by the arrester monitoring system is not just a simple snapshot of electrical parameters, but contains more in-depth information, namely the fluctuation rate vector. This fluctuation rate vector describes the rate of change of the arrester's electrical state between different monitoring segments, which can reflect the dynamic characteristics of the arrester during operation.

[0065] Therefore, in the process of extracting the state mode vector, the arrester monitoring system uses a special technology - Order-Preserving Discrete Cosine Transform (OP-DCT). This conversion method can not only capture the periodic changes in the electrical state snapshot, but also maintain the order of data changes, that is, the generated feature vector is consistent with the original data in the change trend. Through OP-DCT processing, the arrester monitoring system successfully extracted the fluctuation rate vector from the electrical state snapshot of each monitoring segment. These vectors describe the rate of change of the electrical state of the arrester between different monitoring segments, providing an important basis for subsequent performance evaluation.

[0066] In this scenario, the order-preserving property of the order-preserving discrete cosine transform plays a key role. Since the generated fluctuation rate vector needs to be consistent with the actual electrical state change trend, it is very appropriate to use the order-preserving discrete cosine transform. This conversion method ensures that the generated feature vector can accurately reflect the actual working state of the lightning arrester when used for performance evaluation.

[0067] Therefore, the arrester monitoring system collects snapshots of the arrester's electrical status in multiple monitoring sections and uses order-preserving discrete cosine transform to extract the fluctuation rate vector, providing accurate and reliable data support for evaluating the performance change trend of the arrester in a specific monitoring section.

[0068] In a possible implementation, the multiple monitoring segments include an x-th monitoring segment, and x is less than X. The lightning arrester includes Y lightning protection working events, and the electrical state snapshots of the multiple monitoring segments of the lightning arrester include electrical state nodes corresponding to the Y lightning protection working events in the x-th monitoring segment, respectively, and Y is not less than 1. The fluctuation rate vector includes fluctuation rate linkage parameters between the Y lightning protection working events.

[0069] In the present embodiment, in the monitoring center of the power system, the arrester monitoring system continuously collects and analyzes the operating status data of the arrester. In order to evaluate the changing trend of the arrester performance, the arrester monitoring system pays special attention to the electrical state snapshots of multiple monitoring segments, and has determined the lightning protection working events and their corresponding electrical state nodes contained in these electrical state snapshots. These electrical state snapshots contain the instantaneous values ​​of key parameters such as voltage and current of the arrester under different working conditions. Among them, the data of the xth monitoring segment can be understood as an important reference point for evaluating the performance of the arrester. The arrester records the electrical state nodes corresponding to Y lightning protection working events in the electrical state snapshots of multiple monitoring segments. These lightning protection working events are the key actions of the arrester in responding to thunderstorms, and have an important impact on the performance and life of the arrester.

[0070] The fluctuation rate vector is crucial for evaluating the performance of lightning arresters. The fluctuation rate vector describes the rate of change of the electrical state of the lightning arrester between different monitoring segments, and can reflect the dynamic characteristics of the lightning arrester when it is working. In order to extract the fluctuation rate vector, the lightning arrester monitoring system uses order-preserving discrete cosine transform (OP-DCT) to process the electrical state snapshot. The order-preserving discrete cosine transform can capture the periodic changes in the electrical state snapshot and maintain the order of data changes, ensuring that the generated feature vector is consistent with the original data in terms of change trend.

[0071] The performing order-preserving discrete cosine transformation on the electrical state snapshots of the plurality of monitoring segments to generate the fluctuation rate vectors corresponding to the electrical state snapshots of the plurality of monitoring segments respectively includes:

[0072] Step S121, generating a weight vector of the zth lightning protection working event at the i-th fluctuation rate based on the electrical state node of the zth lightning protection working event among the Y lightning protection working events in the x-th monitoring segment and the attention electrical state node of the Y lightning protection working events in the x-th monitoring segment.

[0073] Step S122, based on the weight vector of the zth lightning protection working event on the ith fluctuation rate, and the weight vector of the yth lightning protection working event among the Y lightning protection working events on the ith fluctuation rate, generate the fluctuation rate linkage parameter between the zth lightning protection working event and the yth lightning protection working event, the ith fluctuation rate is a fluctuation rate on the fluctuation rate domain, the i is a positive integer between 0 and X-1, and the fluctuation rate domain is represented by 0 to X-1 fluctuation rates. The z and y are not greater than the Y, and the z is not equal to the y.

[0074] When performing order-preserving discrete cosine transformation, the arrester monitoring system generates a weight vector of the zth lightning protection working event at the ith fluctuation rate based on the electrical state node of the zth lightning protection working event at the xth monitoring segment and the attention electrical state node of the Yth lightning protection working event at the xth monitoring segment. This weight vector reveals the importance of the zth lightning protection working event at a specific fluctuation rate.

[0075] Furthermore, the arrester monitoring system also compares the weight vectors of the zth lightning arrester working event and the yth lightning arrester working event at the same fluctuation rate to generate a fluctuation rate linkage parameter between them. This parameter measures the degree of mutual influence between two different lightning arrester working events under the same fluctuation conditions, providing an important basis for evaluating the performance of the arrester.

[0076] Therefore, the arrester monitoring system generates a complete fluctuation rate linkage parameter matrix by performing the above analysis on all Y lightning arrester working events. This fluctuation rate linkage parameter matrix reveals the dynamic interaction between the working events of the arrester when responding to different thunderstorm events, and provides a comprehensive view of the arrester performance for the arrester monitoring system.

[0077] Based on the fluctuation rate linkage parameter matrix, the arrester monitoring system can evaluate the performance change trend of the arrester in a specific monitoring section (such as the Xth monitoring section), discover potential problems in time and take corresponding maintenance measures.

[0078] In a possible implementation, step S121 may include:

[0079] Step S1211, determining the deviation parameter between the electrical state node of the zth lightning protection working event in the Y lightning protection working events in the xth monitoring segment and the attention electrical state node of the Y lightning protection working events in the xth monitoring segment, and generating the deviation parameter of the zth lightning protection working event in the xth monitoring segment. The deviation parameter reflects the migration order-preserving property of the electrical state node.

[0080] In this embodiment, the arrester monitoring system first focuses on the electrical state node of the zth lightning protection working event (z is a positive integer less than or equal to Y) in the xth monitoring segment. This electrical state node represents the specific electrical state of the arrester under the event, such as the instantaneous values ​​of parameters such as voltage and current. At the same time, the arrester monitoring system also considers the attention electrical state nodes of all Y lightning protection working events in the xth monitoring segment. These attention electrical state nodes may be state points determined by the arrester monitoring system based on experience or previous analysis, which are particularly critical for evaluating the performance of the arrester.

[0081] The arrester monitoring system calculates the deviation parameter between the electrical state node of the zth lightning protection working event and these attention electrical state nodes. This deviation parameter is not only a numerical difference, but more importantly, it reflects the migration order-preserving nature of the electrical state node, that is, the orderliness and trend of the state change. Through calculation, the arrester monitoring system obtains the deviation parameter of the zth lightning protection working event in the xth monitoring section, which provides the basis for subsequent analysis.

[0082] In detail, the electrical state node of each lightning protection working event can be understood as the specific electrical state exhibited by the lightning arrester on the xth monitoring segment when the lightning protection working event occurs. For example, when the lightning arrester encounters a thunderstorm and successfully discharges, the xth monitoring segment may capture the values ​​of parameters such as voltage and current at this time, and these values ​​constitute the electrical state node of the event. The attention electrical state node is particularly critical for evaluating the performance of the lightning arrester. For example, at certain specific voltage or current values, the performance of the lightning arrester will change significantly, so these values ​​will be selected as the attention electrical state nodes.

[0083] When the arrester monitoring system analyzes the zth lightning protection working event, the deviation parameter between the electrical state node of the lightning protection working event and the attention electrical state node can be calculated. This deviation parameter not only represents the difference in numerical value, but more importantly, it reflects the trend and orderliness of the electrical state change. Assume that in the xth monitoring section, the electrical state node of the zth lightning protection working event is a voltage value Vz, and the voltage value of an attention electrical state node is Va. The deviation parameter can be the difference between Vz and Va, that is, ΔV = Vz - Va. But this difference is not just a simple mathematical operation, it also carries the migration order-preserving properties of electrical state changes.

[0084] The migration order-preserving property refers to the electrical state maintaining a certain order and trend during the change process. If the deviation parameters between the electrical state nodes of multiple lightning protection working events and a certain attention electrical state node show a consistent increase or decrease trend, then it can be considered that these events show an orderly migration pattern in the electrical state. For example, if the lightning arrester monitoring system finds that as the thunderstorm intensifies, the voltage value of the lightning arrester in the xth monitoring section gradually increases, and the deviation parameters of each lightning protection working event show a positive growth trend, then the lightning arrester monitoring system can determine that the performance of the lightning arrester is gradually improving when responding to thunderstorms, or at least the changes in its electrical state are orderly and predictable.

[0085] Therefore, the deviation parameter not only reflects the numerical difference between the electrical state node of a single lightning protection working event and the attention electrical state node, but more importantly, it reveals the migration order-preserving nature of the electrical state changes between multiple events, providing an important basis for the arrester monitoring system to evaluate the arrester performance.

[0086] Step S1212, mapping the deviation parameter of the zth lightning protection working event in the xth monitoring section to the ith fluctuation rate in the fluctuation rate domain, and generating a weight vector of the zth lightning protection working event at the ith fluctuation rate.

[0087] Next, the arrester monitoring system needs to map this deviation parameter to a specific fluctuation rate in the fluctuation rate domain. The fluctuation rate domain is a set of multiple fluctuation rates, each of which represents a rate or frequency of electrical state change. The arrester monitoring system selects the i-th fluctuation rate (i is a positive integer between 0 and X-1) as the mapping target. This selection may be based on previous analysis experience or dynamically determined according to the needs of the current analysis. Through a specific mapping function or algorithm, the arrester monitoring system converts the deviation parameter of the z-th lightning protection working event in the x-th monitoring section into a weight vector on the i-th fluctuation rate. This weight vector describes the importance or influence of the event on the fluctuation rate.

[0088] In detail, the fluctuation rate domain is a set of multiple fluctuation rates. Each fluctuation rate represents a rate or frequency of electrical state change. These fluctuation rates can be determined based on experience, historical data or specific analysis requirements.

[0089] The arrester monitoring system needs to map the deviation parameter of the zth lightning arrester working event in the xth monitoring section to the ith fluctuation rate. This mapping process can be achieved through a specific function or algorithm.

[0090] Assume that the deviation parameter is a numerical value that represents the difference between the electrical state node and the attention electrical state node. The fluctuation rate domain is a continuous or discrete rate set. The goal of mapping is to find a suitable mathematical relationship to convert the numerical value of the deviation parameter into a weight value on the i-th fluctuation rate. Through mapping, the lightning arrester monitoring system obtains a weight value of the z-th lightning arrester working event on the i-th fluctuation rate. This weight value represents the importance or influence of the event on the fluctuation rate.

[0091] If the fluctuation rate domain is multidimensional (i.e., contains multiple fluctuation rates), the arrester monitoring system can generate a weight value for each fluctuation rate, thereby forming a weight vector. This weight vector describes the characteristics of the zth lightning arrester working event in the entire fluctuation rate domain.

[0092] In an example, it is assumed that the deviation parameter is a voltage difference ΔV, whose value is 5 kV. The fluctuation rate domain includes 5 fluctuation rates, which respectively represent different electrical state change rates.

[0093] The arrester monitoring system uses a mapping function to map ΔV to the i-th fluctuation rate. This function can be linear, nonlinear, or an interpolation function based on empirical data. Assume that the mapping function is f(ΔV), and f(5kV) = 0.8. Then, at the i-th fluctuation rate, the weight value of the z-th lightning protection working event is 0.8. If the fluctuation rate domain is multidimensional, the arrester monitoring system can calculate a weight value for each fluctuation rate, for example, the weight is 0.2 at the first fluctuation rate, the weight is 0.4 at the second fluctuation rate, and so on. Finally, the arrester monitoring system obtains a weight vector, such as [0.2, 0.4, 0.8, 0.3, 0.1], which describes the characteristics of the z-th lightning protection working event in the entire fluctuation rate domain. By mapping and generating weight vectors, the arrester monitoring system can more comprehensively evaluate the performance of the arrester at different fluctuation rates and provide strong support for the stable operation of the power system.

[0094] In a possible implementation, step S122 may include:

[0095] Step S1221, performing a dot product operation on the weight vector of the zth lightning protection working event at the i-th fluctuation rate and the weight vector of the yth lightning protection working event among the Y lightning protection working events at the i-th fluctuation rate to generate a dot product result.

[0096] In this embodiment, the lightning arrester monitoring system first selects the weight vector of the zth lightning protection working event at the i-th fluctuation rate. This weight vector represents the importance or influence of the lightning protection working event on the fluctuation rate. These weight vectors can be multidimensional, depending on the definition and dimension of the fluctuation rate domain. Then, the lightning arrester monitoring system selects the weight vector of the yth lightning protection working event (y is a positive integer less than or equal to Y, and y is not equal to z) at the i-th fluctuation rate. This weight vector also represents the characteristics of the corresponding event at the fluctuation rate. The lightning arrester monitoring system performs a dot multiplication operation on these two weight vectors. The dot multiplication operation is a vector operation, and the result is a scalar that reflects the similarity or consistency of the two vectors in direction. Through the dot multiplication operation, the lightning arrester monitoring system generates a dot multiplication result. This dot multiplication result is actually a scalar value, which quantitatively describes the correlation between the zth lightning protection working event and the yth lightning protection working event at the i-th fluctuation rate.

[0097] Assume that at the i-th fluctuation rate, the weight vector of the z-th lightning protection work event is [0.5, 0.8], and the weight vector of the y-th lightning protection work event is [0.4, 0.9].

[0098] The arrester monitoring system performs a dot multiplication operation: 0.5 * 0.4 + 0.8 * 0.9 = 0.2 + 0.72 = 0.92.

[0099] The generated dot product result is actually a scalar value of 0.92. This value represents the similarity or consistency of the weight vectors of the z-th lightning protection work event and the y-th lightning protection work event at the i-th fluctuation rate. The closer the value is to 1, the more similar the characteristics of the two events at the fluctuation rate; the closer the value is to -1, the more opposite the characteristics; the closer the value is to 0, the two events have almost no correlation at the fluctuation rate.

[0100] Through dot multiplication operation, the arrester monitoring system can quantify the correlation between different lightning protection working events at the same fluctuation rate, thereby providing an important basis for subsequent arrester performance evaluation and system stability analysis.

[0101] Step S1222: outputting the vector amplitude corresponding to the dot product result as the fluctuation rate linkage parameter between the z-th lightning protection working event and the y-th lightning protection working event.

[0102] The arrester monitoring system then calculates the vector amplitude corresponding to the dot product result. In the case of a scalar, the vector amplitude is actually the absolute value of the scalar value. However, in some contexts, this step may be redundant because the result of the dot product is already a scalar value. The arrester monitoring system outputs this calculated vector amplitude (or dot product result) as a fluctuation rate linkage parameter between the zth lightning protection working event and the yth lightning protection working event. This fluctuation rate linkage parameter measures the degree of mutual influence between two events at the same fluctuation rate and is an important indicator when evaluating the performance of the arrester.

[0103] Assume that the dot product of the z-th lightning protection operation event and the y-th lightning protection operation event at the ith fluctuation rate has been calculated to be 0.92 (as described in the previous example).

[0104] In this case, the dot product value of 0.92 directly reflects the linkage of the two events at the i-th fluctuation rate. Because this value is close to 1, it means that the weight vector directions of the two events at this fluctuation rate are very similar, and they have a strong linkage at this fluctuation rate. Therefore, the lightning arrester monitoring system outputs 0.92 as the fluctuation rate linkage parameter between the z-th lightning arrester working event and the y-th lightning arrester working event. This parameter can be used in subsequent lightning arrester performance evaluation or system stability analysis to help engineers understand the degree of correlation between different lightning arrester working events under the same fluctuation conditions.

[0105] Therefore, the vector amplitude corresponding to the dot product value is output as the fluctuation rate linkage parameter, which is actually to directly use the similarity value obtained by the dot product calculation as a quantitative indicator of the degree of correlation between two lightning protection working events at a specific fluctuation rate.

[0106] In a possible implementation manner, the fluctuation rate vector also includes a fluctuation rate change value of the z-th lightning protection working event. Thus, this embodiment can calculate the Euclidean length square value corresponding to the weight vector of the z-th lightning protection working event on the i-th fluctuation rate to generate the fluctuation rate change value of the z-th lightning protection working event.

[0107] In detail, each lightning protection work event has an associated fluctuation rate vector. This fluctuation rate vector not only includes the weight distribution of the lightning protection work event at different fluctuation rates, but also reflects the fluctuation rate change value of the lightning protection work event. For the zth lightning protection work event, its fluctuation rate vector naturally also includes the weight at the i-th fluctuation rate, as well as the fluctuation rate change value unique to the event.

[0108] In order to generate the fluctuation rate change value of the z-th lightning protection working event, the lightning arrester monitoring system first needs to calculate the square of the Euclidean length of the weight vector of the event at the i-th fluctuation rate. Euclidean length, also known as L2 norm, is a way to measure the size of a vector. The calculation method is to take the sum of the squares of each element in the weight vector. For example, if the weight vector of the z-th lightning protection working event at the i-th fluctuation rate is [0.3, 0.4, 0.5], then its Euclidean length squared value is 0.3² + 0.4² + 0.5² = 0.09 + 0.16 + 0.25 = 0.5.

[0109] The arrester monitoring system then uses the calculated Euclidean length square value to generate a fluctuation rate change value for the zth lightning protection working event. This fluctuation rate change value may be based on some transformation of the Euclidean length square value or directly use this value as an indicator. The specific form of the transformation may depend on the algorithm design of the arrester monitoring system and the requirements for arrester performance evaluation. For example, the arrester monitoring system can convert the Euclidean length square value through a predetermined function to obtain a standardized fluctuation rate change value between 0 and 1. The closer this value is to 1, the greater the weight vector amplitude of the zth lightning protection working event on the i-th fluctuation rate, that is, the more significant the impact of the event on the fluctuation rate.

[0110] Once the fluctuation rate change value of the zth lightning arrester operation event is generated, the arrester monitoring system can use it for various analyses. For example, by comparing the fluctuation rate change values ​​of different events, it can be identified which events are more critical under specific fluctuation conditions.

[0111] In a possible implementation, step S130 may include:

[0112] Step S131, loading the electrical state snapshots of the multiple monitoring segments and the corresponding operating mode vectors into the arrester operating trend estimation network, so that the operating mode vector is transmitted to each layer of the arrester operating trend estimation network in the forward propagation, and the estimated electrical state node of the arrester in the Xth monitoring segment is generated.

[0113] The arrester monitoring system first loads the electrical state snapshots of multiple monitoring segments and their corresponding operation mode vectors into the arrester operation trend estimation network. Under the action of the operation mode vector, these data are forward propagated layer by layer in the network. Each layer of the network will calculate based on the input data and its internal weights, biases and other parameters to generate a new data representation. In this way, the information of the operation mode vector is gradually transmitted and integrated into each layer of the network, affecting the final output result.

[0114] Through the forward propagation of the arrester operation trend estimation network, the arrester monitoring system generates the estimated electrical state nodes of the arrester in the Xth monitoring section. These estimated electrical state nodes are a set of data representations inside the network, which reflect the possible electrical state of the arrester in this monitoring section. The values ​​of these estimated electrical state nodes are automatically calculated by the arrester operation trend estimation network based on the input data and its learned patterns.

[0115] Step S132: generating a snapshot of the estimated electrical state of the lightning arrester in the Xth monitoring section according to the estimated electrical state node of the lightning arrester in the Xth monitoring section.

[0116] Once the estimated electrical state nodes are obtained, the arrester monitoring system can generate an estimated electrical state snapshot of the arrester in the Xth monitoring segment based on these estimated electrical state nodes. This estimated electrical state snapshot is a comprehensive representation of the information of each electrical state node of the arrester in this monitoring segment. The arrester monitoring system may convert the node value into a more intuitive form, such as a numerical value or chart of electrical parameters such as voltage and current, to facilitate engineers to analyze and judge.

[0117] In a possible implementation manner, the operation mode vector includes at least one of a fluctuation rate linkage parameter between the zth lightning protection working event and the yth lightning protection working event among the Y lightning protection working events and a fluctuation rate change value of the zth lightning protection working event.

[0118] The method further comprises:

[0119] Step A110: adding the fluctuation rate linkage parameter between the z-th lightning protection working event and the y-th lightning protection working event in the Y lightning protection working events as a state link vector in the electrical state snapshot.

[0120] And, in step A120, the fluctuation rate change value of the zth lightning protection working event is added as the state member vector in the electrical state snapshot, so that the operation mode vector is transmitted to each layer of the lightning arrester operation trend estimation network in the forward propagation.

[0121] First, this embodiment clarifies that the operation mode vector includes the fluctuation rate linkage parameter between the zth lightning protection working event and the yth lightning protection working event among the Y lightning protection working events, and at least one of the fluctuation rate change values ​​of the zth lightning protection working event. This information reflects the mutual correlation of different lightning protection working events in electrical state and the state change of a single event.

[0122] In order to integrate this important information into the electrical state snapshot, the fluctuation rate linkage parameter between the zth lightning protection working event and the yth lightning protection working event can be added as the state link vector in the electrical state snapshot. That is, a new data field or structure is created in the electrical state snapshot, which is specifically used to store the linkage parameters between the two lightning protection working events, which helps to better capture and utilize the relationship between events in subsequent network processing.

[0123] At the same time, the arrester monitoring system adds the fluctuation rate change value of the zth lightning protection working event as the state member vector in the electrical state snapshot, which also reserves a position for the state change information of a single lightning protection working event in the lightning protection working snapshot, ensuring that this information can be fully considered in subsequent analysis. Once the electrical state snapshot is enhanced with this additional information, the arrester monitoring system can load it into the arrester operation trend estimation network. During the forward propagation process of each layer of the network, the information of the operating mode vector (including the state link vector and the state member vector) is transmitted and processed layer by layer. The neurons and connections in the network will calculate and learn based on these input data to generate more accurate arrester state estimation results.

[0124] Through the above steps, not only the content of the electrical status snapshot is enriched, but also it is ensured that this information can be effectively used in the arrester operation trend estimation network, which helps to improve the accuracy and reliability of state estimation, thereby better supporting the stable operation of the power system and the performance evaluation of the arrester.

[0125] In one possible implementation, the method includes:

[0126] Step S101, obtaining a template electrical state snapshot of a template arrester in multiple monitoring sections. The multiple monitoring sections are monitoring sections located forward of the Xth monitoring section, and the template electrical state snapshot carries the annotated operating trend characteristic data of the arrester working mode of the template arrester in the Xth monitoring section, where X is greater than 1.

[0127] Step S102, extracting state mode vectors from the template electrical state snapshots of the plurality of monitoring segments, generating template operation mode vectors corresponding to the template electrical state snapshots of the plurality of monitoring segments, wherein the template operation mode vectors reflect the cyclic characteristic changes of the arrester working mode of the template arrester.

[0128] Step S103, loading the template electrical state snapshots of the multiple monitoring segments and the corresponding template operation mode vectors to the arrester operation trend estimation network, generating operation trend training data of the template arrester in the arrester operation mode of the Xth monitoring segment.

[0129] Step S104, determining the error parameter of the arrester operation trend estimation network according to the labeled operation trend characteristic data and the operation trend training data.

[0130] Step S105: training the arrester operation trend estimation network according to the error parameter.

[0131] The arrester monitoring system first obtains the template electrical status snapshots of the template arrester in multiple monitoring sections. These monitoring sections are the monitoring sections located in front of the Xth monitoring section, and the template electrical status snapshot carries the annotated operation trend characteristic data of the arrester working mode of the template arrester in the Xth monitoring section. Here, X is an integer greater than 1, indicating that the arrester monitoring system is concerned with the arrester working status after the Xth monitoring section.

[0132] Next, the arrester monitoring system extracts the state mode vectors of the template electrical state snapshots of the acquired multiple monitoring segments. The purpose of this step is to extract information from the snapshots that can reflect the cyclical characteristic changes of the arrester's working mode. For example, the electrical state snapshots can be deeply analyzed to extract the characteristic data therein. These data include electrical parameters such as voltage, current, temperature, and information such as the working state and operating environment of the arrester. By processing and analyzing these data, the arrester monitoring system generates template operation mode vectors corresponding to the template electrical state snapshots of multiple monitoring segments.

[0133] Next, the arrester monitoring system loads the extracted template electrical state snapshot and the corresponding template operation mode vector into the arrester operation trend estimation network. By loading these data, the arrester monitoring system generates the operation trend training data of the arrester operation mode of the template arrester in the Xth monitoring segment. These data are the key inputs for network training and learning. They contain rich information about the arrester under historical working conditions, which helps the network learn an accurate prediction model.

[0134] Finally, the arrester monitoring system determines the error parameters of the arrester operation trend estimation network based on the annotated operation trend feature data and the operation trend training data. These error parameters reflect the difference between the network prediction results and the actual arrester working state. The arrester monitoring system uses these error parameters to train the arrester operation trend estimation network. By continuously adjusting the network's weights and bias parameters, the arrester monitoring system gradually reduces the prediction error and improves the network's prediction accuracy. Finally, the arrester monitoring system obtains a fully trained deep learning model that can accurately predict the arrester operation trend.

[0135] Among them, for the arrester operation trend estimation network, a commonly used loss function, such as Mean Squared Error (MSE), can be used.

[0136] Assume the following variables and symbols:

[0137] (y_i): the actual annotated operating trend characteristic data of the ith sample (i.e. the true value of the arrester working mode)

[0138] (\hat{y}_i): Operation trend training data of the ith sample (i.e., network prediction value of the arrester working mode)

[0139] (N): Total number of samples (i.e., number of electrical state snapshots and template operating mode vectors used for training)

[0140] The formula for mean square error (MSE) can be expressed as:

[0141] [MSE = \frac{1}{N} \sum_{i=1}^{N} (y_i - \hat{y}_i)^2]

[0142] In this formula, the square of the difference between the actual value and the predicted value is calculated for each sample, and then all these squared differences are summed and finally divided by the total number of samples to get the average error.

[0143] In order to optimize the network and reduce this error, gradient descent or its variants (such as stochastic gradient descent, Adam, etc.) are usually used to update the weights and biases of the network. These optimization algorithms adjust the network parameters according to the gradient of the error to minimize the loss function. After the error parameters are determined, the gradient of each parameter in the network can be calculated through the back propagation algorithm, and the network weights can be updated accordingly, thereby training the arrester operation trend estimation network to improve its prediction accuracy.

[0144] Wherein, the template operation mode vector includes a template fluctuation rate vector.

[0145] The state mode vector extraction of the template electrical state snapshots of the multiple monitoring segments to generate the template operation mode vectors corresponding to the template electrical state snapshots of the multiple monitoring segments includes: performing order-preserving discrete cosine transformation on the template electrical state snapshots of the multiple monitoring segments to generate the template fluctuation rate vectors corresponding to the template electrical state snapshots of the multiple monitoring segments. The order-preserving discrete cosine transformation is a discrete cosine transformation with an order-preserving property, and the order-preserving property is used to indicate that the generated feature vector has a property of consistent change effect with the loaded feature vector.

[0146] The multiple monitoring segments include the xth monitoring segment, the template lightning arrester includes Y template lightning arrester working events, and the template electrical state snapshots of the multiple monitoring segments include the template electrical state nodes corresponding to the Y template lightning arrester working events in the xth monitoring segment, and Y is not less than 1. The template fluctuation rate vector includes template fluctuation rate linkage parameters between the Y template lightning arrester working events.

[0147] The step of performing order-preserving discrete cosine transformation on the template electrical state snapshots of the multiple monitoring segments to generate the template fluctuation rate vectors corresponding to the template electrical state snapshots of the multiple monitoring segments includes: generating a template weight vector of the zth template lightning protection working event at the i-th fluctuation rate based on the template electrical state node of the zth template lightning protection working event among the Y template lightning protection working events at the x-th monitoring segment, and the template attention electrical state node of the Y template lightning protection working event at the x-th monitoring segment. Generating the template fluctuation rate linkage parameter between the zth template lightning protection working event and the yth template lightning protection working event based on the template weight vector of the zth template lightning protection working event at the i-th fluctuation rate, and the template weight vector of the yth template lightning protection working event among the Y template lightning protection working events at the i-th fluctuation rate. Wherein, the i-th fluctuation rate is a fluctuation rate on the fluctuation rate domain, i is a positive integer between 0 and X-1, and the fluctuation rate domain is represented by 0 to X-1 fluctuation rates. The z and y are not greater than the Y, and the z is not equal to the y.

[0148] The method of generating a template weight vector of the zth template lightning protection working event at the i-th fluctuation rate based on the template electrical state node of the zth template lightning protection working event among the Y template lightning protection working events in the x-th monitoring segment and the template attention electrical state node of the Y template lightning protection working events in the x-th monitoring segment comprises: determining the deviation parameter between the template electrical state node of the zth template lightning protection working event among the Y template lightning protection working events in the x-th monitoring segment and the template attention electrical state node of the Y template lightning protection working events in the x-th monitoring segment, and generating a template deviation parameter of the zth template lightning protection working event in the x-th monitoring segment. The template deviation parameter reflects the migration order-preserving property of the template electrical state node. The template deviation parameter of the zth template lightning protection working event in the x-th monitoring segment is mapped to the i-th fluctuation rate in the fluctuation rate domain, and a template weight vector of the zth template lightning protection working event at the i-th fluctuation rate is generated.

[0149] The template fluctuation rate linkage parameter between the zth template lightning protection working event and the yth template lightning protection working event is generated based on the template weight vector of the zth template lightning protection working event at the i-th fluctuation rate and the template weight vector of the yth template lightning protection working event among the Y template lightning protection working events at the i-th fluctuation rate, including: performing a dot multiplication operation on the template weight vector of the zth template lightning protection working event at the i-th fluctuation rate and the template weight vector of the yth template lightning protection working event among the Y template lightning protection working events at the i-th fluctuation rate to generate a dot multiplication result. The vector amplitude corresponding to the dot multiplication result is output as the template fluctuation rate linkage parameter between the zth template lightning protection working event and the yth template lightning protection working event. The template fluctuation rate vector also includes the template fluctuation rate change value of the zth template lightning protection working event.

[0150] The method further comprises:

[0151] The Euclidean length square value corresponding to the template weight vector of the z-th template lightning protection working event at the i-th fluctuation rate is calculated to generate the template fluctuation rate change value of the z-th template lightning protection working event.

[0152] The step of loading the template electrical state snapshots of the plurality of monitoring segments and the template operation mode vectors corresponding to the respective monitoring segments into the arrester operation trend estimation network to generate operation trend training data of the template arrester in the arrester operation mode of the Xth monitoring segment includes:

[0153] The template electrical state snapshots of the multiple monitoring segments and the corresponding template operation mode vectors are loaded into the arrester operation trend estimation network, so that the template operation mode vector is transmitted to each layer of the arrester operation trend estimation network in the forward propagation, and the estimated electrical state node of the template arrester in the Xth monitoring segment is generated.

[0154] The operation trend training data of the arrester working mode of the template arrester in the Xth monitoring section is generated according to the estimated electrical state node of the template arrester in the Xth monitoring section.

[0155] It is worth noting that the embodiments of the above training process can refer to the relevant descriptions of the aforementioned embodiments, which will not be repeated here.

[0156] Figure 2 The hardware structure of the arrester monitoring system 100 provided in the embodiment of the present invention for implementing the above-mentioned arrester monitoring method is shown. Figure 2As shown, the arrester monitoring system 100 may include a processor 110 , a machine-readable storage medium 120 , a bus 130 , and a communication unit 140 .

[0157] In an exemplary design concept, the arrester monitoring system 100 can be a single arrester monitoring system or a server group. The server group can be centralized or distributed (for example, the arrester monitoring system 100 can be a distributed system). In an exemplary design concept, the arrester monitoring system 100 can be local or remote. For example, the arrester monitoring system 100 can access data and / or data stored in the machine-readable storage medium 120 via a network. For another example, the arrester monitoring system 100 can be directly connected to the machine-readable storage medium 120 to access the stored data and / or data. In an exemplary design concept, the arrester monitoring system 100 can be implemented on a cloud platform. As an example only, the cloud platform can include a private cloud, a public cloud, a hybrid cloud, a community cloud, a distributed cloud, an internal cloud, a multi-layer cloud, etc. or any combination thereof.

[0158] The machine-readable storage medium 120 can store data and / or instructions. In an exemplary design concept, the machine-readable storage medium 120 can store multidimensional monitoring data obtained from an external terminal. In an exemplary design concept, the machine-readable storage medium 120 can store multidimensional monitoring data and / or instructions used by the arrester monitoring system 100 to execute or use to complete the exemplary method described in the present invention. In an exemplary design concept, the machine-readable storage medium 120 may include a large-capacity memory, a removable memory, a volatile read-write memory, a read-only memory (ROM), etc. or any combination thereof. Exemplary large-capacity memory may include a disk, an optical disk, a solid-state disk, etc. Exemplary removable memory may include a flash drive, a floppy disk, an optical disk, a memory card, a compressed disk, a tape, etc. Exemplary volatile read-write memory may include a random access memory (RAM). Exemplary RAM may include active random access memory (DRAM), double data rate synchronous active random access memory (DDR SDRAM), passive random access memory (SRAM), thyristor random access memory (T-RAM) and zero capacitance random access memory (Z-RAM), etc. Exemplary read-only memories may include mask read-only memories (MROMs), programmable read-only memories (PROMs), erasable programmable read-only memories (PEROMs), electrically erasable programmable read-only memories (EEPROMs), compact disk read-only memories (CD-ROMs), and digital versatile disk read-only memories, etc. In an exemplary design concept, the machine-readable storage medium 120 may be implemented on a cloud platform. As an example only, a cloud platform may include a private cloud, a public cloud, a hybrid cloud, a community cloud, a distributed cloud, an internal cloud, a multi-layer cloud, etc., or any combination thereof.

[0159] During the specific implementation process, at least one processor 110 executes computer executable instructions stored in the machine-readable storage medium 120, so that the processor 110 can execute the lightning arrester monitoring method of the above method embodiment. The processor 110, the machine-readable storage medium 120 and the communication unit 140 are connected through the bus 130, and the processor 110 can be used to control the sending and receiving actions of the communication unit 140.

[0160] The specific implementation process of the processor 110 can refer to the various method embodiments executed by the above-mentioned lightning arrester monitoring system 100. The implementation principles and technical effects are similar, and this embodiment will not be repeated here.

[0161] In addition, an embodiment of the present invention further provides a readable storage medium, in which computer executable instructions are preset. When a processor executes the computer executable instructions, the above-mentioned lightning arrester monitoring method is implemented.

[0162] It should be understood that the above description is for illustrative purposes only and is not intended to limit the scope of the present invention. For those of ordinary skill in the art, various modifications and variations can be made in conjunction with the description of the present invention. However, these modifications and variations will not deviate from the scope of the present invention.

[0163] The basic concepts have been described above. Obviously, for those of ordinary skill in the art who have read this application, the above invention disclosure is only for example and does not constitute a limitation of the present invention. Although not explicitly stated herein, those skilled in the art may make various modifications, improvements and corrections to the present invention. Such modifications, improvements and corrections are suggested in the present invention, so such modifications, improvements and corrections still belong to the spirit and scope of the exemplary embodiments of the present invention.

[0164] At the same time, the present invention uses specific words to describe the embodiments of the present invention. For example, "one embodiment", "an embodiment", and / or "some embodiments" refer to a certain feature, structure or characteristic related to at least one embodiment of the present invention. Therefore, it should be emphasized and noted that "one embodiment" or "an embodiment" or "an alternative embodiment" mentioned twice or more in different positions in this specification does not necessarily refer to the same embodiment. In addition, some features, structures or characteristics in one or more embodiments of the present invention can be appropriately combined.

[0165] In addition, it will be appreciated by those skilled in the art that various aspects of the present invention may be illustrated and described in conjunction with a plurality of patentable categories or situations, including any new and useful process, machine, product, or combination of substances, or any new and useful improvements thereto. Accordingly, various aspects of the present invention may be performed entirely by hardware, entirely by software (including firmware, resident software, microcode, etc.), or by a combination of hardware and software. The above hardware or software may all be referred to as "units," "modules," or "systems." In addition, various aspects disclosed in the present invention may take the form of a computer program product embodied in one or more computer-readable media, wherein computer-readable program code is contained therein.

[0166] A computer readable signal medium may include a propagated data signal containing computer program code, such as in baseband or as part of a carrier wave. Such propagated signals may be in a variety of forms, including electromagnetic, optical, etc., or any suitable combination. A computer readable signal medium may be any computer readable medium other than a computer readable storage medium, which may be coupled to an instruction execution system, device, or apparatus to communicate, propagate, or transmit a program for use. The program code on the computer readable signal medium may be transmitted in conjunction with any suitable medium, including radio, cable, fiber optic cable, RF, or the like, or any combination thereof.

[0167] The computer program code required for the operation of each part of the present invention can be written in any one or more programming languages, including subject-oriented programming languages ​​such as Java, Scala, Smalltalk, Eiffel, JADE, Emerald, C++, C#, VB.NET, Python, etc., conventional procedural programming languages ​​such as C language, Visual Basic, Fortran 2003, Perl, COBOL 2002, PHP, ABAP, active programming languages ​​such as Python, Ruby and Groovy, or other programming languages, etc. The program code can be run entirely on the computer, or run on the computer as an independent software package, or run partially on the computer and partially on the remote computer, or run entirely on the remote computer or server. In the latter case, the remote computer can be connected to the computer in combination with any network form, such as a local area network (LAN) or a wide area network (WAN), or connected to an external computer (e.g., in combination with the Internet), or in a cloud computing environment, or used as a service such as software as a service (SaaS).

[0168] In addition, unless explicitly stated in the claims, the order of the processing elements and sequences described in the present invention, the use of alphanumeric characters, or the use of other names are not intended to limit the order of the processes and methods of the present invention. Although the above disclosure discusses some invention embodiments currently considered useful in conjunction with various examples, it should be understood that such details are only for illustrative purposes, and the attached claims are not limited to the disclosed embodiments. On the contrary, the claims are intended to cover all modifications and equivalent combinations that match the essence and scope of the embodiments of the present invention. For example, although the system components described above can be implemented in conjunction with hardware devices, they can also be implemented in conjunction with software solutions alone, such as installing the described system on an existing server or mobile device.

[0169] Similarly, it should be noted that in order to simplify the description of the present invention disclosure and thus help understand one or more embodiments of the invention, in the foregoing description of the embodiments of the invention, sometimes multiple features are grouped into one embodiment, drawings, or descriptions thereof. Similarly, it should be noted that in order to simplify the description of the present invention disclosure and thus help understand one or more embodiments of the invention, in the foregoing description of the embodiments of the invention, sometimes multiple features are grouped into one embodiment, drawings, or descriptions thereof.

Claims

1. A lightning arrester monitoring method, characterized in that: The method comprises: Obtaining snapshots of electrical states of the arrester in multiple monitoring sections; the multiple monitoring sections are monitoring sections located forward of the Xth monitoring section, the Xth monitoring section is the monitoring section to be evaluated, X is greater than 1, and the forward direction refers to the monitoring section located before the monitoring section to be evaluated in the time series; Extracting state mode vectors from the electrical state snapshots of the plurality of monitoring segments, and generating operation mode vectors corresponding to the electrical state snapshots of the plurality of monitoring segments; the operation mode vectors reflect cyclic characteristic changes of the arrester working mode of the arrester; Loading the electrical state snapshots of the plurality of monitoring segments and the respective corresponding operation mode vectors into the arrester operation trend estimation network to generate an estimated electrical state snapshot of the arrester in the Xth monitoring segment; Generate a performance change trend of the arrester according to the electrical state snapshots of the arrester in the plurality of monitoring segments and the estimated electrical state snapshot in the Xth monitoring segment; The operating mode vector includes a fluctuation rate vector; The extracting state mode vectors of the electrical state snapshots of the plurality of monitoring segments to generate operation mode vectors corresponding to the electrical state snapshots of the plurality of monitoring segments respectively includes: Performing order-preserving discrete cosine transformation on the electrical state snapshots of the multiple monitoring segments to generate the fluctuation rate vectors corresponding to the electrical state snapshots of the multiple monitoring segments; The order-preserving discrete cosine transform is a discrete cosine transform with an order-preserving property, and the order-preserving property is used to indicate that the generated feature vector has a property of having a consistent change effect with the feature vector loaded thereto; The multiple monitoring segments include an x-th monitoring segment, where x is less than X; the arrester includes Y lightning protection working events, and the arrester includes the electrical state nodes corresponding to the Y lightning protection working events in the x-th monitoring segment in the electrical state snapshots of the multiple monitoring segments, where Y is not less than 1; the fluctuation rate vector includes fluctuation rate linkage parameters between the Y lightning protection working events; The performing order-preserving discrete cosine transformation on the electrical state snapshots of the plurality of monitoring segments to generate the fluctuation rate vectors corresponding to the electrical state snapshots of the plurality of monitoring segments respectively includes: Generate a weight vector of the zth lightning protection working event at the i-th fluctuation rate according to the electrical state node of the zth lightning protection working event among the Y lightning protection working events at the x-th monitoring segment and the attention electrical state node of the Y lightning protection working events at the x-th monitoring segment; According to the weight vector of the zth lightning protection working event on the ith fluctuation rate, and the weight vector of the yth lightning protection working event among the Y lightning protection working events on the ith fluctuation rate, the fluctuation rate linkage parameter between the zth lightning protection working event and the yth lightning protection working event is generated, the ith fluctuation rate is a fluctuation rate on the fluctuation rate domain, the i is a positive integer between 0 and X-1, and the fluctuation rate domain is represented by 0 to X-1 fluctuation rates; the z and the y are not greater than the Y, and the z is not equal to the y; The method generates a weight vector of the zth lightning protection working event at the i-th fluctuation rate according to the zth lightning protection working event among the Y lightning protection working events at the electrical state node of the x-th monitoring segment and the attention electrical state node of the Y lightning protection working events at the x-th monitoring segment, including: Determine the deviation parameter between the electrical state node of the zth lightning protection working event among the Y lightning protection working events in the xth monitoring segment and the attention electrical state node of the Y lightning protection working events in the xth monitoring segment, and generate the deviation parameter of the zth lightning protection working event in the xth monitoring segment; the deviation parameter reflects the migration order-preserving property of the electrical state node; Mapping the deviation parameter of the z-th lightning protection working event in the x-th monitoring section to the i-th fluctuation rate in the fluctuation rate domain, and generating a weight vector of the z-th lightning protection working event on the i-th fluctuation rate; The generating of the fluctuation rate linkage parameter between the zth lightning protection working event and the yth lightning protection working event according to the weight vector of the zth lightning protection working event on the ith fluctuation rate and the weight vector of the yth lightning protection working event among the Y lightning protection working events on the ith fluctuation rate includes: Performing a dot product operation on the weight vector of the zth lightning protection working event at the ith fluctuation rate and the weight vector of the yth lightning protection working event among the Y lightning protection working events at the ith fluctuation rate to generate a dot product result; The vector amplitude corresponding to the dot product result is output as the fluctuation rate linkage parameter between the z-th lightning protection working event and the y-th lightning protection working event.

2. The lightning arrester monitoring method according to claim 1, characterized in that: The fluctuation rate vector also includes the fluctuation rate change value of the zth lightning protection working event; The method further comprises: The squared Euclidean length value corresponding to the weight vector of the z-th lightning protection working event on the i-th fluctuation rate is calculated to generate the fluctuation rate change value of the z-th lightning protection working event.

3. The lightning arrester monitoring method according to claim 1, characterized in that: The step of loading the electrical state snapshots of the plurality of monitoring segments and the respective corresponding operation mode vectors into the arrester operation trend estimation network to generate an estimated electrical state snapshot of the arrester in the Xth monitoring segment includes: Loading the electrical state snapshots of the plurality of monitoring segments and the corresponding operation mode vectors to the arrester operation trend estimation network, so that the operation mode vector is transmitted to each layer of the arrester operation trend estimation network in the forward propagation, and generating an estimated electrical state node of the arrester in the Xth monitoring segment; According to the estimated electrical state node of the lightning arrester in the Xth monitoring section, a snapshot of the estimated electrical state of the lightning arrester in the Xth monitoring section is generated.

4. The lightning arrester monitoring method according to claim 3, characterized in that: The operation mode vector includes at least one of a fluctuation rate linkage parameter between the zth lightning protection working event and the yth lightning protection working event among the Y lightning protection working events and a fluctuation rate change value of the zth lightning protection working event; The method further comprises: Adding the fluctuation rate linkage parameter between the z-th lightning protection working event and the y-th lightning protection working event in the Y lightning protection working events as a state link vector in the electrical state snapshot; And, the fluctuation rate change value of the zth lightning protection working event is added as the state member vector in the electrical state snapshot, so that the operation mode vector is transmitted to each layer of the lightning arrester operation trend estimation network in the forward propagation.

5. The lightning arrester monitoring method according to claim 1, characterized in that: The method comprises: Obtaining a template electrical state snapshot of the template arrester in multiple monitoring sections; the multiple monitoring sections are monitoring sections located forward of the Xth monitoring section, the template electrical state snapshot carries the annotated operation trend characteristic data of the arrester working mode of the template arrester in the Xth monitoring section, where X is greater than 1; Extracting state mode vectors from the template electrical state snapshots of the plurality of monitoring segments, generating template operation mode vectors corresponding to the template electrical state snapshots of the plurality of monitoring segments; the template operation mode vectors reflect the cyclic characteristic changes of the arrester working mode of the template arrester; Loading the template electrical state snapshots of the multiple monitoring segments and the template operation mode vectors corresponding to each of the plurality of monitoring segments into the arrester operation trend estimation network, and generating operation trend training data of the template arrester in the arrester operation mode of the Xth monitoring segment; Determining error parameters of the arrester operation trend estimation network based on the annotated operation trend characteristic data and the operation trend training data; Training the arrester operation trend estimation network according to the error parameter; Wherein, the template operation mode vector includes a template fluctuation rate vector; The extracting state mode vectors of the template electrical state snapshots of the multiple monitoring segments to generate template operation mode vectors corresponding to the template electrical state snapshots of the multiple monitoring segments respectively includes: Performing order-preserving discrete cosine transformation on the template electrical state snapshots of the multiple monitoring segments to generate the template fluctuation rate vectors corresponding to the template electrical state snapshots of the multiple monitoring segments; The order-preserving discrete cosine transform is a discrete cosine transform with an order-preserving property, and the order-preserving property is used to indicate that the generated feature vector has a property of having a consistent change effect with the feature vector loaded thereto; The multiple monitoring segments include the xth monitoring segment, the template lightning arrester includes Y template lightning protection working events, the template lightning arrester includes the template electrical state nodes corresponding to the Y template lightning protection working events in the xth monitoring segment in the template electrical state snapshots of the multiple monitoring segments, and Y is not less than 1; the template fluctuation rate vector includes the template fluctuation rate linkage parameters between the Y template lightning protection working events; The performing order-preserving discrete cosine transformation on the template electrical state snapshots of the multiple monitoring segments to generate the template fluctuation rate vectors corresponding to the template electrical state snapshots of the multiple monitoring segments respectively includes: Generate a template weight vector of the zth template lightning protection working event at the i-th fluctuation rate according to the template electrical state node of the zth template lightning protection working event among the Y template lightning protection working events at the x-th monitoring segment and the template attention electrical state node of the Y template lightning protection working events at the x-th monitoring segment; According to the template weight vector of the zth template lightning protection working event at the i-th fluctuation rate, and the template weight vector of the yth template lightning protection working event among the Y template lightning protection working events at the i-th fluctuation rate, the template fluctuation rate linkage parameter between the zth template lightning protection working event and the yth template lightning protection working event is generated; wherein the i-th fluctuation rate is a fluctuation rate on the fluctuation rate domain, the i is a positive integer between 0 and X-1, and the fluctuation rate domain is represented by 0 to X-1 fluctuation rates; the z and the y are not greater than the Y, and the z is not equal to the y; The method of generating a template weight vector of the zth template lightning protection working event at the i-th fluctuation rate based on the zth template lightning protection working event among the Y template lightning protection working events at the template electrical state node of the x-th monitoring segment and the template attention electrical state node of the Y template lightning protection working events at the x-th monitoring segment comprises: Determine the deviation parameter between the template electrical state node of the zth template lightning protection working event in the Y template lightning protection working events in the xth monitoring segment and the template attention electrical state node of the Y template lightning protection working events in the xth monitoring segment, and generate the template deviation parameter of the zth template lightning protection working event in the xth monitoring segment; the template deviation parameter reflects the migration order-preserving property of the template electrical state node; Mapping the template deviation parameter of the zth template lightning protection working event in the xth monitoring section to the i-th fluctuation rate in the fluctuation rate domain, and generating a template weight vector of the zth template lightning protection working event at the i-th fluctuation rate; The template fluctuation rate linkage parameter between the zth template lightning protection working event and the yth template lightning protection working event is generated based on the template weight vector of the zth template lightning protection working event at the i-th fluctuation rate and the template weight vector of the yth template lightning protection working event among the Y template lightning protection working events at the i-th fluctuation rate, including: Performing a dot product operation on the template weight vector of the zth template lightning protection working event at the ith fluctuation rate and the template weight vector of the yth template lightning protection working event among the Y template lightning protection working events at the ith fluctuation rate to generate a dot product result; Outputting the vector amplitude corresponding to the dot product result as the template fluctuation rate linkage parameter between the z-th template lightning protection working event and the y-th template lightning protection working event; The template fluctuation rate vector also includes the template fluctuation rate change value of the zth template lightning protection working event; The method further comprises: Calculate the Euclidean length square value corresponding to the template weight vector of the z-th template lightning protection working event at the i-th fluctuation rate, and generate the template fluctuation rate change value of the z-th template lightning protection working event; The step of loading the template electrical state snapshots of the plurality of monitoring segments and the template operation mode vectors corresponding to the respective monitoring segments into the arrester operation trend estimation network to generate operation trend training data of the template arrester in the arrester operation mode of the Xth monitoring segment includes: Loading the template electrical state snapshots of the multiple monitoring segments and the corresponding template operation mode vectors to the arrester operation trend estimation network, so that the template operation mode vector is transmitted to each layer of the arrester operation trend estimation network in the forward propagation, and the estimated electrical state node of the template arrester in the Xth monitoring segment is generated; The operation trend training data of the arrester working mode of the template arrester in the Xth monitoring section is generated according to the estimated electrical state node of the template arrester in the Xth monitoring section.

6. A lightning arrester monitoring system, characterized in that: The arrester monitoring system comprises a processor and a machine-readable storage medium, wherein the machine-readable storage medium stores machine-executable instructions, and the machine-executable instructions are loaded and executed by the processor to implement the arrester monitoring method of any one of claims 1-5.

Images