Lightning arrester health state monitoring system and method based on big data analysis

Abstract

The invention discloses a lightning arrester health state monitoring system and method based on big data analysis, and relates to the field of lightning arrester detection, the system comprises a data acquisition module, a central processing module, a data storage module and a user interface; the data acquisition module is used for acquiring lightning arrester data; the central processing module comprises data preprocessing, time sequence analysis, fault detection and classification and health state scoring; the data storage module is used for receiving and storing data transmitted by the central processing module and the user interface and constructing a data set; and the user interface is used for displaying the abnormal condition, the abnormal condition type and the health state score. According to the invention, environmental pollution data is incorporated into the model, and the influence of environmental factors on lightning arrester data is displayed; an isolated forest is used for completing identification of abnormal conditions, and an alarm is given out immediately; a weighting method and color coding are adopted, and the health state score of the lightning arrester is visually displayed.

Description

Technical Field

[0001] The present invention relates to the field of lightning arrester detection, and in particular to a lightning arrester health status monitoring system and method based on big data analysis. Background Art

[0002] Lightning arresters are key protective devices in power systems, mainly used to prevent lightning from striking electrical equipment. Traditional lightning arrester detection methods have some defects, such as poor real-time performance, low monitoring accuracy, and insufficient assessment of the impact of environmental pollutants.

[0003] Traditional methods of lightning arrester detection mainly include leakage current test, resistive current test and infrared thermal imaging detection. Although these methods can evaluate the performance of lightning arresters to a certain extent, they usually rely on regular inspections, cannot provide real-time health status of lightning arresters, may miss key fault signals, and do not adequately consider the impact of environmental factors. For example, environmental pollutants such as sulfur oxides, nitrogen oxides and dust may affect the performance of lightning arresters, but these factors are often ignored in traditional detection. In addition, traditional detection methods also have limitations in data processing and analysis, making it difficult to achieve a comprehensive assessment of the health status of lightning arresters. Summary of the invention

[0004] The object of the present invention is to provide a lightning arrester health status monitoring system and method based on big data analysis to solve the problems raised in the prior art.

[0005] To achieve the above object, the present invention provides the following technical solution: a lightning arrester health status monitoring method based on big data analysis, the method comprising the following steps: S1. Install sensors to collect arrester data; S2, preprocessing the collected arrester data; S3, using time series to analyze the changing trend of current data in the arrester data and evaluate the impact of environmental pollution in the arrester data on the arrester performance; S4. Use the isolation forest algorithm to identify whether there are abnormal conditions in the arrester data. If there are abnormal conditions, issue an alarm immediately and analyze the type of abnormal conditions. S5. Calculate the health status score of the arrester by weighted method based on the current data and environmental pollution data; S6. Display the abnormal situation, abnormal situation type and health status score on the user interface.

[0006] According to the above scheme, in step S1, the sensors include a leakage current sensor, a current sensor, a capacitive current sensor, a particle counter, an optical sensor and a temperature sensor; the lightning arrester data includes leakage current, resistive current, capacitive current, particulate matter concentration in the surrounding air, sulfur oxide concentration, nitrogen oxide concentration and ambient temperature.

[0007] According to the above scheme, in step S2, the preprocessing includes data cleaning, data normalization and data smoothing; improving the accuracy of subsequent analysis and models; The data cleaning removes noise, outliers and incomplete data; the data normalization standardizes the data collected by different sensors to ensure that the data are within the same range; the data smoothing uses smoothing technology to reduce random fluctuations in the data and make the trend more obvious.

[0008] According to the above scheme, in step S3, the time series analysis includes: An autoregressive integrated moving average model is constructed to analyze the changing trend of current data in the arrester data; the current data includes leakage current, resistive current and capacitive current; The autoregressive integrated moving average model formula is as follows: ; Among them, A t It is represented by the value of the time series at time point t; c is represented by the constant term; α t Represented as white noise error term; β1, β2, …, β p Expressed as the coefficients of the autoregressive term; θ1, θ2, …, θ q It is represented as the coefficient of the moving average term; p is represented as the order of the autoregressive term; q is represented as the order of the moving average term; The autoregressive integrated moving average model is estimated using historical data to estimate the autoregressive term β, the constant term c and the parameter θ of the moving average term; Using the time series decomposition method, the time series is decomposed into trend components, seasonal components and random components. The formula is as follows: A t =T t +S t +R t ; Among them, T t Expressed as trend component; S t Expressed as seasonal component; R t Represented as random component; Environmental pollution data Z t As an exogenous variable, the model is included to analyze the impact of environmental pollution data on the arrester current data. The model is expressed as: A t =f(Tt +S t +R t +Z t ); the pollution data include the concentration of particulate matter, sulfur oxide concentration, nitrogen oxide concentration and ambient temperature in the surrounding air; The constructed regression integral moving average model is used to predict future trends, and the prediction performance of the regression integral moving average model is evaluated using the prediction error index. The formula for the mean square error is as follows: ; Among them, MSE represents the mean square error value; Y t Expressed as actual value; It is represented as the predicted value; n is represented as the number of samples; The changing trend of the current data in the arrester data is obtained based on the regression integral moving average model results, and the impact of environmental pollution in the arrester data on the arrester performance is evaluated; Through time series analysis, the impact of environmental pollution on the performance of lightning arresters can be analyzed, including that high concentrations of sulfur oxides and nitrogen oxides will accelerate the aging of insulation, and high temperatures will affect capacitor performance; high environmental pollution will also affect current fluctuations and trends, reducing the accuracy of the analysis.

[0009] According to the above scheme, in step S4, the isolation forest algorithm includes the following steps: Constructing an isolation tree: randomly selecting m sample points from the historical data as the root node of the tree; randomly selecting a feature dimension, and randomly selecting a split value b between the maximum and minimum values ​​of the feature, the feature dimensions including leakage current, resistive current, capacitive current, particulate matter concentration in the surrounding air, sulfur oxide concentration, nitrogen oxide concentration and ambient temperature; using the split value b to divide the historical data into two parts, one part contains samples with feature values ​​less than the split value b, and the other part contains samples with feature values ​​greater than or equal to the split value b; recursively repeating the above steps on the divided subsets until the number of samples in the subset is less than a preset threshold or the maximum tree depth is reached, and the isolation tree is constructed; Repeat the steps of constructing isolation trees to construct X isolation trees, which form an isolation forest. For each sample x, calculate its path length h(x) in each tree to get the anomaly score of the isolation forest. The formula is as follows: ; Where s(x) is the anomaly score; E(h(x)) is the average path length of sample point x in all isolation trees; c(n) is the normalization constant; Set a threshold γ. When the anomaly score s(x) ≥ γ, the sample point x is considered to be an outlier and an alarm is issued immediately. Check each characteristic value of the sample point x that is an abnormal value, wherein the characteristic values ​​include leakage current, resistive current, capacitive current, particulate matter concentration in the surrounding air, sulfur oxide concentration, nitrogen oxide concentration, and ambient temperature; determine whether each characteristic value exceeds a normal threshold, and analyze and locate the abnormal situation; the abnormal situation includes leakage current abnormality, resistive current abnormality, capacitive current abnormality, particulate matter concentration in the surrounding air abnormality, sulfur oxide concentration abnormality, nitrogen oxide concentration abnormality, and ambient temperature abnormality; Analyze the abnormal situation type through the abnormal situation, and transmit the abnormal situation type to the user interface; the abnormal situation type includes insulation aging, internal short circuit, capacitance change and environmental pollution; Abnormal leakage current and abnormal resistive current indicate insulation aging; abnormal leakage current and abnormal capacitive current indicate internal short circuit; abnormal capacitive current indicates capacitance change; abnormal concentration of particulate matter, sulfur oxide, nitrogen oxide or abnormal ambient temperature in the surrounding air indicates environmental pollution; The isolation forest algorithm is used for anomaly detection, which can efficiently identify anomalies in the data. The isolation forest has a strong ability to detect anomalies, is suitable for big data processing, and can determine whether there are anomalies by calculating the anomaly score, thereby improving the accuracy and timeliness of the monitoring system.

[0010] According to the above solution, in step S5, the weighting method includes: For the i-th eigenvalue f i , f i The normal range is [f i,min ,f i,max ], where f i,min Expressed as eigenvalue f i The minimum value of f i,max Expressed as eigenvalue f i The characteristic values ​​include leakage current, resistive current, capacitive current, concentration of particulate matter in the surrounding air, concentration of sulfur oxides, concentration of nitrogen oxides and ambient temperature; The formula for feature scoring is as follows: ; Among them, score i Represented as the score of the i-th feature; if the feature value is within the normal range, the score is 1, otherwise the score is 0; The scores of each feature are combined using a weighted method to obtain the health score of the arrester, as follows: ; Where H represents the health status score of the arrester; d represents the number of features; w i Expressed as the weight of the i-th feature, w i ≥0 and the weights of all features add up to 1.

[0011] According to the above scheme, in step S6, the user interface displays abnormal conditions, abnormal condition types and health status scores, and divides the health status into 3 levels, namely excellent, good and poor; when the health status is excellent, a green health status score is displayed, when the health status is good, an orange health status score is displayed, and when the health status is poor, a red health status score is displayed.

[0012] A lightning arrester health status monitoring system based on big data analysis, the system includes: a data acquisition module, a central processing module, a data storage module and a user interface; The data acquisition module is used to collect arrester data and transmit the arrester data to the central processing module; the arrester data includes leakage current, resistive current, capacitive current, particulate matter concentration in the surrounding air, sulfur oxide concentration, nitrogen oxide concentration and ambient temperature; The central processing module includes data preprocessing, time series analysis, fault detection and classification, and health status scoring; the arrester data, the change trend of the current data in the arrester data, the impact of environmental pollution on the arrester performance, abnormal conditions, abnormal condition types, and health status scores in the arrester data are transmitted to the data storage module; The data preprocessing is used to preprocess the arrester data, and the preprocessing includes data cleaning, data normalization and data smoothing; The time series analysis is to perform time series analysis on the pre-processed arrester data, analyze the change trend of the current data in the arrester data and evaluate the impact of the environmental pollution data in the arrester data on the arrester performance; the current data includes leakage current, resistive current and capacitive current; the environmental pollution data includes the concentration of particulate matter, sulfur oxide concentration, nitrogen oxide concentration and ambient temperature in the surrounding air; The fault detection and classification is to perform an isolation forest algorithm on the pre-processed arrester data, and determine whether there is an abnormality in the arrester data through the isolation forest algorithm. If there is an abnormality, an alarm is immediately issued, and the abnormality type is analyzed through the abnormality and time series analysis results, and the abnormality and the abnormality type are transmitted to the user interface; the abnormality includes leakage current abnormality, resistive current abnormality, capacitive current abnormality, abnormal concentration of particulate matter in the surrounding air, abnormal sulfur oxide concentration, abnormal nitrogen oxide concentration and abnormal ambient temperature; the abnormality type includes insulation aging, internal short circuit, capacitance change and environmental pollution; The health status score is calculated based on the arrester data using a weighted method, and the health status is divided into three levels, namely, excellent, good and poor; when the health status is excellent, the health status score is displayed as a green health status score, when the health status is good, the health status score is displayed as an orange health status score, and when the health status is poor, the health status score is displayed as a red health status score; the health status score is transmitted to the user interface; The data storage module is used to receive and store the data transmitted by the central processing module and the user interface to construct a data set; The user interface is used to display abnormal situations, abnormal situation types and health status scores; and transmit the processing process and results of solving abnormal situations to the data storage module.

[0013] According to the above scheme, the historical data stored in the data storage module can be queried by selecting a query type through the user interface, and the query type includes data name, time range, abnormal situation name and abnormal situation type name; the historical data includes lightning arrester data, the changing trend of current data in the lightning arrester data, the impact of environmental pollution in the lightning arrester data on the performance of the lightning arrester, abnormal situations, abnormal situation types and health status scores.

[0014] Compared with the prior art, the present invention has the following beneficial effects: 1. The present invention can immediately issue an alarm when abnormal situations are detected through big data analysis, and analyze the abnormality type; it helps to quickly locate the problem and reduce potential risks and losses; 2. The present invention adopts weighted method and color coding to intuitively display the health status score of the arrester; 3. The present invention incorporates environmental pollution data into the model, demonstrates the impact of environmental factors on arrester data, and improves the comprehensiveness of arrester health monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is a flowchart of the steps of the method for monitoring the health status of a lightning arrester based on big data analysis of the present invention; Figure 2 It is a structural schematic diagram of a lightning arrester health status monitoring system based on big data analysis of the present invention; Figure 3 It is a structural schematic diagram of the central processing module of the arrester health status monitoring system based on big data analysis of the present invention. DETAILED DESCRIPTION

[0016] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0017] Example: Figure 1-Figure 3 As shown, the present invention provides a technical solution for a lightning arrester health status monitoring method based on big data analysis; Installing various sensors on the arrester, the sensors including leakage current sensors, current sensors, capacitive current sensors, particle counters, optical sensors and temperature sensors; Through these sensors, the data of the lightning arrester can be collected in real time, and the data of the lightning arrester includes leakage current, resistive current, capacitive current, concentration of particulate matter in the surrounding air, concentration of sulfur oxides, concentration of nitrogen oxides and ambient temperature.

[0018] Preprocessing the collected arrester data, wherein the preprocessing includes data cleaning, data normalization and data smoothing; The data cleaning removes noise, outliers and incomplete data to ensure data quality; the data normalization standardizes data collected by different sensors to ensure that the data are within the same range; the data smoothing uses smoothing technology to reduce random fluctuations in the data, making trends more obvious and improving the accuracy of subsequent analysis and models.

[0019] Using the time series analysis method, an autoregressive integral moving average model is constructed to analyze the changing trend of the current data in the arrester data; the current data includes leakage current, resistive current and capacitive current; Estimate the parameters of the autoregressive term, constant term and moving average term through historical data; Use time series decomposition method to decompose time series into trend component, seasonal component and random component; Incorporate environmental pollution data as exogenous variables into the model, including particulate matter concentration, sulfur oxide concentration, nitrogen oxide concentration and ambient temperature; analyze their impact on current data; The constructed model is used to predict future trends, and the prediction performance of the model is evaluated using the prediction error index; the changing trend of the current data is obtained based on the model results, and the impact of environmental pollution on the performance of the arrester is evaluated.

[0020] Constructing an isolation tree: randomly selecting m sample points from the historical data as the root node of the tree; randomly selecting a feature dimension, and randomly selecting a split value b between the maximum and minimum values ​​of the feature, the feature dimensions including leakage current, resistive current, capacitive current, particulate matter concentration in the surrounding air, sulfur oxide concentration, nitrogen oxide concentration and ambient temperature; using the split value b to divide the historical data into two parts, one part contains samples with feature values ​​less than the split value b, and the other part contains samples with feature values ​​greater than or equal to the split value b; recursively repeating the above steps on the divided subsets until the number of samples in the subset is less than a preset threshold or the maximum tree depth is reached, and the isolation tree is constructed; The number of randomly selected samples is 100; the feature dimension is leakage current; the segmentation value is 8mA; Repeat the steps of constructing isolation trees to construct X isolation trees, which form an isolation forest. For each sample x, calculate its path length h(x) in each tree to get the anomaly score of the isolation forest. The formula is as follows: ; Where s(x) is the anomaly score; E(h(x)) is the average path length of sample point x in all isolation trees; c(n) is the normalization constant; For sample x, E(h(x))=2.5, n=1000, then c(n)≈0.0014; s(x)=2-2.5 / 0.0014≈0.036; Set a threshold γ. When the anomaly score s(x) ≥ γ, the sample point x is considered to be an outlier and an alarm is issued immediately. Set the threshold γ=0.05, and the anomaly score s(x)=0.036≥γ=0.05; so this sample is not an outlier; If the sample is an abnormal value, check each characteristic value of the sample point x that is an abnormal value, the characteristic values ​​include leakage current, resistive current, capacitive current, particulate matter concentration in the surrounding air, sulfur oxide concentration, nitrogen oxide concentration and ambient temperature; judge whether each characteristic value exceeds the normal threshold, and analyze and locate the abnormal situation; the abnormal situation includes leakage current abnormality, resistive current abnormality, capacitive current abnormality, particulate matter concentration in the surrounding air abnormality, sulfur oxide concentration abnormality, nitrogen oxide concentration abnormality and ambient temperature abnormality; Analyze the abnormal situation type through the abnormal situation, and transmit the abnormal situation type to the user interface; the abnormal situation type includes insulation aging, internal short circuit, capacitance change and environmental pollution; Abnormal leakage current and resistive current indicate insulation aging; abnormal leakage current and capacitive current indicate internal short circuit; abnormal capacitive current indicates capacitance change; abnormal concentration of particulate matter, sulfur oxide, nitrogen oxide or ambient temperature in the surrounding air indicates environmental pollution.

[0021] For each feature value, a score is calculated based on its normal range; the scores of each feature are combined using a weighted method to obtain the health status score of the lightning arrester; the weight reflects the importance of each feature in the overall health status score, ensuring that the weights of all features add up to 1; The weight of each characteristic value is set as follows: leakage current w1=0.2, resistive current w2=0.2, capacitive current w3=0.2, particulate matter concentration w4=0.1, sulfur oxide concentration w5=0.1, nitrogen oxide concentration w6=0.1, ambient temperature w7=0.1; The normal range of each characteristic value is: leakage current f1 = [0, 10], resistive current f2 = [0, 5], capacitive current f3 = [0, 2], particulate matter concentration f4 = [0, 100], sulfur oxide concentration f5 = [0, 20], nitrogen oxide concentration f6 = [0, 15], ambient temperature f7 = [10, 40]; The characteristic values ​​collected are: leakage current f1=8, resistive current f2=3, capacitive current f3=1, particulate matter concentration f4=150, sulfur oxide concentration f5=18, nitrogen oxide concentration f6=10, ambient temperature f7=25; The characteristic scores are: leakage current score1=1, resistive current score2=1, capacitive current score3=1, particulate matter concentration score4=0, sulfur oxide concentration score5=1, nitrogen oxide concentration score6=1, ambient temperature score7=1; The health score of the arrester is obtained as H=100×(0.2×1+0.2×1+0.2×1+0.1×0+0.1×1+0.1×1+0.1×1)=90; The health status is divided into 3 levels: excellent, good and poor; When the health status score H≥90, the health status is excellent, and the health status score is displayed as green; When the health status score is 90>H≥70, the health status is good, and the health status score is displayed in orange; When the health status score is 70>H, the health status is poor, and the health status score is displayed in red; The health status score of the arrester is H=90≥90, and the health status is excellent, which is displayed as a green 90; The user interface displays the anomaly, anomaly type, and health status score.

[0022] Another embodiment of the present invention is a lightning arrester health status monitoring system based on big data analysis; the system includes a data acquisition module, a central processing module, a data storage module and a user interface; The data acquisition module collects arrester data and transmits the arrester data to the central processing module; the arrester data includes leakage current, resistive current, capacitive current, particle concentration in the surrounding air, sulfur oxide concentration, nitrogen oxide concentration and ambient temperature.

[0023] The central processing module includes data preprocessing, time series analysis, fault detection and classification, and health status scoring functions; The data preprocessing module is responsible for preprocessing the collected arrester data, including data cleaning, data normalization and data smoothing; data cleaning is to remove noise, outliers and incomplete data to ensure data quality; data normalization is to standardize the data collected by different sensors to ensure that all data are within the same range; data smoothing uses smoothing technology to reduce random fluctuations in the data and make the trend more obvious; The time series analysis module analyzes the pre-processed data to evaluate the change trend of the current data in the arrester data and analyze the impact of the environmental pollution data in the arrester data on the arrester performance; the current data includes leakage current, resistive current and capacitive current; the environmental pollution data includes particulate matter concentration, sulfur oxide concentration, nitrogen oxide concentration and ambient temperature; By constructing an autoregressive integrated moving average model, the changing trend of current data is analyzed, and environmental pollution data is incorporated into the model as an exogenous variable to evaluate its impact on the performance of the arrester.

[0024] The fault detection and classification module uses the isolation forest algorithm to detect whether there are anomalies in the arrester data; if an anomaly is detected, the system will immediately issue an alarm and analyze the type of anomaly based on the anomaly and time series analysis results; Abnormal conditions include abnormal leakage current, abnormal resistive current, abnormal capacitive current, abnormal particle concentration, abnormal sulfur oxide concentration, abnormal nitrogen oxide concentration and abnormal ambient temperature; abnormal condition types include insulation aging, internal short circuit, capacitance change and environmental pollution; Abnormal leakage current and abnormal resistive current indicate insulation aging; abnormal leakage current and abnormal capacitive current indicate internal short circuit; abnormal capacitive current indicates capacitance change; abnormal concentration of particulate matter, sulfur oxide, nitrogen oxide or abnormal ambient temperature in the surrounding air indicates environmental pollution; The exception and exception type will be transmitted to the user interface.

[0025] The health status scoring module calculates the health status score of the arrester based on the arrester data using a weighted method, and divides the health status into three levels: excellent, good, and poor; the scoring results are displayed through the user interface, and the specific color identification is as follows: when the health status is excellent, the health status score is displayed as a green health status score; when the health status is good, the health status score is displayed as an orange health status score; when the health status is poor, the health status score is displayed as a red health status score; the health status score will also be transmitted to the user interface.

[0026] The data storage module receives and stores data from the central processing module and the user interface to construct a data set; the data includes arrester data, changing trends of current data, the impact of environmental pollution on arrester performance, abnormal conditions, abnormal condition types and health status scores; The data storage module also receives and stores the process and results of solving the abnormal situation fed back by the user interface.

[0027] The user interface is used to display abnormal conditions, abnormal condition types and health status scores, and allows users to query historical data by selecting different query types. The query type includes data name, time range, abnormal condition name and abnormal condition type name. The historical data includes arrester data, trend of current data, impact of environmental pollution on arrester performance, abnormal conditions, abnormal condition types and health status scores. The user interface also feeds back the process and results of resolving abnormal conditions to the data storage module.

[0028] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. A lightning arrester health status monitoring method based on big data analysis, characterized in that: The method comprises the following steps: S1. Install sensors to collect arrester data; S2, preprocessing the collected arrester data; S3, using time series to analyze the changing trend of current data in the arrester data and evaluate the impact of environmental pollution in the arrester data on the arrester performance; S4. Use the isolation forest algorithm to identify whether there are abnormal conditions in the arrester data. If there are abnormal conditions, issue an alarm immediately and analyze the type of abnormal conditions. S5. Calculate the health status score of the arrester by weighted method based on the current data and environmental pollution data; S6. Display the abnormal situation, abnormal situation type and health status score on the user interface.

2. The method for monitoring the health status of a lightning arrester based on big data analysis according to claim 1 is characterized in that: In step S1, the sensors include leakage current sensors, current sensors, capacitive current sensors, particle counters, optical sensors and temperature sensors; the arrester data include leakage current, resistive current, capacitive current, particulate matter concentration in the surrounding air, sulfur oxide concentration, nitrogen oxide concentration and ambient temperature.

3. The method for monitoring the health status of a lightning arrester based on big data analysis according to claim 1 is characterized in that: In step S2, the preprocessing includes data cleaning, data normalization and data smoothing; The data cleaning removes noise, outliers and incomplete data; the data normalization standardizes the data collected by different sensors to ensure that the data are within the same range; and the data smoothing uses smoothing technology to reduce random fluctuations in the data.

4. The method for monitoring the health status of a lightning arrester based on big data analysis according to claim 1 is characterized in that: In step S3, the time series analysis includes: An autoregressive integrated moving average model is constructed to analyze the changing trend of current data in the arrester data; the current data includes leakage current, resistive current and capacitive current; The autoregressive integrated moving average model formula is as follows: ； Among them, A t It is represented by the value of the time series at time point t; c is represented by the constant term; α t Represented as white noise error term; β1, β2, …, β p Expressed as the coefficients of the autoregressive term; θ1, θ2, …, θ q It is represented as the coefficient of the moving average term; p is represented as the order of the autoregressive term; q is represented as the order of the moving average term; The autoregressive integrated moving average model is estimated using historical data to estimate the autoregressive term β, the constant term c, and the parameter θ of the moving average term; Using the time series decomposition method, the time series is decomposed into trend components, seasonal components and random components. The formula is as follows: A t =T t +S t +R t ; Among them, T t Expressed as trend component; S t Expressed as seasonal component; R t Represented as random component; Environmental pollution data Z t As an exogenous variable, the model is included to analyze the impact of environmental pollution data on the arrester current data. The model is expressed as: A t =f(T t +S t +R t +Z t ); the pollution data include the concentration of particulate matter, sulfur oxide concentration, nitrogen oxide concentration and ambient temperature in the surrounding air; The constructed regression integral moving average model is used to predict future trends, and the prediction performance of the regression integral moving average model is evaluated using the prediction error index. The formula for the mean square error is as follows: ； Among them, MSE represents the mean square error value; Y t Expressed as actual value; It is represented as the predicted value; n is represented as the number of samples; According to the results of the regression integral moving average model, the changing trend of the current data in the arrester data is obtained, and the impact of environmental pollution in the arrester data on the arrester performance is evaluated.

5. The method for monitoring the health status of a lightning arrester based on big data analysis according to claim 1 is characterized in that: In step S4, the isolation forest algorithm includes the following steps: Constructing an isolation tree: randomly selecting m sample points from historical data as the root node of the tree; randomly selecting a feature dimension and randomly selecting a split value b between the maximum and minimum values ​​of the feature, the feature dimensions including leakage current, resistive current, capacitive current, particulate matter concentration in the surrounding air, sulfur oxide concentration, nitrogen oxide concentration and ambient temperature; Use the split value b to divide the historical data into two parts, one of which contains samples with characteristic values ​​less than the split value b, and the other contains samples with characteristic values ​​greater than or equal to the split value b; Recursively repeat the above steps on the segmented subsets until the number of samples in the subset is less than the preset threshold or the maximum tree depth is reached, and the isolation tree is constructed; Repeat the steps of constructing isolation trees to construct X isolation trees, which form an isolation forest. For each sample x, calculate its path length h(x) in each tree to get the anomaly score of the isolation forest. The formula is as follows: ； Where s(x) is the anomaly score; E(h(x)) is the average path length of sample point x in all isolation trees; c(n) is the normalization constant; Set a threshold γ. When the anomaly score s(x) ≥ γ, the sample point x is considered to be an outlier and an alarm is issued immediately. Check each characteristic value of the sample point x that is an abnormal value, wherein the characteristic values ​​include leakage current, resistive current, capacitive current, particulate matter concentration in the surrounding air, sulfur oxide concentration, nitrogen oxide concentration, and ambient temperature; determine whether each characteristic value exceeds a normal threshold, and analyze and locate the abnormal situation; the abnormal situation includes leakage current abnormality, resistive current abnormality, capacitive current abnormality, particulate matter concentration in the surrounding air abnormality, sulfur oxide concentration abnormality, nitrogen oxide concentration abnormality, and ambient temperature abnormality; Analyze the abnormal situation type through the abnormal situation, and transmit the abnormal situation type to the user interface; the abnormal situation type includes insulation aging, internal short circuit, capacitance change and environmental pollution; Abnormal leakage current and resistive current indicate insulation aging; abnormal leakage current and capacitive current indicate internal short circuit; abnormal capacitive current indicates capacitance change; abnormal concentration of particulate matter, sulfur oxide, nitrogen oxide or ambient temperature in the surrounding air indicates environmental pollution.

6. The method for monitoring health status of lightning arresters based on big data analysis according to claim 1 is characterized in that: In step S5, the weighting method includes: For the i-th eigenvalue f i , f i The normal range is [f i,min ,f i,max ], where f i,min Expressed as eigenvalue f i The minimum value of f i,max Expressed as eigenvalue f i The characteristic values ​​include leakage current, resistive current, capacitive current, concentration of particulate matter in the surrounding air, concentration of sulfur oxides, concentration of nitrogen oxides and ambient temperature; The formula for feature scoring is as follows: ； Among them, score i Represented as the score of the i-th feature; if the feature value is within the normal range, the score is 1, otherwise the score is 0; The scores of each feature are combined using a weighted method to obtain the health score of the arrester, as follows: ； Where H represents the health status score of the arrester; d represents the number of features; w i Expressed as the weight of the i-th feature, w i ≥0 and the weights of all features add up to 1.

7. The method for monitoring health status of lightning arresters based on big data analysis according to claim 1 is characterized in that: In step S6, the user interface displays abnormal conditions, abnormal condition types and health status scores, and divides the health status into three levels, namely excellent, good and poor; when the health status is excellent, the health status score is displayed as green, when the health status is good, the health status score is displayed as orange, and when the health status is poor, the health status score is displayed as red.

8. The arrester health status monitoring system based on big data analysis is characterized by: The system includes: a data acquisition module, a central processing module, a data storage module and a user interface; The data acquisition module is used to collect arrester data and transmit the arrester data to the central processing module; the arrester data includes leakage current, resistive current, capacitive current, particulate matter concentration in the surrounding air, sulfur oxide concentration, nitrogen oxide concentration and ambient temperature; The central processing module includes data preprocessing, time series analysis, fault detection and classification, and health status scoring; the arrester data, the change trend of the current data in the arrester data, the impact of environmental pollution on the arrester performance, abnormal conditions, abnormal condition types, and health status scores in the arrester data are transmitted to the data storage module; The data preprocessing is used to preprocess the arrester data, and the preprocessing includes data cleaning, data normalization and data smoothing; The time series analysis is to perform time series analysis on the pre-processed arrester data, analyze the change trend of the current data in the arrester data and evaluate the impact of the environmental pollution data in the arrester data on the arrester performance; the current data includes leakage current, resistive current and capacitive current; the environmental pollution data includes the concentration of particulate matter, sulfur oxide concentration, nitrogen oxide concentration and ambient temperature in the surrounding air; The fault detection and classification is to perform an isolation forest algorithm on the pre-processed arrester data, and determine whether there is an abnormality in the arrester data through the isolation forest algorithm. If there is an abnormality, an alarm is immediately issued, and the abnormality type is analyzed through the abnormality and time series analysis results, and the abnormality and the abnormality type are transmitted to the user interface; the abnormality includes leakage current abnormality, resistive current abnormality, capacitive current abnormality, abnormal concentration of particulate matter in the surrounding air, abnormal sulfur oxide concentration, abnormal nitrogen oxide concentration and abnormal ambient temperature; the abnormality type includes insulation aging, internal short circuit, capacitance change and environmental pollution; The health status score is calculated based on the arrester data using a weighted method, and the health status is divided into three levels, namely, excellent, good and poor; when the health status is excellent, the health status score is displayed as a green health status score, when the health status is good, the health status score is displayed as an orange health status score, and when the health status is poor, the health status score is displayed as a red health status score; the health status score is transmitted to the user interface; The data storage module is used to receive and store the data transmitted by the central processing module and the user interface to construct a data set; The user interface is used to display abnormal situations, abnormal situation types and health status scores; and transmit the processing process and results of solving abnormal situations to the data storage module.

9. The arrester health status monitoring system based on big data analysis according to claim 8 is characterized in that: The user interface can be used to select a query type to query historical data stored in the data storage module, wherein the query type includes data name, time range, abnormal situation name, and abnormal situation type name; the historical data includes lightning arrester data, changing trends of current data in lightning arrester data, impact of environmental pollution in lightning arrester data on lightning arrester performance, abnormal situations, abnormal situation types, and health status scores.

Images