Electric energy metering box fault early warning method and system based on data driving

Abstract

The invention relates to the technical field of electric energy metering box management, in particular to an electric energy metering box fault early warning method and system based on data driving, and the method comprises the steps: obtaining the comprehensive monitoring data of an electric energy box; determining a plurality of electric energy box evaluation indexes for evaluating the electric energy box according to the comprehensive monitoring data, and constructing an electric energy box state evaluation system; filtering the monitoring data based on a preset filtering method and an evaluation system to obtain filtered comprehensive monitoring data; calculating model digital characteristics and model certainty according to the filtered comprehensive monitoring data, and calculating an index weight of a normal cloud model to obtain a model index weight; constructing an electric energy metering box evaluation model; and acquiring data of the to-be-predicted electric energy metering box, performing electric arc identification on the to-be-predicted electric energy metering box by utilizing edge calculation, and analyzing actual measurement data of the electric energy box when the electric arc identification passes, thereby obtaining a fault early warning result of the to-be-predicted electric energy metering box. The fault occurrence rate and the maintenance cost of the electric energy metering box can be effectively reduced, and the accuracy of fault early warning of the electric energy metering box is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power metering box management, and in particular, to a data-driven power metering box fault warning method and system. Background Art

[0002] With the continuous advancement of the construction of modern power distribution and utilization networks, power metering boxes, as key components with functions such as power information collection, visual interaction, and electricity consumption behavior collection, play an important role in the stable operation of the power distribution network. Relevant data shows that as of 2017, the cumulative number of households with power metering boxes collected reached 447 million, and the coverage rate reached 99.03%. Although power metering boxes have been extremely popular, the complex working environment has high requirements for their reliable operation. For example, terminal aging, line loss, accumulation combustion, and the temperature and humidity of the air may all have a greater impact on the reliability of the operation of power metering boxes. Abnormal operation of power metering boxes is likely to trigger a chain reaction of power grid failures, and even endanger the safety and stability of the operation of the power distribution and utilization network. Therefore, it is urgent to carry out research on the status assessment and fault warning of power metering boxes.

[0003] Traditional methods for power metering box status assessment mostly use periodic manual inspections, which are costly, inefficient, and have poor timeliness in fault discovery and warning. There are differences in the work levels of inspectors, and the quantification criteria for various assessment indicators are still not perfect. Summary of the Invention

[0004] The present invention provides a data-driven power metering box fault warning method and a computer-readable storage medium, and its main purpose is to reduce the failure rate and maintenance cost of power metering boxes and improve the accuracy of power metering box fault warning.

[0005] To achieve the above object, a data-driven power metering box fault warning method provided by the present invention includes:

[0006] Obtaining historical monitoring data and on-site measured data of the power metering box based on a preset data collection instruction, obtaining power metering historical monitoring data and power metering on-site measured data, and obtaining comprehensive monitoring data of the power metering box according to the power metering historical monitoring data and power metering on-site measured data;

[0007] Determining multiple evaluation indicators for evaluating the power metering box according to the comprehensive monitoring data of the power metering box, obtaining multiple power metering box evaluation indicators, and constructing a status evaluation system for the power metering box according to the multiple power metering box evaluation indicators, obtaining a power metering box status evaluation system;

[0008] Filtering the comprehensive monitoring data of the power metering box based on a preset multi-source data density filtering method and the power metering box status evaluation system to obtain filtered comprehensive monitoring data;

[0009] Calculate the digital features and certainty degrees of a preset normal cloud model based on the filtered comprehensive monitoring data to obtain the model digital features and model certainty degrees, and calculate the index weights of the normal cloud model according to the power box status evaluation system to obtain the model index weights;

[0010] Construct a normal cloud model of the electric energy metering box based on the filtered comprehensive monitoring data and the power box status evaluation system, using the model digital features, model certainty degrees and model index weights to obtain the electric energy metering box evaluation model;

[0011] Based on a preset intelligent air switch, obtain the real-time monitoring data of the electric energy metering box to be predicted to get the actual measured data of the power box. According to the actual measured data of the power box, use a pre-constructed edge computing device to identify the electric arc of the electric energy metering box to be predicted to obtain the identification result. After confirming that the identification result is the preset passing result, use the electric energy metering box evaluation model to analyze the actual measured data of the power box to obtain the fault warning result of the electric energy metering box to be predicted.

[0012] Optionally, filtering the comprehensive monitoring data of the power box based on the preset multi-source data density filtering method and the power box status evaluation system to obtain the filtered comprehensive monitoring data includes:

[0013] Obtain the target indicators in the power status evaluation system, and obtain multiple samples in the comprehensive monitoring data of the power box based on the target indicators in the power box status evaluation system to obtain multiple target samples;

[0014] Calculate the Euclidean distance between every two samples in the multiple target samples to obtain multiple sample Euclidean distances;

[0015] Arrange the multiple sample Euclidean distances in ascending order to obtain a sample ascending queue, and calculate the maximum neighbor distance and local density based on the sample ascending queue to obtain the sample maximum neighbor distance and local density;

[0016] Calculate the local density threshold and maximum neighbor distance threshold of the comprehensive monitoring data of the power box according to the sample maximum neighbor distance and local density, and use the local density threshold and maximum neighbor distance threshold to filter the comprehensive monitoring data of the power box to obtain the filtered comprehensive monitoring data.

[0017] Optionally, calculating the digital features and certainty degrees of a preset normal cloud model based on the filtered comprehensive monitoring data to obtain the model digital features and model certainty degrees includes:

[0018] Construct a data sample based on the filtered comprehensive monitoring data to obtain a monitoring data sample, and calculate the expected value of the monitoring data sample to obtain the sample data expected value;

[0019] Calculate the uncertainty entropy value of the monitored data sample to obtain the monitored sample entropy value, and perform a secondary solution on the monitored sample entropy value to obtain the monitored sample hyperentropy;

[0020] Calculate the certainty degree of the monitored data sample according to the sample data expected value and the monitored sample hyperentropy to obtain the sample certainty degree, and use the sample data expected value, the monitored sample entropy value, and the monitored sample hyperentropy as the model digital characteristics of the normal cloud model, and use the sample certainty degree as the model certainty degree of the normal cloud model.

[0021] Optionally, calculating the index weight of the normal cloud model according to the power box status evaluation system to obtain the model index weight, including:

[0022] Obtain the evaluation indexes in the power box status evaluation system, construct a judgment matrix according to the evaluation indexes in the power box status evaluation system to obtain the evaluation judgment matrix, and calculate the maximum eigenvalue of the evaluation judgment matrix to obtain the evaluation maximum eigenvalue;

[0023] Obtain the weight vector of the normal cloud model according to the preset characteristic equation and the evaluation maximum eigenvalue to obtain the index characteristic weight vector;

[0024] Obtain the problems corresponding to the evaluation indexes and the sample data corresponding to the problems, construct an evaluation matrix according to the sample data in the problems corresponding to the evaluation indexes to obtain the initial evaluation matrix, and perform normalization processing on the initial evaluation matrix to obtain the standardized matrix;

[0025] Calculate the entropy of the evaluation indexes based on the preset entropy weight method to obtain the evaluation index entropy, calculate the entropy weight of the evaluation indexes according to the evaluation index entropy to obtain the evaluation index entropy weight, and obtain the weight vector of the evaluation indexes according to the evaluation index entropy weight to obtain the operation characteristic weight vector;

[0026] Calculate the index weight of the normal cloud model according to the index characteristic weight vector and the operation characteristic weight vector to obtain the model index weight.

[0027] Optionally, according to the filtered comprehensive monitoring data and the power box status evaluation system, using the model digital characteristics, the model certainty degree, and the model index weight to construct a normal cloud model of the power metering box to obtain the power metering box evaluation model, including:

[0028] Calculate the evaluation weight of the power box status evaluation system by using the model digital characteristics, the model certainty degree, and the model index weight to obtain the index evaluation weight;

[0029] Evaluate the cloud centers of the normal cloud model and the preset standard normal cloud model based on the power box status evaluation system to obtain the cloud gravity deviation degree;

[0030] Calculate the deviation degree of the normal cloud model according to the evaluation weight of the said index and the cloud gravity center deviation degree, and obtain the cloud gravity center deviation degree;

[0031] Obtain the status evaluation result of the electric energy metering box based on the electric energy box status evaluation system, and calculate the membership level of the status evaluation result of the electric energy metering box according to the cloud gravity center deviation degree and the preset maximum membership degree principle to obtain the evaluation membership degree;

[0032] Determine the evaluation level corresponding to the electric energy box status evaluation system according to the evaluation membership degree to obtain the membership evaluation level, and construct a normal cloud model of the electric energy metering box according to the membership evaluation level and the filtered comprehensive monitoring data to obtain an electric energy metering box evaluation model.

[0033] Optionally, the constructing a normal cloud model of the electric energy metering box according to the membership evaluation level and the filtered comprehensive monitoring data to obtain an electric energy metering box evaluation model includes:

[0034] Predict the fault degree of the electric energy metering box according to the membership evaluation level to obtain the predicted fault degree;

[0035] Obtain the basic data corresponding to the predicted fault degree from the filtered comprehensive monitoring data to obtain the basic monitoring data;

[0036] Divide the basic monitoring data and the predicted fault degree into a test set and a training set to obtain an evaluation training set and an evaluation test set, and use the evaluation training set to train the normal cloud model, and use the evaluation test set to test the trained normal cloud model to obtain the electric energy metering box evaluation model.

[0037] Optionally, after calculating the digital characteristics and the certainty degree of the preset normal cloud model according to the filtered comprehensive monitoring data to obtain the model digital characteristics and the model certainty degree, the method further includes:

[0038] Obtain the fault label in the filtered comprehensive monitoring data, and calculate the digital characteristics in the model digital characteristics whose correlation with the fault label is greater than the preset correlation threshold to obtain the screened digital characteristics;

[0039] Use the principal component analysis method to reduce the dimension of the screened digital characteristics to obtain the dimension-reduced screened digital characteristics.

[0040] Optionally, the calculation formula of the expected value of the sample data is as follows:

[0041]

[0042] Wherein, is the expected value of the sample data under evaluation index i, x ij is the j-th sample data under the i-th evaluation index, is the sample data set composed of the data samples of evaluation index i, is the set The number of samples included.

[0043] Optionally, the calculation formula of the sample certainty is as follows:

[0044]

[0045] where μ(x ij ) is the sample certainty under the sample data x ij , indicating the membership degree of the sample data x ij to the i-th evaluation index, and the value range is [0, 1], x ij is the j-th sample data under the i-th evaluation index, is the sample data expectation, is the monitoring sample hyperentropy.

[0046] To achieve the above object, the present invention also provides a data-driven power metering box fault warning system, including:

[0047] A comprehensive monitoring data collection module, configured to obtain the historical monitoring data and on-site measured data of the power metering box based on a preset data collection instruction, obtain the electricity meter historical monitoring data and the electricity meter on-site measured data, and obtain the comprehensive monitoring data of the power box according to the electricity meter historical monitoring data and the electricity meter on-site measured data;

[0048] A state evaluation system construction module, configured to determine multiple evaluation indexes for evaluating the power metering box according to the comprehensive monitoring data of the power box, obtain multiple power box evaluation indexes, and construct a state evaluation system of the power metering box according to the multiple power box evaluation indexes, to obtain a power box state evaluation system;

[0049] An evaluation model construction module, configured to filter the comprehensive monitoring data of the power box based on a preset multi-source data density filtering method and the power box state evaluation system, to obtain filtered comprehensive monitoring data;

[0050] Calculate the digital characteristics and certainty of a preset normal cloud model according to the filtered comprehensive monitoring data, obtain model digital characteristics and model certainty, and calculate the index weights of the normal cloud model according to the power box state evaluation system, to obtain model index weights;

[0051] According to the filtered comprehensive monitoring data and the power metering box status evaluation system, a normal cloud model of the power metering box is constructed by using the model digital features, model certainty, and model index weights to obtain a power metering box evaluation model;

[0052] A fault warning analysis module, configured to obtain real-time monitoring data of a power metering box to be predicted based on a preset intelligent air switch, obtain measured data of the power metering box, and based on the measured data of the power metering box, use a pre-constructed edge computing device to perform arc identification on the power metering box to be predicted to obtain an identification result. After confirming that the identification result is a preset passing result, the power metering box evaluation model is used to analyze the measured data of the power metering box to obtain a fault warning result of the power metering box to be predicted.

[0053] To solve the above problems, the present invention further provides an electronic device, which includes:

[0054] A memory that stores at least one instruction; and a processor that executes the instruction stored in the memory to implement the above-mentioned data-driven power metering box fault warning method.

[0055] To solve the above problems, the present invention further provides a computer-readable storage medium, in which at least one instruction is stored, and the at least one instruction is executed by a processor in an electronic device to implement the above-mentioned data-driven power metering box fault warning method.

[0056] To solve the problems described in the background art, the present invention obtains the historical monitoring data and on-site measured data of the electric energy metering box based on a preset data collection instruction, obtaining the historical monitoring data of the electric meter and the on-site measured data of the electric meter. It can be seen that the present invention not only considers the measured data but also the historical data, improving the richness of the data and laying a data foundation for the subsequent construction of the model. And based on the historical monitoring data of the electric meter and the on-site measured data of the electric meter, it obtains the comprehensive monitoring data of the electric energy box; determines multiple evaluation indicators for evaluating the electric energy metering box according to the comprehensive monitoring data of the electric energy box, obtaining multiple electric box evaluation indicators, and constructs a state evaluation system for the electric energy metering box according to the multiple electric box evaluation indicators, obtaining the state evaluation system of the electric energy box. It can be seen that for the fault prediction of the present invention, it is necessary to first evaluate the indicators, providing a basis for subsequent quantitative analysis and model construction; filters the comprehensive monitoring data of the electric energy box based on a preset multi-source data density filtering method and the state evaluation system of the electric energy box, obtaining the filtered comprehensive monitoring data; calculates the digital characteristics and certainty of a preset normal cloud model according to the filtered comprehensive monitoring data, obtaining the model digital characteristics and model certainty, and calculates the index weights of the normal cloud model according to the state evaluation system of the electric energy box, obtaining the model index weights; constructs a normal cloud model of the electric energy metering box according to the filtered comprehensive monitoring data and the state evaluation system of the electric energy box, using the model digital characteristics, model certainty and model index weights, obtaining the evaluation model of the electric energy metering box. It can be seen that the present invention can provide a more accurate quantitative analysis for the state evaluation of the electric energy metering box by calculating the digital characteristics, certainty and index weights; obtains the real-time monitoring data of the electric energy metering box to be predicted based on a preset intelligent air switch, obtaining the measured data of the electric energy box, and according to the measured data of the electric energy box, uses a pre-constructed edge computing device to identify the electric arc of the electric energy metering box to be predicted, obtaining the identification result. After confirming that the identification result is a preset passing result, analyzes the measured data of the electric energy box using the evaluation model of the electric energy metering box, obtaining the fault warning result of the electric energy metering box to be predicted. It can be seen that through the intelligent air switch and edge computing technology, the present invention can quickly and accurately identify and warn of the faults of the electric energy metering box. Therefore, the present invention can reduce the failure rate and maintenance cost of the electric energy metering box and improve the accuracy of the fault warning of the electric energy metering box. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 FIG. is a schematic flow chart of a data-driven electric energy metering box fault warning method provided by an embodiment of the present invention;

[0058] Figure 2 FIG. is a functional module diagram of a data-driven electric energy metering box fault warning system provided by an embodiment of the present invention;

[0059] Figure 3Schematic structural diagram of an electronic device for implementing the data-driven power metering box fault warning method provided by an embodiment of the present invention.

[0060] Explanation of reference numerals:

[0061] 1. Electronic device; 10. Processor; 11. Memory; 12. Bus.

[0062] The implementation, functional characteristics and advantages of the object of the present invention will be further described in conjunction with the embodiments with reference to the drawings. Detailed implementation manners

[0063] It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0064] An embodiment of the present application provides a data-driven power metering box fault warning method. The execution subject of the data-driven power metering box fault warning method includes, but is not limited to, at least one of electronic devices such as a server, a terminal, etc. that can be configured to execute the method provided by the embodiment of the present application. In other words, the data-driven power metering box fault warning method can be executed by software or hardware installed on a terminal device or a server device, and the software can be a blockchain platform. The server includes, but is not limited to: a single server, a server cluster, a cloud server, or a cloud server cluster, etc.

[0065] Referring to Figure 1 As shown, it is a flowchart of a data-driven power metering box fault warning method provided by an embodiment of the present invention. In this embodiment, the data-driven power metering box fault warning method includes:

[0066] S1. Obtain the historical monitoring data and on-site measured data of the power metering box based on a preset data collection instruction to obtain the electricity meter historical monitoring data and the electricity meter on-site measured data, and obtain the comprehensive monitoring data of the power box according to the electricity meter historical monitoring data and the electricity meter on-site measured data.

[0067] It can be understood that the data collection instruction is an instruction issued by a person who predicts the fault of the power metering box, and the person confirms the necessary environment for predicting the fault of the power metering box, where the necessary environment is the fault prediction environment.

[0068] It should be understood that historical monitoring data refers to the data collected and recorded in the past time period, usually from monitoring devices such as smart meters and metering chips inside the power metering box, as well as the monitoring system in the power system background, etc. These devices or systems will regularly or real-time record the operating parameters of the power metering box, such as current, voltage, power, power consumption, etc., and store the data in the database. On-site measured data refers to the data obtained by professional personnel or automated monitoring devices on-site measuring the power metering box and its related power metering devices at a specific time point or time period. It can be the data obtained by power operation and maintenance personnel using portable measuring instruments such as clamp meters and multimeters for on-site measurement, or the data collected by on-site monitoring devices installed in the power metering box. The comprehensive monitoring data of the power box is a dataset that integrates the historical monitoring data of the electricity meter and the on-site measured data of the electricity meter, and can comprehensively and systematically reflect the operating state and related characteristics of the power metering box.

[0069] Furthermore, by combining and analyzing the historical monitoring data and the on-site measured data, the operating state of the power metering box can be understood more comprehensively and accurately. It can not only grasp its long-term operating trend but also promptly discover potential problems that may occur currently.

[0070] It should be understood that after obtaining the comprehensive monitoring data of the power box, it also includes cleaning and screening the comprehensive monitoring data of the power box to eliminate abnormal data and noise for subsequent processing.

[0071] S2. Determine multiple evaluation indicators for evaluating the power metering box according to the comprehensive monitoring data of the power box to obtain multiple power box evaluation indicators, and construct a state evaluation system for the power metering box according to the multiple power box evaluation indicators to obtain a power box state evaluation system.

[0072] It should be explained that evaluation indicators refer to various parameters used to measure the operating state of the power metering box. These indicators can be divided into two categories: power box operating indicators and power box environmental indicators.

[0073] For example, power box operating indicators include: incoming / outgoing line current, incoming / outgoing line voltage, power box power, power box power, mechanical vibration frequency, line loss, etc. Among them, the incoming / outgoing line current reflects the current load situation of the power metering box, and the unit is usually ampere. The line loss reflects the power transmission efficiency of the power metering box, usually expressed as a percentage; power box environmental indicators include: power box temperature, terminal temperature, power box humidity, smoke concentration, magnetic field strength, dust concentration, etc.

[0074] It can be understood that constructing a state evaluation system for the power metering box according to the multiple power box evaluation indicators to obtain a power box state evaluation system clarifies the key elements of the evaluation and provides a basis for subsequent quantitative analysis and model construction.

[0075] S3. Filter the comprehensive monitoring data of the power box based on the preset multi-source data density filtering method and the power box status evaluation system to obtain filtered comprehensive monitoring data.

[0076] It is understandable that the multi-source data density filtering method is a multi-source data density filtering method based on data density identification and screening. It can calculate the Euclidean distance between samples, obtain the maximum neighbor distance and local density, and monitor and eliminate abnormal data in the power metering box collection data through threshold comparison to ensure the effectiveness of the monitoring data.

[0077] It should be explained that filtering the comprehensive monitoring data of the power box based on the preset multi-source data density filtering method and the power box status evaluation system to obtain filtered comprehensive monitoring data includes:

[0078] Obtain the target indicators in the power status evaluation system, and obtain multiple samples in the comprehensive monitoring data of the power box based on the target indicators in the power box status evaluation system to obtain multiple target samples;

[0079] Calculate the Euclidean distance between every two samples among the multiple target samples to obtain multiple sample Euclidean distances;

[0080] Arrange the multiple sample Euclidean distances in ascending order to obtain a sample ascending queue, and calculate the maximum neighbor distance and local density based on the sample ascending queue to obtain the sample maximum neighbor distance and local density;

[0081] Calculate the local density threshold and maximum neighbor distance threshold of the comprehensive monitoring data of the power box according to the sample maximum neighbor distance and local density, and use the local density threshold and maximum neighbor distance threshold to filter the comprehensive monitoring data of the power box to obtain filtered comprehensive monitoring data.

[0082] Furthermore, calculate the Euclidean distance between every two samples among the multiple target samples. For example, the target indicators in the target samples are voltage, current, and power factor, and a total of four target sample data x1=(220, 5, 0.9), x2=(222, 6, 0.92), x2=(218, 4, 0.88), and x3=(225, 7, 0.95) are obtained. Then, the Euclidean distance between each target sample data and the other three target sample data needs to be calculated respectively. Among them, the number of calculated indicators is 3, indicating that the dimension of the calculated vector is 3. The Euclidean distance calculation formula between x1 and x2 is The Euclidean distance calculation formulas between x1 and x3 and between x1 and x4 are similar to the Euclidean distance calculation formula between x1 and x2, so they will not be elaborated here.

[0083] It should be understood that the maximum neighbor distance of the sample refers to the distance from a sample point to its nearest sample point with a higher (or equal) density. The local density represents the number or distribution density of neighboring sample points within a given distance range of a sample point. When calculating, the local density threshold is used as the boundary to count the number of sample points within this range. The local density threshold and the maximum neighbor distance threshold are critical values for filtering data calculated based on the local density and maximum neighbor distance of the sample. By setting reasonable thresholds, noise data and outliers can be removed, and valid data can be retained.

[0084] It can be understood that by calculating the Euclidean distance between samples, the similarity or difference between samples can be quantified. The smaller the distance, the closer the samples are in the feature space and the higher the similarity; conversely, the greater the difference. This helps with subsequent data clustering analysis, anomaly detection, etc.

[0085] Furthermore, sorting the sample Euclidean distances in ascending order and calculating the maximum neighbor distance and local density is to better understand the distribution characteristics and density of the data. Based on the calculated local density threshold and maximum neighbor distance threshold, normal data and abnormal data can be effectively distinguished, achieving precise filtering of the data.

[0086] S4. Calculate the digital features and degree of certainty of the preset normal cloud model based on the filtered comprehensive monitoring data to obtain the model digital features and model degree of certainty, and calculate the index weights of the normal cloud model according to the power box status evaluation system to obtain the model index weights.

[0087] It can be understood that by constructing a normal cloud model, the conversion between qualitative concepts and quantitative values can be carried out. Through digital features (expectation, entropy, hyperentropy) and degree of certainty, randomness and fuzziness are comprehensively considered to process uncertain information. This enables the model to accurately evaluate the operating status of complex systems such as power metering boxes, and through the determination of index weights, comprehensive evaluation of multiple indicators is achieved. Finally, based on the evaluation results of the normal cloud model, potential faults can be detected and warned in a timely manner, ensuring the stable operation of the system and improving the accuracy and reliability of the evaluation.

[0088] Furthermore, the normal cloud model refers to a mathematical model for processing uncertain information. The normal cloud can convert between qualitative concepts and quantitative values, and is particularly suitable for describing concepts with randomness and fuzziness. In the application, the normal cloud model is used for the status evaluation of power metering boxes, that is, to judge the relationship between the fault situation of the power metering box and the existing data.

[0089] It should be explained that the calculation of the digital features and degree of certainty of the preset normal cloud model based on the filtered comprehensive monitoring data to obtain the model digital features and model degree of certainty includes:

[0090] Construct a data sample based on the filtered comprehensive monitoring data to obtain a monitoring data sample, and calculate the expected value of the monitoring data sample to obtain a sample data expected value;

[0091] Calculate the uncertainty entropy value of the monitoring data sample to obtain a monitoring sample entropy value, and perform a secondary solution on the monitoring sample entropy value to obtain a monitoring sample hyperentropy;

[0092] Calculate the certainty degree of the monitoring data sample based on the sample data expected value and the monitoring sample hyperentropy to obtain a sample certainty degree, and use the sample data expected value, the monitoring sample entropy value, and the monitoring sample hyperentropy as the model digital characteristics of the normal cloud model, and use the sample certainty degree as the model certainty degree of the normal cloud model.

[0093] It should be understood that the model certainty degree refers to the membership degree of each cloud droplet (sample) to the qualitative concept, that is, the probability that the cloud droplet belongs to the qualitative concept. The closer the value is to 1, the higher the membership degree, and the more certain that the cloud droplet belongs to the qualitative concept.

[0094] For example, use the normal cloud model to perform fault prediction analysis on an evaluation index (such as temperature) of an electric energy metering box. First, construct a data sample of this index based on the filtered comprehensive monitoring data, and calculate that the sample data expected value is 25°C, indicating the expected temperature value of this index under normal operating conditions. Then, calculate that the monitoring sample entropy value is 2, reflecting the uncertainty degree of the temperature data. Then, perform a secondary solution on the monitoring sample entropy value to obtain a monitoring sample hyperentropy of 0.5, further quantifying the uncertainty of the entropy value. From this, the model certainty degree can be calculated as 0.82, indicating that the probability that this temperature data point belongs to the normal operating state (i.e., the fault-free state) is 82%. In the fault prediction of the electric energy metering box, set a certainty degree threshold, such as 0.7. When the certainty degree of a certain data point is lower than this threshold, it is considered that there is a fault risk for this data point and further inspection and processing are required. For example, if the calculated certainty degree of another temperature data point is 0.6, which is lower than the threshold of 0.7, then it can be warned that the electric energy metering box may have a fault, such as problems such as abnormal temperature rise caused by poor heat dissipation, so as to take measures in advance for maintenance and repair to avoid further expansion of the fault.

[0095] Furthermore, the calculation formula of the sample data expected value is as follows:

[0096]

[0097] Among them, is the sample data expected value under the evaluation index i, x ij is the jth sample data under the ith evaluation index, is the sample data set composed of the data samples of the evaluation index i, For the set The number of samples included

[0098] Furthermore, the calculation formula for monitoring the sample entropy value is as follows:

[0099]

[0100] Wherein, is the monitoring sample entropy value, is the monitoring sample hyperentropy, a is a constant, is the maximum value in the sample data set composed of the data samples of evaluation index i, is the minimum value in the sample data set composed of the data samples of evaluation index i

[0101] It should be understood that according to the entropy calculation principle in information theory, first determine the monitoring sample entropy value, and then calculate its entropy value based on the distribution of the monitoring sample entropy value, and use this as the hyperentropy. And in this application, this hyperentropy is a constant

[0102] It can be understood that the calculation formula for the sample certainty degree is as follows:

[0103]

[0104] Wherein, μ(x ij ) is the sample certainty degree under the sample data x ij , representing the membership degree of the sample data x ij to the i-th evaluation index, and the value range is [0, 1], x ij is the j-th sample data under the i-th evaluation index, is the sample data expectation, is the monitoring sample hyperentropy

[0105] Furthermore, calculating the index weights of the normal cloud model according to the power box status evaluation system to obtain the model index weights includes:

[0106] Obtain the evaluation indexes in the power box status evaluation system, construct a judgment matrix according to the evaluation indexes in the power box status evaluation system to obtain an evaluation judgment matrix, and calculate the maximum eigenvalue of the evaluation judgment matrix to obtain the evaluation maximum eigenvalue;

[0107] Obtain the weight vector of the normal cloud model according to the preset characteristic equation and the evaluation maximum eigenvalue to obtain the index characteristic weight vector;

[0108] Obtain the problems corresponding to the evaluation indicators and the sample data corresponding to the problems. Construct an evaluation matrix based on the sample data in the problems corresponding to the evaluation indicators to obtain an initial evaluation matrix, and perform normalization processing on the initial evaluation matrix to obtain a standardized matrix;

[0109] Calculate the entropy of the evaluation indicators based on the preset entropy weight method to obtain the evaluation indicator entropy. Calculate the entropy weight of the evaluation indicators based on the evaluation indicator entropy to obtain the evaluation indicator entropy weight, and obtain the weight vector of the evaluation indicators based on the evaluation indicator entropy weight to obtain the operation characteristic weight vector;

[0110] Calculate the index weight of the normal cloud model based on the index characteristic weight vector and the operation characteristic weight vector to obtain the model index weight.

[0111] Furthermore, the decision matrix represents the data obtained by pairwise comparison of each data under two evaluation indicators in the power box status evaluation system. The problems corresponding to the evaluation indicators refer to what kind of project problems this indicator belongs to. Multiple evaluation indicators can correspond to one project problem. For example, evaluation indicators such as power box temperature, terminal temperature, power box humidity, smoke concentration, magnetic field strength, and dust concentration can all belong to safety problems.

[0112] It should be explained that the normalization processing of the initial evaluation matrix can be achieved by performing a linear transformation on each indicator in the initial evaluation matrix and mapping it to a specific interval, such as [0, 1].

[0113] Furthermore, the calculation formula for the index weight of the normal cloud model is as follows:

[0114] ω = aω e + bω0

[0115] where ω e is the index characteristic weight vector, ω0 is the operation characteristic weight vector, and a and b represent the weight coefficients of the index characteristic weight vector and the operation characteristic weight vector.

[0116] It should be understood that after calculating the digital characteristics and certainty of the preset normal cloud model based on the filtered comprehensive monitoring data to obtain the model digital characteristics and the model certainty, the method further includes:

[0117] Obtain the fault labels in the filtered comprehensive monitoring data, and calculate the digital characteristics in the model digital characteristics whose correlation with the fault labels is greater than the preset correlation threshold to obtain the screened digital characteristics;

[0118] Use the principal component analysis method to perform dimensionality reduction on the screened digital characteristics to obtain the dimensionality-reduced screened digital characteristics.

[0119] It is understandable that by using methods such as correlation analysis and principal component analysis to evaluate and select the extracted features, the features most influential for the fault diagnosis of the electricity meter can be retained, the data dimension can be reduced, and the training efficiency and accuracy of the model can be improved.

[0120] S5. According to the filtered comprehensive monitoring data and the electricity meter box status evaluation system, use the model digital features, model certainty, and model index weights to construct a normal cloud model of the electricity metering box, and obtain an electricity metering box evaluation model.

[0121] It should be understood that the step of constructing a normal cloud model of the electricity metering box according to the filtered comprehensive monitoring data and the electricity meter box status evaluation system, using the model digital features, model certainty, and model index weights to obtain an electricity metering box evaluation model includes:

[0122] Calculate the evaluation weight of the electricity meter box status evaluation system using the model digital features, model certainty, and model index weights to obtain an index evaluation weight;

[0123] Based on the electricity meter box status evaluation system, evaluate the cloud centers of the normal cloud model and the preset standard normal cloud model to obtain a cloud gravity center deviation degree;

[0124] Calculate the deviation degree of the normal cloud model according to the index evaluation weight and the cloud gravity center deviation degree to obtain a cloud gravity center deviation degree;

[0125] Based on the electricity meter box status evaluation system, obtain the status evaluation result of the electricity metering box, and calculate the membership grade of the status evaluation result of the electricity metering box according to the cloud gravity center deviation degree and the preset maximum membership degree principle to obtain an evaluation membership degree;

[0126] Determine the evaluation grade corresponding to the electricity meter box status evaluation system according to the evaluation membership degree to obtain a membership evaluation grade, and construct a normal cloud model of the electricity metering box according to the membership evaluation grade and the filtered comprehensive monitoring data to obtain an electricity metering box evaluation model.

[0127] It should be explained that the cloud center represents the data point in the normal cloud model that can best represent the model, that is, the evaluation index that can best represent the model, and is usually related to the expected values of each evaluation index in the power box status evaluation system for constructing the normal cloud model. The status evaluation result is a comprehensive judgment on the overall operation status of the electric energy metering box. Based on the power box status evaluation system, considering multiple evaluation indexes and the mutual relationship between the evaluation indexes, a conclusion about the current status of the power box is obtained by using the evaluation method based on the normal cloud model. The evaluation membership degree is used to describe the degree to which the status of the electric energy metering box belongs to a certain evaluation level. For example, the membership evaluation levels of the electric energy metering box are divided into three levels: "good", "general", and "poor". It is calculated that the membership degree of the power box status belonging to the "good" level is 0.8, the membership degree belonging to the "general" level is 0.15, and the membership degree belonging to the "poor" level is 0.05. This indicates that the status of the electric energy metering box has an 80% probability of belonging to the "good" level.

[0128] It should be understood that based on the standard operation status of the electric energy metering box, membership degree thresholds belonging to good, medium, poor, and pathological conditions are set for each index of the power box status evaluation. When the operation status of the electric energy metering box is in the poor and pathological conditions, the fault warning mechanism is started, and it is respectively judged whether the membership degrees of each index belonging to good, medium, poor, and pathological conditions exceed the corresponding thresholds. If the corresponding membership degree threshold is exceeded, the index is an abnormal index.

[0129] Furthermore, constructing the normal cloud model of the electric energy metering box according to the membership evaluation level and the filtered comprehensive monitoring data to obtain the power box evaluation model includes:

[0130] Predicting the fault degree of the electric energy metering box according to the membership evaluation level to obtain the predicted fault degree;

[0131] Obtaining the basic data corresponding to the predicted fault degree from the filtered comprehensive monitoring data to obtain the basic monitoring data;

[0132] Dividing the basic monitoring data and the predicted fault degree into a test set and a training set to obtain an evaluation training set and an evaluation test set, and using the evaluation training set to train the normal cloud model, and using the evaluation test set to test the trained normal cloud model to obtain the power box evaluation model.

[0133] It should be noted that when the status evaluation level of the electric energy metering box is "good" obtained through the membership degree evaluation level, the predicted fault degree is no obvious fault risk. At this time, only regular inspections need to be carried out on the electric energy metering box according to the regular schedule. The basic data refers to the voltage data, current data, power data, etc. of the electric energy metering box. The basic data is related to the predicted fault degree. For example, when the status evaluation level of the electric energy metering box is "poor", the predicted fault degree is regarded as a serious fault, which may cause damage to the electric energy metering box and affect the normal operation of the power grid. At this time, the power data may show abnormal and drastic changes. For example, the active power drops significantly instantaneously, while the reactive power rises sharply. This may indicate a short circuit or open circuit fault inside the electric energy metering box. For example, when the active power drops by 20% within 10 minutes, it is regarded as a significant instantaneous drop in active power, and when the reactive power rises by more than 15% within 10 minutes, it is regarded as a sharp rise in reactive power. At this time, there may be a short circuit or open circuit fault inside the electric energy metering box.

[0134] It should be understood that after the operation status evaluation result of the electric energy metering box is generated, the abnormal index data corresponding to the operation status evaluation result is uploaded to the control center and the staff is notified. Based on the received data, the staff can understand the operation status of each electric energy metering box and the possible operation hidden dangers, and arrange the maintenance work immediately.

[0135] S6. Based on the preset intelligent air switch, obtain the real-time monitoring data of the electric energy metering box to be predicted to get the measured data of the electric energy box. According to the measured data of the electric energy box, use the pre-constructed edge computing device to identify the arc of the electric energy metering box to be predicted to get the identification result. After confirming that the identification result is the preset passing result, use the electric energy metering box evaluation model to analyze the measured data of the electric energy box to obtain the fault warning result of the electric energy metering box to be predicted.

[0136] In the embodiment of the present invention, the intelligent air switch can also be called an intelligent circuit breaker, which is a new type of device integrating modern electronic technology, communication technology and computer technology. It not only has the basic functions of traditional air switches, such as on-off control of circuits, overload protection, short circuit protection, etc., but also adds intelligent functions such as remote control, fault diagnosis, data analysis, etc.

[0137] Among them, through various built-in sensors, the intelligent air switch can monitor parameters such as current, voltage, and temperature in the circuit in real time to construct electric energy metering box data.

[0138] It should be explained that the arc can be identified by monitoring the waveform changes of current and voltage. During normal operation, the waveforms of current and voltage have certain regularities, while when an arc occurs, the waveforms of its current and voltage will show characteristics such as mutation and high-frequency oscillation. For example, when an arc occurs, the current waveform may show spike pulses, and the voltage waveform may show a drop phenomenon.

[0139] Furthermore, the recognition result is whether there is an arc or not, and the passing result is a preset judgment criterion. For example, the passing result is that no arc occurs in the power metering box to be predicted within 24 hours.

[0140] It should be understood that in the fault prediction of the power metering box, the arc phenomenon is a very critical fault indication. When an arc phenomenon occurs in the power metering box, there will be a greater risk, which may not only cause equipment damage but also trigger safety accidents such as fires. Therefore, it is necessary to make full use of the advantage of rapid edge computing processing for rapid identification to avoid risks. Specifically, based on the preset intelligent air switch, the real-time monitoring data of the power metering box to be predicted is obtained to get the measured data of the power metering box. According to the measured data of the power metering box, edge computing is used to identify the arc of the power metering box to be predicted. The edge computing device can process at the source of data generation, greatly reducing the time and cost of data transmission and improving the processing efficiency. When the recognition passes and the recognition result is that an arc phenomenon is confirmed to exist, the measured data of the power metering box is analyzed using the power metering box evaluation model. Through this rapid response mechanism, potential faults can be discovered in a timely manner, and corresponding measures can be taken, such as cutting off the power supply and sending out an alarm, to prevent the further expansion of the fault. This method not only improves the safety and reliability of the power metering box but also reduces the power outage time and maintenance cost caused by faults.

[0141] To solve the problems described in the background art, the present invention obtains the historical monitoring data and on-site measured data of the electric energy metering box based on a preset data collection instruction, obtaining the historical monitoring data of the electric meter and the on-site measured data of the electric meter. It can be seen that the present invention not only considers the measured data but also the historical data, improving the richness of the data and laying a data foundation for the subsequent construction of the model. And according to the historical monitoring data of the electric meter and the on-site measured data of the electric meter, the comprehensive monitoring data of the electric energy box is obtained; according to the comprehensive monitoring data of the electric energy box, a number of evaluation indicators for evaluating the electric energy metering box are determined, obtaining a number of electric box evaluation indicators, and according to the number of electric box evaluation indicators, a state evaluation system of the electric energy metering box is constructed, obtaining an electric energy box state evaluation system. It can be seen that for the fault prediction of the present invention, it is necessary to first evaluate the indicators, providing a basis for subsequent quantitative analysis and model construction; based on a preset multi-source data density filtering method and the electric energy box state evaluation system, the comprehensive monitoring data of the electric energy box is filtered, obtaining filtered comprehensive monitoring data; according to the filtered comprehensive monitoring data, the digital characteristics and certainty of a preset normal cloud model are calculated, obtaining model digital characteristics and model certainty, and according to the electric energy box state evaluation system, the index weights of the normal cloud model are calculated, obtaining model index weights; according to the filtered comprehensive monitoring data and the electric energy box state evaluation system, using the model digital characteristics, model certainty and model index weights, a normal cloud model of the electric energy metering box is constructed, obtaining an electric energy metering box evaluation model. It can be seen that in the present invention, by calculating the digital characteristics, certainty and index weights, more accurate quantitative analysis can be provided for the state evaluation of the electric energy metering box; based on a preset intelligent air switch, the real-time monitoring data of the electric energy metering box to be predicted is obtained, obtaining the measured data of the electric energy box. According to the measured data of the electric energy box, an edge computing device pre-constructed is used to identify the electric arc of the electric energy metering box to be predicted, obtaining an identification result. After confirming that the identification result is a preset passing result, the electric energy metering box evaluation model is used to analyze the measured data of the electric energy box, obtaining a fault warning result of the electric energy metering box to be predicted. It can be seen that through the intelligent air switch and edge computing technology, the present invention can quickly and accurately identify and warn of the faults of the electric energy metering box. Therefore, the present invention can reduce the failure rate and maintenance cost of the electric energy metering box and improve the accuracy of the fault warning of the electric energy metering box.

[0142] As Figure 2 shown, it is a functional module diagram of a data-driven electric energy metering box fault warning system provided by an embodiment of the present invention.

[0143] The data-driven power metering box fault warning system 100 according to the present invention can be installed in an electronic device. According to the implemented functions, the data-driven power metering box fault warning system 100 may include a comprehensive monitoring data collection module 101, a status evaluation system construction module 102, an evaluation model construction module 103, and a fault warning analysis module 104. The modules in the present invention may also be referred to as units, which refer to a series of computer program segments that can be executed by a processor of an electronic device and can complete fixed functions, and are stored in the memory of the electronic device.

[0144] The comprehensive monitoring data collection module 101 is configured to obtain historical monitoring data and on-site measured data of the power metering box based on a preset data collection instruction, obtain the electricity meter historical monitoring data and the electricity meter on-site measured data, and obtain the comprehensive monitoring data of the power box according to the electricity meter historical monitoring data and the electricity meter on-site measured data;

[0145] The status evaluation system construction module 102 is configured to determine a plurality of evaluation indicators for evaluating the power metering box according to the comprehensive monitoring data of the power box, obtain a plurality of power box evaluation indicators, and construct a status evaluation system of the power metering box according to the plurality of power box evaluation indicators, so as to obtain a power box status evaluation system;

[0146] The evaluation model construction module 103 is configured to filter the comprehensive monitoring data of the power box based on a preset multi-source data density filtering method and the power box status evaluation system to obtain filtered comprehensive monitoring data;

[0147] Calculate the digital features and certainty degrees of a preset normal cloud model according to the filtered comprehensive monitoring data, obtain model digital features and model certainty degrees, and calculate the index weights of the normal cloud model according to the power box status evaluation system to obtain model index weights;

[0148] Construct a normal cloud model of the power metering box according to the filtered comprehensive monitoring data and the power box status evaluation system, using the model digital features, model certainty degrees and model index weights, to obtain a power metering box evaluation model;

[0149] The fault warning analysis module 104 is configured to obtain real-time monitoring data of the power metering box to be predicted based on a preset intelligent air switch, obtain the measured data of the power box, and perform arc identification on the power metering box to be predicted by using a pre-constructed edge computing device according to the measured data of the power box to obtain an identification result. After confirming that the identification result is a preset passing result, analyze the measured data of the power box by using the power metering box evaluation model to obtain a fault warning result of the power metering box to be predicted.

[0150] Specifically, when the modules in the data-driven power metering box fault warning system 100 according to the embodiments of the present invention are used, they adopt the same technical means as those in the Figure 1 data-driven power metering box fault warning method described above, and can produce the same technical effects, which will not be elaborated here.

[0151] As Figure 3 shown, it is a schematic structural diagram of an electronic device for implementing the data-driven power metering box fault warning method provided by an embodiment of the present invention.

[0152] The electronic device 1 may include a processor 10, a memory 11, and a bus 12, and may further include a computer program stored in the memory 11 and executable on the processor 10, such as a data-driven power metering box fault warning method program.

[0153] Among them, the memory 11 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, mobile hard disk, multimedia card, card-type memory (such as SD or DX memory, etc.), magnetic memory, magnetic disk, optical disk, etc. The memory 11 may be an internal storage unit of the electronic device 1 in some embodiments, such as the mobile hard disk of the electronic device 1. The memory 11 may also be an external storage device of the electronic device 1 in other embodiments, such as a plug-in mobile hard disk, a Smart Media Card (SMC), a Secure Digital (SD) card, a Flash Card, etc. equipped on the electronic device 1. Further, the memory 11 also includes the internal storage unit of the electronic device 1 and the external storage device. The memory 11 can be used not only to store application software installed on the electronic device 1 and various types of data, such as the code of the data-driven power metering box fault warning method program, but also to temporarily store data that has been output or will be output.

[0154] In some embodiments, the processor 10 may be composed of an integrated circuit. For example, it may be composed of a single packaged integrated circuit, or may be composed of multiple packaged integrated circuits with the same or different functions, including a combination of one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and various control chips. The processor 10 is the control core (Control Unit) of the electronic device, connecting various components of the entire electronic device through various interfaces and circuits, and by running or executing programs or modules stored in the memory 11 (such as a program for early warning of power metering box faults based on data driving), and calling data stored in the memory 11, to perform various functions of the electronic device 1 and process data.

[0155] The bus 12 may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. This bus 12 can be divided into an address bus, a data bus, a control bus, etc. The bus 12 is set to enable connection and communication between the memory 11 and at least one processor 10, etc.

[0156] Figure 3 Only the electronic device with components is shown. Those skilled in the art can understand that Figure 3 the shown structure does not constitute a limitation on the electronic device 1, and it may include fewer or more components than shown, or combine certain components, or have a different component arrangement.

[0157] For example, although not shown, the electronic device 1 may further include a power source (such as a battery) for supplying power to each component. Preferably, the power source can be logically connected to the at least one processor 10 through a power management device, so as to implement functions such as charge management, discharge management, and power consumption management through the power management device. The power source may also include any components such as one or more DC or AC power sources, a recharge device, a power failure detection circuit, a power converter or inverter, and a power status indicator. The electronic device 1 may also include various sensors, a Bluetooth module, a Wi-Fi module, etc., which will not be elaborated here.

[0158] Furthermore, the electronic device 1 may further include a network interface. Optionally, the network interface may include a wired interface and / or a wireless interface (such as a WI-FI interface, a Bluetooth interface, etc.), which is usually used to establish a communication connection between the electronic device 1 and other electronic devices.

[0159] Optionally, the electronic device 1 may further include a user interface, which may be a display, an input unit (such as a keyboard), and optionally, the user interface may also be a standard wired interface or a wireless interface. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch liquid crystal display, and an OLED (Organic Light-Emitting Diode) toucher, etc. Among them, the display may also be appropriately referred to as a display screen or a display unit, which is used to display the information processed in the electronic device 1 and to display a visual user interface.

[0160] The program of the data-driven power metering box fault warning method stored in the memory 11 of the electronic device 1 is a combination of multiple instructions, and when running in the processor 10, it can implement:

[0161] Obtain the historical monitoring data and on-site measured data of the power metering box based on a preset data collection instruction, obtain the electricity meter historical monitoring data and the electricity meter on-site measured data, and obtain the comprehensive monitoring data of the power box according to the electricity meter historical monitoring data and the electricity meter on-site measured data;

[0162] Determine multiple evaluation indicators for evaluating the power metering box according to the comprehensive monitoring data of the power box, obtain multiple power box evaluation indicators, and construct a state evaluation system for the power metering box according to the multiple power box evaluation indicators, obtain the power box state evaluation system;

[0163] Filter the comprehensive monitoring data of the power box based on a preset multi-source data density filtering method and the power box state evaluation system to obtain filtered comprehensive monitoring data;

[0164] Calculate the digital characteristics and certainty of a preset normal cloud model according to the filtered comprehensive monitoring data, obtain the model digital characteristics and the model certainty, and calculate the index weights of the normal cloud model according to the power box state evaluation system to obtain the model index weights;

[0165] Construct a normal cloud model of the power metering box according to the filtered comprehensive monitoring data and the power box state evaluation system, using the model digital characteristics, the model certainty, and the model index weights, to obtain the power metering box evaluation model;

[0166] Based on the preset intelligent air switch, obtain the real-time monitoring data of the power metering box to be predicted to get the measured data of the power box. According to the measured data of the power box, use the pre-constructed edge computing device to identify the arc of the power metering box to be predicted to obtain the identification result. After confirming that the identification result is the preset passing result, use the power metering box evaluation model to analyze the measured data of the power box to obtain the fault warning result of the power metering box to be predicted.

[0167] Specifically, for the specific implementation method of the above instructions by the processor 10, reference can be made to Figures 1 to 3 the description of the relevant steps in the corresponding embodiment, which will not be elaborated here.

[0168] Furthermore, if the modules / units integrated in the electronic device 1 are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. The computer-readable storage medium can be volatile or non-volatile. For example, the computer-readable medium can include: any entity or system capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disc, computer memory, read-only memory (ROM, Read-Only Memory).

[0169] The present invention also provides a computer-readable storage medium, and the readable storage medium stores a computer program. When the computer program is executed by the processor of the electronic device, it can implement:

[0170] Based on the preset data collection instructions, obtain the historical monitoring data and on-site measured data of the power metering box to get the historical monitoring data of the power meter and the on-site measured data of the power meter, and obtain the comprehensive monitoring data of the power box according to the historical monitoring data of the power meter and the on-site measured data of the power meter;

[0171] Determine multiple evaluation indicators for evaluating the power metering box according to the comprehensive monitoring data of the power box to obtain multiple power box evaluation indicators, and construct a state evaluation system for the power metering box according to the multiple power box evaluation indicators to obtain a power box state evaluation system;

[0172] Based on the preset multi-source data density filtering method and the power box state evaluation system, filter the comprehensive monitoring data of the power box to obtain filtered comprehensive monitoring data;

[0173] Calculate the digital characteristics and certainty degrees of the preset normal cloud model according to the filtered comprehensive monitoring data to obtain model digital characteristics and model certainty degrees, and calculate the index weights of the normal cloud model according to the power box state evaluation system to obtain model index weights;

[0174] According to the above-mentioned filtering comprehensive monitoring data and power box status evaluation system, a normal cloud model of the electric energy metering box is constructed by using the model digital features, model certainty, and model index weights, and an electric energy metering box evaluation model is obtained.

[0175] Based on the preset intelligent air switch, the real-time monitoring data of the to-be-predicted electric energy metering box is obtained to get the measured data of the power box. According to the measured data of the power box, the pre-constructed edge computing device is used to identify the arc of the to-be-predicted electric energy metering box to obtain the identification result. After confirming that the identification result is the preset passing result, the electric energy metering box evaluation model is used to analyze the measured data of the power box to obtain the fault warning result of the to-be-predicted electric energy metering box.

[0176] In several embodiments provided by the present invention, it should be understood that the disclosed devices, systems, and methods can be implemented in other ways. For example, the system embodiments described above are only illustrative, and there may be other division methods in actual implementation.

[0177] The modules described as separate components may or may not be physically separated. The components shown as modules may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0178] In addition, in each embodiment of the present invention, the functional modules can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit. The above-mentioned integrated units can be implemented in the form of hardware or in the form of a combination of hardware and software functional modules.

[0179] For those skilled in the art, it is obvious that the present invention is not limited to the details of the above-mentioned exemplary embodiments, and without departing from the spirit or basic characteristics of the present invention, the present invention can be implemented in other specific forms.

[0180] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A data-driven fault warning method for electric energy metering boxes, characterized in that, The method includes: Obtaining historical monitoring data and on-site measured data of the electric energy metering box based on a preset data collection instruction to obtain historical monitoring data of the electric meter and on-site measured data of the electric meter, and obtaining comprehensive monitoring data of the electric energy box according to the historical monitoring data of the electric meter and the on-site measured data of the electric meter; Determining a plurality of evaluation indexes for evaluating the electric energy metering box according to the comprehensive monitoring data of the electric energy box to obtain a plurality of electric box evaluation indexes, and constructing a state evaluation system of the electric energy metering box according to the plurality of electric box evaluation indexes to obtain an electric energy box state evaluation system; Filtering the comprehensive monitoring data of the electric energy box based on a preset multi-source data density filtering method and the electric energy box state evaluation system to obtain filtered comprehensive monitoring data; Calculating digital characteristics and certainty degrees of a preset normal cloud model according to the filtered comprehensive monitoring data to obtain model digital characteristics and model certainty degrees, and calculating index weights of the normal cloud model according to the electric energy box state evaluation system to obtain model index weights; Constructing a normal cloud model of the electric energy metering box by using the model digital characteristics, model certainty degrees and model index weights according to the filtered comprehensive monitoring data and the electric energy box state evaluation system to obtain an electric energy metering box evaluation model; Obtaining real-time monitoring data of the to-be-predicted electric energy metering box based on a preset intelligent air switch to obtain measured data of the electric energy box, identifying an arc of the to-be-predicted electric energy metering box by using a pre-constructed edge computing device according to the measured data of the electric energy box to obtain an identification result, and after confirming that the identification result is a preset passing result, analyzing the measured data of the electric energy box by using the electric energy metering box evaluation model to obtain a fault warning result of the to-be-predicted electric energy metering box.

2. The data-driven power metering box fault warning method according to claim 1, characterized in that, The filtering the comprehensive monitoring data of the electric energy box based on a preset multi-source data density filtering method and the electric energy box state evaluation system to obtain filtered comprehensive monitoring data includes: Obtaining a target index in the electric energy state evaluation system, and obtaining a plurality of samples in the comprehensive monitoring data of the electric energy box based on the target index in the electric energy box state evaluation system to obtain a plurality of target samples; Calculating the Euclidean distance between every two samples among the plurality of target samples to obtain a plurality of sample Euclidean distances; Sorting the plurality of sample Euclidean distances in ascending order to obtain a sample ascending queue, and calculating the maximum neighbor distance and local density based on the sample ascending queue to obtain the sample maximum neighbor distance and local density; Calculating a local density threshold and a maximum neighbor distance threshold of the comprehensive monitoring data of the electric energy box according to the sample maximum neighbor distance and local density, and filtering the comprehensive monitoring data of the electric energy box by using the local density threshold and the maximum neighbor distance threshold to obtain filtered comprehensive monitoring data.

3. The data-driven power metering box fault warning method according to claim 1, characterized in that The calculating digital characteristics and certainty degrees of a preset normal cloud model according to the filtered comprehensive monitoring data to obtain model digital characteristics and model certainty degrees includes: Constructing a data sample according to the filtered comprehensive monitoring data to obtain a monitoring data sample, and calculating an expected value of the monitoring data sample to obtain a sample data expected value; Calculate the uncertainty entropy value of the monitored data sample to obtain the monitored sample entropy value, and perform a secondary solution on the monitored sample entropy value to obtain the monitored sample hyper-entropy; Calculate the certainty degree of the monitored data sample according to the sample data expected value and the monitored sample hyper-entropy to obtain the sample certainty degree, and use the sample data expected value, the monitored sample entropy value, and the monitored sample hyper-entropy as the model digital characteristics of the normal cloud model, and use the sample certainty degree as the model certainty degree of the normal cloud model.

4. The data-driven power metering box fault warning method according to claim 1, characterized in that, The calculating the index weight of the normal cloud model according to the power box status evaluation system to obtain the model index weight includes: Obtain the evaluation indexes in the power box status evaluation system, construct a judgment matrix according to the evaluation indexes in the power box status evaluation system to obtain an evaluation judgment matrix, and calculate the maximum eigenvalue of the evaluation judgment matrix to obtain the evaluation maximum eigenvalue; Obtain the weight vector of the normal cloud model according to the preset characteristic equation and the evaluation maximum eigenvalue to obtain the index characteristic weight vector; Obtain the problems corresponding to the evaluation indexes and the sample data corresponding to the problems, construct an evaluation matrix according to the sample data in the problems corresponding to the evaluation indexes to obtain an initial evaluation matrix, and perform a normalization process on the initial evaluation matrix to obtain a standardized matrix; Calculate the entropy of the evaluation indexes based on the preset entropy weight method to obtain the evaluation index entropy, calculate the entropy weight of the evaluation indexes according to the evaluation index entropy to obtain the evaluation index entropy weight, and obtain the weight vector of the evaluation indexes according to the evaluation index entropy weight to obtain the operation characteristic weight vector; Calculate the index weight of the normal cloud model according to the index characteristic weight vector and the operation characteristic weight vector to obtain the model index weight.

5. The data-driven power metering box fault warning method according to claim 1, characterized in that The constructing the normal cloud model of the electric energy metering box according to the filtered comprehensive monitored data and the power box status evaluation system, using the model digital characteristics, the model certainty degree, and the model index weight to obtain the electric energy metering box evaluation model includes: Calculate the evaluation weight of the power box status evaluation system by using the model digital characteristics, the model certainty degree, and the model index weight to obtain the index evaluation weight; Evaluate the cloud centers of the normal cloud model and the preset standard normal cloud model based on the power box status evaluation system to obtain the cloud center deviation degree; Calculate the deviation degree of the normal cloud model according to the index evaluation weight and the cloud center deviation degree to obtain the cloud center deviation degree; Obtain the status evaluation result of the electric energy metering box based on the power box status evaluation system, and calculate the membership level of the status evaluation result of the electric energy metering box according to the cloud center deviation degree and the preset maximum membership degree principle to obtain the evaluation membership degree; Determine the evaluation level corresponding to the power box status evaluation system according to the evaluation membership degree to obtain the membership evaluation level, and construct the normal cloud model of the electric energy metering box according to the membership evaluation level and the filtered comprehensive monitored data to obtain the electric energy metering box evaluation model.

6. The data-driven power metering box fault warning method according to claim 5, characterized in that, The constructing the normal cloud model of the electric energy metering box according to the membership evaluation level and the filtered comprehensive monitored data to obtain the electric energy metering box evaluation model includes: Predict the fault degree of the electric energy metering box according to the subordinate evaluation level to obtain the predicted fault degree; Obtain the basic data corresponding to the predicted fault degree from the filtered comprehensive monitoring data to obtain the basic monitoring data; Divide the basic monitoring data and the predicted fault degree into a test set and a training set to obtain an evaluation training set and an evaluation test set, and use the evaluation training set to train the normal cloud model, and use the evaluation test set to test the trained normal cloud model to obtain the electric energy metering box evaluation model.

7. The data-driven power metering box fault warning method according to claim 1, wherein After calculating the digital characteristics and certainty of the preset normal cloud model according to the filtered comprehensive monitoring data to obtain the model digital characteristics and model certainty, the method further includes: Obtain the fault label in the filtered comprehensive monitoring data, and calculate the digital characteristics in the model digital characteristics whose correlation with the fault label is greater than the preset correlation threshold to obtain the screened digital characteristics; Use the principal component analysis method to reduce the dimension of the screened digital characteristics to obtain the dimension-reduced screened digital characteristics.

8. The method for early warning of faults in an electric energy metering box based on data driving according to claim 3, wherein The calculation formula of the expected value of the sample data is as follows: Among them, is the expected value of the sample data under evaluation index i, x ij is the j-th sample data under the i-th evaluation index, is the sample data set composed of the sample data of evaluation index i, is the set The number of samples included.

9. The data-driven power metering box fault warning method according to claim 3, characterized in that The calculation formula of the sample certainty is as follows: Among them, μ(x ij ) is the sample certainty under the sample data x ij , representing the membership degree of the sample data x ij to the i-th evaluation index, with a value range of [0, 1], and x ij is the j-th sample data under the i-th evaluation index, is the sample data expectation, is the monitoring sample hyperentropy.

10. A data-driven power metering box fault warning system, characterized in that, The system includes: A comprehensive monitoring data collection module, configured to obtain the historical monitoring data and on-site measured data of the electric energy metering box based on a preset data collection instruction to obtain the electric meter historical monitoring data and the electric meter on-site measured data, and obtain the electric energy box comprehensive monitoring data according to the electric meter historical monitoring data and the electric meter on-site measured data; A state evaluation system construction module, configured to determine multiple evaluation indicators for evaluating the electric energy metering box according to the electric energy box comprehensive monitoring data to obtain multiple electric box evaluation indicators, and construct a state evaluation system of the electric energy metering box according to the multiple electric box evaluation indicators to obtain the electric energy box state evaluation system; An evaluation model construction module, configured to filter the electric energy box comprehensive monitoring data based on a preset multi-source data density filtering method and the electric energy box state evaluation system to obtain the filtered comprehensive monitoring data; Calculate the digital characteristics and certainty of the preset normal cloud model according to the filtered comprehensive monitoring data to obtain the model digital characteristics and model certainty, and calculate the index weight of the normal cloud model according to the electric energy box state evaluation system to obtain the model index weight; Construct a normal cloud model of the electric energy metering box according to the filtered comprehensive monitoring data and the electric energy box state evaluation system, using the model digital characteristics, model certainty and model index weight to obtain the electric energy metering box evaluation model; A fault warning analysis module, configured to obtain the real-time monitoring data of the electric energy metering box to be predicted based on a preset intelligent air switch to obtain the electric energy box measured data, and according to the electric energy box measured data, use a pre-constructed edge computing device to identify the electric arc of the electric energy metering box to be predicted to obtain the identification result, and after confirming that the identification result is a preset passing result, use the electric energy metering box evaluation model to analyze the electric energy box measured data to obtain the fault warning result of the electric energy metering box to be predicted.

Images