Electric energy meter precision degradation evaluation system and equipment based on sampling unit failure characteristics

Abstract

The application discloses a sampling unit fault feature-based electric energy meter precision degradation evaluation system and equipment, relates to the technical field of electric energy meters, and comprises the following: a reading module extracts the sampling log of a target sampling unit, and obtains current voltage input and response data; an identification module extracts features and identifies impact modes such as surges, short circuits and high-frequency overloads, and establishes a response feature mapping; a first prediction module inputs the impact modes and the response features into a time sequence channel, and generates a preliminary fault prediction result; a second prediction module constructs a test database based on the preliminary result, performs fault verification, and outputs a final prediction result; and a degradation evaluation module fuses working state data and the prediction result, and completes multi-factor precision degradation evaluation. Furthermore, the technical effects of multi-dimensional prediction fusion, improved prediction accuracy and optimized prediction efficiency are achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of electric energy meter, and particularly relates to an electric energy meter precision degradation evaluation system and equipment based on sampling unit fault characteristics. BACKGROUND

[0002] In the field of electric energy meter precision degradation evaluation, traditional methods often rely on simple data reading and preliminary analysis, such as only monitoring part of the data of the electric energy meter under normal working conditions, which is difficult to comprehensively and deeply obtain detailed information of the sampling unit under various working conditions.

[0003] At the same time, for the fault characteristic impact mode that may occur in the sampling unit, such as surge impact, short circuit impact and high frequency overload, there are obvious deficiencies in identification and analysis, lack of effective feature extraction and accurate identification means, and thus the response characteristics of the impact mode cannot be clearly determined, affecting the efficiency and accuracy of fault identification and analysis. SUMMARY

[0004] The present application provides an electric energy meter precision degradation evaluation system and equipment based on sampling unit fault characteristics to solve the technical problems of one-sided prediction, low prediction accuracy and poor efficiency in the prior art, and achieves the technical effects of multi-dimensional prediction fusion, improved prediction accuracy and optimized prediction efficiency.

[0005] In a first aspect, the present application provides an electric energy meter precision degradation evaluation system based on sampling unit fault characteristics, wherein the electric energy meter precision degradation evaluation system based on sampling unit fault characteristics comprises:

[0006] A reading module is configured to read the sampling log of a target sampling unit, extract a collection data set, and the collection data set comprises an input data set of current and voltage and a response data set of the target sampling unit.

[0007] An identification module is configured to identify a characteristic impact mode after feature extraction of the collection data set, the characteristic impact mode comprises surge impact, short circuit impact and high frequency overload, and establish a response feature of impact mode mapping.

[0008] A first prediction module is configured to input the characteristic impact mode and the response feature as mapping input data into a time sequence fault prediction channel, and establish a first fault prediction result.

[0009] A second prediction module is configured to establish a test database based on the first fault prediction result, perform fault verification of the target sampling unit, and establish a second fault prediction result.

[0010] A degradation evaluation module is configured to establish working state data of the electric energy meter according to the collection data set, perform multi-factor fusion precision degradation evaluation on the working state data and the second fault prediction result, and generate a precision degradation evaluation result.

[0011] In a feasible implementation, the time sequence fault prediction channel in the first prediction module comprises:

[0012] An independent prediction layer is configured to read the feature extraction result of the feature impact mode, perform independent impact prediction on the corresponding feature impact mode according to the feature extraction result, and establish an independent impact prediction result, wherein the feature extraction result comprises voltage, current fluctuation frequency, fluctuation amplitude, and duration.

[0013] A time sequence compensation layer is configured to, after receiving the independent impact prediction result, perform time sequence linkage compensation according to the time sequence relationship of the independent impact prediction result, and establish a time sequence linkage compensation result.

[0014] A hidden danger verification layer is configured to read the time sequence linkage compensation result and the response feature, perform real response verification, and establish the first fault prediction result according to the real response verification result.

[0015] In a feasible implementation, the second prediction module comprises:

[0016] A global constraint unit is configured to obtain basic unit information of a target sampling unit, and configure a globally covered basic test database according to the basic unit information.

[0017] A local enhancement unit is configured to generate a test focus according to the first fault prediction result, perform database focus enhancement on the basic test database according to the test focus, and establish the test database.

[0018] In a feasible implementation, the degradation evaluation module comprises:

[0019] A state feature extraction unit is configured to perform feature extraction on the working state data, such as voltage fluctuation rate, current fluctuation rate, working temperature, power load, and power consumption behavior mode, and establish a state feature set.

[0020] A multi-factor fusion evaluation unit is configured to take the state feature set and the second fault prediction result as multi-dimensional evaluation factors, perform fusion precision degradation impact analysis after adaptive weight allocation, and establish a precision degradation evaluation result.

[0021] In a feasible implementation, after the adaptive weight allocation, the fusion precision degradation impact analysis is performed, and the precision degradation evaluation result is established, and the execution step comprises:

[0022] The historical data set of the electric energy meter is acquired, a precision influence correlation degree analysis of multi-dimensional evaluation factors is performed according to the historical data set, and a first dynamic weight is established according to a result of the precision influence correlation degree analysis.

[0023] The basic information of the electric energy meter is taken as a matching feature, a similar matching of the experience database is performed, a weight self-adaptive analysis of heuristic rules is performed according to a result of the similar matching, and a second dynamic weight is established.

[0024] After the first dynamic weight and the second dynamic weight are verified, the adaptive allocation weight is completed.

[0025] In a feasible implementation manner, the reading module comprises:

[0026] An abnormal data monitoring unit is configured to perform self-fluctuation authentication of a data stream of the sampling log after the sampling log is read, establish a self-fluctuation anomaly, call a simulation environment to perform simulation testing of the data stream, and establish an auxiliary fluctuation anomaly according to a result of the simulation testing.

[0027] A correction unit is configured to perform abnormality identification correction according to the self-fluctuation anomaly and the auxiliary fluctuation anomaly, and extract a collection data set according to a result of the abnormality identification correction.

[0028] In a feasible implementation manner, the electric energy meter precision degradation evaluation system based on a sampling unit fault feature further comprises:

[0029] A warning threshold setting module is configured to acquire a task database of the electric energy meter, and establish a fixed precision threshold according to the task database.

[0030] A real-time warning module is configured to perform trigger authentication of a precision degradation evaluation result according to the fixed precision threshold, establish a trigger authentication result, and report a hierarchical warning signal by using the trigger authentication result.

[0031] In a feasible implementation manner, the real-time warning module is further configured to:

[0032] Perform precision degradation trend identification on the precision degradation evaluation result.

[0033] Generate a degradation rate index according to a result of the precision degradation trend identification, and correct the trigger authentication result according to the degradation rate index.

[0034] Report the hierarchical warning signal according to the corrected trigger authentication result.

[0035] In a feasible implementation manner, the electric energy meter precision degradation evaluation system based on a sampling unit fault feature further comprises:

[0036] The self-updating module is configured to receive the precision degradation evaluation result, synchronously acquire real measurement feedback, evaluate and verify the precision degradation evaluation result according to the real measurement feedback, establish a self-updating database, and update the degradation evaluation module by using the self-updating database.

[0037] In a second aspect, the present application further provides a device comprising a memory configured to store executable instructions, and a processor configured to execute the executable instructions stored in the memory to implement the precision degradation evaluation system for electric energy meters based on sampling unit fault features.

[0038] The present application discloses a precision degradation evaluation system for electric energy meters based on sampling unit fault features and a device thereof, which comprises a reading module configured to read a sampling log of a target sampling unit, extract a collected data set, and collect an input data set of current and voltage and a response data set of the target sampling unit; an identification module configured to perform feature extraction on the collected data set, identify a feature impact mode, and establish a mapping relationship of response features based on the impact mode, wherein the feature impact mode comprises a surge impact, a short-circuit impact and a high-frequency overload; a first prediction module configured to input the feature impact mode and the response features as mapping inputs into a time sequence fault prediction channel, and establish a first fault prediction result; a second prediction module configured to establish a test database based on the first fault prediction result, perform fault verification on the target sampling unit, and obtain a second fault prediction result; and a degradation evaluation module configured to establish an electric energy meter working state data according to the collected data set, combine the second fault prediction result, perform precision degradation evaluation by using a multi-factor fusion method, and generate a precision degradation evaluation result. BRIEF DESCRIPTION OF DRAWINGS

[0039] Figure 1 FIG. 1 is a structural schematic diagram of the precision degradation evaluation system for electric energy meters based on sampling unit fault features according to the present application.

[0040] Figure 2 FIG. 2 is a structural schematic diagram of an exemplary device according to the present application.

[0041] FIG. 1 is a structural schematic diagram of the precision degradation evaluation system for electric energy meters based on sampling unit fault features according to the present application. DETAILED DESCRIPTION

[0042] The above technical solutions will be described in detail below in combination with the accompanying drawings and specific embodiments, so as to better understand the above technical solutions. Obviously, the described embodiments are only part of the embodiments of the present application, not all embodiments of the present application, and it should be understood that the present application is not limited to the example embodiments for explaining the present application. Based on the embodiments of the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the scope of protection of the present application. In addition, it should be noted that, for the convenience of description, only the parts related to the present application are shown in the drawings, not all.

[0043] Embodiment one, as Figure 1 The flowchart of the electric energy meter precision degradation evaluation system based on the sampling unit fault characteristics of the present application, wherein the electric energy meter precision degradation evaluation system based on the sampling unit fault characteristics comprises:

[0044] The reading module 11 is used for reading the sampling log of the target sampling unit, extracting the collection data set, and the collection data set comprises the input data set of current and voltage and the response data set of the target sampling unit.

[0045] Specifically, the target sampling unit refers to a specific component in the electric energy meter for acquiring current, voltage and other key parameters; the sampling log refers to a collection of current, voltage and other data recorded at certain time intervals during the operation of the electric energy meter; the collection data set comprises the input data set of current and voltage (i.e. the actual received current and voltage values of the electric energy meter during operation) and the response data set of the target sampling unit (i.e. the output response of the target sampling unit after receiving the input signal, such as voltage drop, current change, etc.), which is used to provide original data support for subsequent fault identification and prediction.

[0046] In some embodiments, the reading module comprises:

[0047] The abnormal data monitoring unit is used for self-fluctuation authentication of the data stream of the sampling log after reading the sampling log, establishing self-fluctuation anomaly, calling the simulation environment to perform simulation test of the data stream, and establishing auxiliary fluctuation anomaly according to the simulation test result; the correction unit is used for abnormal identification correction according to the self-fluctuation anomaly and the auxiliary fluctuation anomaly, and extracting the collection data set according to the abnormal identification correction result.

[0048] Specifically, the data stream corresponds to the current, voltage and other data continuously generated in the operation process of the electric energy meter according to certain time sequence; the self-fluctuation authentication refers to automatic detection and verification of the fluctuation in the data stream, to judge whether it exceeds the normal fluctuation range; accordingly, the self-fluctuation anomaly refers to the data fluctuation that exceeds the normal fluctuation range.

[0049] Specifically, if a self-wave fluctuation anomaly is found, a simulation environment is called to perform simulation testing, the simulation environment simulates the fluctuation of the data stream, and auxiliary wave fluctuation anomaly data (such as an abnormal pattern of the self-wave fluctuation anomaly) is generated to further verify and supplement the analysis result of the self-wave fluctuation anomaly.

[0050] Specifically, the main function of the correction unit is to correct the self-wave fluctuation anomaly and the auxiliary wave fluctuation anomaly identified by the abnormal data monitoring unit. The correction algorithm and logical judgment of the abnormal data are involved, for example, a smoothing algorithm is used to filter out abnormal fluctuations caused by external interference or a prediction algorithm is used to correct data deviation caused by internal faults, and finally accurate data sets are extracted to provide a reliable data basis for subsequent fault identification and prediction.

[0051] Through the introduction of the abnormal data monitoring unit and the correction unit, a high-quality data basis can be provided for the precision degradation evaluation system of the electric energy meter.

[0052] The identification module 12 is configured to identify a feature impact pattern after feature extraction on the collected data set, the feature impact pattern including surge impact, short-circuit impact, and high-frequency overload, and establish a response feature of the impact pattern mapping.

[0053] Specifically, the feature impact pattern refers to an abnormal energy feature pattern caused by sudden load change, power grid disturbance, or equipment failure within a specific time window, which has characteristics such as short-time high amplitude, high frequency, or nonlinear fluctuation; the response feature refers to the characteristic response index (and corresponding numerical range or combination) of the system under the impact pattern, which includes but is not limited to instantaneous current rise rate, voltage drop amplitude, frequency deviation, and harmonic distortion rate.

[0054] Specifically, the feature impact pattern includes but is not limited to: surge impact, which typically occurs at the moment of device startup or power recovery, and is characterized by a short-time sharp rise in current; short-circuit impact, which occurs when a line short-circuit or ground fault occurs, and is characterized by a sudden voltage drop and current saturation; high-frequency overload, which is characterized by an increase in high-frequency components, which may be accompanied by heat rise and harmonic distortion.

[0055] Specifically, the collected data set is processed by sliding window segmentation, and the statistical features (such as mean, variance, skewness), frequency domain features (such as FFT spectrum energy, harmonic ratio), and time-frequency features (such as STFT, CWT coefficient) of each segment of data are extracted; then, based on a pre-set rule or a pre-trained model (such as a convolutional neural network, a random forest, or a support vector machine), the extracted feature vectors are classified to identify whether there is an impact event; if there is, the type of impact pattern (such as surge, short-circuit, or high-frequency overload) to which it belongs is further identified.

[0056] Further, for each identified impact mode, the corresponding response characteristic indicator is recorded to construct a mapping table or mapping function between impact mode and response characteristic, which is used for subsequent prediction, early warning or compensation control.

[0057] Through the impact mode identification and response characteristic mapping method described above, the classification of impact modes is realized, and different impact modes can trigger different response strategies, such as load switching, filter activation or path reconstruction. The classification results can provide a basis for subsequent dynamic compensation and strategy selection.

[0058] The first prediction module 13 is configured to input the feature impact mode and the response characteristic as mapping input data into a time sequence fault prediction channel to establish a first fault prediction result.

[0059] Specifically, the mapping input data is an input vector formed by combining the feature impact mode and the response characteristic, which is the input basis of the prediction model; the first fault prediction result refers to the type, risk level or probability of possible faults in a future time window predicted based on time sequence modeling and response verification.

[0060] In some embodiments, the time sequence fault prediction channel in the first prediction module includes:

[0061] An independent prediction layer is configured to read the feature extraction result of the feature impact mode, perform independent impact prediction of the corresponding feature impact mode according to the feature extraction result, establish an independent impact prediction result, and the feature extraction result includes voltage, current fluctuation frequency, fluctuation amplitude and duration; a time sequence compensation layer is configured to receive the independent impact prediction result, perform time sequence linkage compensation according to the time sequence relationship of the independent impact prediction result, and establish a time sequence linkage compensation result; a hidden danger verification layer is configured to read the time sequence linkage compensation result and the response characteristic to perform real response verification, and establish the first fault prediction result according to the real response verification result.

[0062] Specifically, the time sequence fault prediction channel includes the following three functional layers:

[0063] An independent prediction layer is configured to read the feature extraction result of the feature impact mode; to perform independent modeling and impact prediction for each impact mode, and to output an independent impact prediction result; preferably, this layer can use a lightweight neural network (such as 1D-CNN, GRU) to model each impact type separately.

[0064] Temporal compensation layer, configured to receive independent impact prediction results, identify linkage relationship thereof in time dimension, establish causal relationship and cumulative effect on time sequence, and output temporal linkage compensation results; preferably, the layer can capture time sequence dependency by using structures such as temporal attention and bidirectional LSTM.

[0065] Hidden danger verification layer, configured to read temporal linkage compensation results and response feature data, perform similarity matching and deviation analysis by comparing historical real response curves, and further generate first fault prediction results in combination with real historical data; preferably, the layer can perform response verification by using a variational autoencoder (VAE) or a residual matching network (Residual Matching Network).

[0066] Through the above-described temporal fault prediction channel based on feature impact mapping, potential coupling relationship between impact patterns can be identified, and time precision and structural accuracy of fault prediction can be improved; through a real verification mechanism of response features, false positives and overfitting can be eliminated, and prediction credibility can be improved.

[0067] The second prediction module 14 is configured to establish a test database based on the first fault prediction results, perform fault verification of a target sampling unit, and establish second fault prediction results.

[0068] Specifically, the first fault prediction results are initial prediction conclusions output by the temporal fault prediction channel, and exemplarily include fault types, risk levels, prediction time windows, and the like; the test database is a structured verification database constructed based on the first fault prediction results as indexes, and is configured to collect real-time or backtracking data related to prediction events; through the test database, comparative analysis based on measured data and prediction data can be implemented, and then corrected prediction results (i.e., second fault prediction results) can be established.

[0069] In some embodiments, the second prediction module includes:

[0070] A global constraint unit is configured to obtain basic unit information of the target sampling unit, and configure a globally covered basic test database according to the basic unit information; and a local enhancement unit is configured to generate test attention according to the first fault prediction results, perform database attention enhancement on the basic test database according to the test attention, and establish the test database.

[0071] Specifically, the basic unit information refers to the basic attribute information of the target sampling unit, such as the physical structure, deployment location, function type, and historical running state, which can come from the production unit corresponding to the target sampling unit and the equipment selection database. The basic test database is a preset universal test data set covering global regions and multiple types of equipment, including historical sampling data, typical fault samples, and environmental parameters.

[0072] Specifically, the test focus is the focus content derived from the first fault prediction result, which is used to indicate the key time period, key indicator, key location, or key mode involved in the current prediction event. Based on the test focus, the basic test database can be filtered, reweighted, labeled, or deep completed (i.e., the process of focus enhancement) to generate a more targeted test database.

[0073] For example, if the prediction is high-frequency harmonic interference, the focus indicators generated are frequency offset and harmonic energy. If the prediction is thermal anomaly, the focus is on temperature rise curve and infrared image features. Further, the following enhancement operations are performed on the basic test database: dimension filtering, which only retains the indicator dimensions related to the test focus; label reweighting, which improves the sample weight related to the current focus content; data completion, which schedules adjacent nodes or historical data for completion if the focus dimension data is missing; context expansion, which introduces data in the time period before and after the prediction event to enhance the temporal context. Finally, the enhanced test database is generated as the data support for the current verification task.

[0074] Further, the sampling data is compared and analyzed with the enhanced test database (such as similarity calculation, trend fitting, and anomaly detection). Optionally, if the verification is consistent, the first prediction result is confirmed; if the deviation is significant, the second fault prediction result is output, and the model parameters are updated.

[0075] The above-mentioned database enhancement based on test focus and second fault prediction result establishment method avoids redundant calling of global data through the test focus mechanism, improves the database targeting and verification efficiency, and further helps to enhance the verifiability and credibility of the prediction result.

[0076] The degradation evaluation module 15 is configured to establish an electric energy meter working state data based on the collected data set, perform multi-factor fusion precision degradation evaluation on the working state data and the second fault prediction result, and generate a precision degradation evaluation result.

[0077] Specifically, the electric energy meter working state data is a structured state vector extracted based on a collection data set, which can reflect the current and historical operation state of the electric energy meter; the multi-factor fusion model is an analysis model for fusing the electric energy meter state data and the second fault prediction result, which can be implemented in a manner such as weighted scoring, Bayesian inference, and graph neural network, to obtain an accuracy degradation evaluation result indicating the accuracy degradation trend of the electric energy meter in the current or future period of time. Optionally, the accuracy degradation evaluation result is in the form of degradation level, accuracy offset, residual life, etc.

[0078] Through the above multi-factor fusion-based electric energy meter accuracy degradation evaluation method, quantitative perception of the state of the metering device can be achieved, and more comprehensive accuracy degradation evaluation can be formed. The use of prior information provided by the second fault prediction result helps to improve the forward-looking and accuracy of the evaluation model.

[0079] In some embodiments, the degradation evaluation module comprises:

[0080] a state feature extraction unit configured to extract features of voltage fluctuation rate, current fluctuation rate, working temperature, power load, and power consumption behavior mode from the working state data, and establish a state feature set; and a multi-factor fusion evaluation unit configured to take the state feature set and the second fault prediction result as multi-dimensional evaluation factors, perform fusion accuracy degradation influence analysis after adaptive weight allocation, and establish an accuracy degradation evaluation result.

[0081] Specifically, the voltage fluctuation rate refers to the standard deviation or coefficient of variation of voltage change per unit time, reflecting power supply stability; the current fluctuation rate refers to the current change amplitude per unit time, reflecting load dynamic characteristics; the working temperature refers to temperature information near the inside or outside of the electric energy meter, which can represent thermal aging risk; the power load refers to active power or apparent power per unit time, reflecting user power consumption intensity; the power consumption behavior mode refers to periodic or sudden characteristics extracted through time series clustering or frequency domain analysis, such as diurnal load distribution, holiday power consumption change, etc.; and the state feature set is a vector set composed of the above-mentioned multiple features, which is used as an evaluation input.

[0082] Specifically, the fusion accuracy degradation influence analysis is a joint modeling of the state feature set and the second fault prediction result to comprehensively quantify the affected state of the electric energy meter accuracy. First, the state feature set and the second fault prediction result are taken as multi-dimensional evaluation factors, adaptive weight allocation is performed on the multi-dimensional evaluation factors, and optionally, the adaptive weight allocation includes: generating a first dynamic weight based on feature correlation analysis of historical data; generating a second dynamic weight based on similarity matching of electric energy meter basic information and an experience database, combined with heuristic rules; and performing weight verification and normalization processing on the first dynamic weight and the second dynamic weight to complete the final weight allocation.

[0083] Further, after completing the weight distribution, a precision degradation evaluation model is constructed to perform fusion precision degradation impact analysis; the precision degradation evaluation model can be constructed using neural networks, logistic regression, fuzzy reasoning, empirical fitting, etc., wherein the model inputs include: state vectors (such as temperature rise rate, voltage fluctuation rate), fault prediction factors (such as prediction confidence, impact indicators); the evaluation model outputs are: current precision degradation level (such as normal, slight degradation, severe degradation), precision offset (such as ±0.5%), remaining useful life prediction value (such as 120 days), etc. The precision degradation evaluation model is trained and verified through historical degradation samples.

[0084] In the above process, the accuracy of the evaluation is improved by modeling the precision degradation trend based on multi-source data; the forward-looking and adaptability are enhanced by fusing state characteristics and prediction results.

[0085] In some implementations, after the adaptive weight distribution, the fusion precision degradation impact analysis is performed, and the precision degradation evaluation result is established, and the execution step includes:

[0086] The historical data set of the electric energy meter is obtained, the precision influence correlation degree analysis of the multi-dimensional evaluation factors is performed according to the historical data set, the first dynamic weight is established according to the precision influence correlation degree analysis result; the basic information of the electric energy meter is taken as the matching feature, the similar matching of the experience database is performed, the weight adaptive analysis of the heuristic rule is performed according to the similar matching result, and the second dynamic weight is established; after the weight verification of the first dynamic weight and the second dynamic weight, the adaptive weight distribution is completed.

[0087] Specifically, first, the historical operation data set of the target electric energy meter is obtained, which includes state data and actual precision offset records in multiple historical periods; then, the precision influence correlation degree analysis is performed on each dimension evaluation factor (such as voltage fluctuation rate, current fluctuation rate, temperature rise rate, etc.) in the historical data set, for example: the correlation coefficient (such as Pearson coefficient, mutual information, grey correlation degree, etc.) between the evaluation factor and the measurement error is calculated to identify the dominant factor affecting the precision degradation and its dynamic change trend, and then according to the analysis result, a preliminary weight is assigned to each evaluation factor to form a first dynamic weight set.

[0088] Specifically, the basic information of the target electric energy meter is obtained, including but not limited to device model, production batch, installation environment, running time, etc.; the basic information is taken as the matching feature, and the similar matching is performed in the pre-constructed experience database; optionally, K-neighbor matching, hash index, feature vector similarity, etc. are used to extract the historical degradation patterns of several similar devices as the similar matching result; then, heuristic rule analysis is performed based on the matching result, for example:

[0089] If the high-temperature environment in similar devices is frequently marked as a degradation inducer, the weight of the working temperature factor is increased; if the current fluctuation rate sensitivity exists universally in a certain type of device, the weight of the corresponding factor is adjusted accordingly. Finally, a second dynamic weight set is generated.

[0090] Further, the first dynamic weight and the second dynamic weight are fused, where the fusion method can include weighted average method, confidence weighting method, fuzzy logic fusion, etc.; then, a weight verification mechanism is executed, including deviation checking with historical evaluation results and cross-validation with artificially annotated samples, if the deviation of any one exceeds a preset threshold, the weight fine-tuning mechanism is triggered, until neither exceeds the threshold, then it is considered that the adaptive weight distribution is completed, which is used for subsequent fusion precision degradation influence analysis.

[0091] Through the above process, the weight generation based on the history of itself and the history of similar individuals and adaptive fusion help to improve the orderliness and accuracy of weight generation, ensuring that the obtained weight conforms to the actual device state, and the verification of the weight further ensures the accuracy of the fused weight.

[0092] In some embodiments, further comprising:

[0093] The early warning threshold setting module is configured to obtain a task database of the electric energy meter, and establish a fixed precision threshold according to the task database; the real-time early warning module is configured to perform trigger authentication of the precision degradation evaluation result according to the fixed precision threshold, establish a trigger authentication result, and output a hierarchical early warning signal by using the trigger authentication result.

[0094] Specifically, to realize the rapid response and active intervention to the precision degradation state of the electric energy meter, the system further comprises a warning threshold setting module and a real-time early warning module, which are configured to set a fixed precision threshold based on the precision requirements of different types of electric energy meters in the task database, and trigger an authentication mechanism according to the precision degradation evaluation result during operation, finally realizing the output of a hierarchical early warning signal.

[0095] Specifically, the early warning threshold setting module is configured to obtain a task database of the electric energy meter, and the task database includes: the corresponding metering precision grade of different types of electric energy meters, the application scenarios (such as residential electricity, industrial metering, high-voltage metering, etc.) to which they belong, and the upper and lower limits of the precision required by the relevant standards. By traversing the task database with the basic information (such as model, purpose, installation location) of the electric energy meter as the index, the corresponding fixed precision threshold can be established.

[0096] Optionally, the fixed precision threshold can include: a first threshold value: mild degradation, prompting observation; a second threshold value: moderate degradation, prompting repair; and a third threshold value: severe degradation, prompting replacement.

[0097] Specifically, the real-time early warning module is configured to, after generating the precision degradation evaluation result, call the corresponding fixed precision threshold, execute the trigger authentication, including comparing the current evaluation result with the fixed threshold, judging whether the boundary is crossed, if the boundary is crossed, establishing the trigger authentication result according to the exceeding degree, and outputting the graded early warning signal. Exemplarily, the graded early warning signal includes but is not limited to: a text prompt (such as "the current precision degradation has reached a moderate level"), a graphical level indication (such as a red-yellow-green three-color indicator), starting a remote alarm mechanism or reporting to the master station system.

[0098] By introducing the task database driven fixed threshold setting and the real-time trigger authentication mechanism, the individualized precision early warning of different types of electric energy meters and the real-time monitoring and dynamic response to the precision degradation trend can be realized. Through the graded early warning mechanism, the operation and maintenance efficiency can be improved and the manual intervention cost can be reduced.

[0099] In some implementations, the real-time early warning module is further configured to:

[0100] perform precision degradation trend identification on the precision degradation evaluation result, generate a degradation rate index according to the precision degradation trend identification result, correct the trigger authentication result according to the degradation rate index, and output the graded early warning signal according to the corrected trigger authentication result.

[0101] Specifically, to further improve the forward-looking and response sensitivity of the precision degradation early warning, the real-time early warning module further includes a precision degradation trend identification and degradation rate correction mechanism, which is configured to identify the dynamic evolution trend of the precision degradation of the electric energy meter, and correct the trigger authentication result according to the degradation rate, so as to realize the graded early warning with higher timeliness and accuracy.

[0102] Specifically, first, a time series modeling is performed on the precision degradation evaluation result to extract continuous evaluation values in a period of time, and then a trend analysis method such as linear regression trend fitting, moving average change rate, exponential smoothing or Kalman filtering is used to identify the direction of the degradation trend (such as stable, slow degradation, rapid degradation), and output the trend identification result.

[0103] Further, a degradation rate index is calculated according to the trend identification result, such as the average degradation rate in a preset time window or the average degradation rate in multiple time windows weighted by a weighting method, and then the degradation rate index is combined with the trigger authentication result to reflect the risk degree of the imminent boundary crossing.

[0104] Exemplarily, the authentication level is corrected according to the degradation rate index, including: if the current evaluation value has not crossed the boundary but the degradation rate is higher than the threshold, triggering a higher level of early warning in advance; if the current evaluation value has crossed the boundary but the degradation rate is slowing down, the early warning level can be lowered or the response can be delayed; and the corrected trigger authentication result is output.

[0105] The above process introduces a degradation trend identification and rate correction mechanism, which can not only judge the current state, but also predict future risks, provide early warning when the trend is deteriorating without crossing the boundary, improve system safety, and avoid false positives or misjudgments caused by short-term fluctuations.

[0106] In some embodiments, further comprising:

[0107] The self-updating module is configured to receive the precision degradation evaluation result, synchronously acquire real measurement feedback, evaluate and verify the precision degradation evaluation result according to the real measurement feedback, establish a self-updating database, and use the self-updating database to optimize and update the degradation evaluation module.

[0108] Specifically, to improve the long-term accuracy and adaptability of the precision degradation evaluation model of the electric energy meter, the system further comprises a self-updating module configured to synchronously receive real measurement feedback from manual calibration, on-site re-measurement or standard instruments after obtaining the precision degradation evaluation result, use the feedback to verify the current evaluation result, establish a self-updating database, and continuously optimize and update the degradation evaluation module based on the database.

[0109] Specifically, the real measurement feedback includes measured error values from manual review or on-site calibration, comparison data from standard measurement instruments or precision deviations in periodic verification reports; wherein the feedback data can be actual error values or deviation values from the system evaluation result.

[0110] Specifically, the current precision degradation evaluation result is compared with the real feedback to calculate error offset, residual distribution, systematic deviation trend and other evaluation indexes to determine whether the current evaluation model has problems such as deviation, drift or overfitting. At the same time, each evaluation result, real feedback and error analysis result are packaged and stored in the self-updating database, wherein the data structure can include: time stamp; electric energy meter type and configuration; evaluation value, real value, deviation, environmental variable, etc. The database is used for subsequent optimization and update of the degradation evaluation module.

[0111] Specifically, the historical data in the self-updating database is used to periodically perform model retraining, parameter fine-tuning, feature weight update, abnormal pattern recognition rule correction and other operations on the degradation evaluation module to realize self-learning and adaptive update of the degradation evaluation module. The effect of continuously improving evaluation accuracy and enhancing system adaptability is achieved.

[0112] In summary, the electric energy meter precision degradation evaluation system based on sampling unit fault features provided by the present application has the following technical effects:

[0113] The sampling log of the target sampling unit is read through the reading module, and a collection data set is extracted, the collection data set including an input data set of current and voltage and a response data set of the target sampling unit; the identification module performs feature extraction on the collection data set, identifies a feature impact mode, the feature impact mode including a surge impact, a short-circuit impact and a high-frequency overload, and establishes a mapping relationship of response features based on the impact mode; the first prediction module takes the feature impact mode and the response features as mapping inputs, inputs a time sequence fault prediction channel, and establishes a first fault prediction result; the second prediction module establishes a test database based on the first fault prediction result, performs fault verification of the target sampling unit, and obtains a second fault prediction result; the degradation evaluation module establishes an electric energy meter working state data according to the collection data set, combines the second fault prediction result, adopts a multi-factor fusion method to perform precision degradation evaluation, and generates a precision degradation evaluation result, so that the technical effects of multi-dimensional prediction fusion, improved prediction accuracy and optimized prediction efficiency are realized.

[0114] In the second embodiment, as Figure 2 The structure diagram of the exemplary device provided by the present application is shown, which shows the block diagram of the exemplary device suitable for realizing the embodiments of the present application. Figure 2 The device shown is only an example, and should not bring any limitation to the function and use range of the embodiments of the present application. As Figure 2 As shown, the device includes a processor 31, a memory 32, an input device 33 and an output device 34; the number of processors 31 in the device can be one or more, Figure 2 In the foregoing, the processor 31 in the device, the memory 32, the input device 33 and the output device 34 can be connected through a bus or other means, Figure 2 In the foregoing, the connection through the bus is taken as an example.

[0115] It should be understood that the embodiments mentioned in the specification focus on their differences from other embodiments, and the specific embodiments in the foregoing embodiment one are also applicable to the electric energy meter precision degradation evaluation device based on the sampling unit fault features described in the second embodiment, and for the sake of brevity of the specification, no further expansion is made here.

[0116] It should be understood that the embodiments disclosed in the present application and the above description can enable those skilled in the art to implement the present application. Meanwhile, the present application is not limited to the part of the embodiments mentioned above, and it should be understood that those skilled in the art can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement to some technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application.

Claims

1. An electric energy meter precision degradation evaluation system based on sampling unit fault characteristics, characterized in that, The method comprises the following steps: reading the sampling log of the target sampling unit, extracting the collection data set, which comprises an input data set of current and voltage and a response data set of the target sampling unit; after feature extraction on the collection data set, identifying the feature impact mode, which comprises surge impact, short circuit impact and high frequency overload, and establishing the response feature of the impact mode mapping; the first prediction module uses the feature impact mode and the response feature as mapping input data to input into the time sequence fault prediction channel to establish the first fault prediction result; the second prediction module uses the first fault prediction result to establish a test database and performs fault verification on the target sampling unit to establish the second fault prediction result; the degradation evaluation module uses the collection data set to establish the working state data of the electric energy meter, performs multi-factor fusion precision degradation evaluation on the working state data and the second fault prediction result, and generates the precision degradation evaluation result; the time sequence fault prediction channel in the first prediction module comprises: an independent prediction layer for reading the feature extraction result of the feature impact mode, performing independent impact prediction on the corresponding feature impact mode according to the feature extraction result, establishing an independent impact prediction result, and the feature extraction result comprising voltage, current fluctuation frequency, fluctuation amplitude and duration; a time sequence compensation layer for receiving the independent impact prediction result and performing time sequence linkage compensation according to the time sequence relationship of the independent impact prediction result to establish a time sequence linkage compensation result; a hidden danger verification layer for reading the time sequence linkage compensation result and the response feature to perform real response verification and establishing the first fault prediction result according to the real response verification result.

2. The system for evaluating the precision degradation of an electric energy meter based on the fault features of sampling units according to claim 1, characterized in that, The second prediction module comprises: a global constraint unit for obtaining the basic unit information of the target sampling unit and configuring a globally covered basic test database according to the basic unit information; a local enhancement unit for generating a test focus according to the first fault prediction result, enhancing the database focus of the basic test database according to the test focus, and establishing the test database.

3. The system for evaluating the precision degradation of an electric energy meter based on the fault features of sampling units according to claim 1, characterized in that, The degradation evaluation module comprises: a state feature extraction unit for extracting the features of voltage fluctuation rate, current fluctuation rate, working temperature, power load and power consumption behavior mode from the working state data to establish a state feature set; a multi-factor fusion evaluation unit for using the state feature set and the second fault prediction result as multi-dimensional evaluation factors, performing fusion precision degradation impact analysis after adaptive weight allocation, and establishing the precision degradation evaluation result.

4. The electric energy meter precision degradation evaluation system based on sampling unit fault features of claim 3, wherein, After adaptive weight allocation, the fusion precision degradation impact analysis is performed to establish the precision degradation evaluation result, and the execution steps comprise: obtaining the historical data set of the electric energy meter, performing precision impact correlation analysis of the multi-dimensional evaluation factors according to the historical data set, establishing the first dynamic weight according to the precision impact correlation analysis result; using the basic information of the electric energy meter as the matching feature to perform similar matching of the experience database, performing weight adaptive analysis of the heuristic rules according to the similar matching result, and establishing the second dynamic weight; After the first dynamic weight and the second dynamic weight are verified, the adaptive allocation weight is completed.

5. The system for evaluating the precision degradation of an electric energy meter based on the fault features of sampling units according to claim 1, characterized in that, The reading module comprises: An abnormal data monitoring unit is configured to, after reading the sampling log, perform self-fluctuation authentication on the data flow of the sampling log, establish a self-fluctuation anomaly, call a simulation environment to perform simulation testing on the data flow, and establish an auxiliary fluctuation anomaly according to the simulation testing result; A correction unit is configured to perform anomaly identification correction according to the self-fluctuation anomaly and the auxiliary fluctuation anomaly, and extract a collection data set according to the anomaly identification correction result.

6. The system for evaluating the precision degradation of an electric energy meter based on the fault features of sampling units according to claim 1, characterized in that, Further comprising: A pre-warning threshold setting module is configured to obtain a task database of the electric energy meter, and establish a fixed precision threshold according to the task database; A real-time pre-warning module is configured to perform trigger authentication of the precision degradation evaluation result according to the fixed precision threshold, establish a trigger authentication result, and report a hierarchical pre-warning signal by using the trigger authentication result.

7. The sampling unit fault feature based electric energy meter precision degradation evaluation system according to claim 6, characterized in that, The real-time pre-warning module is further configured to: perform precision degradation trend identification on the precision degradation evaluation result; generate a degradation rate index according to the precision degradation trend identification result, and correct the trigger authentication result according to the degradation rate index; report the hierarchical pre-warning signal according to the corrected trigger authentication result.

8. The system for evaluating the precision degradation of an electric energy meter based on the fault features of sampling units according to claim 1, characterized in that, Further comprising: An automatic updating module is configured to, after receiving the precision degradation evaluation result, synchronously obtain real measurement feedback, perform evaluation verification on the precision degradation evaluation result according to the real measurement feedback, establish an automatic updating database, and perform optimization and updating of the degradation evaluation module by using the automatic updating database.

Images