Intelligent analysis system for operation failure of electric energy meter based on sensor monitoring

Abstract

This invention relates to the field of electricity meter operation monitoring technology, specifically to an intelligent analysis system for electricity meter operation faults based on sensor monitoring. The system includes: an operation status acquisition module, which collects the voltage, current, active power, reactive power, and power factor of the electricity meter and obtains the operation status at each moment; a periodic analysis module, which uses active power as a load parameter and obtains a basic period based on the load parameters within a preset time period, then segments the preset time period to obtain different time periods and obtains the period fit at each moment; and a fault monitoring module, which obtains the comprehensive rationality of the load linkage deviation at each moment; and uses the comprehensive rationality of the load linkage deviation at a moment and the period fit to correct the squared Mahalanobis distance corresponding to the data points composed of voltage, current, reactive power, and power factor at that moment, and uses the corrected squared Mahalanobis distance for fault monitoring. This application can effectively monitor the operation faults of electricity meters.

Description

Technical Field

[0001] This invention relates to the field of energy meter operation monitoring technology, specifically to an intelligent analysis system for energy meter operation faults based on sensor monitoring. Background Technology

[0002] With the development of smart grids and the increasing complexity of electricity consumption and load types, electricity meters have gradually transformed from traditional metering devices into intelligent devices integrating metering, monitoring, and communication. The requirements for their operational reliability and data accuracy are constantly increasing. However, the operating status of existing electricity meters largely relies on monitoring single electrical parameters or simple threshold judgments, making it difficult to promptly identify hidden and progressive faults caused by environmental changes, equipment aging, and complex load fluctuations, easily leading to misjudgments or missed diagnoses. Therefore, it is necessary to achieve real-time monitoring, fault diagnosis, and early warning of electricity meter operating status through multi-source sensor data acquisition and fusion analysis, combined with anomaly detection and intelligent identification algorithms, thereby improving the reliability and intelligence level of the electricity metering system.

[0003] Existing methods for analyzing electricity meter malfunctions typically employ multivariate Z-score anomaly detection based on Mahalanobis distance. This method maps multidimensional variables collected from the electricity meter onto a unified statistical space, detecting anomalies by measuring the distance between sample points and the center of normal distribution, thus identifying malfunctions in electricity meter operation. However, this method relies on the assumption of relatively stable statistical characteristics of the data, namely, that electricity meter operating data has an approximately constant mean and fluctuation range within a certain time range. In practical applications, however, electricity meter operating data is affected by user load behavior, diurnal cycles, and environmental factors, exhibiting significant non-stationary characteristics. Its mean and variance change dynamically over time, and the data distribution varies significantly across different time periods. Anomaly detection methods based on global statistical characteristics struggle to accurately characterize the dynamic range of normal operating conditions, easily misjudging normal load fluctuations as anomalies, leading to a high false alarm rate. Furthermore, for slow-changing anomalies caused by equipment aging or poor contact, their characteristics are often masked by periodic fluctuations, easily resulting in missed alarms. Summary of the Invention

[0004] To address the aforementioned technical problems, the present invention aims to provide an intelligent analysis system for the operation faults of electricity meters based on sensor monitoring. The specific technical solution adopted is as follows:

[0005] One embodiment of the present invention provides an intelligent analysis system for the operation faults of electricity meters based on sensor monitoring. The system includes:

[0006] The operation status acquisition module is used to collect various parameters of the energy meter, such as voltage, current, reactive power, and power factor; it uses a preset time length of one parameter for each time moment to form a pre-sequence of that parameter; and it obtains the operation status of that time moment based on the changes in the pre-sequence of various parameters.

[0007] The periodic analysis module is used to take the active power of the electricity meter as the load parameter; obtain the basic cycle based on the load parameter within the preset time period and divide the preset time period into segments to obtain different time periods; and obtain the cycle adaptability at a given moment by using the difference between the operating status at a given moment and the operating status at the corresponding moment in each time period.

[0008] The fault monitoring module is used to obtain the load linkage degree of a parameter at a given time based on the similarity between the preceding sequence of a parameter and the preceding sequence of the load parameter at that time; to obtain the allowable fluctuation range of the load linkage of a parameter at each time based on the load linkage degree of a parameter; to obtain the reasonableness of the load linkage deviation of a parameter at a given time based on the load linkage degree of a parameter and the allowable fluctuation range of the load linkage of a parameter at that time; to obtain the comprehensive reasonableness of the load linkage deviation at a given time using the load linkage degree of various parameters and the reasonableness of the load linkage deviation; to correct the squared Mahalanobis distance corresponding to the data points composed of voltage, current, reactive power, and power factor at that time using the comprehensive reasonableness of the load linkage deviation at a given time and the cycle adaptability; and to use the corrected squared Mahalanobis distance for fault monitoring.

[0009] Preferably, the operating state at a given moment is obtained based on the changes in the preceding sequence of various parameters, including:

[0010] The variance of the preceding sequence of a parameter at a given time moment is obtained to determine the fluctuation level of that parameter at that time moment. The absolute value of the difference between the fluctuation level of that parameter at that time moment and the mean of the fluctuation level of that parameter at each time moment within a preset time period is divided by the mean of the fluctuation level of that parameter at each time moment within the preset time period, and then multiplied by the fluctuation level of that parameter at that time moment to obtain the fluctuation characteristic value of that parameter at that time moment. The mean of the fluctuation characteristic values ​​of all parameters at that time moment is taken as the operating state at that time moment.

[0011] Preferably, the basic cycle is obtained based on the load parameters within a preset time period, and the preset time period is segmented to obtain different time periods, including:

[0012] Perform Fourier transform on the load parameters within the preset time period and extract the period corresponding to the main frequency as the base period; use the length of the base period to divide the preset time period into different time periods in chronological order.

[0013] Preferably, the periodicity of a given moment is obtained by comparing its operational state at a given moment with the operational states at corresponding moments in various time periods, including:

[0014] By using an exponential function with the natural constant as the base, a negative correlation is made between the absolute value of the difference between the operating state at a given moment and the operating state at the corresponding moment within a time period to obtain the similarity of the operating states for that time period; the periodicity of that moment is obtained by calculating the similarity of the operating states for each time period.

[0015] Preferably, the load linkage degree of a parameter at a given time is obtained based on the similarity between the preceding sequence of a parameter at a given time and the preceding sequence of the load parameter at that time, including:

[0016] The cosine similarity between the preceding sequence of a parameter at a given time and the preceding sequence of the load parameter at that time is used as the load linkage degree of that parameter at that time.

[0017] Preferably, the allowable fluctuation range of load linkage for a certain parameter is obtained based on the load linkage degree of that parameter at each time point, including:

[0018] The mean of the absolute values ​​of the differences in load linkage degree of a parameter between any two adjacent moments within a preset time period is obtained and recorded as the average linkage degree difference. The allowable fluctuation range of load linkage for that parameter is obtained by multiplying the difference between the maximum and minimum values ​​of the load linkage degree of a parameter at each moment within the preset time period with the average linkage degree difference.

[0019] Preferably, the reasonableness of the load linkage deviation of a parameter at a given time is obtained based on the load linkage degree of a parameter at a given time and the allowable fluctuation range of the load linkage of that parameter, including:

[0020] The difference between the load linkage degree of a parameter at a certain moment and the mean value of the load linkage degree of the same parameter at all moments within a preset time period is divided by the allowable fluctuation range of the load linkage of that parameter and normalized to obtain a normalized value. The normalized value is then negatively correlated with an exponential function with the natural constant as the base to obtain the rationality of the load linkage deviation of that parameter at that moment.

[0021] Preferably, the comprehensive rationality of the load linkage deviation at a given moment is obtained by utilizing the load linkage degree and load linkage deviation rationality of various parameters at that moment, including:

[0022] The load linkage degree of each parameter at a given time is normalized to obtain the weight of the reasonableness of the load linkage deviation of each parameter at that time. The reasonableness of the load linkage deviation of each parameter at that time is obtained by weighting and averaging the reasonableness of the load linkage deviation of each parameter at that time.

[0023] Preferably, the squared Mahalanobis distance corresponding to the data points composed of voltage, current, reactive power, and power factor at a given moment is corrected using the comprehensive rationality of the load linkage deviation at a given moment and the cycle fit, including:

[0024] The correction factor for a given moment is obtained by multiplying the load linkage deviation at a given moment by the comprehensive rationality and cycle adaptability. The voltage, current, reactive power, and power factor at that moment are used to form the data point for that moment. The squared Mahalanobis distance corresponding to the data point at that moment is multiplied by the difference between the first preset value and the correction factor for that moment to obtain the squared Mahalanobis distance of the data point at that moment after correction.

[0025] The embodiments of the present invention have at least the following beneficial effects: This application collects various parameters such as voltage, current, active power, reactive power, and power factor of the electricity meter, and then obtains the operating status at each moment; then, the active power is used as the load parameter, and a basic cycle is obtained according to the load parameters within a preset time period. The preset time period is then segmented using the basic cycle to obtain different time periods; the cycle adaptability at a given moment is obtained by comparing the operating status at a given moment with the operating status at the corresponding moment in each time period; furthermore, by analyzing the linkage relationship between parameters and load fluctuations under different operating states, the load linkage degree of each parameter at each moment, the allowable fluctuation range of load linkage for each parameter, and the load linkage deviation of each parameter at each moment are obtained. The system comprehensively assesses the rationality of load linkage deviation at each moment, effectively separating random disturbances from real anomalies. It utilizes multi-parameter coupling characteristics to identify difficult-to-detect hidden and progressive faults, improving the sensitivity and discrimination capability of anomaly feature extraction. Finally, it uses the comprehensive rationality of load linkage deviation at a given moment and cycle adaptability to correct the squared Mahalanobis distance corresponding to the data points composed of voltage, current, reactive power, and power factor at that moment. The corrected squared Mahalanobis distance is then used for fault monitoring, and dynamic correction of the squared Mahalanobis distance effectively reduces false alarms and missed alarms under non-stationary data conditions. This enables accurate identification and intelligent analysis of electricity meter operation faults, significantly improving the system's reliability and engineering application value under complex operating conditions. Attached Figure Description

[0026] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0027] Figure 1 This is a system block diagram of an intelligent analysis system for power meter operation faults based on sensor monitoring, provided as an embodiment of the present invention. Detailed Implementation

[0028] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a sensor-based intelligent analysis system for electricity meter operation faults proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0029] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0030] The following description, in conjunction with the accompanying drawings, details a specific scheme for an intelligent analysis system for energy meter operation faults based on sensor monitoring, provided by the present invention.

[0031] In this embodiment, the main application scenario of the present invention is as follows: the operating data of electricity meters has obvious non-stationary characteristics, the boundary between normal operation and abnormal state is unclear, and false alarms and missed alarms are prone to occur, which increases the difficulty of identifying faults in electricity meter operation. This application solves the problem of difficulty in accurately identifying abnormalities of electricity meters under non-stationary load conditions by analyzing multiple parameters and load linkage and introducing an adaptive correction mechanism, and realizes reliable detection of hidden and progressive faults.

[0032] Please see Figure 1 The diagram illustrates a system block diagram of an intelligent analysis system for energy meter operation faults based on sensor monitoring, provided by an embodiment of the present invention. The system includes the following modules:

[0033] The operating status acquisition module is used to collect various parameters of the energy meter, such as voltage, current, reactive power, and power factor; it uses a preset time length of one parameter for each time moment to form a pre-sequence of that parameter; and it obtains the operating status of that time moment based on the changes in the pre-sequence of various parameters.

[0034] During the operation of an electricity meter, various key parameters are collected in real time through built-in or external sensors, including the meter's voltage, current, active power, reactive power, and power factor. These parameters are continuously recorded according to a unified time base, forming multi-dimensional time series data. Voltage reflects power supply quality and line stability; current characterizes load size and changes; reactive power describes the energy exchange characteristics in the power grid and can reflect changes in inductive or capacitive loads and power quality issues; the power factor measures the phase relationship between voltage and current and can be used to identify wiring abnormalities or changes in load characteristics. Therefore, these parameters are all relatively important.

[0035] Different sensors have different sampling frequencies. Time alignment and resampling are performed on the collected parameters to unify all parameters to the same time scale. Simultaneously, the raw data is preprocessed, including outlier removal, missing value imputation, and noise filtering (such as moving average or low-pass filtering) to reduce the impact of random interference on the analysis results.

[0036] To eliminate the influence of different types of parameter dimensions, various parameters are normalized or standardized so that multidimensional data can be directly analyzed and compared on a unified scale.

[0037] The fluctuations in electrical parameters of electricity meters vary significantly under different operating conditions. Under high loads or heavy operation, the normal fluctuation range of parameters increases significantly, making them easily misjudged as abnormalities. Therefore, it is necessary to comprehensively analyze the current operating status based on the changing characteristics of multiple parameters of the electricity meter. By analyzing the dynamic changes in multi-dimensional parameters such as voltage, current, and power, the real-time operating status of the electricity meter can be identified, thus providing state constraints for subsequent anomaly detection and avoiding misjudging normal fluctuations under high load conditions as faults.

[0038] The core function of an electricity meter is to accurately measure electricity consumption, while actual electricity consumption exhibits significant peak-valley characteristics. Affected by load changes, the fluctuations in various parameters (such as voltage, current, and power) vary significantly across different time periods. When multiple parameters simultaneously exhibit large fluctuations within the same time period, and the overall level is significantly higher than its historical average operating level, it can be determined that the current period is a peak electricity consumption period, at which time the electricity meter is operating more actively.

[0039] For a parameter at a given time, it is necessary to analyze its changes in conjunction with the parameters in its neighborhood. The preceding sequence of the parameter at that time is composed of the parameters of that type with a preset time length including the parameter at that time. That is, the preceding sequence of the parameter at that time is composed of the parameters of that type with a preset time length up to the time at that time. In this application, the time length is represented by the number of data. Preferably, the preset time length in this application is 10, that is, 10 parameters. For example, for a parameter at time t, the preceding sequence of the parameter at time t is composed of the parameters of that type from t-9 to time t.

[0040] Furthermore, the operational state at a given moment is obtained based on the changes in the preceding sequences of various parameters. Specifically, the variance of the preceding sequence of a parameter at a given moment is obtained to determine the fluctuation level of that parameter at that moment; the absolute value of the difference between the fluctuation level of that parameter at that moment and the mean of the fluctuation levels of that parameter at all moments within a preset time period is divided by the mean of the fluctuation levels of that parameter at all moments within the preset time period, and then multiplied by the fluctuation level of that parameter at that moment to obtain the fluctuation characteristic value of that parameter at that moment; the mean of the fluctuation characteristic values ​​of various parameters at that moment is taken as the operational state at that moment.

[0041] The specific calculation model for the running state at a given moment is as follows:

[0042] ,

[0043] in, The operating state at time j represents the operating state of the energy meter at time j; N represents the number of parameter types. It represents the variance of the preceding sequence of the i-th parameter at time j, which is also the degree of fluctuation of the parameter at that time. This represents the average fluctuation of the parameter at each moment within a preset time period. Preferably, the preset time period is one week prior to the j-th moment (the implementer can adjust the preset time period according to the actual situation, and the preset time period includes the j-th moment). This represents the difference in fluctuation between the i-th parameter at time j and the historical average. The larger this expression is, the more active the energy meter's operating status is as reflected by this parameter at time j. Let be the fluctuation characteristic value of the i-th parameter at time j. From this, the operating state at each time point can be obtained.

[0044] The periodic analysis module is used to take the active power of the electricity meter as the load parameter; obtain the basic cycle based on the load parameter within a preset time period and divide the preset time period into segments to obtain different time periods; and obtain the cycle adaptability at a given moment by using the difference between the operating status at a given moment and the operating status at the corresponding moment in each time period.

[0045] When detecting anomalies in electricity meter operation, abnormal signals are typically extracted based on the correlation constraints between various meter parameters and the load. However, in actual operation, the coupling relationship between parameters and load changes dynamically within a certain range under different operating conditions. For example, during peak electricity consumption periods, the load level increases significantly, and parameters such as current and power increase and fluctuate more. Although there is a certain deviation in the relationship between parameters and load at this time, it is still within a reasonable range and does not represent an anomaly. Therefore, if correlation constraints obtained from global statistical laws are used, it is easy to misjudge such normal fluctuations as anomalies. Therefore, by analyzing the linkage between electricity meter parameters and load fluctuations, and establishing a reasonable fluctuation range in combination with differences in operating conditions, it is possible to effectively distinguish between normal and abnormal states, thereby improving the accuracy of anomaly identification of electricity meters under complex and non-stationary operating conditions.

[0046] The load condition needs to be represented using active power, which requires collecting the active power of the electricity meter and using the active power as a load parameter, as active power is a highly representative parameter. Then, the basic cycle is obtained based on the load parameters within a preset time period, and the preset time period is segmented to obtain different time periods.

[0047] Specifically, a Fourier transform is performed on the load parameters within a preset time period to extract the period corresponding to the dominant frequency as the base period. The preset time period is then continuously divided into different time segments according to the length of the base period in chronological order. Here, the base period obtained from the load parameters serves as the basis for segmenting the time corresponding to the multidimensional parameters, enabling effective analysis of the correlation between various parameters and load changes in subsequent analyses.

[0048] For each time period within a preset time frame, and the time period corresponding to a given moment, the operating status of the electricity meter is analyzed for similarity to the operating status at the same time position in historical time periods. If the operating status at that moment shows a high similarity to the states of each time period in the multi-dimensional feature space, it indicates that the operating status at that moment has good consistency with the historical time period pattern, i.e., a high periodicity, suggesting that its operating behavior conforms to existing periodic patterns. Therefore, the periodicity of a given moment is obtained by using the difference between the operating status at a given moment and the operating status at the corresponding moments in each time frame.

[0049] Specifically, the similarity of the operating states at a given moment is obtained by negatively mapping the absolute value of the difference between the operating state at a given moment and the operating state at the corresponding moment within a time period using an exponential function with the natural constant as the base. The periodicity of the operating states at each time period is then calculated to obtain the similarity of the operating states at that moment.

[0050] The specific calculation model for periodicity fitness is as follows:

[0051] ,

[0052] in, The periodicity of the running state at time j is represented, which is also the periodicity of time j; M represents the number of other time periods besides the time period containing time j within the preset time period; This indicates the operating status of the electricity meter at time j; This represents the operating status of the electricity meter at the corresponding time point j within the m-th time period. Here, "corresponding" means that the two times are in the same position within their respective time periods, for example, both being the second time point within their respective time segments. Exp represents an exponential function with the natural constant as its base. Let be the similarity of the operating states corresponding to the m-th time period. From this, the periodicity of each time period can be obtained.

[0053] The fault monitoring module is used to obtain the load linkage degree of a parameter at a given time based on the similarity between the preceding sequence of a parameter and the preceding sequence of the load parameter at that time; to obtain the allowable fluctuation range of the load linkage of a parameter at each time based on the load linkage degree of a parameter; to obtain the reasonableness of the load linkage deviation of a parameter at a given time based on the load linkage degree of a parameter and the allowable fluctuation range of the load linkage of a parameter at that time; to obtain the comprehensive reasonableness of the load linkage deviation at a given time using the load linkage degree of various parameters and the reasonableness of the load linkage deviation; to correct the squared Mahalanobis distance corresponding to the data points composed of voltage, current, reactive power, and power factor at that time using the comprehensive reasonableness of the load linkage deviation at a given time and the cycle adaptability; and to use the corrected squared Mahalanobis distance for fault monitoring.

[0054] The core function of an electricity meter is to measure users' electricity consumption, and fluctuations in users' electricity consumption directly affect changes in electrical parameters such as voltage, current, and power. Therefore, it is necessary to analyze the dynamic fluctuation relationship between various parameters of the electricity meter and load parameters, where the load is represented by active power. By comparing the changing trends of each parameter and active power over time, a high degree of consistency between their curves indicates a strong correlation between the parameter and the load.

[0055] The load linkage degree of a parameter at a given time is obtained by comparing its preceding sequence with the preceding sequence of the load parameter at that time. Specifically, the cosine similarity between the preceding sequence of a parameter at a given time and the preceding sequence of the load parameter at that time is taken as the load linkage degree of that parameter at that time. The load linkage degree is denoted as... , representing the load linkage degree of the i-th parameter at time j. This allows us to obtain the load linkage degree of each parameter at each time point. Load linkage degree refers to the linkage between various parameters and load parameters.

[0056] Under different operating conditions, the fluctuation intensity of various parameters of the electricity meter and the load varies significantly. Especially during peak electricity consumption periods, both parameters and load exhibit strong fluctuations. At this time, the linkage between parameters and load may deviate to some extent, but this deviation is within a reasonable range. Therefore, it is necessary to analyze the allowable range of the linkage deviation. The load linkage degree of each parameter at each moment was calculated above, and its variation characteristics over historical time series were analyzed. When the maximum and minimum values ​​of the linkage degree between a certain parameter and the load differ significantly in history, and the average level of the load linkage degree difference between adjacent moments is high, it indicates that the linkage between this parameter and the load has strong volatility, and its allowable deviation range is relatively wide.

[0057] The allowable fluctuation range of load linkage for a parameter is obtained based on the load linkage degree of a parameter at each time point. Specifically, the average absolute value of the difference between the load linkage degrees of a parameter at every two adjacent times within a preset time period is obtained and recorded as the average linkage degree difference; the difference between the maximum and minimum values ​​of the load linkage degrees of a parameter at each time point within the preset time period is multiplied by the average linkage degree difference to obtain the allowable fluctuation range of load linkage for that parameter.

[0058] The specific calculation model for the allowable fluctuation range of load linkage is as follows:

[0059] ,

[0060] in, Let be the allowable fluctuation range of load linkage for the i-th parameter, and represent the allowable fluctuation range of the deviation between the linkage relationship between the i-th parameter and the load. and These represent the maximum and minimum values ​​of the load linkage degree of the i-th parameter at each time point within the preset time period, respectively. The difference between the two values ​​can represent the span of the load linkage degree. It is the average of the absolute values ​​of the differences in load linkage degree between the i-th parameter at any two adjacent times within a preset time period, which is also the average linkage degree difference.

[0061] The load linkage degree calculated at a given moment is compared with the historical average level to obtain the relationship deviation, which is then matched with the corresponding allowable fluctuation range. The smaller the deviation is within the allowable range, the more reasonable the deviation is, indicating that the change in the linkage relationship under the operating state at that moment is still within the historically acceptable range. Conversely, if the deviation significantly exceeds or deviates from the allowable range, it indicates that the linkage relationship may be subject to abnormal disturbances, posing a potential risk of operational failure.

[0062] Therefore, the rationality of the load linkage deviation of a parameter at a given moment can be obtained based on the load linkage degree of a parameter at a given moment and the allowable fluctuation range of the load linkage of that parameter.

[0063] Specifically, the difference between the load linkage degree of a parameter at a certain moment and the mean value of the load linkage degree of the same parameter at various moments within a preset time period is divided by the allowable fluctuation range of the load linkage of that parameter and normalized to obtain a normalized value. The normalized value is then negatively correlated with an exponential function with the natural constant as the base to obtain the rationality of the load linkage deviation of that parameter at that moment.

[0064] The specific calculation model for the rationality of load linkage deviation is as follows:

[0065] ,

[0066] in, The reasonableness of the load linkage deviation of the i-th parameter at time j represents the reasonableness of the linkage relationship deviation between the i-th parameter and the load at time j. This represents the load linkage degree of the i-th parameter at time j; This represents the average load linkage degree of the i-th parameter at each time point within a preset time period; This represents the allowable fluctuation range of the load linkage for the i-th parameter. Here, is the normalized value, and norm represents the normalization function. It represents the ratio of the deviation of the load linkage degree of the i-th parameter at time j from the historical average level to the allowable fluctuation range of the load linkage of the i-th parameter. This reflects the degree of matching between the relationship deviation and the allowable fluctuation range. The smaller this formula is, the greater the reasonableness of the relationship deviation, indicating that it is within the allowable fluctuation range. Therefore, the reasonableness of the load linkage deviation for each parameter at each time point can be obtained.

[0067] Because different types of parameters in an electricity meter have varying degrees of coupling with the load, parameters more closely related to load fluctuations are more sensitive to changes in operating status, and changes in their correlation have a higher weight in fault diagnosis. Therefore, multi-parameter fusion is necessary to perform a weighted analysis of the correlation between each parameter and the load. When the deviation of the correlation between key parameters strongly correlated with the load remains within a reasonable range, i.e., the deviation is relatively reasonable, the current operating status can be determined to be relatively normal.

[0068] Therefore, the overall rationality of the load linkage deviation at a given moment can be obtained by using the load linkage degree and load linkage deviation rationality of various parameters at that moment.

[0069] Specifically, the load linkage degree of various parameters at a given time is normalized to obtain the weight of the reasonableness of the load linkage deviation of various parameters at that time; the reasonableness of the load linkage deviation of various parameters at that time is obtained by weighting and averaging the reasonableness of the load linkage deviation of various parameters at that time using the weight of the reasonableness of the load linkage deviation of various parameters at that time.

[0070] The specific calculation model for the comprehensive rationality of load linkage deviation is as follows:

[0071] ,

[0072] in, The comprehensive rationality of the load linkage deviation at time j represents the comprehensive real-time rationality of the linkage relationship deviation between multiple parameters of the energy meter and the load at time j; N represents the number of types of parameters of the energy meter. The table represents the load linkage degree of the i-th parameter at time j. After normalization, the weights of the reasonableness of the load linkage deviation of various parameters at that time are obtained. The load linkage deviation of the i-th parameter at time j is considered reasonable; norm represents the normalization function.

[0073] Because there are certain dynamic deviations in the relationship between the parameters of the electricity meter and the load under different operating states, if the cycle adaptability of the current operating state is high, it means that the current operating behavior conforms to the historical normal electricity consumption cycle pattern and is close to the overall trend and average characteristics. Local abnormal fluctuations of key parameters may be masked under the macro cycle. At the same time, the operating state of the electricity meter involves the coupling relationship between multiple parameters. Under this condition, if the deviation of the relationship between the parameters of the electricity meter and the load is still within an acceptable range, that is, the real-time rationality is high, it can be determined that the current operating state is relatively normal, and the correction factor is large.

[0074] Therefore, by using the load linkage deviation at a certain moment to comprehensively consider the rationality and cycle fit, the squared Mahalanobis distance corresponding to the data points composed of voltage, current, reactive power and power factor at that moment is corrected.

[0075] Specifically, the correction factor for a given moment is obtained by multiplying the load linkage deviation at a given moment by the comprehensive rationality and cycle adaptability. The voltage, current, reactive power, and power factor at that moment are used to form the data point for that moment. The squared Mahalanobis distance corresponding to the data point at that moment is multiplied by the difference between the first preset value and the correction factor for that moment to obtain the squared Mahalanobis distance of the data point at that moment after correction.

[0076] The revised calculation model for the squared Mahalanobis distance is as follows:

[0077] ,

[0078] in, The squared Mahalanobis distance corresponding to the data point consisting of voltage, current, reactive power, and power factor at time j is represented. This represents the correction factor at time j; the first preset value is 1. This represents the squared Mahalanobis distance of the data point at time j after correction. The squared Mahalanobis distance represents the distance from the data point composed of various parameters at time j to the center of the normal data distribution. The acquisition of this distance is a well-known technique and will not be elaborated upon here. In multivariate energy meter operating data, the center of the normal data distribution is used to characterize the typical feature location of the equipment under healthy operating conditions. It is represented by the statistical mean vector of historical normal operating data in a multidimensional feature space. This distribution center comprehensively reflects the joint variation law of multiple parameters such as voltage, current, and reactive power under normal operating conditions, representing the "baseline operating state" of the system under fault-free conditions.

[0079] Under different operating conditions of electricity meters, multivariate Z-score anomaly detection methods based on Mahalanobis distance typically rely on correlation constraints between parameters to extract abnormal signals. However, since the relationship between electricity meter parameters and load changes dynamically with operating conditions, normal fluctuations may be misjudged as anomalies. Therefore, by analyzing the multi-parameter linkage relationship between the electricity meter and the load, a reasonable fluctuation range of linkage deviation under different operating conditions is established. The anomaly detection results are then dynamically corrected based on the current operating state of the electricity meter, thereby reducing false alarms and missed alarms and improving the accuracy and reliability of electricity meter fault identification. To avoid temporary disruption of the linkage relationship between parameters and load due to short-term drastic fluctuations during peak electricity consumption periods, a correction factor is introduced by analyzing the current operating state in real time to dynamically correct the anomaly detection results. When the linkage deviation between the electricity meter parameters and the load under the current operating state remains within a reasonable range, the correction factor will correspondingly reduce the abnormal index value, thus preventing normal fluctuations from being misjudged as faults.

[0080] The multivariate Z-score anomaly detection method based on Mahalanobis distance is a statistical approach for standardizing and identifying anomalies in electricity meter operating data, taking into account the correlation of multiple parameters. This method first calculates the mean vector and covariance matrix based on historical normal data, mapping the multidimensional observation data into a unified statistical space. Mahalanobis distance is then used to measure the deviation of the current data point from the center of the normal distribution, thus quantifying the overall anomaly degree of the multivariate data. Compared to the traditional univariate Z-score method, this method considers the independent deviation of each parameter and introduces a covariance structure to reflect the correlation between parameters, enabling more accurate anomaly detection under multidimensional coupling conditions. When the Mahalanobis distance exceeds a statistical threshold determined based on the chi-square distribution, the current data point is determined to deviate from the normal distribution, thus identifying an anomaly.

[0081] Finally, by analyzing the linkage between various parameters and load during the operation of the electricity meter, and combining the reasonable deviation range under different operating conditions, the anomaly detection results are dynamically corrected. Based on the multivariate Z-score anomaly detection algorithm based on Mahalanobis distance, a correction factor for the current operating state is introduced to obtain the real-time corrected distance (the squared Mahalanobis distance of a data point at a given moment), which characterizes the degree of deviation of the data point at that moment from the center of the normal distribution in the multidimensional feature space. When the corrected distance at a given moment is very large, exceeding the statistically permissible range (measured by the chi-square distribution threshold), it indicates that the data point deviates too much from the normal distribution, thus determining an anomaly and identifying an electricity meter malfunction.

[0082] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0083] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0084] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A smart analysis system for power meter operation faults based on sensor monitoring, characterized in that, The system includes: The operation status acquisition module is used to collect various parameters of the energy meter, such as voltage, current, reactive power, and power factor; it uses a preset time length of one parameter for each time moment to form a pre-sequence of that parameter; and it obtains the operation status of that time moment based on the changes in the pre-sequence of various parameters. The periodic analysis module is used to take the active power of the electricity meter as the load parameter; obtain the basic cycle based on the load parameter within the preset time period and divide the preset time period into segments to obtain different time periods; and obtain the cycle adaptability at a given moment by using the difference between the operating status at a given moment and the operating status at the corresponding moment in each time period. The fault monitoring module is used to obtain the load linkage degree of a parameter at a given time based on the similarity between the preceding sequence of a parameter and the preceding sequence of the load parameter at that time; to obtain the allowable fluctuation range of the load linkage of a parameter at each time based on the load linkage degree of a parameter; to obtain the reasonableness of the load linkage deviation of a parameter at a given time based on the load linkage degree of a parameter and the allowable fluctuation range of the load linkage of a parameter at that time; to obtain the comprehensive reasonableness of the load linkage deviation at a given time using the load linkage degree of various parameters and the reasonableness of the load linkage deviation; to correct the squared Mahalanobis distance corresponding to the data points composed of voltage, current, reactive power, and power factor at that time using the comprehensive reasonableness of the load linkage deviation at a given time and the cycle adaptability; and to use the corrected squared Mahalanobis distance for fault monitoring.

2. The intelligent analysis system for power meter operation faults based on sensor monitoring according to claim 1, characterized in that, The process of obtaining the running state at a given moment based on the changes in the preceding sequence of various parameters includes: The variance of the preceding sequence of a parameter at a given time moment is obtained to determine the fluctuation level of that parameter at that time moment. The absolute value of the difference between the fluctuation level of that parameter at that time moment and the mean of the fluctuation level of that parameter at each time moment within a preset time period is divided by the mean of the fluctuation level of that parameter at each time moment within the preset time period, and then multiplied by the fluctuation level of that parameter at that time moment to obtain the fluctuation characteristic value of that parameter at that time moment. The mean of the fluctuation characteristic values ​​of all parameters at that time moment is taken as the operating state at that time moment.

3. The intelligent analysis system for power meter operation faults based on sensor monitoring according to claim 1, characterized in that, The step of obtaining the basic cycle based on load parameters within a preset time period and segmenting the preset time period to obtain different time periods includes: Perform Fourier transform on the load parameters within the preset time period and extract the period corresponding to the main frequency as the base period; use the length of the base period to divide the preset time period into different time periods in chronological order.

4. The intelligent analysis system for power meter operation faults based on sensor monitoring according to claim 1, characterized in that, The method of obtaining the periodicity adaptation at a given moment by comparing the operating state at a given moment with the operating states at corresponding moments in each time period includes: By using an exponential function with the natural constant as the base, a negative correlation is made between the absolute value of the difference between the operating state at a given moment and the operating state at the corresponding moment within a time period to obtain the similarity of the operating states for that time period; the periodicity of that moment is obtained by calculating the similarity of the operating states for each time period.

5. The intelligent analysis system for power meter operation faults based on sensor monitoring according to claim 1, characterized in that, The step of obtaining the load linkage degree of a parameter at a given time based on the similarity between the preceding sequence of a parameter at a given time and the preceding sequence of the load parameter at that time includes: The cosine similarity between the preceding sequence of a parameter at a given time and the preceding sequence of the load parameter at that time is used as the load linkage degree of that parameter at that time.

6. The intelligent analysis system for power meter operation faults based on sensor monitoring according to claim 1, characterized in that, The method of obtaining the allowable fluctuation range of load linkage for a parameter based on the load linkage degree at each time point includes: The mean of the absolute values ​​of the differences in load linkage degree of a parameter between any two adjacent moments within a preset time period is obtained and recorded as the average linkage degree difference. The allowable fluctuation range of load linkage for that parameter is obtained by multiplying the difference between the maximum and minimum values ​​of the load linkage degree of a parameter at each moment within the preset time period with the average linkage degree difference.

7. The intelligent analysis system for power meter operation faults based on sensor monitoring according to claim 1, characterized in that, The method of obtaining the rationality of the load linkage deviation of a parameter at a given time based on the load linkage degree of a parameter at a given time and the allowable fluctuation range of the load linkage of that parameter includes: The difference between the load linkage degree of a parameter at a certain moment and the mean value of the load linkage degree of the same parameter at all moments within a preset time period is divided by the allowable fluctuation range of the load linkage of that parameter and normalized to obtain a normalized value. The normalized value is then negatively correlated with an exponential function with the natural constant as the base to obtain the rationality of the load linkage deviation of that parameter at that moment.

8. The intelligent analysis system for power meter operation faults based on sensor monitoring according to claim 1, characterized in that, The method of obtaining the comprehensive rationality of the load linkage deviation at a given moment by utilizing the load linkage degree and load linkage deviation rationality of various parameters at that moment includes: The load linkage degree of each parameter at a given time is normalized to obtain the weight of the reasonableness of the load linkage deviation of each parameter at that time. The reasonableness of the load linkage deviation of each parameter at that time is obtained by weighting and averaging the reasonableness of the load linkage deviation of each parameter at that time.

9. The intelligent analysis system for power meter operation faults based on sensor monitoring according to claim 1, characterized in that, The method of using the load linkage deviation at a certain moment to comprehensively consider the rationality and cycle adaptability to correct the squared Mahalanobis distance corresponding to the data points composed of voltage, current, reactive power, and power factor at that moment includes: The correction factor for a given moment is obtained by multiplying the load linkage deviation at a given moment by the comprehensive rationality and cycle adaptability. The voltage, current, reactive power, and power factor at that moment are used to form the data point for that moment. The squared Mahalanobis distance corresponding to the data point at that moment is multiplied by the difference between the first preset value and the correction factor for that moment to obtain the squared Mahalanobis distance of the data point at that moment after correction.

Images