Harmonic analysis-based acquisition terminal line life prediction method and system

Abstract

This invention relates to the field of power grid monitoring, and more specifically, to a method and system for predicting the lifespan of data acquisition terminal lines based on harmonic analysis. The method includes: first, adaptively classifying and identifying different historical power consumption patterns, and calculating the accelerating temperature coefficient corresponding to the harmonic current under each pattern; then, calculating the existing lifespan loss based on historical temperature rise, and predicting future current; next, using the accelerating temperature coefficient and combining it with the matching degree between future current and each historical power consumption pattern, weighted calculation of future lifespan loss; finally, combining the initial lifespan, historical loss, and future loss to obtain a high-precision predicted remaining lifespan. This invention overcomes the problem of large prediction deviations in complex harmonic environments using traditional methods, significantly improving the accuracy and reliability of remaining lifespan assessment, providing accurate and timely decision-making basis for preventative maintenance of meter lines, and effectively preventing safety accidents caused by insulation aging.

Description

Technical Field

[0001] This invention relates to the field of power grid monitoring. More specifically, this invention relates to a method and system for predicting the lifespan of acquisition terminals based on harmonic analysis. Background Technology

[0002] With the rapid popularization of nonlinear loads such as electric vehicle charging piles, photovoltaic inverters, and various frequency conversion equipment, harmonic pollution in the user-side power grid is becoming increasingly prominent. In particular, the presence of high-order harmonics can easily lead to significant eddy current losses and skin effect in current transformers, causing local overheating in key parts such as metering circuits and terminals, thereby accelerating the aging of insulation materials and seriously affecting the actual service life of electricity metering devices.

[0003] Currently, when assessing the thermal life of equipment under the influence of harmonic environments, conventional methods typically treat the additional losses generated by harmonics of different frequencies as equivalent or superimposed, and predict temperature rise and lifespan accordingly. However, this method has limitations in its applicability under actual complex harmonic operating conditions, and its assessment results deviate significantly from the actual aging state of the equipment. It is difficult to accurately reflect the remaining lifespan of metering devices in increasingly severe harmonic environments, which not only affects metering accuracy but also poses a hidden danger to the long-term safe and stable operation of the user-side power grid. Summary of the Invention

[0004] To address the technical problem of insufficient accuracy of existing evaluation methods in complex harmonic environments, the present invention provides solutions in the following aspects.

[0005] In the first aspect, the method for predicting the lifespan of a data acquisition terminal line based on harmonic analysis includes: Obtain historical current timing, historical temperature rise timing, and initial life parameters of the meter circuit; The historical current time series is divided into multiple current sequences. For each current sequence, time series features and frequency domain features are extracted, and the comprehensive difference distance between any two current sequences is calculated based on the time series features and frequency domain features. All current sequences are clustered based on the comprehensive difference distance to obtain all historical power consumption pattern clusters. For each historical power consumption pattern cluster, the acceleration temperature coefficient of the current at each frequency under the corresponding historical power consumption pattern cluster is determined. Historical lifetime loss is calculated based on historical temperature rise time series and Arrenius model; future current time series is predicted based on historical current time series. Based on the frequency domain characteristics of the future current time series, the acceleration temperature coefficient corresponding to each historical power consumption pattern cluster, and the comprehensive difference distance between the future current time series and the center of each historical power consumption pattern cluster, the future lifetime loss corresponding to each historical power consumption pattern cluster is calculated. By combining initial life parameters, historical life loss, and future life loss, the predicted remaining life of the meter circuit is calculated and output.

[0006] Preferably, when the predicted remaining lifespan of the output meter line is less than or equal to a preset lifespan threshold, an early warning signal is generated and maintenance or replacement is arranged.

[0007] Preferably, the historical temperature rise time series is the difference time series obtained by collecting the internal temperature time series of the meter and the ambient temperature time series, and subtracting the corresponding elements of the internal temperature time series of the meter and the ambient temperature time series.

[0008] Preferably, obtaining the comprehensive difference distance includes: Calculate the sum of the absolute values ​​of the differences between the frequency amplitudes in the spectrum graphs corresponding to the two current sequences, and use this as the first distance between the two current sequences; The sum of the dynamic time-warped distances between the components of two current sequences after empirical mode decomposition is calculated as the second distance between the two current sequences. The sum of the first distance and the second distance is taken as the comprehensive difference distance between the two current sequences.

[0009] Preferably, the step of dividing the historical current time series into multiple current sequences includes obtaining current sequences under different division schemes by setting multiple unequal length division schemes; After clustering current sequences of various lengths, the effective period score of each clustering result is calculated. The current sequence with the largest effective period score is selected as the final current sequence to participate in the clustering, and then the final historical electricity consumption pattern cluster is obtained. Specifically, for each historical electricity consumption pattern cluster under a clustering result, the average comprehensive difference distance between each current sequence within the cluster and all current sequences outside the cluster is calculated as the inter-cluster distance; the average comprehensive difference distance between each current sequence within the cluster and other current sequences within its own cluster is calculated as the intra-cluster distance; and the effective cycle score is calculated based on the inter-cluster distance and intra-cluster distance of all current sequences within the cluster of all historical electricity consumption pattern clusters.

[0010] Preferably, obtaining the accelerating temperature coefficient includes: The historical temperature rise time series is divided into multiple temperature rise sequences according to the historical current time series division pattern. The RMS of each temperature rise sequence is calculated, and the RMS of all temperature rise sequences are constructed into the first effective value sequence. Based on any frequency in the frequency domain characteristics as the target frequency, the first amplitude sequence of the target frequency is calculated. The Pearson correlation coefficient between the first effective value sequence and the first amplitude sequence is used as the first temperature coefficient of the target frequency. Based on any historical electricity consumption pattern cluster as the target cluster, calculate the second effective value sequence of the target cluster and the second temperature coefficient of the target frequency within the target cluster; The average of the first and second temperature coefficients is used as the acceleration temperature coefficient of the target frequency under the target cluster.

[0011] Preferably, obtaining the accelerating temperature coefficient includes: The historical temperature rise time series is divided into multiple temperature rise sequences according to the historical current time series division pattern. The RMS of each temperature rise sequence is calculated, and the RMS of all temperature rise sequences are constructed into the first effective value sequence. Based on any frequency in the frequency domain characteristics as the target frequency, the first amplitude sequence of the target frequency is calculated. The Pearson correlation coefficient between the first effective value sequence and the first amplitude sequence is used as the first temperature coefficient of the target frequency. Based on any historical electricity consumption pattern cluster as the target cluster, calculate the second effective value sequence of the target cluster and the second temperature coefficient of the target frequency within the target cluster; Using the comprehensive difference distance between the target cluster center and other historical power consumption pattern cluster centers as the weight, the second temperature coefficient of all other historical power consumption pattern clusters is weighted and summed to obtain a correction term. The average of the sum of the correction term and the first temperature coefficient is used as the acceleration temperature coefficient of the target frequency under the target cluster.

[0012] Preferably, the acquisition of future lifespan loss includes: The future current time series is subjected to Fourier transform to obtain the spectrum; based on any historical power consumption pattern cluster as the target cluster, the sum of the products of the amplitude of each frequency in the spectrum and its acceleration temperature coefficient under the target cluster is calculated to obtain the predicted equivalent temperature rise. Based on the predicted equivalent temperature rise, the future lifetime loss is calculated using the Arrenius model.

[0013] Preferably, the acquisition of the predicted remaining lifespan includes: The predicted remaining lifetime is obtained by subtracting the sum of historical lifetime loss and weighted future lifetime loss from the initial lifetime parameter, and then dividing by the mean of the predicted acceleration factor corresponding to each historical power consumption pattern cluster calculated by the Arrenius model. The weighted future lifetime loss is obtained by weighting and summing the future lifetime losses corresponding to each historical power consumption pattern cluster by calculating the normalized weight of the comprehensive difference distance between the future current time series and the center of each historical power consumption pattern cluster.

[0014] Secondly, a data acquisition terminal line life prediction system based on harmonic analysis includes: a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the data acquisition terminal line life prediction method based on harmonic analysis described in any one of the claims is implemented.

[0015] The present invention has the following beneficial effects: 1. This invention constructs a comprehensive difference distance (third distance) that simultaneously captures the frequency domain similarity and temporal morphological similarity of power consumption behavior by fusing spectral amplitude differences (first distance) with the dynamic time warping distance of empirical mode decomposition components (second distance). Based on this, the optimal segmentation scheme is dynamically determined by maximizing inter-cluster differences and minimizing intra-cluster differences using an effective periodicity fraction index, thereby ensuring that each cluster precisely corresponds to an independent power consumption pattern. This fundamentally solves the problem of power consumption pattern aliasing and feature dilution caused by fixed window truncation, laying a precise data foundation for subsequent harmonic thermal analysis based on pattern differentiation and improving the initial accuracy of lifetime assessment.

[0016] 2. This invention analyzes the broad correlation between harmonics and temperature rise at a global level (first temperature coefficient), then delves into each subdivided power consumption mode cluster to calculate the close correlation between frequency and temperature rise under specific operating conditions (second temperature coefficient). Finally, it introduces a cross-cluster correction term weighted by the distance between mode clusters to comprehensively generate an accelerating temperature coefficient that reflects the actual heating capacity of the frequency under a specific power consumption mode. This makes the aging rate calculation based on the Arrenius model closer to physical reality, thereby significantly improving the prediction accuracy of the remaining life of the line, providing a reliable basis for preventive maintenance, and reducing the safety risks caused by uncontrolled insulation aging from the source. Attached Figure Description

[0017] Figure 1 This is a flowchart of steps S1-S4 in the method for predicting the lifespan of a data acquisition terminal line based on harmonic analysis according to an embodiment of the present invention.

[0018] Figure 2 This is a schematic diagram of the structure of the acquisition terminal line life prediction system based on harmonic analysis according to an embodiment of the present invention. Detailed Implementation

[0019] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments.

[0020] This invention is primarily applied to electricity metering devices in user-side power grids, particularly smart meters and their connecting lines (including terminals, current transformers, and internal PCB circuitry). These devices are exposed to complex harmonic environments generated by nonlinear loads such as electric vehicle charging stations, photovoltaic inverters, and frequency converters, leading to significantly accelerated insulation aging. This invention aims to accurately predict the remaining service life of the internal circuitry by analyzing the current and temperature data collected by the meter, providing a scientific basis for preventative maintenance by power grid operation and maintenance departments, and avoiding metering inaccuracies or safety accidents caused by line aging and failure.

[0021] Furthermore, from the user-side power grid, the electricity meters and their associated lines that require life prediction are selected as the objects of implementation of this invention.

[0022] Reference Figure 1 The method for predicting the lifespan of a data acquisition terminal line based on harmonic analysis includes steps S1-S4, as detailed below: S1: Obtain the historical current timing, historical temperature rise timing, and initial life parameters of the meter circuit.

[0023] The lifespan degradation of electricity meter circuits is the result of the long-term cumulative effect of current thermal effects, influenced by internal heating, environmental heat dissipation, and the inherent properties of the materials themselves. Existing prediction methods often rely solely on current or single temperature data, failing to establish a comprehensive data foundation that reflects the aging drivers, thus limiting prediction accuracy.

[0024] In this embodiment of the invention, historical current timing data from the date of installation and commissioning to the present time is collected via a smart fusion terminal connected to the meter or directly from the meter's communication module. Simultaneously, internal temperature timing data (via a built-in temperature sensing chip) and ambient temperature timing data (via an additional temperature sensor) are also collected. Furthermore, the initial design life (in hours) of the meter's circuitry and the normal operating temperature defined in the product specifications are obtained from the equipment management system.

[0025] In addition, in order to isolate environmental factors, the difference between the corresponding elements of the internal temperature time series of the meter and the ambient temperature time series is used as the historical temperature rise time series, thereby directly reflecting the current-induced heating effect.

[0026] S2: Based on historical current time series, multiple different historical power consumption pattern clusters are divided through time-frequency feature fusion analysis and clustering. For each historical power consumption pattern cluster, the correlation between the amplitude of the fundamental wave and each harmonic current and the temperature rise of the meter line is analyzed to determine the acceleration temperature coefficient under the corresponding historical power consumption pattern cluster.

[0027] User electricity consumption behavior is periodic and phased, with significant differences in load characteristics at different times (such as working during the day and charging at night), resulting in drastically different harmonic components and thermal shock modes to the lines. Traditional methods use fixed time windows to segment current sequences, which easily breaks down single electricity consumption patterns or mixes different patterns, leading to distortion of the subsequent analysis and dilution of pattern characteristics.

[0028] In this embodiment of the invention, various unequal-length segmentation schemes are tried on the historical current time series collected above to obtain a set of current sequences divided by different segmentation schemes.

[0029] For all current sequences under a single segmentation scheme, in order to measure the overall similarity of any two current sequences in terms of electricity consumption behavior, time-series features and frequency-domain features are extracted, and the comprehensive difference distance between any two current sequences is calculated based on the time-series features and frequency-domain features.

[0030] In this process, one of any two selected current sequences is taken as the first sequence, and the other is taken as the second sequence. The specific process for obtaining the above-mentioned comprehensive difference distance is as follows: First, fast Fourier transforms are performed on the first and second sequences respectively to obtain their respective spectra, which contain amplitude information of each frequency component. The sum of the absolute values ​​of the amplitude differences at all corresponding frequency points in the spectra of the first and second sequences is calculated as the first distance between the first and second sequences, to measure the similarity of the current harmonic components in the two time periods.

[0031] Next, empirical mode decomposition is performed on the first and second sequences respectively to obtain two sets of IMFs representing oscillations at different time scales. The dynamic time warping distance of the corresponding order components in the two sets of IMFs is then calculated. Finally, the dynamic time warping distances of all corresponding components are summed to form the second distance between the first and second sequences, in order to measure the similarity in the current fluctuation pattern and rhythm between the two time periods.

[0032] Finally, the sum of the first distance and the second distance is taken as the comprehensive difference distance between the first sequence and the second sequence. The smaller the comprehensive difference distance value, the more similar the electricity consumption behavior represented by the first sequence and the second sequence is.

[0033] Furthermore, based on the above operations, the comprehensive difference distance between any two current sequences under a single segmentation scheme can be obtained. Then, based on this comprehensive difference distance, the DBSCAN algorithm (Density-Based Spatial Clustering of Applications with Noise) is used to cluster all current sequences, initially obtaining several historical power consumption pattern clusters. Similarly, historical power consumption pattern clusters under all segmentation schemes can be obtained.

[0034] Different clustering results of varying quality were obtained through the above different segmentation schemes. In order to objectively select the clustering result that best describes the actual electricity consumption pattern, an effective cycle score is defined as the evaluation criterion.

[0035] The process of obtaining the effective periodicity score for a clustering result is as follows: First, calculate the average comprehensive difference distance between each current sequence within any historical electricity consumption pattern cluster under the clustering result and all current sequences outside the cluster, as the inter-cluster distance (the larger the value, the more obvious the difference between the current sequence within the cluster and the electricity consumption patterns of other clusters outside the cluster); calculate the average comprehensive difference distance between each current sequence within the cluster and other current sequences within its own cluster, as the intra-cluster distance (the smaller the value, the more similar the current sequence within the cluster is to the electricity consumption patterns of its peers within the cluster).

[0036] Then, the inter-cluster distance calculated above is divided by a natural exponential function value with intra-cluster distance as the index to obtain a quality contribution ratio for evaluating a single current sequence. The quality contribution ratios of all current sequences within a historical power consumption pattern cluster are arithmetically averaged to obtain the overall quality score of the historical power consumption pattern cluster. This overall quality score comprehensively reflects the average cohesion and exponentiation level of the corresponding historical power consumption pattern cluster across all clusters.

[0037] Finally, the overall quality scores of all historical electricity consumption pattern clusters are averaged again to obtain the effective period scores of the corresponding clustering results.

[0038] It should be noted that in calculating the quality contribution ratio, the denominator is designed as an exponential function. The primary purpose is to avoid the case where the denominator is 0. Secondly, compared with the conventional practice of adding a small constant value to the denominator, using an exponential function is a more natural and concise mathematical approach. It does not require the introduction of additional hyperparameters, simplifies the evaluation process, and also ensures the continuity and smoothness of the evaluation function.

[0039] Finally, the current sequence with the longest effective period fraction is selected as the final current sequence to participate in clustering, and then the final historical electricity consumption pattern cluster is obtained.

[0040] Under different electricity usage modes, the harmonic components flowing through the circuit vary. Due to differences in the skin effect, proximity effect, and core eddy current loss mechanism, harmonics of different orders contribute significantly to the temperature rise of local hot spots in the circuit, even if they produce the same additional losses. Traditional methods treat all harmonics as equivalent, ignoring this fundamental physical difference, leading to distortion of the thermal model.

[0041] Therefore, it is necessary to calculate an accelerating temperature coefficient for each historical power consumption pattern cluster, and for each characteristic frequency (fundamental and odd harmonics) within that cluster, to quantify the efficiency of the current at that frequency in inducing temperature rise under this specific power consumption pattern. The specific operation is as follows: Based on the aforementioned final historical current time series division model, the historical temperature rise time series is synchronously divided to obtain multiple temperature rise sequences. The RMS (Root Mean Square) of each temperature rise sequence is calculated, and the RMS of all temperature rise sequences are used to construct the first effective value sequence.

[0042] Furthermore, a target frequency, such as the fundamental frequency, the third harmonic, the fifth harmonic, etc., is selected, and the amplitude of the target frequency is extracted from all current sequences to construct the first amplitude sequence.

[0043] Furthermore, the Pearson correlation coefficient between the first effective value sequence and the first amplitude sequence is calculated as the first temperature coefficient of the target frequency, which reflects the broad correlation between the target frequency and the overall temperature rise of the meter circuit.

[0044] Similarly, a specific historical electricity consumption pattern cluster obtained by clustering the current sequences is taken as the target cluster. All current sequences and corresponding temperature rise sequences belonging to the target cluster are identified. The RMS of all temperature rise sequences in the target cluster is calculated and constructed as the second effective value sequence. The amplitude of the target frequency in all current sequences in the target cluster is calculated and the second amplitude sequence is constructed. The Pearson correlation coefficient between the second effective value sequence and the second amplitude sequence is calculated and used as the second temperature coefficient of the target frequency within the target cluster. This second temperature coefficient reflects the local close correlation between the amplitude change of the target frequency current and the temperature rise change under this specific electricity consumption pattern.

[0045] Based on the correlation coefficients of the above two levels (i.e., the first temperature coefficient and the second temperature coefficient), the acceleration temperature coefficient of the target frequency under the target cluster is calculated. In this embodiment of the invention, one of the following fusion strategies is adopted: Basic fusion: The average of the first and second temperature coefficients is used as the acceleration temperature coefficient of the target frequency under the target cluster.

[0046] Weighted Enhanced Fusion: Using the comprehensive difference distance between the target cluster center and other historical power consumption pattern cluster centers as the weight, the second temperature coefficient of all other historical power consumption pattern clusters is weighted and summed to obtain a correction term; the mean of the sum of the correction term and the first temperature coefficient is used as the acceleration temperature coefficient of the target frequency under the target cluster, thereby improving the robustness of the acceleration temperature coefficient during the transition between power consumption patterns.

[0047] For example, the acceleration temperature coefficient obtained by the above weighted enhanced fusion is expressed by the following formula: In the formula, For the above-mentioned accelerating temperature coefficient, This refers to the second temperature coefficient mentioned above. For the target frequency in the th excluding the target cluster The second temperature coefficient within a cluster of historical electricity consumption patterns This represents the total number of historical power consumption pattern clusters excluding the target cluster. For the target cluster center and the first The overall difference distance between the centers of historical electricity consumption pattern clusters This is the sum of the overall difference distances between the target cluster center and other historical power consumption pattern cluster centers.

[0048] S3: Calculate historical lifetime loss based on historical temperature time series and initial lifetime parameters; simultaneously, predict future current time series based on historical current time series; and calculate the future lifetime loss corresponding to each power consumption mode cluster based on the frequency domain characteristics and acceleration temperature coefficient of the future current time series.

[0049] The remaining lifespan of an electricity meter's circuit equals its initial lifespan minus incurred losses and future losses. Incurred losses must be accurately assessed based on its actual historical temperature rise trajectory. Meanwhile, future user electricity consumption behavior (especially the possibility of new nonlinear loads) is highly uncertain, and simply assuming future harmonic levels will be the same as historical levels (static extrapolation) carries significant risks. Therefore, it is essential to predict the meter's potential future current conditions based on its historical data.

[0050] First, based on the RMS (converted to absolute temperature) of each temperature rise sequence obtained above, and combined with the normal operating temperature of its meter circuit, the acceleration factor of the time period corresponding to each temperature rise sequence is calculated using the Arrenius model (a widely used thermal aging assessment model in the prior art). Then, the products of the acceleration factors of all temperature rise sequences and their durations are accumulated to obtain the historical lifespan loss of the meter circuit.

[0051] Then, using time series prediction models such as long short-term memory networks, the current time series of the electricity meter in the near future (such as the past 24 hours) is used as input to predict the future current time series in the future (such as the next 24 hours).

[0052] Next, a Fast Fourier Transform is performed on the predicted future current time series to analyze its spectrum. For each historical electricity consumption pattern cluster obtained above, the following calculations are performed: The predicted equivalent temperature rise is obtained by multiplying the amplitude of each frequency in the spectrum of the future current time series by the aforementioned accelerating temperature coefficient and then summing the results.

[0053] Using the Arrenius model, the future lifespan loss caused by this electricity consumption pattern is calculated based on the predicted equivalent temperature rise. The calculated acceleration factor is labeled as the predicted acceleration factor.

[0054] Through the above operations, not only were historical losses assessed based on accurate thermal models (Arrhenius model and accelerating temperature coefficient), but more importantly, by predicting future current and combining it with historical power consumption patterns to estimate future losses, coverage of different power consumption scenarios (including new harmonic characteristics) that may occur in the future was achieved. This breaks the limitations of static extrapolation and enables predictions to dynamically respond to changes in user behavior.

[0055] S4: Calculate and output the predicted remaining lifespan of the meter line by combining the initial lifespan parameters, historical lifespan loss, and future lifespan loss corresponding to each power consumption mode cluster.

[0056] Although the above calculation yields a future lifespan loss, future electricity consumption behavior may not perfectly match a certain historical electricity consumption pattern, but is more likely to be a mixture or transition of multiple electricity consumption patterns.

[0057] In this embodiment of the invention, the comprehensive difference distance between the future current time series and the center of each historical power consumption pattern cluster is calculated and normalized to obtain a normalized weight (i.e., the probability weight of the future current time series belonging to each historical power consumption pattern cluster), and then the future lifetime loss corresponding to each historical power consumption pattern cluster is weighted and summed to obtain the weighted future lifetime loss.

[0058] After obtaining the weighted future lifetime loss, the sum of the historical lifetime loss and the weighted future lifetime loss is subtracted from the initial lifetime parameter, and then divided by the mean of the prediction acceleration factor corresponding to each historical power consumption pattern cluster calculated by the Arrenius model (used to standardize the loss time) to obtain the predicted remaining lifetime.

[0059] Finally, the predicted remaining lifespan is compared with a preset threshold (which can be set to 10% to 20% of the initial lifespan parameter). If the predicted remaining lifespan is less than or equal to the threshold, an early warning signal is generated and the maintenance process is triggered; otherwise, the remaining lifespan information is output.

[0060] This invention also provides a system for predicting the lifespan of a data acquisition terminal line based on harmonic analysis. For example... Figure 2 As shown, the system includes a processor and a memory. The memory stores computer program instructions, which, when executed by the processor, implement the method for predicting the lifespan of a data acquisition terminal line based on harmonic analysis according to the first aspect of the present invention.

[0061] The system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces, the settings and functions of which are known in the art and will not be described in detail here.

[0062] It should be noted that those skilled in the art can make various modifications and improvements without departing from the inventive concept, and these all fall within the scope of protection of this invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

Claims

1. A method for predicting the lifespan of a data acquisition terminal line based on harmonic analysis, characterized in that, include: Obtain historical current timing, historical temperature rise timing, and initial life parameters of the meter circuit; The historical current time series is divided into multiple current sequences. For each current sequence, time series features and frequency domain features are extracted, and the comprehensive difference distance between any two current sequences is calculated based on the time series features and frequency domain features. All current sequences are clustered based on the comprehensive difference distance to obtain all historical power consumption pattern clusters. For each historical power consumption pattern cluster, the acceleration temperature coefficient of the current at each frequency under the corresponding historical power consumption pattern cluster is determined. Historical lifetime loss was calculated based on historical temperature rise time series and Arrenius model; Predicting future current timing based on historical current timing; Based on the frequency domain characteristics of the future current time series, the acceleration temperature coefficient corresponding to each historical power consumption pattern cluster, and the comprehensive difference distance between the future current time series and the center of each historical power consumption pattern cluster, the future lifetime loss corresponding to each historical power consumption pattern cluster is calculated. By combining initial life parameters, historical life loss, and future life loss, the predicted remaining life of the meter circuit is calculated and output.

2. The method for predicting the lifespan of a data acquisition terminal line based on harmonic analysis according to claim 1, characterized in that, When the predicted remaining lifespan of the output meter line is less than or equal to the preset lifespan threshold, an early warning signal is generated and maintenance or replacement is arranged.

3. The method for predicting the lifespan of a data acquisition terminal line based on harmonic analysis according to claim 1, characterized in that, The historical temperature rise time series is obtained by collecting the internal temperature time series of the electricity meter and the ambient temperature time series, and subtracting the corresponding elements of the internal temperature time series of the electricity meter and the ambient temperature time series.

4. The method for predicting the lifespan of a data acquisition terminal line based on harmonic analysis according to claim 1, characterized in that, The acquisition of the comprehensive difference distance includes: Calculate the sum of the absolute values ​​of the differences between the frequency amplitudes in the spectrum graphs corresponding to the two current sequences, and use this as the first distance between the two current sequences; The sum of the dynamic time-warped distances between the components of two current sequences after empirical mode decomposition is calculated as the second distance between the two current sequences. The sum of the first distance and the second distance is taken as the comprehensive difference distance between the two current sequences.

5. The method for predicting the lifespan of a data acquisition terminal line based on harmonic analysis according to claim 1, characterized in that, The process of dividing the historical current time series to obtain multiple current sequences includes obtaining current sequences under different division schemes by setting multiple unequal length division schemes. After clustering current sequences of various lengths, the effective period score of each clustering result is calculated. The current sequence with the largest effective period score is selected as the final current sequence to participate in the clustering, and then the final historical electricity consumption pattern cluster is obtained. Specifically, for each historical electricity consumption pattern cluster under a clustering result, the average comprehensive difference distance between each current sequence within the cluster and all current sequences outside the cluster is calculated as the inter-cluster distance; the average comprehensive difference distance between each current sequence within the cluster and other current sequences within its own cluster is calculated as the intra-cluster distance; and the effective cycle score is calculated based on the inter-cluster distance and intra-cluster distance of all current sequences within the cluster of all historical electricity consumption pattern clusters.

6. The method for predicting the lifespan of a data acquisition terminal line based on harmonic analysis according to claim 1 or 5, characterized in that, The acquisition of the acceleration temperature coefficient includes: The historical temperature rise time series is divided into multiple temperature rise sequences according to the historical current time series division pattern. The RMS of each temperature rise sequence is calculated, and the RMS of all temperature rise sequences are constructed into the first effective value sequence. Based on any frequency in the frequency domain characteristics as the target frequency, the first amplitude sequence of the target frequency is calculated. The Pearson correlation coefficient between the first effective value sequence and the first amplitude sequence is used as the first temperature coefficient of the target frequency. Based on any historical electricity consumption pattern cluster as the target cluster, calculate the second effective value sequence of the target cluster and the second temperature coefficient of the target frequency within the target cluster; The average of the first and second temperature coefficients is used as the acceleration temperature coefficient of the target frequency under the target cluster.

7. The method for predicting the lifespan of a data acquisition terminal line based on harmonic analysis according to claim 1 or 5, characterized in that, The acquisition of the acceleration temperature coefficient includes: The historical temperature rise time series is divided into multiple temperature rise sequences according to the historical current time series division pattern. The RMS of each temperature rise sequence is calculated, and the RMS of all temperature rise sequences are constructed into the first effective value sequence. Based on any frequency in the frequency domain characteristics as the target frequency, the first amplitude sequence of the target frequency is calculated. The Pearson correlation coefficient between the first effective value sequence and the first amplitude sequence is used as the first temperature coefficient of the target frequency. Based on any historical electricity consumption pattern cluster as the target cluster, calculate the second effective value sequence of the target cluster and the second temperature coefficient of the target frequency within the target cluster; Using the comprehensive difference distance between the target cluster center and other historical power consumption pattern cluster centers as the weight, the second temperature coefficient of all other historical power consumption pattern clusters is weighted and summed to obtain a correction term. The average of the sum of the correction term and the first temperature coefficient is used as the acceleration temperature coefficient of the target frequency under the target cluster.

8. The method for predicting the lifespan of a data acquisition terminal line based on harmonic analysis according to claim 1, characterized in that, The acquisition of the future lifespan loss includes: The future current time series is subjected to Fourier transform to obtain the spectrum; based on any historical power consumption pattern cluster as the target cluster, the sum of the products of the amplitude of each frequency in the spectrum and its acceleration temperature coefficient under the target cluster is calculated to obtain the predicted equivalent temperature rise. Based on the predicted equivalent temperature rise, the future lifetime loss is calculated using the Arrenius model.

9. The method for predicting the lifespan of a data acquisition terminal line based on harmonic analysis according to claim 8, characterized in that, The acquisition of the predicted remaining lifetime includes: The predicted remaining lifetime is obtained by subtracting the sum of historical lifetime loss and weighted future lifetime loss from the initial lifetime parameter, and then dividing by the mean of the predicted acceleration factor corresponding to each historical power consumption pattern cluster calculated by the Arrenius model. The weighted future lifetime loss is obtained by weighting and summing the future lifetime losses corresponding to each historical power consumption pattern cluster by calculating the normalized weight of the comprehensive difference distance between the future current time series and the center of each historical power consumption pattern cluster.

10. A data acquisition terminal line life prediction system based on harmonic analysis, characterized in that, include: A processor and a memory, wherein the memory stores computer program instructions that, when executed by the processor, implement the method for predicting the lifespan of a data acquisition terminal line based on harmonic analysis according to any one of claims 1-9.

Images