Dynamic metering calibration method and system for electric energy meter

Abstract

The invention provides an electric energy meter dynamic metering calibration method and system, and relates to the technical field of electric energy meter calibration, and the method comprises the steps: collecting the multi-source metering data of an electric energy meter, constructing an error propagation topology network according to the error influence relation of the multi-source metering data of the electric energy meter, carrying out the quantification of the error of each node in a node set, and carrying out the calculation of the error propagation topology network. Acquiring an error propagation energy field; the method comprises the following steps: constructing an error propagation energy field model, generating an error propagation energy evolution sequence based on time evolution, extracting an energy evolution feature vector, identifying a current evolution state, generating a calibration trigger time window if the current evolution state is detected to be in a critical evolution state, and executing metering calibration on the electric energy meter. The technical problem that dynamic change of errors is difficult to reflect in real time due to the fact that the electric energy meter is calibrated by adopting a fixed period or a fixed time node in the prior art is solved. The technical effects of dynamically determining the calibration time node and improving the metering precision and the resource utilization efficiency are achieved.

Description

Technical Field

[0001] This invention relates to the field of electricity meter calibration technology, specifically to a dynamic metering calibration method and system for electricity meters. Background Technology

[0002] As a key metering device in the power system, the accuracy of electricity meters directly affects the accuracy of electricity measurement. In existing technologies, electricity meter calibration typically employs periodic calibration or fixed-time-node calibration, meaning calibration is performed on the meter at preset time intervals or fixed points. However, in actual operation, the metering error of electricity meters is affected by various factors, including voltage sampling deviation, current sampling drift, load fluctuations, and electromagnetic interference. Different factors have significantly different impacts on the error, and the rate of change and evolution of the error also differ significantly at different operating stages.

[0003] Current technologies, which use a uniform time point for calibration, struggle to accurately reflect the dynamic changes in electricity meter errors. Calibration when errors are still stable may waste calibration resources, while failure to calibrate promptly when errors enter a rapid evolution phase can lead to significant metering deviations, impacting meter accuracy and power system operational safety. Therefore, a dynamic metering calibration method for electricity meters is needed. This method should comprehensively analyze the propagation relationships and evolution of errors from multiple data sources to dynamically determine calibration time points, enabling real-time error monitoring and intelligent calibration, thereby improving the metering accuracy and reliability of electricity meters.

[0004] In summary, existing technologies suffer from the problem of using fixed cycles or fixed time points for electricity meter calibration, which makes it difficult to reflect dynamic changes in error in real time, resulting in wasted calibration resources or increased metering deviations. Summary of the Invention

[0005] The purpose of this application is to provide a dynamic metering calibration method and system for electricity meters, which solves the technical problem that existing technologies use fixed periods or fixed time nodes for electricity meter calibration, making it difficult to reflect the dynamic changes of errors in real time, resulting in wasted calibration resources or increased metering deviations.

[0006] In view of the above problems, this application provides a method and system for dynamic metering calibration of electricity meters.

[0007] The first aspect of this application provides a dynamic metering calibration method for electricity meters, comprising: collecting multi-source metering data of the electricity meter through an internal sampling module; constructing an error propagation topology network based on the error influence relationship of the multi-source metering data; quantifying the error of each node in the node set based on the error propagation topology network to obtain an error propagation energy field; constructing an error propagation energy field model based on the error propagation energy field; generating a time-evolutionary error propagation energy evolution sequence based on the error propagation energy field model; and extracting an energy evolution feature vector from the error propagation energy evolution sequence; wherein, the current evolution state is identified based on the energy evolution feature vector; if the current evolution state is detected to be in a critical evolution state, a calibration trigger time window is generated; and metering calibration is performed on the electricity meter within the calibration trigger time window.

[0008] Optionally, the energy evolution feature vector includes energy growth rate, energy growth acceleration, and energy fluctuation amplitude; the energy growth rate, energy growth acceleration, and energy fluctuation amplitude are used as input variables to construct an evolutionary state determination function, which includes multiple defined evolutionary states and multiple energy evolution feature vector sample groups corresponding to the multiple evolutionary states; wherein, the multiple evolutionary states include a stable energy evolution state, a slow energy growth state, a rapid energy growth state, and a sudden energy change state; the energy evolution feature vector is input into the evolutionary state determination function to calculate feature vector similarity, and the current evolutionary state is output based on the feature vector similarity calculation result.

[0009] Optionally, a calibration trigger time window is generated if the current evolutionary state is detected to be in a critical evolutionary state. The critical evolutionary state includes the energy rapid growth state and the energy mutation state among the plurality of evolutionary states.

[0010] Optionally, the data sources and real-time measurement data of each node are read from the error propagation topology network; the error data set of each node based on the real-time measurement data and theoretical measurement data is calculated; the propagation weights of the error propagation topology network are analyzed to obtain the error propagation weight matrix; the node error propagation energy set is calculated based on the error propagation weight matrix and the error data set; and the node error propagation energy set is mapped to the error propagation topology network to obtain the error propagation energy field.

[0011] Optionally, a time evolution parameter is defined, and an initial decay function is introduced to perform spatial diffusion evolution of the error propagation energy field based on the time evolution parameter, resulting in an error propagation diffusion evolution energy field. The error propagation diffusion evolution energy field is compared with the theoretical error propagation diffusion energy field, and the decay coefficient of the decay function is iteratively optimized based on the evolution accuracy obtained from the comparison, until an optimized decay function with an evolution accuracy reaching a preset threshold is obtained. The error propagation energy field model constructed based on the optimized decay function is then based on the time-evolved error propagation energy evolution sequence.

[0012] Optionally, a calibration time tolerance window is set; based on the calibration time tolerance window, the time node when in the critical evolution state is extended to obtain the calibration trigger time window.

[0013] Optionally, the error evolution rate of each data type is calculated based on the error evolution feature vector; the calibration sensitivity index of each data type is obtained based on the error evolution rate; the current data source type corresponding to the calibration request is obtained, and the size of the calibration time tolerance window is set inversely proportional to the calibration sensitivity index corresponding to the current data source type.

[0014] Optionally, an error propagation energy evolution sequence based on time evolution is generated using the error propagation energy field model, wherein the evolution time window of the time evolution is larger than the initial calibration time window; wherein the initial calibration time window is the initial calibration time window in which the energy meter can be used to perform metering calibration.

[0015] Optionally, the calibration trigger time window is a sub-window of the initial calibration time window.

[0016] A second aspect of this application provides a dynamic metering calibration system for electricity meters, comprising: a data acquisition component for acquiring multi-source metering data of the electricity meter through an internal sampling module; an error quantization component for constructing an error propagation topology network based on the error influence relationship of the multi-source metering data, quantifying the error of each node in the node set based on the error propagation topology network, and obtaining an error propagation energy field; an evolution sequence generation component for constructing an error propagation energy field model based on the error propagation energy field, generating a time-evolutionary error propagation energy evolution sequence based on the error propagation energy field model, and extracting the energy evolution feature vector of the error propagation energy evolution sequence; and a metering calibration component for identifying the current evolution state based on the energy evolution feature vector, generating a calibration trigger time window when the current evolution state is detected to be in a critical evolution state, and performing metering calibration on the electricity meter within the calibration trigger time window.

[0017] One or more technical solutions provided in this application have at least the following technical effects or advantages: The method provided in this application collects multi-source metering data from an electricity meter through an internal sampling module; constructs an error propagation topology network based on the error influence relationship of the multi-source metering data; quantifies the error of each node in the node set based on the error propagation topology network to obtain the error propagation energy field; constructs an error propagation energy field model based on the error propagation energy field; generates a time-evolutionary error propagation energy evolution sequence using the error propagation energy field model; extracts the energy evolution feature vector of the error propagation energy evolution sequence; identifies the current evolution state based on the energy evolution feature vector; and generates a calibration trigger time window if the current evolution state is detected to be in a critical evolution state, and performs metering calibration on the electricity meter within the calibration trigger time window. This achieves the technical effect of comprehensively analyzing the error propagation relationship of multi-source data and dynamically identifying the error evolution state, thereby dynamically determining the calibration time node and improving metering accuracy and resource utilization efficiency.

[0018] The above description is merely an overview of the technical solution of this application. To better understand the technical means of this application and to facilitate its implementation according to the description, and to make the above and other objects, features, and advantages of this application more apparent, specific embodiments of this application are described below. It should be understood that the content described in this section is not intended to identify key or important features of the embodiments of this application, nor is it intended to limit the scope of this application. Other features of this application will become readily apparent through the following description. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in this application 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 merely exemplary. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0020] Figure 1 A flowchart illustrating the dynamic metering calibration method for electricity meters provided in this application.

[0021] Figure 2 A schematic diagram of the structure of the dynamic metering calibration system for electricity meters provided in this application.

[0022] Figure labeling: Data acquisition component 11, error quantization component 12, evolution sequence generation component 13, metrological calibration component 14. Detailed Implementation

[0023] This application provides a dynamic metering calibration method and system for electricity meters, addressing the technical problem that existing technologies use fixed periods or fixed time points for electricity meter calibration, leading to difficulties in reflecting dynamic error changes in real time, wasted calibration resources, or increased metering deviations. It achieves the technical effect of comprehensively analyzing the error propagation relationship of multi-source data and dynamically identifying the error evolution state, thereby dynamically determining the calibration time point and improving metering accuracy and resource utilization efficiency.

[0024] The technical solutions of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. It should be understood that the present invention is not limited to the exemplary embodiments described herein. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention. It should also be noted that, for ease of description, only the parts related to the present invention are shown in the accompanying drawings, not all of them.

[0025] Example 1, as Figure 1 As shown, this application provides a dynamic metering calibration method for electricity meters, which includes: Multi-source metering data of the electricity meter is collected through the internal sampling module of the electricity meter.

[0026] Specifically, a sampling module installed inside the electricity meter collects multi-source metering data in real time during the meter's operation. This sampling module consists of a high-precision analog-to-digital converter (ADC), signal conditioning circuitry, and a clock synchronization unit. It converts various analog electrical signals within the electricity meter into processable digital signals. Under the control of a unified sampling clock, the sampling module synchronously samples voltage signals, current signals, and auxiliary operating parameters in key metering stages of the electricity meter, thereby ensuring a consistent time reference across different data sources.

[0027] During the data acquisition process, a voltage transformer or a resistor divider sampling circuit is connected to the grid voltage port to proportionally reduce the high-voltage signal from the grid side before inputting it to the signal conditioning circuit. After filtering and isolation, the signal is sent to the analog-to-digital converter for digital sampling, thereby obtaining the digital voltage sampling value at the corresponding moment. At the same time, the current sampling device is connected in series in the current loop through a current transformer or a shunt resistor sampling circuit to convert the current signal flowing through the energy meter into a voltage signal proportional to the current. After amplification and filtering, the voltage signal is input to the analog-to-digital converter for sampling to obtain the digital current sampling value.

[0028] After obtaining the synchronized voltage sampling sequence U(t) and current sampling sequence I(t), the metering processing unit performs calculations based on the basic relationship of electricity metering. Instantaneous power is obtained through the instantaneous power calculation formula P(t)=U(t)×I(t), which refers to the instantaneous energy change rate represented by the product of voltage and current at a certain sampling moment. Active power is obtained by averaging the instantaneous power sequence over a power frequency cycle. Active power represents the effective power actually consumed by the load. Reactive power is obtained by calculating the phase difference between voltage and current. Reactive power represents the power exchanged between inductor or capacitor elements. The active power is integrated over time, E=∫P(t)dt, to obtain the cumulative electricity value, forming the multi-source metering data of the electricity meter.

[0029] By collecting multi-source metering data from electricity meters with a unified time reference, the metering status and its changing trends under different operating conditions can be comprehensively reflected, thereby ensuring that the entire dynamic metering calibration method can accurately identify error evolution trends and trigger calibration in a timely manner.

[0030] An error propagation topology network is constructed based on the error influence relationship of the multi-source metering data of the electricity meter. The error of each node in the node set is quantified based on the error propagation topology network to obtain the error propagation energy field.

[0031] Furthermore, based on the error propagation topology network, the error of each node in the node set is quantified to obtain the error propagation energy field. The method includes: reading the data source and real-time measurement data of each node from the error propagation topology network; calculating the error data set of each node based on the real-time measurement data and theoretical measurement data; analyzing the propagation weights of the error propagation topology network to obtain the error propagation weight matrix; calculating the node error propagation energy set based on the error propagation weight matrix and the error data set; and mapping the node error propagation energy set to the error propagation topology network to obtain the error propagation energy field.

[0032] Specifically, an error propagation topology network is constructed based on the error impact relationships of multi-source metering data from electricity meters. These error impact relationships refer to the correlation between the error of one metering parameter and the transmission or amplification effect on other metering parameters. For example, voltage measurement errors are transmitted to active power through the power calculation formula, and current measurement errors affect power calculation. Based on these error impact relationships, various data items in the multi-source metering data of the electricity meter are used as nodes, such as voltage nodes, current nodes, power nodes, and energy nodes. The error impact paths between different nodes are used as edges, thus constructing an error propagation topology network. This error propagation topology network is a directed graph structure that describes the transmission relationship of errors between different metering parameters, where nodes represent error sources or metering variables, and edges represent error propagation directions.

[0033] The data sources and real-time metering data for each node are read from the constructed error propagation topology network. Data sources include voltage node data from voltage transformer sampling, current node data from shunt sampling, and real-time metering data such as real-time voltage and current values. High-precision calibration equipment, such as a standard source, is used to acquire corresponding theoretical metering data, such as standard voltage, theoretical power, or reference energy values. The difference between the real-time metering data and the theoretical metering data for each node is calculated to obtain the error data set for each node, where the error value is represented as e. i =x ir -x it , where e i Let x represent the error value of the i-th node. ir Indicates real-time metering data, x it Theoretical measurement data. Based on the computational dependencies or historical statistical correlations between various measurement parameters, the propagation weights of the error propagation topology are analyzed, and a corresponding propagation weight is assigned to each propagation path in the error propagation topology. For example, the influence weight of voltage error on power error can be calculated using the partial derivative of the power formula: wU→P=| P / U|=I |cosΦ|, where Φ is the power factor angle. Similarly, the propagation weights between other parameters are calculated. The propagation weights reflect the degree of influence of the error as it propagates from one node to another. Based on the propagation weights of the error propagation topology network, an error propagation weight matrix W = [w ij The error propagation weight matrix is ​​a matrix structure used to describe the error propagation relationship between nodes, where the matrix elements w ij This indicates the strength of the impact of the error at node i on node j.

[0034] The error propagation weight matrix is ​​used to perform matrix operations with the node error data set to calculate the error propagation energy value of each node. This error propagation energy characterizes the overall impact of a node's error on the entire network and is calculated by combining the error magnitude and propagation weights, for example, E. i =∑(w ij ×e i 2 ), where E i The error propagation energy of node i is represented by the square operation e. i 2This process transforms error values ​​into non-negative quantities, reflecting only the magnitude of the error without being affected by its direction. This allows the error propagation energy to objectively reflect the error intensity. Furthermore, the resulting set of error propagation energy values ​​for each node is mapped back into the error propagation topology network. This ensures that each node in the network has a corresponding energy value distribution, forming an error propagation energy field that describes the distribution of errors in the network space. This energy field, quantified in energy form on the error propagation topology, directly reflects the concentrated areas and propagation trends of the errors.

[0035] For example, at a certain moment, the following multi-source metering data of the electricity meter is collected: voltage U=221V, current I=5.0A, active power P=1110W, and cumulative energy E=1500kWh, corresponding to theoretical values ​​of U0=220V, I0=5.0A, and P0=1100W, respectively. Through calculation, the errors of each node are: voltage error eU=1V, current error eI=0A, and power error eP=10W. Through calculation, the propagation weight of voltage to power in the error propagation weight matrix is ​​0.6, and the propagation weight of current to power is 0.4. Then, the error propagation energy of the power node is expressed as EP=0.6×(1) 2 +0.4×(0) 2 =0.6, while the propagation energy of the voltage node's own error is EU=1. 2 =1, thus obtaining the node error propagation energy set {EU=1,EI=0,EP=0.6}. Then, marking this energy value on the corresponding node will form the error propagation energy field at the current moment, thus intuitively reflecting that the voltage node is the main energy source for the current error propagation.

[0036] By constructing an error propagation topology network and calculating the error propagation energy field, the error information that was originally scattered in different metering parameters is modeled in a unified manner. This allows the error to not only be quantified but also to reflect the error propagation relationship and the scope of influence, thereby enabling more accurate identification of potential metering deviation risks and improving the effectiveness and accuracy of dynamic metering calibration of electricity meters.

[0037] An error propagation energy field model based on the aforementioned error propagation energy field is constructed. An error propagation energy evolution sequence based on time evolution is generated using the error propagation energy field model, and the energy evolution feature vector of the error propagation energy evolution sequence is extracted.

[0038] Furthermore, an error propagation energy field model based on the aforementioned error propagation energy field is constructed. The method includes: defining time evolution parameters; introducing an initial decay function to perform spatial diffusion evolution of the error propagation energy field based on the time evolution parameters, obtaining an error propagation diffusion evolution energy field; comparing the error propagation diffusion evolution energy field with the theoretical error propagation diffusion energy field; iteratively optimizing the decay coefficient of the decay function based on the evolution accuracy obtained from the comparison, until an optimized decay function with an evolution accuracy reaching a preset threshold is obtained; and constructing an error propagation energy field model based on the optimized decay function, using a time-evolved error propagation energy evolution sequence.

[0039] Specifically, a time evolution parameter t is defined to describe the dynamic change of error propagation energy over time, including the time step Δt and the total evolution time T. An initial decay function f0(t) is also introduced to simulate the diffusion and decay effects of the error during propagation over time and space: f0(t) = e -λt Where λ is the initial decay coefficient, which can be set based on historical data statistics, the average magnitude of node errors, or system experience. For example, the exponential decay coefficient corresponding to an energy decay of approximately 10% to 20% within the initial evolution step can be used as the starting point. Based on the time evolution parameters and the decay function, the error propagation energy field is spatially diffused and evolved. That is, the error energy of each node is simulated to diffuse to neighboring nodes in the network topology over time, while its own energy gradually weakens according to the decay function, thus obtaining the error propagation and diffusion evolution energy field. The error propagation and diffusion evolution energy field E at time t is... t =f0(t) E0+(1-f0(t)) W E t-Δt Where E0 is the initial error propagation energy field, W is the error propagation weight matrix, and E t-Δt This represents the error propagation energy field at the previous moment. The obtained error propagation and diffusion evolution energy field is compared with the theoretically constructed error propagation and diffusion theoretical energy field to calculate the evolution accuracy and evaluate the fitting effect of the current decay function on the energy evolution trend. The evolution accuracy is used to measure the simulated error propagation and diffusion evolution energy field E. t Theoretical error propagation and diffusion theory energy field The degree of fit between them, evolutionary accuracy = .

[0040] Based on the evolution accuracy obtained from the comparison, the decay coefficient λ of the decay function is iteratively optimized until the evolution accuracy reaches a preset threshold. This is achieved by comparing the evolution accuracy obtained from each simulation with the preset threshold, for example, 0.95. If the evolution accuracy is lower than the preset threshold, it indicates that the decay coefficient of the current decay function is too large or too small, causing the energy diffusion rate to deviate from the theoretical trend. The decay coefficient is then adjusted according to the error direction; for example, if the simulated energy decay is too fast, the decay coefficient is decreased, and vice versa. The decay function is reconstructed based on the updated decay coefficient, and a new round of evolution simulation is performed to calculate the new evolution accuracy. These steps are repeated until the evolution accuracy reaches or exceeds the preset threshold, ultimately yielding an optimized decay function. This ensures that the simulated error propagation evolution sequence accurately reflects the error diffusion trend and provides a reliable basis for subsequent feature extraction and evolution state judgment.

[0041] The final error propagation energy field model is constructed using the optimized decay function, generating a time-evolution-based error propagation energy evolution sequence. This sequence, with time as the axis, describes the dynamic change of error energy within the network over time. After obtaining the error propagation energy evolution sequence, feature extraction is performed. Key parameters representing energy change trends, such as energy growth rate, energy growth acceleration, and energy fluctuation amplitude, are extracted from the sequence. Specifically, the energy growth rate is obtained by calculating the difference between adjacent time steps in the error propagation energy evolution sequence, quantifying the rate of energy change per unit time. The energy growth acceleration is obtained by subtracting the energy growth rate sequence again, reflecting the acceleration or deceleration of the energy change trend. Furthermore, the variance of the error propagation energy evolution sequence within a given time window is calculated as the energy fluctuation amplitude, reflecting the stability or volatility of energy over time, forming an energy evolution feature vector.

[0042] By converting the static error energy field into a dynamic evolution sequence through time evolution modeling, it is possible to analyze and judge the development trend of error over time. Furthermore, through feature vector extraction, complex time series information can be quantified into determinable indicators, providing a key basis for evolution state identification and calibration triggering mechanisms. This enables dynamic monitoring and intelligent calibration of electricity meter errors.

[0043] The current evolution state is identified based on the energy evolution feature vector. If the current evolution state is detected to be in a critical evolution state, a calibration trigger time window is generated, and the electricity meter is subjected to metering calibration within the calibration trigger time window.

[0044] Furthermore, the method for identifying the current evolutionary state based on the energy evolution feature vector further includes: the energy evolution feature vector including energy growth rate, energy growth acceleration, and energy fluctuation amplitude; constructing an evolutionary state determination function using the energy growth rate, energy growth acceleration, and energy fluctuation amplitude as input variables; the evolutionary state determination function including multiple defined evolutionary states and multiple energy evolution feature vector sample groups corresponding to the multiple evolutionary states; wherein, the multiple evolutionary states include a stable energy evolution state, a slow energy growth state, a rapid energy growth state, and a sudden energy change state; inputting the energy evolution feature vector into the evolutionary state determination function to calculate the feature vector similarity, and outputting the current evolutionary state based on the feature vector similarity calculation result.

[0045] Furthermore, if a calibration trigger time window is generated when the current evolutionary state is detected to be in a critical evolutionary state, the critical evolutionary state includes the energy rapid growth state and the energy mutation state among the plurality of evolutionary states.

[0046] Specifically, the extracted energy evolution feature vectors, including energy growth rate, energy growth acceleration, and energy fluctuation amplitude, are used as input variables to construct an evolution state judgment function. This function maps current energy evolution features to predefined evolution states through feature vector similarity calculation. It includes multiple defined evolution states and multiple energy evolution feature vector sample sets corresponding to these states. The multiple evolution states include a stable energy evolution state, a slow energy growth state, a rapid energy growth state, and a sudden energy change state. The multiple energy evolution feature vector sample sets corresponding to these states are sample sets of four evolution states pre-generated using historical data or simulation. For example, by collecting long-term operating data from electricity meters, feature vectors are calculated based on the energy evolution sequence and then categorized according to the actual error development pattern. The energy evolution sequence under various evolutionary states is simulated by classifying it into the corresponding evolutionary state or by using an error diffusion model. Feature vectors are calculated and sample groups are formed. Each state can contain multiple sample vectors. The stable energy evolution state is a state with small energy changes and low fluctuations, that is, the energy growth rate is about 0, the energy growth acceleration is about 0, and the energy fluctuation amplitude is small. The slow energy growth state is a state with an energy growth rate greater than 0, an energy growth acceleration of about 0, and a medium energy fluctuation amplitude. The rapid energy growth state is a state with a high energy growth rate, an energy growth acceleration greater than 0, and a medium to high energy fluctuation amplitude. The abrupt energy change state is a state with an extremely high energy growth rate, an abrupt change in energy growth acceleration, and a high energy fluctuation amplitude. For example, the sample feature vectors (growth rate, growth acceleration, fluctuation amplitude) corresponding to the stable state of energy evolution include [0.01, 0.0, 0.02], [0.02, 0.01, 0.03], the sample feature vectors corresponding to the slow energy growth state include [0.05, 0.01, 0.04], [0.06, 0.0, 0.05], the sample feature vectors corresponding to the rapid energy growth state include [0.12, 0.05, 0.08], [0.10, 0.04, 0.07], and the sample feature vectors corresponding to the sudden energy change state include [0.18, 0.12, 0.12], [0.15, 0.10, 0.15].

[0047] The energy evolution feature vector is input into the evolution state determination function. Euclidean distance is used to quantify the matching degree between the current energy evolution feature vector and each sample group, obtaining the feature vector similarity calculation result. After calculating the feature vector similarity between the current energy evolution feature vector and all state samples, the category closest to the current evolution state is determined based on the feature vector similarity calculation result; that is, the state corresponding to the smallest distance is taken as the current evolution state. For example, the current feature vector [0.11, 0.05, 0.08] and the rapidly growing sample vector [0.12, 0.05, 0.08] have an Euclidean distance of approximately 0.01, which is much smaller than the distance with other state samples, therefore it is determined to be a rapidly growing state.

[0048] When the current energy performance state is detected to be in a critical evolution state, i.e., the result is determined to be a state of rapid energy growth or energy mutation, a calibration trigger time window is automatically generated. The trigger time window defines the allowable time range from the detection of the critical state to the execution of calibration, and can be determined based on the system's allowable error accumulation, data sampling rate, and calibration response time. For example: T w =T b +1 / S s , among which, T b For the initial calibration time window, such as 10 minutes, S s The calibration sensitivity index for the current data source is used; the faster the error grows, the higher the sensitivity and the shorter the window, ensuring that the metering deviation has not accumulated further while also considering the feasibility of the operation. Within the calibration trigger time window, the metering calibration operation is automatically performed on the electricity meter. By adjusting internal metering parameters or updating error compensation coefficients, for example, by using the standard source injection method or the virtual load method to perform online calibration of the electricity meter, dynamic and accurate correction of the electricity meter is achieved, thereby reducing the metering deviation caused by error propagation.

[0049] By transforming quantified energy evolution information into operable evolution state judgments and triggering automatic calibration based on the evolution state, and dynamically determining calibration time nodes, not only can error accumulation trends be detected in a timely manner, but intervention can also be carried out before the error reaches a critical level. This enables the electricity meter to achieve dynamic adaptive calibration, improving metering accuracy and the reliability of electricity meter operation.

[0050] Furthermore, if a calibration trigger time window is generated when the current evolutionary state is detected to be in a critical evolutionary state, the method further includes: setting a calibration time tolerance window; and expanding the time node in the critical evolutionary state based on the calibration time tolerance window to obtain the calibration trigger time window.

[0051] Furthermore, the calibration time tolerance window is set by means of: calculating the error evolution rate of change for each data type based on the error evolution feature vector; obtaining the calibration sensitivity index for each data type based on the error evolution rate of change; obtaining the current data source type corresponding to the calibration request; and setting the size of the calibration time tolerance window inversely proportional to the calibration sensitivity index corresponding to the current data source type.

[0052] Specifically, for each data type, such as voltage, current, power, and electrical energy, the sequence e representing the time-varying error of the corresponding node is extracted from the error propagation energy evolution sequence. i(t). Using the energy growth rate and energy growth acceleration in the error evolution eigenvector, combined with the current node error, the error evolution rate of change for each data type is calculated to reflect the rate of change of the corresponding error type per unit time. The error evolution rate of change is calculated as the ratio of the change in error value within adjacent time windows to the time interval. The error evolution rate of change R is... i =(e i (t k )-e i (t k-1 )) / Δt, where R i Let e ​​be the rate of change of error evolution for the i-th data type. i (t k ) represents the data type at the current time t. k The error value, Δt=t k - tk-1 This is the time interval. For example, the error of the power node error at two consecutive sampling time points is e. P (t k-1 )=10W, e P (t k With a power data error evolution rate of 15W and a sampling interval of 5 minutes, the power data error evolution rate is: R P = (15-10) / 5 = 1W / minute. Similarly, the error evolution rate of voltage, current and electrical energy can be calculated.

[0053] Based on the error evolution rate of each data type, a corresponding calibration sensitivity index is constructed. This index is a quantitative measure of the impact of an error of a certain data type on metering accuracy. When the error evolution rate of a certain type of data is high, its sensitivity index is also high, indicating that the error of that type of data has a more significant impact on the metering accuracy, thus requiring faster calibration. Based on historical statistical values ​​of the error evolution rate, such as the historical mean, variance, and maximum historical error evolution rate, a calibration sensitivity index S is defined. i S i =α × Mean of historical error evolution rate of change + β × Variance of historical error evolution rate of change + γ × Maximum historical error evolution rate of change, where α, β, and γ are weighting coefficients, which can be determined by the entropy weight method or expert experience, reflecting the impact of the mean rate of change, volatility, and extreme values ​​on sensitivity. When a calibration request is triggered, it is determined which data type dominated the error trigger, i.e., the current data source type corresponding to the calibration request is determined. Specifically, this is done by calculating the proportion of error evolution rate of change or error propagation energy at the critical node for each data type, and selecting the data type with the largest proportion or the fastest changing data type as the current data source type. The calibration sensitivity index S corresponding to the selected data source type is then used. iSet the calibration time tolerance window T in an inverse proportional relationship. t The higher the sensitivity, the shorter the allowable delay time. The formula is: T t =T b / S i , among which, T b This is the base calibration time window. For example, if the current critical state is triggered by a power node, then the current data source type is power data, sensitivity Sp=2.0, and the base time window T... b If the time tolerance is 10 minutes, then the calibration time tolerance window is 20 / 2.0 = 10 minutes.

[0054] Then, taking the time node at which the critical evolution state is detected as the center, the time node at the critical evolution state is extended based on the calibration time tolerance window. The time tolerance window is extended to form a complete calibration trigger time window, for example, from 5 minutes before the trigger point to 5 minutes after the trigger point, for a total of 10 minutes. Automatic metrological calibration operation is performed within the calibration trigger time window to ensure that the error is corrected in time before it spreads further.

[0055] By introducing a calibration time tolerance window mechanism, calibration triggering is no longer fixed at a fixed time, but dynamically adjusted according to the rate of change of different error sources. This makes the calibration strategy more flexible and accurate, thus avoiding the waste of resources caused by frequent calibration when the error changes slowly, and shortening the calibration response time when the error spreads rapidly. This effectively improves the overall stability and accuracy of electricity meter calibration.

[0056] Furthermore, an error propagation energy evolution sequence based on time evolution is generated using the error propagation energy field model, wherein the evolution time window of the time evolution is larger than the initial calibration time window; wherein the initial calibration time window is the initial calibration time window in which the energy meter can be used to perform metering calibration.

[0057] Specifically, an evolution time window is defined as the length of time used to continuously observe the change of error propagation energy over time. The evolution time window is longer than the initial calibration time window to ensure the complete error evolution process is obtained, capturing potential energy growth, acceleration changes, or fluctuations. The initial calibration time window refers to the time range within which the electricity meter can perform basic metering calibration operations, typically determined by the meter's hardware response capability or standard calibration requirements. For example, if a certain model of electricity meter has a sampling rate of once per second, completing a full calibration operation, including data acquisition, error calculation, and parameter adjustment, requires approximately 300 samples, or 5 minutes. In this case, setting the initial calibration time window to 5 minutes satisfies both the hardware response capability and the standard calibration requirements. Then, the evolution time window is determined by multiplying the initial calibration time window by an amplification factor, typically between 2 and 5. Setting the amplification factor to 3 results in an evolution time window of 3 × 5 = 15 minutes, ensuring sufficient observation data for dynamic error analysis and feature extraction.

[0058] Within the evolution time window, a continuous energy evolution sequence is obtained by iteratively calculating the diffusion and decay of the error propagation energy field over time. The energy value at each time point in the energy evolution sequence is determined jointly by the error propagation energy field model and the optimized decay function. Key features, such as energy growth rate, energy growth acceleration, and energy fluctuation amplitude, can be further extracted from the energy evolution sequence to form an energy evolution feature vector.

[0059] By setting the evolution time window to be larger than the initial calibration time window, sufficient dynamic data can be ensured when the energy meter makes dynamic calibration decisions, avoiding insufficient judgment of error evolution trends or delayed calibration triggering due to an excessively short observation window. By generating a time-evolution-based error propagation energy sequence, the variation of error over time can be observed, capturing error accumulation trends and sudden changes. This provides complete data input for dynamic evolution state identification and intelligent calibration, further improving the accuracy of energy meter calibration and increasing resource utilization efficiency.

[0060] Furthermore, the calibration trigger time window is a sub-window of the initial calibration time window.

[0061] Specifically, when the energy meter is identified as being in a critical evolutionary state through evolutionary status determination, a calibration trigger time window is generated to schedule metering calibration operations. This calibration trigger time window is a sub-window based on the initial calibration time window; that is, its duration and location are both within the range of the initial calibration time window. The length of the calibration trigger time window is dynamically adjusted based on the current error evolution feature vector and calibration sensitivity index, but it never exceeds the initial calibration time window, ensuring that the calibration operation is completed within the equipment's allowed time range and improving resource utilization efficiency.

[0062] First, determine the start and end times and duration of the initial calibration time window. Then, calculate the critical state trigger point based on the error evolution feature vector. Use this trigger point as the center or starting point of the calibration trigger sub-window. Then, expand it forward and backward according to the calibration time tolerance window to form a complete calibration trigger time window. The calibration trigger time window refers to the effective interval within which automatic calibration operations are allowed to be performed, ensuring that calibration is completed before the error accumulates further, while also taking into account operational feasibility.

[0063] By setting the calibration trigger time window as a sub-window of the initial calibration time window, the timeliness of the calibration operation is ensured, and the calibration time range that can be performed by the device is also ensured. This allows the electricity meter to perform calibration operations dynamically and in a timely manner while ensuring metering accuracy, thereby reducing the risk of error accumulation, improving the effectiveness and accuracy of dynamic metering calibration of the electricity meter, and improving resource utilization efficiency.

[0064] Example 2, based on the same inventive concept as the dynamic metering calibration method for electricity meters in the foregoing examples, such as... Figure 2 As shown, this application provides a dynamic metering calibration system for electricity meters, wherein the dynamic metering calibration system for electricity meters includes: The data acquisition component 11 is used to acquire multi-source metering data of the electricity meter through the internal sampling module of the electricity meter; the error quantization component 12 is used to construct an error propagation topology network based on the error influence relationship of the multi-source metering data of the electricity meter, quantify the error of each node in the node set based on the error propagation topology network, and obtain the error propagation energy field; the evolution sequence generation component 13 is used to construct an error propagation energy field model based on the error propagation energy field, generate an error propagation energy evolution sequence based on time evolution using the error propagation energy field model, and extract the energy evolution feature vector of the error propagation energy evolution sequence; the metering calibration component 14 is used to identify the current evolution state based on the energy evolution feature vector, generate a calibration trigger time window if the current evolution state is detected to be in a critical evolution state, and perform metering calibration on the electricity meter within the calibration trigger time window.

[0065] Furthermore, the metrology calibration component 14 is also used for: the energy evolution feature vector including energy growth rate, energy growth acceleration, and energy fluctuation amplitude; constructing an evolution state determination function using the energy growth rate, energy growth acceleration, and energy fluctuation amplitude as input variables, the evolution state determination function including multiple defined evolution states and multiple energy evolution feature vector sample groups corresponding to the multiple evolution states; wherein, the multiple evolution states include a stable energy evolution state, a slow energy growth state, a rapid energy growth state, and a sudden energy change state; inputting the energy evolution feature vector into the evolution state determination function to calculate the feature vector similarity, and outputting the current evolution state based on the feature vector similarity calculation result.

[0066] Furthermore, the metrology calibration component 14 is also used to generate a calibration trigger time window when the current evolutionary state is detected to be in a critical evolutionary state, wherein the critical evolutionary state includes the energy rapid growth state and the energy mutation state among the plurality of evolutionary states.

[0067] Furthermore, the error quantization component 12 is also used to: read the data sources and real-time measurement data of each node from the error propagation topology network; calculate the error data set of each node based on the real-time measurement data and theoretical measurement data; analyze the propagation weights of the error propagation topology network to obtain the error propagation weight matrix; calculate the node error propagation energy set based on the error propagation weight matrix and the error data set; and map the node error propagation energy set to the error propagation topology network to obtain the error propagation energy field.

[0068] Furthermore, the evolution sequence generation component 13 is also used to: define time evolution parameters, introduce an initial decay function to perform spatial diffusion evolution of the error propagation energy field based on the time evolution parameters, and obtain an error propagation diffusion evolution energy field; compare the error propagation diffusion evolution energy field with the error propagation diffusion theoretical energy field, and iteratively optimize the decay coefficient of the decay function according to the evolution accuracy obtained from the comparison, until an optimized decay function with an evolution accuracy reaching a preset threshold is obtained; and construct an error propagation energy field model based on the optimized decay function based on a time-evolution error propagation energy evolution sequence.

[0069] Furthermore, the metrology calibration component 14 is also used to: set a calibration time tolerance window; and expand the time node when in a critical evolution state based on the calibration time tolerance window to obtain a calibration trigger time window.

[0070] Furthermore, the metrology calibration component 14 is also used to: calculate the error evolution rate of change for each data type based on the error evolution feature vector; obtain the calibration sensitivity index for each data type based on the error evolution rate of change; obtain the current data source type corresponding to the calibration request; and set the size of the calibration time tolerance window inversely proportional to the calibration sensitivity index corresponding to the current data source type.

[0071] Furthermore, the evolution sequence generation component 13 is also used to: generate a time-evolution-based error propagation energy evolution sequence using the error propagation energy field model, wherein the evolution time window of the time evolution is larger than the initial calibration time window; wherein the initial calibration time window is the initial calibration time window in which the energy meter can be used to perform metering calibration.

[0072] Furthermore, the metrology calibration component 14 is also configured such that the calibration trigger time window is a sub-window of the initial calibration time window.

[0073] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The dynamic metering calibration method and specific examples of the electricity meter in the aforementioned embodiment one are also applicable to the dynamic metering calibration system of the electricity meter in this embodiment. Through the foregoing detailed description of the dynamic metering calibration method of the electricity meter, those skilled in the art can clearly understand the dynamic metering calibration system of the electricity meter in this embodiment. Therefore, for the sake of brevity, it will not be described in detail here.

[0074] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0075] Obviously, those skilled in the art can make several improvements and modifications to this application without departing from the principles of this application, and these improvements and modifications also fall within the protection scope of this application.

Claims

1. A dynamic metering calibration method for electricity meters, characterized in that, The method includes: Multi-source metering data of the electricity meter is collected through the internal sampling module of the electricity meter; An error propagation topology network is constructed based on the error influence relationship of the multi-source metering data of the electricity meter. The error of each node in the node set is quantified based on the error propagation topology network to obtain the error propagation energy field. Construct an error propagation energy field model based on the error propagation energy field, generate a time-evolutionary error propagation energy evolution sequence using the error propagation energy field model, and extract the energy evolution feature vector of the error propagation energy evolution sequence; The current evolution state is identified based on the energy evolution feature vector. If the current evolution state is detected to be in a critical evolution state, a calibration trigger time window is generated, and the electricity meter is subjected to metering calibration within the calibration trigger time window.

2. The dynamic metering calibration method for electricity meters as described in claim 1, characterized in that, The method for identifying the current evolutionary state based on the energy evolution feature vector further includes: The energy evolution feature vector includes energy growth rate, energy growth acceleration, and energy fluctuation amplitude. An evolutionary state determination function is constructed using the energy growth rate, energy growth acceleration, and energy fluctuation amplitude as input variables. The evolutionary state determination function includes multiple defined evolutionary states and multiple energy evolution feature vector sample groups corresponding to the multiple evolutionary states. The multiple evolutionary states include a stable energy evolution state, a slow energy growth state, a rapid energy growth state, and a sudden energy change state. The energy evolution feature vector is input into the evolution state determination function to calculate the feature vector similarity, and the current evolution state is output based on the feature vector similarity calculation result.

3. The dynamic metering calibration method for electricity meters as described in claim 2, characterized in that, If a calibration trigger time window is generated when the current evolutionary state is detected to be in a critical evolutionary state, the critical evolutionary state includes the energy rapid growth state and the energy mutation state among the plurality of evolutionary states.

4. The dynamic metering calibration method for electricity meters as described in claim 1, characterized in that, The error propagation energy field is obtained by quantifying the error of each node in the node set based on the aforementioned error propagation topology network. The method includes: Read the data sources and real-time measurement data of each node from the error propagation topology network; Calculate the error data set of each node based on the real-time measurement data and the theoretical measurement data; The propagation weights of the error propagation topology network are analyzed to obtain the error propagation weight matrix; Calculate the node error propagation energy set based on the error propagation weight matrix and the error data set; The node error propagation energy set is mapped to the error propagation topology network to obtain the error propagation energy field.

5. The dynamic metering calibration method for an electricity meter as described in claim 1, characterized in that, The method for constructing an error propagation energy field model based on the aforementioned error propagation energy field includes: Define time evolution parameters, introduce an initial decay function, and perform spatial diffusion evolution of the error propagation energy field based on the time evolution parameters to obtain the error propagation diffusion evolution energy field; The energy field of error propagation and diffusion evolution is compared with the energy field of error propagation and diffusion theory. The decay coefficient of the decay function is iteratively optimized based on the evolution accuracy obtained from the comparison until an optimized decay function with an evolution accuracy reaching a preset threshold is obtained. The error propagation energy field model constructed based on the optimized decay function is based on the time-evolved error propagation energy evolution sequence.

6. The dynamic metering calibration method for an electricity meter as described in claim 1, characterized in that, If a calibration trigger time window is generated when the current evolutionary state is detected to be in a critical evolutionary state, the method also includes: Set the calibration time tolerance window; The calibration time tolerance window is used to extend the time node when the process is in a critical evolution state to obtain the calibration trigger time window.

7. The dynamic metering calibration method for an electricity meter as described in claim 6, characterized in that, Methods for setting the calibration time tolerance window include: Calculate the error evolution rate of change for each data type based on the error evolution feature vector; Based on the error evolution rate, the calibration sensitivity index for each data type is obtained; Obtain the current data source type corresponding to the calibration request, and set the size of the calibration time tolerance window inversely proportional to the calibration sensitivity index corresponding to the current data source type.

8. The dynamic metering calibration method for an electricity meter as described in claim 1, characterized in that, The error propagation energy field model is used to generate a time-evolution-based error propagation energy evolution sequence, wherein the evolution time window of the time evolution is larger than the initial calibration time window; The initial calibration time window is the initial calibration time window during which the energy meter can be used to perform metering calibration.

9. The dynamic metering calibration method for an electricity meter as described in claim 8, characterized in that, The calibration trigger time window is a sub-window of the initial calibration time window.

10. A dynamic metering calibration system for electricity meters, characterized in that, The step of implementing the dynamic metering calibration method for an electricity meter according to any one of claims 1 to 9, wherein the dynamic metering calibration system for the electricity meter comprises: The data acquisition component is used to acquire multi-source metering data from the electricity meter through the internal sampling module of the electricity meter. An error quantization component is used to construct an error propagation topology network based on the error influence relationship of the multi-source metering data of the electricity meter, and to quantify the error of each node in the node set based on the error propagation topology network to obtain the error propagation energy field. An evolution sequence generation component is used to construct an error propagation energy field model based on the error propagation energy field, generate a time-evolution-based error propagation energy evolution sequence using the error propagation energy field model, and extract the energy evolution feature vector of the error propagation energy evolution sequence. The metering calibration component is used to identify the current evolution state based on the energy evolution feature vector. If the current evolution state is detected to be in a critical evolution state, a calibration trigger time window is generated, and metering calibration is performed on the energy meter within the calibration trigger time window.

Images