Electric energy meter state perception method and device based on robust solution, equipment and medium

Abstract

This application relates to a robust solution-based method, apparatus, computer device, computer-readable storage medium, and computer program product for assessing the state of electricity meters. The method includes: acquiring total electricity meter data for a target area and sub-electricity meter data for each sub-electricity meter located within the target area; constructing an energy difference matrix based on the total electricity meter data and the sub-electricity meter data; performing variance stabilization processing on the energy difference matrix to obtain a target energy difference matrix; calculating the product of the sub-electricity value matrix constructed from the sub-electricity meter data and the energy deviation coefficient to be solved; constructing an objective function based on the difference between the product and the target energy difference matrix; obtaining a normal equation with the objective function as the goal; robustly solving the normal equation to determine the solved energy deviation coefficient; the solved energy deviation coefficient is used to assess the state of the total electricity meter and each sub-electricity meter, improving the effectiveness of electricity meter state assessment.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to a method, apparatus, computer device, computer-readable storage medium, and computer program product for sensing the state of an electricity meter based on robust solution. Background Technology

[0002] With the rapid development of computer technology, large-scale deployment of smart meters has emerged. This large-scale deployment enables distribution substations to continuously collect time-series data such as electricity consumption, voltage, and current. This data has significant application potential, including substation line loss calculation, energy balance verification, and meter status assessment. In related technologies, simple preprocessing of the meter readings is typically used to ensure basic data integrity. Then, an energy conservation equation is constructed on a substation basis, and the ordinary least squares method is used to solve for the meter-related deviation coefficients, such as meter error. Based on the solved meter error, the meter status is then predicted.

[0003] However, in related technologies, the solution of ordinary least squares is extremely sensitive to noise, the error estimation of electricity meters is prone to jitter and cannot be reproduced, making it difficult to accurately predict the error of electricity meters, resulting in poor performance in assessing the condition of electricity meters. Summary of the Invention

[0004] Therefore, it is necessary to provide a robust solution-based method, device, computer equipment, computer-readable storage medium, and computer program product for assessing the state of electricity meters, which can improve the effectiveness of electricity meter state evaluation.

[0005] Firstly, this application provides a robust solution-based method for sensing the state of an energy meter, including:

[0006] Obtain total energy meter data for the target area, as well as sub-energy meter data for each sub-energy meter located within the target area;

[0007] Based on the total energy meter data and the data of each sub-energy meter, an energy difference matrix is ​​constructed. After variance stabilization processing of the energy difference matrix, the target energy difference matrix is ​​obtained.

[0008] Calculate the product of the energy value matrix constructed from the data of each energy meter and the energy deviation coefficient to be solved. Based on the difference between the product and the target energy difference matrix, construct the objective function.

[0009] With the objective function as the goal, a normal equation is obtained. The normal equation is then robustly solved to determine the solved energy deviation coefficient. The solved energy deviation coefficient is used to evaluate the status of the total energy meter and each sub-energy meter.

[0010] Secondly, this application also provides a state sensing device for an energy meter based on robust solution, comprising:

[0011] The data acquisition module is used to acquire total energy meter data for the target area and sub-energy meter data for each sub-energy meter located within the target area;

[0012] The matrix determination module is used to construct an energy difference matrix based on the total energy meter data and the data of each sub-energy meter, and to obtain the target energy difference matrix after performing variance stabilization processing on the energy difference matrix.

[0013] The function construction module is used to calculate the product of the energy value matrix constructed from the data of each energy meter and the energy deviation coefficient to be solved, and to construct the objective function based on the difference between the product and the target energy difference matrix.

[0014] The coefficient solving module is used to obtain the normal equation with the goal of minimizing the objective function, perform robust solution on the normal equation, and determine the solved energy deviation coefficient. The solved energy deviation coefficient is used to evaluate the status of the total energy meter and each sub-energy meter.

[0015] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0016] Obtain total energy meter data for the target area, as well as sub-energy meter data for each sub-energy meter located within the target area;

[0017] Based on the total energy meter data and the data of each sub-energy meter, an energy difference matrix is ​​constructed. After variance stabilization processing of the energy difference matrix, the target energy difference matrix is ​​obtained.

[0018] Calculate the product of the energy value matrix constructed from the data of each energy meter and the energy deviation coefficient to be solved. Based on the difference between the product and the target energy difference matrix, construct the objective function.

[0019] With the objective function as the goal, a normal equation is obtained. The normal equation is then robustly solved to determine the solved energy deviation coefficient. The solved energy deviation coefficient is used to evaluate the status of the total energy meter and each sub-energy meter.

[0020] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the following steps:

[0021] Obtain total energy meter data for the target area, as well as sub-energy meter data for each sub-energy meter located within the target area;

[0022] Based on the total energy meter data and the data of each sub-energy meter, an energy difference matrix is ​​constructed. After variance stabilization processing of the energy difference matrix, the target energy difference matrix is ​​obtained.

[0023] Calculate the product of the energy value matrix constructed from the data of each energy meter and the energy deviation coefficient to be solved. Based on the difference between the product and the target energy difference matrix, construct the objective function.

[0024] With the objective function as the goal, a normal equation is obtained. The normal equation is then robustly solved to determine the solved energy deviation coefficient. The solved energy deviation coefficient is used to evaluate the status of the total energy meter and each sub-energy meter.

[0025] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, performs the following steps:

[0026] Obtain total energy meter data for the target area, as well as sub-energy meter data for each sub-energy meter located within the target area;

[0027] Based on the total energy meter data and the data of each sub-energy meter, an energy difference matrix is ​​constructed. After variance stabilization processing of the energy difference matrix, the target energy difference matrix is ​​obtained.

[0028] Calculate the product of the energy value matrix constructed from the data of each energy meter and the energy deviation coefficient to be solved. Based on the difference between the product and the target energy difference matrix, construct the objective function.

[0029] With the objective function as the goal, a normal equation is obtained. The normal equation is then robustly solved to determine the solved energy deviation coefficient. The solved energy deviation coefficient is used to evaluate the status of the total energy meter and each sub-energy meter.

[0030] The aforementioned robust solution-based method, device, computer equipment, computer-readable storage medium, and computer program product for sensing the state of electricity meters acquires total electricity meter data for a target area, as well as data from individual sub-meters within that area. Based on this data, an energy difference matrix is ​​constructed. After variance stabilization processing, a target energy difference matrix is ​​obtained to stabilize the variance and ensure its accuracy and effectiveness. Then, the product of the sub-energy value matrix constructed from the sub-meter data and the energy deviation coefficient to be solved is calculated. Based on the difference between this product and the target energy difference matrix, an objective function is constructed. By minimizing this objective function, a normal equation is obtained. Robustly solving this equation determines the solved energy deviation coefficients, ensuring their accuracy. This allows for a more accurate assessment of the state of the total electricity meter and each sub-meter, significantly improving the effectiveness of electricity meter state assessment. Attached Figure Description

[0031] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0032] Figure 1 This is an application environment diagram of a robust solution-based energy meter state sensing method in one embodiment;

[0033] Figure 2 This is a flowchart illustrating a robust solution-based energy meter state sensing method in one embodiment.

[0034] Figure 3 This is a structural block diagram of a power meter state sensing device based on robust solution in one embodiment;

[0035] Figure 4 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0036] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0037] It should be noted that the terms "first," "second," etc., used in this application can be used to describe various elements, but these elements are not limited by these terms. These terms are only used to distinguish the first element from the second element. The terms "comprising" and "having," and any variations thereof, used in this application, are intended to cover non-exclusive inclusion. The term "multiple" used in this application refers to two or more. The term "and / or" used in this application refers to one of the embodiments, or any combination of multiple embodiments.

[0038] The energy meter state sensing method based on robust solution provided in this application embodiment can be applied to, for example... Figure 1 In the application environment shown, terminal 102 communicates with server 104 via a network. A data storage system can store the data that server 104 needs to process. The data storage system can be integrated onto server 104, or it can be located in the cloud or on another network server. The robust solution-based electricity meter status sensing method provided in this application embodiment can be executed independently by terminal 102 or server 104, or it can be executed collaboratively by terminal 102 and server 104.

[0039] In some embodiments, after acquiring total energy meter data for the target area and sub-energy meter data for each sub-energy meter within the target area, terminal 102 sends the acquired total energy meter data for the target area and sub-energy meter data for each sub-energy meter within the target area to server 104. Server 104 constructs an energy difference matrix based on the total energy meter data and the sub-energy meter data. After performing variance stabilization processing on the energy difference matrix, a target energy difference matrix is ​​obtained. Server 104 calculates the product of the sub-energy value matrix constructed from the sub-energy meter data and the energy deviation coefficient to be solved. Based on the difference between the product and the target energy difference matrix, an objective function is constructed. Server 104 obtains a normal equation with the objective function as the goal, performs robust solution on the normal equation, and determines the solved energy deviation coefficient. The solved energy deviation coefficient is used to evaluate the state of the total energy meter and each sub-energy meter.

[0040] In other embodiments, the total energy meter and each sub-meter in the target area may communicate with the server 104. In this case, the total energy meter in the target area directly sends its data to the server 104, and each sub-meter directly sends its own data to the server 104. Then, the server 104 returns the energy difference matrix constructed based on the total energy meter data and the data from each sub-meter. After variance stabilization processing of the energy difference matrix, the target energy difference matrix is ​​obtained, and execution continues. Alternatively, in other embodiments, the total energy meter and each sub-energy meter in the target area may communicate with the terminal 102. After obtaining the total energy meter data of the total energy meter in the target area and the sub-energy meter data of each sub-energy meter located in the target area, the terminal 102 constructs an energy difference matrix based on the total energy meter data and the data of each sub-energy meter. After performing variance stabilization processing on the energy difference matrix, the target energy difference matrix is ​​obtained. The terminal 102 calculates the product of the sub-energy value matrix constructed from the data of each sub-energy meter and the energy deviation coefficient to be solved. Based on the difference between the product and the target energy difference matrix, an objective function is constructed. The terminal 102 obtains the normal equation with the goal of minimizing the objective function, performs robust solution on the normal equation, and determines the solved energy deviation coefficient. The solved energy deviation coefficient is used to evaluate the state of the total energy meter and each sub-energy meter.

[0041] Terminal 102 can be a device that communicates with the main electricity meter and the sub-electricity meters. Terminal 102 can be, but is not limited to, various personal computers, laptops, smartphones, tablets, etc. Server 104 can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides cloud computing services.

[0042] In one exemplary embodiment, such as Figure 2 As shown, a robust solution-based method for sensing the state of an energy meter is provided, which can be applied to computer equipment (which may be...). Figure 1 Terminal 102 in the middle can also be Figure 1 Taking server 104 as an example, the explanation includes the following steps S202 to S208. Wherein:

[0043] Step S202: Obtain the total energy meter data for the target area and the sub-energy meter data for each sub-energy meter located within the target area.

[0044] The main energy meter refers to the energy meter for the transformer substation where the target area is located. It can be considered the substation's master meter, installed on the low-voltage side of the distribution network, used to measure the total electricity consumption of the entire substation (i.e., all the electricity output from the transformer). In other words, it measures the overall energy value of the target area supplied by the substation. The main energy meter data reflects the energy value of the target area recorded by the main energy meter within a preset historical period. Individual energy meters are energy meters installed in each electricity consumption unit within the target area, i.e., installed on the user side (e.g., in a residential household). Each individual energy meter measures the energy value of its respective electricity consumption unit. The individual energy meter data reflects the energy value of the corresponding electricity consumption unit recorded by the corresponding individual energy meter within a preset historical period. The energy value can be considered as the amount of electricity consumed.

[0045] For example, after receiving a data processing request from an electricity meter, the computer device parses the request, determines the target area, and obtains the total electricity meter data for the target area, as well as the sub-electricity meter data for each sub-electricity meter located within the target area.

[0046] Step S204: Based on the total energy meter data and the data of each sub-energy meter, construct an energy difference matrix. After performing variance stabilization processing on the energy difference matrix, obtain the target energy difference matrix.

[0047] The energy difference matrix reflects the differences between the total energy table data and the sub-energy table data. Variance stabilization is used to stabilize the variance of the energy difference matrix. In some embodiments, variance stabilization can be achieved through maximum likelihood estimation, such as the Box-Cox transform method.

[0048] In some embodiments, the total energy meter data includes the historical total energy value for each historical time window; the sub-energy meter data includes the historical sub-energy value for each historical time window. The historical time window is obtained by dividing a preset historical period into segments, and the length of each historical time window may be equal or unequal. The sum of the lengths of all historical time windows equals the length of the preset historical period.

[0049] In view of this, in some embodiments, the sub-meter data includes the historical sub-meter values ​​for each historical time window. Based on the total meter data and the sub-meter data, an energy difference matrix is ​​constructed, including: for each historical time window, determining the energy difference corresponding to the historical time window based on the historical total energy value of the total meter in the corresponding historical time window and the historical sub-meter values ​​for each historical time window; and fusing the energy differences corresponding to each historical time window to obtain the energy difference matrix.

[0050] Among them, the energy difference can be understood as the difference in electrical energy.

[0051] For example, for each historical time window, the computer device obtains the historical total energy value of the historical time window from the total energy meter data and the historical partial energy value of the historical time window from the data of each sub-energy meter. The server determines the energy difference corresponding to each historical time window based on the historical total energy value of the total energy meter in the corresponding historical time window and the historical partial energy values ​​of each sub-energy meter in the corresponding historical time window; according to the order of the historical time windows, an energy difference matrix is ​​obtained based on the energy difference corresponding to each historical time window, with the energy difference corresponding to each historical time window in the same row of the energy difference matrix.

[0052] In the above embodiments, for each historical time window, the energy difference corresponding to the historical time window is determined based on the historical total energy value of the total energy meter in the corresponding historical time window and the historical partial energy value of each sub-energy meter in the corresponding historical time window; the energy difference values ​​corresponding to each historical time window are fused to obtain an energy difference matrix that reflects the difference between the historical total energy value and the historical partial energy value, ensuring the effectiveness of subsequent solutions.

[0053] In some embodiments, the energy difference corresponding to a historical time window is determined based on the historical total energy value of the total energy meter in the corresponding historical time window and the historical partial energy value of each sub-energy meter in the corresponding historical time window. This includes: for each historical time window, superimposing the historical partial energy values ​​of each sub-energy meter in the corresponding historical time window to obtain a superimposed value; and using the difference between the superimposed value and the historical total energy value of the total energy meter in the corresponding historical time window as the energy difference corresponding to the historical time window.

[0054] In view of this, for example, the master equation for energy conservation (time window k) is first constructed as shown in the following formula (1):

[0055] (1)

[0056] in, The total energy value is the historical total energy value of the total energy meter within time window k; The historical energy value of the i-th energy meter within time window k; Let be the sub-meter error (i.e., metering error) of the i-th sub-meter. r is the line loss coefficient (also known as the proportional loss term) that is proportional to the energy value of the total energy meter; f is the fixed loss (such as transformer no-load loss) that is independent of the load within the time window k. Among them, the energy deviation coefficient to be solved includes the sub-meter error, the line loss coefficient and the fixed loss. The sub-meter error, the line loss coefficient and the fixed loss are all unknown quantities to be solved. The sub-meter error, the line loss coefficient and the fixed loss, i.e. the unknown quantities, are moved to the left side, and the known quantities (historical total energy value, historical sub-energy value) are moved to the right side to obtain a form suitable for regression, as shown in the following formula (2):

[0057] (2)

[0058] Then, the N time windows are superimposed and matrix-assembled to obtain the following formula (3):

[0059] (3)

[0060] Where the unknown parameter vector is: The design matrix and right-hand side of the k-th row are shown in the following formulas (4)-(5):

[0061] (4)

[0062] (5)

[0063] It should be noted that for more refined line loss modeling, such as adding current squared related terms and line loss coefficients, one can... The corresponding loss pattern column and the coefficient to be estimated are added to the solution, while the solution process remains unchanged.

[0064] In view of this, based on formula (5), the energy difference corresponding to time window k can be calculated.

[0065] In the above embodiments, by superimposing the historical energy values ​​of each sub-energy meter in the corresponding historical time window, a superimposed value is obtained; the difference between the superimposed value and the historical total energy value of the total energy meter in the corresponding historical time window can accurately determine the energy difference (which can be understood as the energy difference) corresponding to the historical time window, so as to accurately assess the energy difference between the total energy meter and all sub-energy meters.

[0066] In some embodiments, before determining the energy difference matrix, the total energy meter data and the individual energy meter data can be preprocessed, and the energy difference matrix can be constructed based on the preprocessed total energy meter data and the individual energy meter data.

[0067] For example, firstly, time alignment is performed to unify the sampling interval and caliber (e.g., kWh). Clock differences spanning days or weeks are mapped to a unified time axis before aggregation, ensuring that historical time windows are consistent with corresponding electricity consumption periods. Next, missing and outlier data points are repaired, with short missing table code time-series data using local linear interpolation, i.e. .in, This is the interpolated electrical energy value, which can also be understood as the estimated value of the point to be interpolated. t is the time point where interpolation is needed, located at a known time point. and between, yes The electrical energy value, yes The electrical energy value.

[0068] Then, long missing values ​​can be smoothed using Kalman filtering; extreme spikes are limited using IQR (interquartile range) and Huber soft limiting (a limiting method that combines the characteristics of hard and soft limiting) to avoid biasing the regression. Next, topology verification is performed to correctly handle situations such as paralleling, reverse power supply, and meter consolidation, ensuring that the pre-processed total energy meter data and the data from each individual energy meter belong to the same distribution transformer area.

[0069] In some embodiments, after variance stabilization processing of the energy difference matrix, a target energy difference matrix is ​​obtained, including: acquiring a plurality of preset transformation parameters; for each transformation parameter, transforming each energy difference in the energy difference matrix according to a transformation function matching the transformation parameter to obtain a transformed energy difference; calculating the corresponding variance based on each transformed energy difference, and determining the log-likelihood value based on the variance and each transformed energy difference; selecting the transformation parameter corresponding to the largest log-likelihood value as the target transformation parameter, and determining the target energy difference matrix based on each transformed energy difference corresponding to the target transformation parameter.

[0070] The transformation parameters are the parameters used to perform the Box-Cox transformation. The transformation parameters can be zero or non-zero values. The transformation function is the function used to implement the Box-Cox transformation.

[0071] For example, the transformation of each energy difference under each transformation parameter is performed using the following formula (6):

[0072] (6)

[0073] in, Let be the transformation parameters. For each transformation parameter and each energy difference, if the transformation parameter is 0, then for that energy difference y, the transformed energy difference is... If the transformation parameter is not zero, then for the energy difference y, the transformed energy difference is... .

[0074] It should be noted that if the energy difference y is non-positive, the entire value is first shifted, that is, a positive number c is superimposed on y, and the transformation is performed based on the superimposed energy difference; if necessary, the Yeo-Johnson method (a power transformation method) is used to handle zero and non-negative values.

[0075] For each transformation parameter, after obtaining the energy difference after each transformation, let Calculate the corresponding variance: Ignoring the constant term, we perform log-likelihood calculation to obtain the log-likelihood value corresponding to the transformation parameters, as shown in the following formula (7):

[0076] (7)

[0077] Then, the transformation parameter corresponding to the largest log-likelihood value is selected as the target transformation parameter, i.e. Obtain the energy differences after each transformation based on the target transformation parameters, construct a matrix, and obtain the target energy difference matrix.

[0078] Of course, it can also be: using time rolling window estimation and EWMA (exponential weighted moving average) smoothing of the target transformation parameters to reduce time series jitter. When the sample is insufficient, grid search and nearest neighbor smoothing are used. It is determined whether the smoothed change parameter is zero. Based on the judgment result, the change of each energy difference is calculated according to formula (6) to obtain the changed energy difference, so as to obtain the target energy difference matrix.

[0079] In other embodiments, after determining the target transformation parameters, an intermediate energy difference matrix is ​​obtained based on the energy differences of each transformed value corresponding to the target transformation parameters. The intermediate energy difference matrix is ​​then subjected to consistency processing, and only the intermediate energy difference matrix and the residual are subjected to Box-Cox transformation again. Finally, the transformation effect is verified by Shapiro-Wilk normality test and Breusch-Pagan homoscedasticity test. After the verification is passed, the sentence obtained by the second transformation is used as the target energy difference matrix.

[0080] In the above embodiments, for each transformation parameter, each energy difference in the energy difference matrix is ​​transformed according to a transformation function that matches the transformation parameter to obtain the transformed energy difference; based on each transformed energy difference, the corresponding variance is calculated; based on the variance and each transformed energy difference, the log-likelihood value is determined; the transformation parameter corresponding to the largest log-likelihood value is selected as the target transformation parameter; based on each transformed energy difference corresponding to the target transformation parameter, the target energy difference matrix is ​​determined. In this way, skewness and heteroscedasticity can be weakened, making the error closer to the assumption of "zero mean and near-constant variance", thereby improving the robustness and threshold controllability of the subsequent solution process.

[0081] Step S206: Calculate the product of the energy value matrix constructed from the data of each energy meter and the energy deviation coefficient to be solved. Based on the difference between the product and the target energy difference matrix, construct the objective function.

[0082] For example, the steps for determining the sub-energy value matrix are as follows: For each sub-energy meter, obtain the historical energy value of each historical time window from the corresponding sub-energy meter data. Based on the historical energy value of each sub-energy meter in each historical time window, construct the sub-energy value matrix. For example, refer to formulas (1)-(4) above to obtain the sub-energy value matrix. .

[0083] For example, the sub-meter error of each sub-meter and the line loss coefficient corresponding to the total energy meter, which are to be solved, are as described above: .

[0084] For example, a basic model is constructed based on the partial energy value matrix, the energy deviation coefficient to be solved, and the target energy difference matrix, as shown in the following formula (8):

[0085] (8)

[0086] in, For the target energy difference matrix, For the energy value matrix, The energy deviation coefficient to be solved is... The model parameters of the basic model follow a normal distribution. The weight matrix is, i.e. ;in, ;in, , These are confidence factors for equipment level and communication quality, respectively, ranging from 0 to 1. This is an estimate of the variance within a historical time window k. During fault replacement, planned maintenance, and power outage periods, the corresponding historical time window can be reduced or directly blocked.

[0087] Therefore, based on the basic model, the objective function is determined through regularization, as shown in the following formula (9):

[0088] (9)

[0089] in, This is the standard Ridge regression; if it is necessary to encourage parameter smoothing between adjacent branches or similar devices, then... It is a first-order difference (or graph Laplace) operator; For function coefficients, And control the bias-variance tradeoff to avoid As the condition number increases further, the bias increases slightly but the variance decreases significantly, and the condition number decreases. Data-driven automatic selection can be achieved using the following two methods: 1) in a logarithmic grid Search for the smallest GCV ;2) In Select the bend point on the curve as In actual engineering, the results of the two calculation methods are often close. The more conservative one can be used, and cross-validation with leave one method is necessary when needed.

[0090] Step S208: With the objective function as the goal, the normal equation is obtained. The normal equation is solved robustly to determine the solved energy deviation coefficient. The solved energy deviation coefficient is used to evaluate the status of the total energy meter and each sub-energy meter.

[0091] For example, the derivative of the objective function is calculated, i.e., the gradient is obtained, resulting in the calculated function. Setting the calculated function to 0 yields the normal equation. For example: The normal equation is obtained as follows: ;in, ; .in, It is a weighted and regularized Hessian matrix (or the kernel matrix of the generalized inverse), and is a symmetric positive definite matrix. It is a weighted projection of the data.

[0092] For example, after determining the normal equation, the Neumann approximation inverse and equivalent iteration are first performed. Specifically: take Greater than the largest eigenvalue, and defined : , ;

[0093] Therefore, the following can be satisfied: , To find the truncated Neumann approximate inverse (Kth order) and approximate solution, refer to the following formulas (10)-(11):

[0094] (10)

[0095] (11)

[0096] The equivalent non-power iteration, that is, using the following formulas (12)-(13) to achieve multiple iterations, is more efficient and facilitates early stopping:

[0097] (12)

[0098] (13)

[0099] The iteration stops when the stopping criterion shown in formula (14) is met:

[0100] (14)

[0101] Will meet the stopping criteria As a solved energy deviation coefficient, the solved energy deviation coefficient is used to sense (assess) the status of the main energy meter and each sub-meter. Based on engineering application experience, it can be estimated using power iteration. This ensures convergence while avoiding excessively slow convergence.

[0102] In other embodiments, the method further includes: as k increases, the residual norm decreases while the solution norm increases. When the solution norm increases much faster than the residual norm decreases, it indicates that noise is being fitted—the inflection point of the iterative L-curve at this point is the optimal K*, i.e., the following formulas (15)-(16):

[0103] (15)

[0104] (16)

[0105] Engineering implementation can be achieved when the unit iteration revenue threshold is reached. Stop at time; relative change threshold Alternatively, a small holdout set can be set, and the K value with the smallest verification error can be selected. Usually, K∈[0,8] can satisfy the trade-off between accuracy and stability.

[0106] To facilitate the construction of statistical thresholds, an upper bound on the covariance is set: Therefore, the energy deviation coefficient corresponding to the optimal K* is output as the solved energy deviation coefficient. At this time, the table error is output. Point estimates and variance calibers for line loss coefficient r and fixed loss f; snapshots of suspected and traceable parameters by transformer area. Then, statistical thresholds are set. ,in This is the quantile of the standard normal distribution. The engineering threshold is superimposed with historical quantiles (e.g., P95 / P99) and equipment level conditions, and a technical alarm is only triggered if the threshold is exceeded for m consecutive windows. For example: A series of m historical time windows satisfy .

[0107] As a result, the on-site verification results are fed back to the time trajectories of λ*, α, and K (the reliability can be adjusted up and down), triggering recalibration and retraining conditions, thus forming a long-term stable adaptive system.

[0108] It should be noted that the k mentioned in this step is an order in the order K, not the historical time window k mentioned earlier. The two have different meanings. For ease of distinction, the order K mentioned in this step can be replaced with order P, and the k mentioned in this step can be replaced with p.

[0109] It should be noted that, in the robust solution-based energy meter state sensing method involved in this application, on the data distribution side, Box–Cox(λ) is used to transform skewed and heteroscedastic residuals into near-Gaussian homoscedasticity, and robust limiting and weight W are superimposed to suppress outlier effects; secondly, on the numerical ill-conditioned side, Tikhonov(α) (the process of determining the objective function mentioned above) is used to significantly reduce The condition number, combined with the Neumann approximate inverse (β,K), replaces direct inversion with controllable early stopping and provides a convergence bound, avoiding solution jitter and noise amplification. Thirdly, on the engineering practicality side, an automatic selection and reinjection mechanism for λ, α, and K is established, forming an interpretable, traceable, and recalcible end-to-end process. This achieves stable estimation of low-voltage meter errors, line loss coefficients, and fixed losses, improving alarm accuracy and result consistency, and maintaining convergence and robustness under multiple operating conditions, complex topologies, and reverse power transmission scenarios.

[0110] It should be noted that in the above process, the Box-Cox algorithm (including the rolling MLE selection and positive domain guarantee with λ) stabilizes the variance, and then uses the biased Tikhonov solution with Neumann series truncation to estimate the meter error and the threshold criterion in energy balance. The parameterization of the Neumann approximate inverse is implemented, and the algorithm framework and its engineering implementation details are shown, including spectral radius constraints, truncation order, and K-controlling numerical stability and estimation bias. A detection closed loop is constructed, forming a discrimination and reinjection mechanism based on the dual-domain residual inspection in the transform domain / original domain and the linkage of suspicious meter technology thresholds.

[0111] In the robust solution-based energy meter state sensing method described above, total energy meter data and individual energy meter data for each sub-meter within the target area are acquired. Based on these data, an energy difference matrix is ​​constructed. After variance stabilization processing, a target energy difference matrix is ​​obtained to stabilize the variance and ensure its accuracy and effectiveness. Then, the product of the sub-energy value matrix constructed from the individual energy meter data and the energy deviation coefficient to be solved is calculated. Based on the difference between this product and the target energy difference matrix, an objective function is constructed. By minimizing the objective function, a normal equation is obtained. Robust solution to this normal equation determines the solved energy deviation coefficients, ensuring their accuracy. Therefore, based on these solved energy deviation coefficients, the state of the total energy meter and each sub-meter can be more accurately assessed, significantly improving the effectiveness of energy meter state assessment.

[0112] In some embodiments, the solved energy deviation coefficient includes the sub-meter error of each sub-meter and the line loss coefficient corresponding to the total energy meter. The method further includes: verifying whether the total energy meter is in an abnormal state based on the line loss coefficient and a preset line loss coefficient threshold; and verifying whether the sub-meter is in an abnormal state based on the sub-meter error and a preset error threshold for each sub-meter.

[0113] For example, if the line loss coefficient is greater than the line loss coefficient threshold, the total energy meter is determined to be in an abnormal state; if the line loss coefficient is less than or equal to the line loss coefficient threshold, the total energy meter is determined to be in a normal state. For example, when the total energy meter is in an abnormal state, the difference between the line loss coefficient and the line loss coefficient threshold is calculated. Based on the difference, the abnormality level is determined, and an alarm message matching the abnormality level is generated to implement corresponding handling measures. The larger the difference, the higher the abnormality level.

[0114] For example, for the sub-meter error, if the sub-meter error is greater than a preset error threshold, the sub-meter is determined to be in an abnormal state; if the sub-meter error is less than or equal to the preset error threshold, the sub-meter is determined to be in a normal state. For example, when a sub-meter is in an abnormal state, the difference between the sub-meter error and the error threshold is calculated. Based on the difference, the abnormality level is determined, and an alarm message matching the abnormality level is generated for that sub-meter, so that corresponding handling measures can be taken. The larger the difference, the higher the abnormality level.

[0115] In this embodiment, based on the sub-meter error of each sub-meter in the solved energy deviation coefficient and the line loss coefficient corresponding to the total energy meter, the status of the total energy meter or each sub-meter can be accurately evaluated, which greatly improves the evaluation effect of the energy meter status.

[0116] In some embodiments, the solved energy deviation coefficient includes a fixed loss coefficient, a line loss coefficient corresponding to the total energy meter, and the sub-meter error of each sub-meter. The method further includes: obtaining the target sub-energy value of each sub-meter within the target time period; for each sub-meter, determining the corresponding corrected sub-energy value based on the target sub-energy value and the corresponding sub-meter error, and superimposing the corrected sub-energy values ​​of each sub-meter to obtain a superimposed value; and predicting the energy value of the total energy meter within the target time period based on the superimposed value, the fixed loss coefficient, and the line loss coefficient.

[0117] The target period is the future period.

[0118] For example, after obtaining the target energy values ​​of each sub-meter within the target time period, for each sub-meter, calculate the first difference between the unit value (e.g., value 1) and the sub-meter error, calculate the product of the target energy value of the sub-meter and the first difference to obtain the corresponding corrected energy value, and superimpose the corrected energy values ​​of each sub-meter to obtain the superimposed value. Add the superimposed value to the fixed loss coefficient to obtain the summed value. Calculate the second difference between the unit value and the line loss coefficient, and use the ratio of the summed value to the second difference as the energy value of the total energy meter within the target time period. For example, the target energy values ​​of each sub-meter within the target time period, the sub-meter errors of each sub-meter included in the solved energy deviation coefficient, the line loss coefficient, and the fixed loss coefficient can be substituted into formula (1) to calculate the energy value of the total energy meter within the target time period.

[0119] In this embodiment, after solving for the sub-meter error, line loss coefficient, and fixed loss coefficient of each sub-meter, it is possible to accurately predict the total energy value of the meter for any future target time period.

[0120] In one specific embodiment, the specific steps are as follows:

[0121] First, the computer equipment acquires total energy meter data for the target area, as well as sub-energy meter data for each sub-energy meter located within the target area. The total energy meter data includes historical total energy values ​​for each historical time window; the sub-energy meter data includes historical sub-energy values ​​for each historical time window.

[0122] Secondly, for each historical time window, the computer equipment superimposes the historical energy values ​​of each sub-energy meter in the corresponding historical time window to obtain the superimposed value; the difference between the superimposed value and the historical total energy value of the total energy meter in the corresponding historical time window is used as the energy difference value corresponding to the historical time window; the energy difference values ​​corresponding to each historical time window are merged to obtain the energy difference matrix.

[0123] Then, the computer device acquires multiple preset transformation parameters; for each transformation parameter, it transforms each energy difference in the energy difference matrix according to the transformation function that matches the transformation parameter to obtain the transformed energy difference; based on each transformed energy difference, it calculates the corresponding variance; based on the variance and each transformed energy difference, it determines the log-likelihood value; it selects the transformation parameter corresponding to the largest log-likelihood value as the target transformation parameter; based on each transformed energy difference corresponding to the target transformation parameter, it determines the target energy difference matrix.

[0124] Then, the computer equipment calculates the product of the energy value matrix constructed from the data of each sub-meter and the energy deviation coefficient to be solved. Based on the difference between the product and the target energy difference matrix, an objective function is constructed. The normal equation is obtained by minimizing the objective function. A robust solution to the normal equation is performed to determine the solved energy deviation coefficient. The solved energy deviation coefficient is used to evaluate the status of the total energy meter and each sub-meter. The solved energy deviation coefficient includes the line loss coefficient and fixed loss coefficient corresponding to the total energy meter, and the sub-meter error of each sub-meter.

[0125] Finally, the computer equipment verifies whether the main energy meter is in an abnormal state based on the line loss coefficient and a preset line loss coefficient threshold; for the sub-meter error of each sub-meter, it verifies whether the sub-meter is in an abnormal state based on the sub-meter error and a preset error threshold. Alternatively: the computer equipment obtains the target energy value of each sub-meter within the target time period; for each sub-meter, based on the target energy value and the corresponding sub-meter error, it determines the corresponding corrected energy value, and adds the corrected energy values ​​of each sub-meter to obtain a superimposed value; based on the superimposed value, the fixed loss coefficient, and the line loss coefficient, it predicts the energy value of the main energy meter within the target time period.

[0126] In the above embodiments, total energy meter data for the target area and data for each sub-energy meter within the target area are acquired. Based on the total energy meter data and the data for each sub-energy meter, an energy difference matrix is ​​constructed. After variance stabilization processing of the energy difference matrix, a target energy difference matrix is ​​obtained to stabilize the variance and ensure the accuracy and effectiveness of the target energy difference matrix. Then, the product of the sub-energy value matrix constructed from the data of each sub-energy meter and the energy deviation coefficient to be solved is calculated. Based on the difference between the product and the target energy difference matrix, an objective function is constructed. Thus, by minimizing the objective function, a normal equation is obtained. Robustly solving the normal equation determines the solved energy deviation coefficient, ensuring the accuracy of the energy deviation coefficient solution. Therefore, based on the solved energy deviation coefficient, the state of the total energy meter and each sub-energy meter can be more accurately assessed, greatly improving the effectiveness of energy meter state assessment.

[0127] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages in other steps. It is understood that the steps in different embodiments can be freely combined as needed, and all non-contradictory solutions formed by such combinations are within the scope of protection of this application.

[0128] Based on the same inventive concept, this application also provides a robust solution-based energy meter state sensing device for implementing the robust solution-based energy meter state sensing method described above. The solution provided by this device is similar to the solution described in the above method. Therefore, the specific limitations of one or more robust solution-based energy meter state sensing device embodiments provided below can be found in the limitations of the robust solution-based energy meter state sensing method described above, and will not be repeated here.

[0129] In one exemplary embodiment, such as Figure 3 As shown, a robust solution-based energy meter state sensing device 300 is provided, comprising: a data acquisition module 302, a matrix determination module 304, a function construction module 306, and a coefficient solving module 308, wherein:

[0130] The data acquisition module 302 is used to acquire total energy meter data of the total energy meter of the target area, as well as the sub-energy meter data of each sub-energy meter located in the target area;

[0131] The matrix determination module 304 is used to construct an energy difference matrix based on the total energy meter data and the data of each sub-energy meter, and to obtain the target energy difference matrix after performing variance stabilization processing on the energy difference matrix.

[0132] The function construction module 306 is used to calculate the product of the energy value matrix constructed from the data of each energy meter and the energy deviation coefficient to be solved, and to construct the objective function based on the difference between the product and the target energy difference matrix.

[0133] The coefficient solving module 308 is used to obtain the normal equation with the goal of minimizing the objective function, perform robust solution on the normal equation, and determine the solved energy deviation coefficient. The solved energy deviation coefficient is used to evaluate the status of the total energy meter and each sub-energy meter.

[0134] In some embodiments, the total energy meter data includes the historical total energy value of each historical time window; the sub-energy meter data includes the historical sub-energy value of each historical time window; the matrix determination module 304 is used to determine the energy difference corresponding to each historical time window based on the historical total energy value of the total energy meter in the corresponding historical time window and the historical sub-energy value of each sub-energy meter in the corresponding historical time window for each historical time window; and to fuse the energy differences corresponding to each historical time window to obtain an energy difference matrix.

[0135] In some embodiments, the matrix determination module 304 is used to superimpose the historical energy values ​​of each sub-energy meter in the corresponding historical time window for each historical time window to obtain a superimposed value; and to use the difference between the superimposed value and the historical total energy value of the total energy meter in the corresponding historical time window as the energy difference value corresponding to the historical time window.

[0136] In some embodiments, the matrix determination module 304 is used to obtain a plurality of preset transformation parameters; for each transformation parameter, transform each energy difference in the energy difference matrix according to a transformation function that matches the transformation parameter to obtain a transformed energy difference; calculate the corresponding variance based on each transformed energy difference; determine the log-likelihood value based on the variance and each transformed energy difference; select the transformation parameter corresponding to the largest log-likelihood value as the target transformation parameter; and determine the target energy difference matrix based on each transformed energy difference corresponding to the target transformation parameter.

[0137] In some embodiments, the solved energy deviation coefficient includes the sub-meter error of each sub-meter and the line loss coefficient corresponding to the total energy meter. The device also includes a state determination module, which is used to verify whether the total energy meter is in an abnormal state based on the line loss coefficient and a preset line loss coefficient threshold; and for the sub-meter error of each sub-meter, to verify whether the sub-meter is in an abnormal state based on the sub-meter error and a preset error threshold.

[0138] In some embodiments, the solved energy deviation coefficient includes the line loss coefficient, fixed loss coefficient, and sub-meter error of each sub-meter corresponding to the total energy meter. The device also includes an energy prediction module for obtaining the target energy value of each sub-meter within the target time period; for each sub-meter, based on the target energy value and the corresponding sub-meter error, determining the corresponding corrected energy value, and superimposing the corrected energy values ​​of each sub-meter to obtain a superimposed value; based on the superimposed value, the fixed loss coefficient, and the line loss coefficient, predicting the energy value of the total energy meter within the target time period.

[0139] The modules in the aforementioned robust solution-based energy meter state sensing device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of a computer device in software form, so that the processor can call and execute the corresponding operations of each module.

[0140] In one exemplary embodiment, a computer device is provided, which may be a server or a terminal, and its internal structure diagram may be as follows. Figure 4 As shown, this computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs stored in the non-volatile storage media. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communicating with external terminals via a network connection. When the computer program is executed by the processor, it implements a robust solution-based method for sensing the state of an energy meter.

[0141] Those skilled in the art will understand that Figure 4 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0142] In an exemplary embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to perform the following steps: acquiring total energy meter data for a target area and sub-energy meter data for each sub-energy meter located within the target area; constructing an energy difference matrix based on the total energy meter data and the sub-energy meter data, and obtaining a target energy difference matrix after variance stabilization processing of the energy difference matrix; calculating the product of the sub-energy value matrix constructed from the sub-energy meter data and the energy deviation coefficient to be solved, and constructing an objective function based on the difference between the product and the target energy difference matrix; obtaining a normal equation with the objective function as the goal, robustly solving the normal equation, and determining the solved energy deviation coefficient, which is used to evaluate the state of the total energy meter and each sub-energy meter.

[0143] In one embodiment, the total energy meter data includes the historical total energy value for each historical time window; the sub-energy meter data includes the historical sub-energy value for each historical time window. When the processor executes the computer program, it also implements the following steps: for each historical time window, based on the historical total energy value of the total energy meter in the corresponding historical time window and the historical sub-energy value of each sub-energy meter in the corresponding historical time window, the energy difference value corresponding to the historical time window is determined; the energy difference values ​​corresponding to each historical time window are fused to obtain an energy difference matrix.

[0144] In one embodiment, when the processor executes the computer program, it further implements the following steps: for each historical time window, superimpose the historical energy values ​​of each sub-energy meter in the corresponding historical time window to obtain a superimposed value; and use the difference between the superimposed value and the historical total energy value of the total energy meter in the corresponding historical time window as the energy difference value corresponding to the historical time window.

[0145] In one embodiment, when the processor executes the computer program, it further performs the following steps: obtaining a plurality of preset transformation parameters; for each transformation parameter, transforming each energy difference in the energy difference matrix according to a transformation function that matches the transformation parameter to obtain a transformed energy difference; calculating the corresponding variance based on each transformed energy difference, and determining the log-likelihood value based on the variance and each transformed energy difference; selecting the transformation parameter corresponding to the largest log-likelihood value as the target transformation parameter, and determining the target energy difference matrix based on each transformed energy difference corresponding to the target transformation parameter.

[0146] In one embodiment, the solved energy deviation coefficient includes the sub-meter error of each sub-meter and the line loss coefficient corresponding to the total energy meter. When the processor executes the computer program, it also implements the following steps: based on the line loss coefficient and a preset line loss coefficient threshold, it verifies whether the total energy meter is in an abnormal state; for the sub-meter error of each sub-meter, based on the sub-meter error and a preset error threshold, it verifies whether the sub-meter is in an abnormal state.

[0147] In one embodiment, the solved energy deviation coefficient includes the line loss coefficient, fixed loss coefficient, and sub-meter error of each sub-meter corresponding to the total energy meter. When the processor executes the computer program, it also performs the following steps: obtaining the target energy value of each sub-meter within the target time period; for each sub-meter, determining the corresponding corrected energy value based on the target energy value and the corresponding sub-meter error, and superimposing the corrected energy values ​​of each sub-meter to obtain the superimposed value; and predicting the energy value of the total energy meter within the target time period based on the superimposed value, the fixed loss coefficient, and the line loss coefficient.

[0148] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When executed by a processor, the computer program performs the following steps: acquiring total energy meter data for a target area and sub-energy meter data for each sub-energy meter located within the target area; constructing an energy difference matrix based on the total energy meter data and the sub-energy meter data, and obtaining a target energy difference matrix after variance stabilization processing of the energy difference matrix; calculating the product of the sub-energy value matrix constructed from the sub-energy meter data and the energy deviation coefficient to be solved, and constructing an objective function based on the difference between the product and the target energy difference matrix; obtaining a normal equation with the objective function as the goal, performing robust solution on the normal equation, and determining the solved energy deviation coefficient, which is used to evaluate the state of the total energy meter and each sub-energy meter.

[0149] In one embodiment, the total energy meter data includes the historical total energy value for each historical time window; the sub-energy meter data includes the historical sub-energy value for each historical time window. When the computer program is executed by the processor, it further implements the following steps: for each historical time window, based on the historical total energy value of the total energy meter in the corresponding historical time window and the historical sub-energy value of each sub-energy meter in the corresponding historical time window, the energy difference value corresponding to the historical time window is determined; the energy difference values ​​corresponding to each historical time window are fused to obtain an energy difference matrix.

[0150] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: for each historical time window, superimpose the historical energy values ​​of each sub-energy meter in the corresponding historical time window to obtain a superimposed value; and use the difference between the superimposed value and the historical total energy value of the total energy meter in the corresponding historical time window as the energy difference value corresponding to the historical time window.

[0151] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: obtaining a plurality of preset transformation parameters; for each transformation parameter, transforming each energy difference in the energy difference matrix according to a transformation function matching the transformation parameter to obtain a transformed energy difference; calculating the corresponding variance based on each transformed energy difference, and determining the log-likelihood value based on the variance and each transformed energy difference; selecting the transformation parameter corresponding to the largest log-likelihood value as the target transformation parameter, and determining the target energy difference matrix based on each transformed energy difference corresponding to the target transformation parameter.

[0152] In one embodiment, the solved energy deviation coefficient includes the sub-meter error of each sub-meter and the line loss coefficient corresponding to the total energy meter. When the computer program is executed by the processor, it also performs the following steps: based on the line loss coefficient and a preset line loss coefficient threshold, verify whether the total energy meter is in an abnormal state; for the sub-meter error of each sub-meter, based on the sub-meter error and a preset error threshold, verify whether the sub-meter is in an abnormal state.

[0153] In one embodiment, the solved energy deviation coefficient includes the line loss coefficient, fixed loss coefficient, and sub-meter error of each sub-meter corresponding to the total energy meter. When the computer program is executed by the processor, it also performs the following steps: obtaining the target energy value of each sub-meter within the target time period; for each sub-meter, determining the corresponding corrected energy value based on the target energy value and the corresponding sub-meter error, and superimposing the corrected energy values ​​of each sub-meter to obtain the superimposed value; and predicting the energy value of the total energy meter within the target time period based on the superimposed value, the fixed loss coefficient, and the line loss coefficient.

[0154] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, performs the following steps: acquiring total energy meter data for a target area and sub-energy meter data for each sub-energy meter located within the target area; constructing an energy difference matrix based on the total energy meter data and the sub-energy meter data, and obtaining a target energy difference matrix after variance stabilization processing of the energy difference matrix; calculating the product of the sub-energy value matrix constructed from the sub-energy meter data and the energy deviation coefficient to be solved, and constructing an objective function based on the difference between the product and the target energy difference matrix; obtaining a normal equation with the objective function as the goal, performing robust solution to the normal equation, and determining the solved energy deviation coefficient, which is used to evaluate the state of the total energy meter and each sub-energy meter.

[0155] In one embodiment, the total energy meter data includes the historical total energy value for each historical time window; the sub-energy meter data includes the historical sub-energy value for each historical time window. When the computer program is executed by the processor, it further implements the following steps: for each historical time window, based on the historical total energy value of the total energy meter in the corresponding historical time window and the historical sub-energy value of each sub-energy meter in the corresponding historical time window, the energy difference value corresponding to the historical time window is determined; the energy difference values ​​corresponding to each historical time window are fused to obtain an energy difference matrix.

[0156] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: for each historical time window, superimpose the historical energy values ​​of each sub-energy meter in the corresponding historical time window to obtain a superimposed value; and use the difference between the superimposed value and the historical total energy value of the total energy meter in the corresponding historical time window as the energy difference value corresponding to the historical time window.

[0157] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: obtaining a plurality of preset transformation parameters; for each transformation parameter, transforming each energy difference in the energy difference matrix according to a transformation function matching the transformation parameter to obtain a transformed energy difference; calculating the corresponding variance based on each transformed energy difference, and determining the log-likelihood value based on the variance and each transformed energy difference; selecting the transformation parameter corresponding to the largest log-likelihood value as the target transformation parameter, and determining the target energy difference matrix based on each transformed energy difference corresponding to the target transformation parameter.

[0158] In one embodiment, the solved energy deviation coefficient includes the sub-meter error of each sub-meter and the line loss coefficient corresponding to the total energy meter. When the computer program is executed by the processor, it also performs the following steps: based on the line loss coefficient and a preset line loss coefficient threshold, verify whether the total energy meter is in an abnormal state; for the sub-meter error of each sub-meter, based on the sub-meter error and a preset error threshold, verify whether the sub-meter is in an abnormal state.

[0159] In one embodiment, the solved energy deviation coefficient includes the line loss coefficient, fixed loss coefficient, and sub-meter error of each sub-meter corresponding to the total energy meter. When the computer program is executed by the processor, it also performs the following steps: obtaining the target energy value of each sub-meter within the target time period; for each sub-meter, determining the corresponding corrected energy value based on the target energy value and the corresponding sub-meter error, and superimposing the corrected energy values ​​of each sub-meter to obtain the superimposed value; and predicting the energy value of the total energy meter within the target time period based on the superimposed value, the fixed loss coefficient, and the line loss coefficient.

[0160] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0161] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0162] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0163] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for sensing the state of an energy meter based on robust solution, characterized in that, The method includes: Obtain total energy meter data for the target area, as well as sub-energy meter data for each sub-energy meter located within the target area; Based on the total energy meter data and the data of each sub-energy meter, an energy difference matrix is ​​constructed. After variance stabilization processing of the energy difference matrix, the target energy difference matrix is ​​obtained. Calculate the product of the energy value matrix constructed from the data of each energy meter and the energy deviation coefficient to be solved. Based on the difference between the product and the target energy difference matrix, construct the objective function. With the objective function as the goal, a normal equation is obtained. The normal equation is then robustly solved to determine the solved energy deviation coefficient. The solved energy deviation coefficient is used to evaluate the status of the total energy meter and each sub-energy meter.

2. The method according to claim 1, characterized in that, The total energy meter data includes historical total energy values ​​for each historical time window; the sub-energy meter data includes historical sub-energy values ​​for each historical time window; the step of constructing an energy difference matrix based on the total energy meter data and the sub-energy meter data includes: For each historical time window, the energy difference corresponding to the historical time window is determined based on the historical total energy value of the total energy meter in the corresponding historical time window and the historical partial energy value of each sub-energy meter in the corresponding historical time window. By integrating the energy differences corresponding to each historical time window, an energy difference matrix is ​​obtained.

3. The method according to claim 2, characterized in that, The determination of the energy difference value corresponding to the historical time window based on the historical total energy value of the total energy meter in the corresponding historical time window and the historical partial energy value of each sub-energy meter in the corresponding historical time window includes: For each historical time window, the historical energy values ​​of each energy meter in the corresponding historical time window are superimposed to obtain the superimposed value; The difference between the superimposed value and the historical total energy value of the total energy meter in the corresponding historical time window is used to determine the energy difference value corresponding to the historical time window.

4. The method according to claim 1, characterized in that, The process of performing variance stabilization on the energy difference matrix to obtain the target energy difference matrix includes: Obtain multiple preset transformation parameters; For each transformation parameter, each energy difference in the energy difference matrix is ​​transformed according to a transformation function that matches the transformation parameter to obtain the transformed energy difference. Based on the energy difference after each transformation, the corresponding variance is calculated, and based on the variance and the energy difference after each transformation, the log-likelihood value is determined. The transformation parameter corresponding to the largest log-likelihood value is selected as the target transformation parameter. Based on the energy difference values ​​after each transformation corresponding to the target transformation parameter, the target energy difference matrix is ​​determined.

5. The method according to any one of claims 1 to 4, characterized in that, The solved energy deviation coefficient includes the sub-meter error of each sub-meter and the line loss coefficient corresponding to the total energy meter. The method also includes: Based on the line loss coefficient and the preset line loss coefficient threshold, verify whether the total energy meter is in an abnormal state; For the sub-meter error, based on the sub-meter error and a preset error threshold, it is verified whether the sub-meter is in an abnormal state.

6. The method according to any one of claims 1 to 4, characterized in that, The solved energy deviation coefficient includes the line loss coefficient, fixed loss coefficient, and sub-meter error of each sub-meter corresponding to the total energy meter. The method further includes: Obtain the target energy value of each sub-meter within the target time period; For each sub-meter, based on the target sub-meter value and the corresponding sub-meter error, a corresponding corrected sub-meter value is determined, and the corrected sub-meter values ​​of each sub-meter are superimposed to obtain the superimposed value. Based on the superimposed value, the fixed loss coefficient, and the line loss coefficient, the total energy value of the electricity meter within the target time period is predicted.

7. A state sensing device for an electricity meter based on robust solution, characterized in that, The device includes: The data acquisition module is used to acquire total energy meter data for the target area and sub-energy meter data for each sub-energy meter located within the target area; The matrix determination module is used to construct an energy difference matrix based on the total energy meter data and the data of each sub-energy meter, and to obtain the target energy difference matrix after performing variance stabilization processing on the energy difference matrix. The function construction module is used to calculate the product of the energy value matrix constructed from the data of each energy meter and the energy deviation coefficient to be solved, and to construct the objective function based on the difference between the product and the target energy difference matrix. The coefficient solving module is used to obtain the normal equation with the goal of minimizing the objective function, perform robust solution on the normal equation, and determine the solved energy deviation coefficient. The solved energy deviation coefficient is used to evaluate the status of the total energy meter and each sub-energy meter.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images