Decoupling calculation method, device and equipment for distribution area line loss influencing factors

Abstract

The application relates to the technical field of power system data analysis, and provides a power distribution substation line loss influencing factor decoupling calculation method, device and equipment. The method obtains historical electrical quantity data and real-time electrical quantity data, performs data identification, missing value completion and multi-sampling rate time synchronization alignment processing on the real-time electrical quantity data, obtains data configuration and real-time operation data, determines a calculation loss initial value through the data configuration, then constructs a third-order symmetric coupling coefficient matrix to quantize multi-factor interaction effects, and iteratively calculates a target function combined with multiple constraints to obtain each factor loss component meeting a convergence condition, the calculation precision of additional loss is improved, fine calculation of substation additional loss is realized, data basis is provided for quantitative decision-making of targeted loss reduction, and the technical problems of insufficient calculation precision and poor adaptability of the existing fine calculation method of low-voltage substation additional loss are solved.

Description

Technical Field

[0001] This application relates to the field of power system data analysis technology, and in particular to a decoupling calculation method, apparatus and equipment for the influencing factors of distribution substation line loss. Background Technology

[0002] In the context of building a new power system, the lean management of line losses in low-voltage distribution transformer areas has become a core technical means to measure the economic efficiency of distribution network operation and support precise loss reduction decisions. The accuracy of low-voltage distribution transformer area line loss calculation fundamentally depends on the accurate characterization of three-phase imbalance, harmonic distortion, and reactive power deficiency, as well as their coupled interactive effects. According to current research statistics, the additional losses generated by the coupled interaction of these three factors can account for 30% to 50% of the total line loss. With the large-scale connection of distributed photovoltaic power, electric vehicle charging piles, and variable frequency household loads to the distribution network, the coupled superposition of these three factors—three-phase imbalance, harmonic distortion, and reactive power deficiency—is becoming increasingly significant, becoming a key focus and challenge in transformer area loss reduction management.

[0003] The field of line loss is highly specialized. The generation mechanism of additional losses has a clear physical relationship with electrical quantities, rather than a simple statistical fitting relationship. To overcome the insufficient accuracy of the simple superposition method, existing research and engineering practices have made many improvements. However, these methods still have significant limitations when facing the actual operating scenarios of low-voltage distribution areas. Currently, the calculation of additional losses in distribution areas mainly adopts three technical routes: The first is the engineering-practical simple superposition method, which calculates the three types of losses—three-phase imbalance, harmonic distortion, and reactive power deficiency—independently and then sums them directly. The principle is simple and easy to implement in engineering. However, its calculation logic is based on the assumption that the factors are independent of each other, completely ignoring the coupling amplification effect. According to actual measurements, the system error is usually between 15% and 30%. However, the coupling coefficient of this type of method is mostly a fixed empirical value. It does not consider the differences in load characteristics of different distribution areas, nor does it model the higher-order coupling effect of the three factors. It is difficult to meet the accuracy requirements of lean line loss management in distribution areas. This makes it difficult to have a reliable quantitative basis for refined loss reduction management, often resulting in the situation of "calculated losses, but no cause found on site". The second type is the data-driven black-box method, which uses machine learning models to output the bus loss rate end-to-end. The overall fitting accuracy is high, but it can only provide a single total result and does not have the ability to attribute and decompose the loss components of factors, so it cannot support targeted governance decisions. However, this type of method uses blind source separation methods such as independent component analysis to decompose losses. However, this type of method directly applies general algorithms in the field of signal processing without considering the physical constraints of line loss calculation. It often produces unreasonable results such as negative loss, non-conservation of total loss, and confusion of the physical meaning of components. This makes loss reduction governance a blind state of "knowing what but not why", unable to clarify the specific contribution of various power quality problems to losses, and difficult to guide the selection of targeted governance measures. This often leads to mismatch in loss reduction investment. On-site, there are often problems of "investing in reactive power compensation but not much reduction in line loss", resulting in a waste of investment with a return on investment of less than 40%. The third category is that existing refined loss calculation methods, whether analytical methods based on physical models or numerical methods based on simulations, all heavily rely on complete transformer substation topology parameters and equipment parameters, and generally assume that the substation has full-configuration measurement data. This deficiency makes it difficult to promote and apply these methods on a large scale in real-world engineering environments where the accuracy of low-voltage substation topology files is less than 70% and the measurement configuration gradient varies greatly. They can only be applied in a small number of demonstration substations, covering less than 10% of the total area, and many low-configuration substations cannot enjoy the technical benefits of refined analysis. For example, the electromagnetic transient simulation method, which performs numerical calculations by building detailed component models, has the highest theoretical accuracy, but it is highly dependent on complete substation topology parameters. However, the accuracy of low-voltage substation topology files is less than 70%, and the large amount of computation makes it difficult to apply online. Summary of the Invention

[0004] This application provides a decoupling calculation method, apparatus, and equipment for the influencing factors of distribution transformer area line loss, which solves the technical problems of insufficient calculation accuracy and poor adaptability of existing refined calculation methods for additional losses in low-voltage distribution transformer areas.

[0005] To achieve the above objectives, this application provides the following technical solution: On the one hand, a decoupled calculation method for the influencing factors of distribution substation line loss is provided, including the following steps: Historical electrical quantity data and real-time electrical quantity data of the distribution network area are obtained according to the sampling time. Data identification, missing value completion and multi-sampling rate time synchronization alignment processing are performed on the real-time electrical quantity data to obtain data configuration and real-time operation data. Ideal operating condition data is selected from the historical electrical quantity data, and univariate linear regression is used to fit the ideal operating condition data to obtain the comprehensive equivalent resistance and no-load loss of the transformer area; based on the real-time operating data, the comprehensive equivalent resistance of the transformer area and the no-load loss, the actual total loss and fundamental reference loss of the transformer area are calculated. Based on the data configuration and the real-time operating data, the equivalent frequency coefficient is determined; based on the real-time operating data, the comprehensive equivalent resistance of the transformer area, the equivalent frequency coefficient, and the fundamental reference loss, the initial values ​​of unbalanced additional loss, harmonic additional loss, and reactive power additional loss are determined. Obtain the third-order symmetric coupling coefficient matrix. Based on the third-order symmetric coupling coefficient matrix, the real-time operating data, the comprehensive equivalent resistance of the transformer area, the fundamental reference loss, the initial value of the unbalanced additional loss, the initial value of the harmonic additional loss, and the initial value of the reactive power additional loss, determine the initial values ​​of the coupling interaction loss and the total loss of the transformer area. Based on the initial values ​​of the total loss of the transformer area, the fundamental reference loss, the initial values ​​of the unbalanced additional loss, the initial values ​​of the harmonic additional loss, and the initial values ​​of the reactive power additional loss, the observation signal matrix is ​​determined; based on the observation signal matrix, the objective function is constructed using the dynamic penalty function method; based on the actual total loss of the transformer area, the objective function is iteratively calculated using multiple constraints until the loss components of each factor that satisfy the convergence condition are obtained. The loss components include fundamental frequency reference loss, coupling interaction loss, unbalanced additional loss, harmonic additional loss, and reactive power additional loss. The convergence condition includes the number of iterations reaching the iteration number threshold or the difference between the calculation results of two adjacent iterations being less than the difference threshold.

[0006] Optionally, the initial values ​​of unbalanced additional loss, harmonic additional loss, and reactive power additional loss are determined based on the real-time operating data, the equivalent resistance of the transformer area, the equivalent frequency coefficient, and the fundamental reference loss, including: The negative sequence current is calculated based on the three-phase current from the real-time operating data. The initial value of unbalanced additional loss is calculated based on the comprehensive equivalent resistance of the transformer area, the negative sequence current, and the neutral line current and neutral line equivalent resistance of the real-time operating data. Obtain the skin effect engineering correction coefficient, and calculate the initial value of the harmonic additional loss based on the skin effect engineering correction coefficient, the equivalent frequency coefficient and the fundamental reference loss; Obtain the target economic operating power factor, and calculate the initial value of reactive power additional loss based on the fundamental reference loss, the target economic operating power factor, and the actual operating power factor of the real-time operating data.

[0007] Optionally, determining the equivalent frequency coefficient based on the data configuration and the real-time operating data includes: If the data is configured as an enhanced configuration level or the total harmonic distortion rate of the current in the real-time running data is greater than the harmonic distortion rate threshold, then the equivalent frequency coefficient is calculated based on the harmonic order, secondary harmonic current and fundamental current of the real-time running data. If the data is configured as a standard configuration or the total harmonic distortion rate of the current in the real-time running data is not greater than the harmonic distortion rate threshold, then the skin effect correction coefficient is obtained, and the equivalent frequency coefficient is calculated based on the skin effect correction coefficient and the total harmonic distortion rate of the current in the real-time running data.

[0008] Optionally, the initial values ​​of coupling interaction loss and total loss of the transformer area are determined based on the third-order symmetric coupling coefficient matrix, the real-time operating data, the comprehensive equivalent resistance of the transformer area, the fundamental reference loss, the initial value of the unbalanced additional loss, the initial value of the harmonic additional loss, and the initial value of the reactive power additional loss, including: Based on the three-phase current of the real-time operating data, the negative sequence current and the three-phase average current are calculated, and the negative sequence current is used as the three-phase unbalance characteristic current; based on all the subharmonic currents of the real-time operating data or the total harmonic distortion rate and fundamental current of the real-time operating data, the harmonic characteristic current is calculated; based on the three-phase average current and the actual operating power factor of the real-time operating data, the reactive characteristic current is calculated. The three-factor high-order nonlinear amplification coefficient is obtained, and the coupling interaction loss is calculated based on the initial values ​​of the unbalance additional loss, the harmonic additional loss, the reactive power additional loss, the three-phase unbalance characteristic current, the harmonic characteristic current, the reactive power characteristic current, the three-factor high-order nonlinear amplification coefficient, and the coupling coefficients of the three factors in the third-order symmetric coupling coefficient matrix. The initial value of the total loss of the transformer area is calculated based on the coupling interaction loss, the fundamental reference loss, the initial value of the unbalanced additional loss, the initial value of the harmonic additional loss, and the initial value of the reactive power additional loss.

[0009] Optionally, obtaining the third-order symmetric coupling coefficient matrix includes: Based on the type of distribution network area access, determine the typical values ​​of the coupling coefficients of the three pairwise coupling factors in the third-order symmetric coupling coefficient matrix; Based on the actual total loss of the transformer area, the fundamental reference loss, the initial value of the unbalanced additional loss, the initial value of the harmonic additional loss, the initial value of the reactive power additional loss, the real-time operating data, and the comprehensive equivalent resistance of the transformer area, the coupling coefficient calibration values ​​of the pairwise coupling factors of the three factors in the third-order symmetrical coupling coefficient matrix are obtained. The coupling coefficient calibration values ​​include the first coupling calibration coefficient between the three-phase unbalance and harmonics, the second coupling calibration coefficient between the three-phase unbalance and reactive power, and the third coupling calibration coefficient between the harmonics and reactive power. Based on the three-phase imbalance, total harmonic distortion of current, and actual operating power factor of the real-time operating data, determine the coupling coefficient calibration value; and compare the coupling coefficient calibration value with the corresponding typical value of the coupling coefficient to determine the coupling coefficient difference. If the difference in the coupling coefficients is greater than the difference threshold, then the calibration value of the coupling coefficients of the pairwise coupling factors of the three factors in the third-order symmetric coupling coefficient matrix is ​​used as the coupling coefficient of the pairwise coupling of the three factors in the third-order symmetric coupling coefficient matrix. If the difference in the coupling coefficients is not greater than the difference threshold, then the typical value of the coupling coefficient of the pairwise coupling factors of the three factors in the third-order symmetric coupling coefficient matrix is ​​taken as the coupling coefficient of the pairwise coupling of the three factors in the third-order symmetric coupling coefficient matrix.

[0010] Optionally, the objective function is: ; ; ; In the formula, J(w) is the objective function, x is the observed signal matrix, T is the transpose of the matrix, and P... total0 Let P be the initial value of the total loss in the transformer area. base For the fundamental reference loss, P unb0 For the initial value of the unbalanced additional loss, P har0 P is the initial value for the additional harmonic loss. rea0Let d be the initial value of reactive power loss, d be the eigenvalue diagonal matrix corresponding to the observed signal matrix, E be the eigenvector matrix corresponding to the observed signal matrix, z be the whitened matrix, V be the whitening matrix, G() be the contrast function of the original fast independent component analysis, λ' be the dynamic penalty coefficient of the dynamic penalty function method, and y be the initial value of reactive power loss. i Let E{·} be the factor loss component of the i-th factor, E{·} be the mathematical expectation operator, and w be the separation weight vector matrix of the dynamic penalty function method.

[0011] On the other hand, a decoupling calculation device for the influencing factors of distribution area line loss is provided, comprising: a field data layer, an edge computing layer and a cloud platform layer connected in sequence; The field data layer is used to acquire historical and real-time electrical quantity data of the distribution network area according to the sampling time. The edge computing layer is used to obtain the fundamental reference loss, coupling interaction loss and loss components of each factor of the distribution network area access based on the historical electrical quantity data and the real-time electrical quantity data according to the decoupling calculation method of the distribution area line loss influencing factors described above. The cloud platform layer is used to generate a governance decision scheme based on the fundamental frequency reference loss, the coupling interaction loss, and the loss components of each factor.

[0012] Optionally, the edge computing layer includes a measurement classification adaptation module, a multi-source data processing module, a reference loss calculation module, a factor loss preliminary calculation module, and a coupled component decoupling attribution module connected in sequence. The measurement classification and adaptation module is used to determine the operating level of the data source of the real-time electrical quantity data to obtain the data configuration; and to perform missing estimation processing on the real-time electrical quantity data according to the data configuration to obtain standardized real-time data. The multi-source data processing module is used to perform data identification, missing value completion, and multi-sampling rate time synchronization alignment processing on the standardized real-time data to obtain real-time running data; The reference loss calculation module is used to filter ideal operating condition data from the historical electrical quantity data, and to perform univariate linear regression fitting on the ideal operating condition data to obtain the comprehensive equivalent resistance and no-load loss of the transformer area; based on the real-time operating data, the comprehensive equivalent resistance of the transformer area and the no-load loss, the module calculates the actual total loss of the transformer area and the fundamental reference loss. The factor loss initial calculation module is used to determine the equivalent frequency coefficient based on the data configuration and the real-time operating data; and to determine the initial values ​​of unbalanced additional loss, harmonic additional loss and reactive power additional loss based on the real-time operating data, the comprehensive equivalent resistance of the transformer area, the equivalent frequency coefficient and the fundamental reference loss. The coupling component decoupling attribution module is used to obtain a third-order symmetric coupling coefficient matrix. Based on the third-order symmetric coupling coefficient matrix, the real-time operating data, the comprehensive equivalent resistance of the transformer area, the fundamental reference loss, the initial value of the unbalanced additional loss, the initial value of the harmonic additional loss, and the initial value of the reactive power additional loss, it determines the coupling interaction loss and the initial value of the total loss of the transformer area. Based on the initial value of the total loss of the transformer area, the fundamental reference loss, the initial value of the unbalanced additional loss, the initial value of the harmonic additional loss, and the initial value of the reactive power additional loss, it determines the observation signal matrix. Based on the observation signal matrix, it constructs a function using a dynamic penalty function method to obtain the objective function. Based on the actual total loss of the transformer area, iteratively calculates the objective function using multiple constraints until the loss components of each factor satisfy the convergence condition are obtained. The loss components include fundamental frequency reference loss, coupling interaction loss, unbalanced additional loss, harmonic additional loss, and reactive power additional loss. The convergence condition includes the number of iterations reaching the iteration number threshold or the difference between the calculation results of two adjacent iterations being less than the difference threshold.

[0013] Optionally, the objective function is: ; ; ; In the formula, J(w) is the objective function, x is the observed signal matrix, T is the transpose of the matrix, and P... total0 Let P be the initial value of the total loss in the transformer area. base For the fundamental reference loss, P unb0 For the initial value of the unbalanced additional loss, P har0 P is the initial value for the additional harmonic loss. rea0 Let d be the initial value of reactive power loss, d be the eigenvalue diagonal matrix corresponding to the observed signal matrix, E be the eigenvector matrix corresponding to the observed signal matrix, z be the whitened matrix, V be the whitening matrix, G() be the contrast function of the original fast independent component analysis, λ' be the dynamic penalty coefficient of the dynamic penalty function method, and y be the initial value of reactive power loss. i Let E{·} be the factor loss component of the i-th factor, E{·} be the mathematical expectation operator, and w be the separation weight vector matrix of the dynamic penalty function method.

[0014] On the other hand, a terminal device is provided, including a processor and a memory; The memory is used to store program code and transmit the program code to the processor; The processor is used to execute the above-described decoupling calculation method for the influencing factors of distribution substation line loss according to the instructions in the program code.

[0015] This invention discloses a decoupled calculation method, apparatus, and equipment for the influencing factors of line loss in a distribution transformer area. The method includes: acquiring historical and real-time electrical quantity data of the distribution network transformer area according to sampling time; performing data identification, missing value completion, and multi-sampling rate time synchronization alignment processing on the real-time electrical quantity data to obtain data configuration and real-time operating data; filtering ideal operating condition data from the historical electrical quantity data; applying univariate linear regression fitting processing to the ideal operating condition data to obtain the transformer area's comprehensive equivalent resistance and no-load loss; calculating the actual total loss and fundamental frequency reference loss of the transformer area based on the real-time operating data, the transformer area's comprehensive equivalent resistance, and no-load loss; determining the equivalent frequency coefficient based on the data configuration and real-time operating data; determining the initial values ​​of unbalanced additional loss, harmonic additional loss, and reactive power additional loss based on the real-time operating data, the transformer area's comprehensive equivalent resistance, equivalent frequency coefficient, and fundamental frequency reference loss; and obtaining the third-order... The symmetric coupling coefficient matrix is ​​used to determine the initial values ​​of coupling interaction loss and total loss of the transformer area based on the third-order symmetric coupling coefficient matrix, real-time operating data, the comprehensive equivalent resistance of the transformer area, the fundamental reference loss, the initial values ​​of unbalanced additional loss, the initial values ​​of harmonic additional loss, and the initial values ​​of reactive power additional loss. The observation signal matrix is ​​then determined based on these initial values. A dynamic penalty function method is used to construct the objective function based on the observation signal matrix. Multiple constraints are applied to iteratively calculate the objective function based on the actual total loss of the transformer area until the loss components satisfying the convergence condition are obtained. These loss components include the fundamental reference loss, coupling interaction loss, unbalanced additional loss, harmonic additional loss, and reactive power additional loss. The convergence condition includes the number of iterations reaching an iteration threshold or the difference between the results of two adjacent iterations being less than a difference threshold.

[0016] As can be seen from the above technical solutions, this application has the following advantages: The decoupling calculation method for the influencing factors of distribution transformer area line loss determines the initial value of the calculated loss through data configuration, then constructs a third-order symmetric coupling coefficient matrix to quantify the interaction effect of multiple factors, and combines multiple constraints to iteratively calculate the objective function to obtain the loss components of each factor that meet the convergence condition, thereby improving the calculation accuracy of additional loss, realizing the refined calculation of additional loss in the transformer area, providing a data basis for targeted loss reduction, and solving the technical problems of insufficient calculation accuracy and poor adaptability of existing refined calculation methods for additional loss in low-voltage transformer areas.

[0017] This decoupling calculation device for distribution transformer area line loss influencing factors constructs an edge-cloud collaborative deployment architecture through a field data layer, an edge computing layer, and a cloud platform layer. It uses data configuration to determine the initial value of the calculated loss, then constructs a third-order symmetric coupling coefficient matrix to quantify the interaction effect of multiple factors, and iteratively calculates the objective function in combination with multiple constraints to obtain the loss components of each factor that meet the convergence condition, thereby improving the calculation accuracy of additional loss and realizing refined calculation of additional loss in the transformer area; and generates governance decision schemes through the cloud platform layer. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of 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 only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating the steps of the decoupling calculation method for the influencing factors of distribution substation line loss described in the embodiments of this application; Figure 2 This is a schematic diagram of the framework of the decoupling calculation device for the influencing factors of distribution substation line loss according to another embodiment of this application; Figure 3 This is a flowchart illustrating the decoupling calculation device for the influencing factors of distribution substation line loss described in an embodiment of this application. Figure 4 This is a schematic diagram of the terminal device described in an embodiment of this application. Detailed Implementation

[0020] To make the inventive objectives, features, and advantages of this application more apparent and understandable, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described below are only some embodiments of this application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0021] In the description of the embodiments of this application, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of the embodiments of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0022] In the embodiments of this application, unless otherwise explicitly specified and limited, the terms "installation," "connection," "linking," "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. Those skilled in the art can understand the specific meaning of the above terms in the embodiments of this application according to the specific circumstances.

[0023] Patent terminology used in this application: Linear interpolation refers to an interpolation method where the interpolation function is a first-order polynomial, and the interpolation error at the interpolation nodes is zero.

[0024] Branch monitoring units (commonly known as LTUs, or low-voltage branch monitoring units or low-voltage branch monitoring terminals) are key devices in smart distribution networks. They are mainly used for real-time acquisition of the operating status of low-voltage distribution lines (400V), fault detection, and data uploading. They are typically installed on the outgoing side of prefabricated substations, community distribution rooms, pole-mounted transformers, or branch boxes.

[0025] Coefficient of Determination (R) 2 The coefficient of determination (COD) is a core statistical indicator that measures the goodness of fit of a regression model to observed data, reflecting the proportion of the variance in the dependent variable that the independent variable explains.

[0026] This application provides a decoupled calculation method, apparatus, and equipment for the influencing factors of distribution transformer substation line losses, solving the technical problems of insufficient calculation accuracy and poor adaptability in existing refined calculation methods for additional losses in low-voltage distribution transformer substations. The purpose of this decoupled calculation method, apparatus, and equipment for the influencing factors of distribution transformer substation line losses is to construct a decoupled attribution method with low data dependence. It obtains the comprehensive equivalent resistance of the substation through regression fitting of quasi-ideal operating condition data, thereby eliminating dependence on distribution network topology parameters. Simultaneously, it designs a hierarchical measurement configuration mechanism to be compatible with substations with different measurement configurations, significantly improving the method's engineering applicability and its widespread application.

[0027] Example 1: Figure 1 This is a flowchart illustrating the steps of the decoupling calculation method for the influencing factors of distribution substation line loss described in this application embodiment.

[0028] like Figure 1 As shown in the figure, this application provides a decoupling calculation method for the influencing factors of distribution substation line loss, including the following steps: S1. Obtain historical and real-time electrical quantity data of the distribution network area according to the sampling time. Perform data identification, missing value completion and multi-sampling rate time synchronization alignment on the real-time electrical quantity data to obtain data configuration and real-time operation data.

[0029] It should be noted that in the decoupling calculation of the factors affecting line loss in the distribution substation, step S1 can be understood as: obtaining historical and real-time electrical quantity data of the distribution substation from the database of the distribution network, providing basic data for the subsequent steps of decoupling the calculation of the factors affecting line loss in the distribution substation. Secondly, using preset rules, data identification, missing value completion, and multi-sampling rate time synchronization alignment are performed on the real-time electrical quantity data to obtain data configuration and real-time operating data, providing the data required for analysis in subsequent steps. In this embodiment, data configuration includes standard configuration level, enhanced configuration level, and full configuration level. Real-time operating data includes at least three-phase current, three-phase voltage, harmonic order, subharmonic current of each harmonic, fundamental current, total harmonic distortion rate, three-phase unbalance, active power, reactive power, actual operating power factor, neutral current, and neutral equivalent resistance. Historical electrical quantity data can be the 15-minute time-series operating data of the distribution substation over the past 30 days. The sampling time can be selected as 15 minutes, in which case the historical electrical quantity data and the real-time electrical quantity data are time-series operation data.

[0030] In this embodiment, the process of performing data identification, missing value completion, and multi-sampling rate time synchronization alignment on real-time electrical quantity data to obtain data configuration and real-time operating data includes: determining the operating level of the data source of the real-time electrical quantity data to obtain data configuration; performing missing value estimation on the real-time electrical quantity data according to the data configuration to obtain standardized real-time data; and performing data identification, missing value completion, and multi-sampling rate time synchronization alignment on the standardized real-time data to obtain real-time operating data.

[0031] It should be noted that during the data configuration acquisition process, a gradient measurement configuration adaptive operation mechanism can be set to determine the operating level of the data source for real-time electrical quantities, thereby obtaining the data configuration. Based on the data configuration, a data missing versus normal judgment rule is used to perform missing estimation processing on the real-time electrical quantities, resulting in standardized real-time data. For the standardized real-time data, electrical rationality rules and a hierarchical completion strategy can be used for data identification, missing value completion, and multi-sampling rate time synchronization alignment processing, resulting in real-time operating data. In this embodiment, the decoupling calculation method for distribution substation line loss influencing factors can be compatible with substations with different measurement configurations through the set gradient measurement configuration adaptive operation mechanism, making the method widely applicable and facilitating large-scale promotion.

[0032] In this embodiment of the application, the adaptive operation mechanism for gradient measurement configuration includes: If the real-time electrical quantity data accessed by the distribution network area only includes data from the smart converged terminal (such as the edge gateway) of the area, then the data configuration is determined to be at the standard configuration level. If the real-time electrical quantity data includes data from the intelligent converged terminal of the distribution area (such as the edge gateway) and the branch harmonic monitoring device (such as the harmonic analyzer), then the data configuration is determined to be at the enhanced configuration level; If the real-time electrical quantity data includes data from smart fusion terminals (such as edge gateways), branch harmonic monitoring devices (such as harmonic analyzers), branch monitoring units, and smart meters, then the data configuration is determined to be at the full configuration level; this full configuration level is used to calculate the equivalent frequency coefficient based on the enhanced configuration level type. If the real-time electrical quantity data includes data from smart converged terminals (such as edge gateways), branch monitoring units, and smart meters (without harmonics), then the data configuration is determined to be at the enhanced configuration level. However, the real-time electrical quantity data at this enhanced configuration level has the defect of insufficient data. If the real-time electrical quantity data includes data from smart converged terminals (such as edge gateways), branch monitoring units, and smart meters, then the data configuration is determined to be at the standard configuration level. This standard configuration level of real-time electrical quantity data has the defect of insufficient data. If the real-time electrical quantity data does not include data from intelligent converged terminals (such as edge gateways), it means that the acquired real-time electrical quantity data does not meet the data required for the decoupling calculation of the factors affecting line loss in the distribution substation, and the data needs to be acquired again.

[0033] It should be noted that in the process of determining the data configuration through the gradient measurement configuration adaptive operation mechanism, the data access of the distribution area's intelligent fusion terminal is a prerequisite for operation. Without the data access of the intelligent fusion terminal, the decoupling calculation method for the line loss influencing factors of the distribution area cannot be implemented and an alarm will be output. In this embodiment, the standard configuration level is the access to the three-phase voltage, three-phase current, active power, reactive power, and total harmonic distortion rate of the current obtained by the intelligent fusion terminal according to the sampling time (e.g., 15 minutes). On this basis, the enhanced configuration level is the access to the data of the harmonic monitoring device (e.g., the effective values ​​of each harmonic current, fundamental current, etc.). The full configuration level is the access to the minute-level data of the branch monitoring unit (e.g., branch three-phase current, three-phase voltage, active power, reactive power, etc.) and the minute-level data of the smart energy meter (e.g., user current, voltage, active power, reactive power, and daily frozen electricity, etc.). If either the branch monitoring unit or the smart meter is missing, the branch user-level traceability function will not be enabled, and only the substation-level component will be output; if the data from the branch smart meter is connected but there is no harmonic data, it will be judged as full configuration level.

[0034] In this embodiment, the decoupled calculation method for the influencing factors of distribution transformer area line loss uses a gradient measurement configuration adaptive operation mechanism to set three levels of data configuration: standard configuration level, enhanced configuration level, and full configuration level. This enables the obtained real-time operation data to support automatic degradation when the data source is interrupted and automatic upgrade when the data is restored. When switching between upgrade and downgrade, the accuracy level identifier is output. During downgrade operation, the corresponding replacement algorithm is used to estimate the missing data, and it is compatible with all transformer areas from low configuration to high configuration.

[0035] In this embodiment, the rules for judging missing and normal data can be understood as follows: if the data obtained from a single sampling point is missing, the actual data received within the cross-section is less than 80% of the expected amount, the timestamp deviation exceeds ±5 minutes, or the value fails the electrical rationality check, then the electrical quantity data of that sampling point is determined to be missing; if the data obtained from a single sampling point is 80% of the actual data received within the cross-section, the timestamp deviation is within ±5 minutes, or the value passes the electrical rationality check, then the electrical quantity data of that sampling point is determined to be normal. If the real-time electrical quantity data is judged to be missing for 4 consecutive sampling points (1 hour), a downgrade is triggered. After recovery, if all 24 consecutive sampling points (6 hours) are normal, an upgrade is triggered. The asymmetry of the upgrade and downgrade thresholds creates hysteresis to avoid frequent switching of boundary states. If the real-time electrical quantity data is judged to be missing for less than 4 consecutive sampling points, no downgrade is performed. Then, interpolation can be used to complete the real-time electrical quantity data that is not downgraded, resulting in standardized real-time data.

[0036] It should be noted that the intelligent fusion terminal uses a sampling time segment as one sampling point, and minute-level data is first aggregated to the sampling time segment. Missing data is estimated using corresponding substitution algorithms. For example, when there is no neutral current data, the three-phase current phasor sum is used for estimation; when there is no harmonic data, the components are estimated based on the total harmonic distortion rate and the typical spectrum distribution of the 3rd / 5th / 7th harmonic current ratio of 7:2:1 in a typical residential transformer area. Specifically, during data-triggered degraded operation, the corresponding substitution algorithm is used to estimate the missing data type. The correspondence between the substitution algorithm and the data type includes: if neutral current data is missing, the effective value of the neutral current is estimated using the three-phase current phasor sum synthesis method; if harmonic current data is missing, the harmonic current is estimated based on the total harmonic distortion rate and the typical harmonic spectrum distribution; if minute-level data from branch monitoring units and smart meters is missing, no numerical estimation is performed, the branch user-level loss tracing function is directly turned off, and only the five types of loss components at the transformer area level are calculated and output. The alternative algorithm is only enabled during degraded operation and is automatically disabled after the data source recovers and triggers an upgrade. If the distribution network is a low-voltage three-phase four-wire system, the neutral current is the phasor sum of the three-phase currents, i.e. In the formula, For neutral line current, , , These represent the three-phase currents A, B, and C, all in amperes (A). The intelligent fusion terminal provides phase information for each phase current, using the A-phase voltage as a reference. The phasors of each phase current are constructed based on their measured effective values ​​and phase angles. When there is no harmonic monitoring data (i.e., data from smart meters), the total harmonic distortion rate and fundamental current measured by the intelligent fusion terminal in the transformer substation are used, combined with the typical harmonic spectrum distribution of the residential transformer substation, to equivalently estimate the harmonic currents. The estimated harmonic currents include: The first step is to calculate the total harmonic current. Defined by the total harmonic distortion rate (THC), the total harmonic current is: In the formula, The total harmonic current is the effective value, in amperes (A). The total harmonic distortion of the current is dimensionless and is measured by the intelligent fusion terminal of the transformer substation. The current is the fundamental frequency, measured in amperes (A), and was obtained from the smart fusion terminal of the distribution area.

[0037] The second step is to allocate the harmonic currents according to the typical spectrum distribution; for residential transformer areas, the 3rd, 5th, and 7th harmonics are the main ones, with a typical current content ratio of 7:2:1. It is the third harmonic current. It is the 5th harmonic current. If it is the 7th harmonic current, then and Therefore, due to the law of conservation of energy, we have The harmonic currents can be estimated as follows: , and .

[0038] In this embodiment of the application, to provide high-quality synchronous data for subsequent calculations, the standardized real-time data can be processed by using electrical rationality rules and hierarchical completion strategies for data identification, missing value completion, and multi-sampling rate time synchronization alignment to obtain real-time running data. This process can be understood as follows: based on electrical rationality rules, abnormal data is identified in the standardized real-time data, including obvious erroneous data such as voltage exceeding limits, negative current, and power non-conservation. Hierarchical completion strategies are used to fill missing data of different importance levels using methods such as linear interpolation, statistical mean, and nearest median. All data are uniformly resampled into the sampling time section to form real-time running data of a time-aligned high-quality data matrix.

[0039] It should be noted that during the acquisition of real-time operational data, various measurement data of the transformer area are uniformly organized into an n×m dimensional time-series data matrix X, where n is the number of cross-sections with a sampling time (e.g., 15 minutes) (96 cross-sections per day, i.e., n = 96 per day); and m is the number of data channels. In this embodiment, taking the enhanced configuration level transformer area as an example, m = 16, and the obtained real-time operational data is denoted as X, i.e. In the formula, the three-phase voltages A, B, and C are respectively: The three-phase currents are as follows: The active power is P, the reactive power is Q, the total harmonic distortion of the current is THDI, and the neutral current is I. N The fundamental current is I l The 3rd, 5th, and 7th harmonic currents are The actual operating power factor is If the data configuration is at the full configuration level, in addition to the real-time operating data at the enhanced configuration level, data from branch monitoring units and smart meters collected from branches / users will be added.

[0040] S2. Select ideal operating condition data from historical electrical quantity data, and use univariate linear regression to fit the ideal operating condition data to obtain the comprehensive equivalent resistance and no-load loss of the transformer area; calculate the actual total loss and fundamental reference loss of the transformer area based on real-time operating data, comprehensive equivalent resistance and no-load loss of the transformer area.

[0041] It should be noted that step S2 can be understood as follows: First, the historical electrical quantity data obtained in step S1 is filtered according to preset rules or set conditions to select an ideal operating condition data of a sampling point as a quasi-ideal operating condition sample of the sampling point; second, univariate linear regression fitting is used to fit all quasi-ideal operating condition samples to obtain the comprehensive equivalent resistance and no-load loss of the distribution area; finally, the actual total loss and fundamental frequency reference loss of the distribution area are calculated using preset formulas based on real-time operating data, comprehensive equivalent resistance and no-load loss of the distribution area, providing basic data for the decoupling calculation of distribution area line loss influencing factors in subsequent steps. In this embodiment, the decoupling calculation method of distribution area line loss influencing factors obtains the comprehensive equivalent resistance, no-load loss, actual total loss and fundamental frequency reference loss of the distribution area through the parameterless equivalent resistance identification in step S2. It does not rely on topology parameters, and can obtain the physically meaningful comprehensive equivalent resistance, no-load loss, actual total loss and fundamental frequency reference loss of the distribution area using only measurement data, thus achieving independence from the dependence on distribution network topology parameters.

[0042] In the embodiments of this application, the preset rules or setting conditions for screening ideal operating condition data are that the three-phase current imbalance is less than the imbalance threshold (e.g., 3%), the total harmonic distortion rate of the current is less than the ideal threshold for harmonic distortion rate (e.g., 3%), and the actual operating power factor is greater than the power factor threshold (e.g., 0.95).

[0043] It should be noted that the unbalance threshold can be determined according to the GB / T15543 standard "Power Quality - Three-phase Voltage Unbalance". The harmonic distortion rate threshold can be determined according to the GB / T14549 standard "Power Quality - Harmonics in Public Power Grids", and the power factor threshold can be determined according to the industry standard for economic operation of distribution networks.

[0044] In this embodiment, the ideal operating condition data is fitted using univariate linear regression to obtain the comprehensive equivalent resistance and no-load loss of the transformer substation. This can be understood as: for each quasi-ideal operating condition sample, the square of the three-phase average current I is calculated. avg 2 As the independent variable, it corresponds to the total power loss P in the transformer area. loss As the dependent variable, the univariate linear least squares method was used to fit the data of all quasi-ideal working condition samples, resulting in the linear equation P. loss =K×I avg 2 +B, the equivalent resistance R of the transformer area can be obtained based on the slope K and intercept B. eq and no-load loss P T0 Then, regression fitting verification is performed until the slope K and intercept B that pass the verification are used as the equivalent resistance R of the transformer area. eq and no-load loss P T0 .

[0045] It should be noted that the regression fitting verification includes the following: the coefficient of determination of the regression fitting must be greater than 0.9. Otherwise, the calculated comprehensive equivalent resistance of the transformer area is considered unstable and should be handled according to the following procedure: First, the sample window is expanded from 30 days to 60 days for re-screening; if it still does not meet the requirement, the migration calibration mechanism is activated. The equivalent resistance of the benchmark transformer area with the same power supply area, capacity, and wire type is used as the initial value, and then the top 20% of the samples that are closest to the ideal operating condition in this transformer area are used for fine-tuning until the comprehensive equivalent resistance and no-load loss of the transformer area that meet the coefficient of determination of the regression fitting are obtained.

[0046] In this embodiment, the preset formulas for calculating the actual total loss and fundamental reference loss of the transformer area based on real-time operating data, the comprehensive equivalent resistance of the transformer area, and the no-load loss are as follows: ; ; In the formula, P loss The actual total loss in the transformer area is expressed in kilowatts (kW); I avg The effective value of the three-phase average current (referring to the arithmetic mean of the three-phase currents A, B, and C within a sampling time segment), in amperes (A). R eq The equivalent resistance of the transformer substation is expressed in ohms (Ω); P T0P represents the no-load loss of the transformer in the distribution area, expressed in kilowatts (kW). base For the fundamental reference loss, I l This is the fundamental current.

[0047] S3. Determine the equivalent frequency coefficient based on the data configuration and real-time operating data; determine the initial values ​​of unbalanced additional loss, harmonic additional loss and reactive power additional loss based on the real-time operating data, the comprehensive equivalent resistance of the transformer area, the equivalent frequency coefficient and the fundamental reference loss.

[0048] It should be noted that step S3 can be understood as follows: First, based on the data configuration and real-time operating data obtained in step S1, a formula for calculating the equivalent frequency coefficient is determined using preset rules, thereby determining the equivalent frequency coefficient; then, based on step S2 and the equivalent frequency coefficient, the initial values ​​of unbalanced additional loss, harmonic additional loss, and reactive power additional loss are calculated using preset formulas, providing basic calculation data for subsequent steps to construct the observation signal matrix and solve for the loss components of each factor. In this embodiment, the decoupling calculation method for distribution substation line loss influencing factors provides physical anchor points for subsequent decoupling by calculating the initial values ​​of the three single-factor additional losses.

[0049] S4. Obtain the third-order symmetric coupling coefficient matrix. Based on the third-order symmetric coupling coefficient matrix, real-time operating data, the comprehensive equivalent resistance of the transformer area, the fundamental reference loss, the initial value of the unbalanced additional loss, the initial value of the harmonic additional loss, and the initial value of the reactive power additional loss, determine the initial values ​​of the coupling interaction loss and the total loss of the transformer area.

[0050] It should be noted that step S4 can be understood as first obtaining the third-order symmetric coupling coefficient matrix, and then calculating the initial values ​​of coupling interaction loss and total distribution area loss using a preset formula based on the third-order symmetric coupling coefficient matrix, real-time operating data, the comprehensive equivalent resistance of the distribution area, the fundamental reference loss, the initial value of unbalanced additional loss, the initial value of harmonic additional loss, and the initial value of reactive power additional loss. In this embodiment, the decoupling calculation method for distribution area line loss influencing factors quantifies the multi-factor interaction effect of distribution network loss through the third-order symmetric coupling coefficient matrix in step S4, breaking the independent assumption of the traditional simple superposition method and significantly improving the calculation accuracy; the third-order symmetric coupling coefficient matrix realizes the quantification and interpretable decomposition of coupling interaction loss.

[0051] In this embodiment, during the process of obtaining the third-order symmetric coupling coefficient matrix, the typical values ​​of the measured pairwise coupling coefficients of the three factors can be obtained by referring to the type of distribution network access area, and the single-factor self-action coefficients are used to construct the third-order symmetric coupling coefficient matrix. The expression for the third-order symmetric coupling coefficient matrix is: ; In the formula, the diagonal elements C of the third-order symmetric coupling coefficient matrix uu Chh C rr These are the single-factor self-actualization coefficients for three categories: three-phase imbalance, harmonics, and reactive power, respectively, all set to 1. The off-diagonal elements represent typical values ​​of the pairwise coupling coefficients, derived from statistical measurements in the demonstration area. For example, the typical value range for residential areas is the typical value C of the coupling coefficient between three-phase imbalance and harmonics. uh =0.15~0.30, typical value of coupling coefficient C between three-phase imbalance and reactive power. ur =0.10~0.20, typical value of coupling coefficient C between harmonics and reactive power. hr =0.05~0.12, satisfying symmetry C uh =C hu C ur =C ru C hr =C rh r represents reactive power, h represents harmonics, and u represents three-phase imbalance. This can be adjusted according to the actual load type of the distribution network area: For distribution network areas connected to centralized charging pile areas, C... uh =Take the upper limit, i.e., C uh =0.30; The type of distribution network area access is industrial load area, C hr Take the upper limit, i.e., C hr =0.12.

[0052] It should be noted that the constructed third-order symmetric coupling coefficient matrix is ​​quantified by using the current product form to quantify the pairwise cross-loss terms of the three-phase imbalance, harmonics, and reactive power factors and the high-order coupling amplification effect of the three factors. The coupling coefficients are determined by a two-stage strategy of initializing the typical values ​​of the coupling coefficients and unsupervised self-calibration under extreme conditions.

[0053] S5. Determine the observation signal matrix based on the initial values ​​of total loss, fundamental frequency reference loss, unbalanced additional loss, harmonic additional loss, and reactive power additional loss in the transformer area; construct the objective function using the dynamic penalty function method based on the observation signal matrix; iterate the objective function using multiple constraints based on the actual total loss in the transformer area until the loss components of each factor satisfy the convergence condition are obtained; the factor loss components include fundamental frequency reference loss, coupling interaction loss, unbalanced additional loss, harmonic additional loss, and reactive power additional loss, and the convergence condition includes the number of iterations reaching the iteration number threshold or the difference between the calculation results of two adjacent iterations being less than the difference threshold.

[0054] It should be noted that the iteration threshold can be 100 times, and the difference threshold can be 1×e. -6Step S5 can be understood as follows: Based on the initial values ​​of total loss, fundamental reference loss, unbalanced additional loss, harmonic additional loss, and reactive power additional loss obtained from steps S2 to S4, an observation signal matrix is ​​first constructed. Then, based on the observation signal matrix, a function is constructed using the dynamic penalty function method to obtain the objective function. Finally, based on the actual total loss of the transformer area, a fast independent component analysis algorithm is used to apply three constraints: non-negativity constraint, energy conservation constraint, and proportional rationality constraint. The objective function is iteratively calculated until the estimated values ​​of each loss component that meet the convergence condition are obtained as the corresponding loss components of each factor. This decomposes the actual total loss of the transformer area into five interpretable components with clear physical meaning: fundamental reference loss, unbalanced loss, harmonic loss, reactive power loss, and coupling interaction loss.

[0055] In this embodiment of the application, the objective function is: ; ; ; In the formula, J(w) is the objective function, x is the observed signal matrix, T is the transpose of the matrix, and P... total0 Let P be the initial value of the total loss in the transformer area. base For the fundamental reference loss, P unb0 For the initial value of the unbalanced additional loss, P har0 P is the initial value for the additional harmonic loss. rea0 Let d be the initial value of reactive power loss, d be the eigenvalue diagonal matrix corresponding to the observed signal matrix, E be the eigenvector matrix corresponding to the observed signal matrix, z be the whitened matrix, V be the whitening matrix, G() be the contrast function of the original fast independent component analysis, λ' be the dynamic penalty coefficient of the dynamic penalty function method, and y be the initial value of reactive power loss. i Let E{·} be the factor loss component of the i-th factor, E{·} be the mathematical expectation operator, and w be the separation weight vector matrix of the dynamic penalty function method. Here, E{·} represents the statistical average, which in engineering implementation is the sample mean of n cross-sections (real-time electrical quantity data from n sampling points) within the window, i.e.: ; In the formula, t is the sampling time for acquiring data, G(x) = logcosh(x) is the contrast function of the original fast independent component analysis, w is the separation weight vector matrix, i.e., w = [w1, w2, w3, w4, w5], and is the unknown to be solved iteratively. This represents the weighted sum of the observation vectors at the t-th cross-section. In this embodiment, the objective function J(w) is the improved function. Applying a non-negativity constraint to the Fast Independent Component Analysis (FAC) algorithm can be understood as the dynamic penalty coefficient λ' of the objective function increasing linearly with the number of iterations. The initial value of the dynamic penalty coefficient λ' can be set to 0.05, increasing by 0.05 with each iteration, ensuring both initial convergence speed and non-negativity of later results. i To improve the estimate of the i-th loss component obtained by the independent component analysis algorithm.

[0056] It should be noted that in obtaining the eigenvector matrix and the eigenvalue diagonal matrix, the observed signal matrix is ​​centered to obtain a centered matrix; the covariance matrix is ​​constructed based on the centered matrix; the eigenvalues ​​and eigenvectors of the covariance matrix are solved to obtain several eigenvalues ​​and several eigenvectors; the eigenvalue diagonal matrix d is constructed from these eigenvalues, i.e. `diag()` represents a diagonal matrix, and `λ1`, `λ2`, `λ3`, `λ4`, and `λ5` are the five eigenvalues ​​of the covariance matrix. The eigenvector matrix `E` is composed of several eigenvectors. e1, e2, e3, e4, and e5 are the eigenvectors corresponding to the five eigenvalues ​​λ1, λ2, λ3, λ4, and λ5, respectively. In this embodiment, the expression for the centering matrix is: ; The expression for the covariance matrix is: ; In the formula, x is the observed signal matrix. μ Let x be the mean vector of all elements in the observed signal matrix. For a centered matrix, C x Let be the covariance matrix.

[0057] In this embodiment, the energy conservation constraint expression applied by the Fast Independent Component Analysis algorithm is: ; In the formula, y represents the final value of the i-th loss component after energy conservation projection correction, in kilowatts (kW), and is also used as the value for calculating the observation signal matrix in the next iteration; i Let ΔP be the estimated value of the i-th loss component calculated in the iteration before projection correction, in kilowatts (kW); ΔP is the total loss residual, in kilowatts (kW), which is the difference between the actual total loss of the transformer area and the sum of the estimated values ​​of all loss components, i.e., ΔP = P loss -∑y j ;∑y jThe sum of all current loss component estimates, in kilowatts (kW), with subscript j iterating through the loss component estimates of the five categories: fundamental reference loss, unbalanced loss, harmonic loss, reactive power loss, and coupling interaction loss.

[0058] In this embodiment, the proportional rationality constraint imposed by the fast independent component analysis algorithm can be understood as follows: the proportional deviation between the estimated value of each loss component obtained by iterative calculation and the corresponding initial loss calculation does not exceed 50%, to avoid unreasonable physical meaning of the decomposition result. After decomposition, the physical meaning of each component is matched by the correlation coefficient, and the stability is verified by 100 resampling cycles using the bootstrap method. When the coefficient of variation of the decomposition result is greater than 15%, the time window is expanded and the calculation is recalculated.

[0059] It should be noted that the fast independent component analysis algorithm in step S5 of the decoupling calculation method for line loss in distribution substations is not a simple transfer of the independent component analysis algorithm, but a targeted improvement based on the specific physical characteristics of the line loss calculation scenario. General independent component analysis, as a common blind source separation algorithm in the field of signal processing, only optimizes statistical independence and does not consider the physical constraints of the application domain. Directly applying it to the line loss decomposition scenario will result in three inherent defects. The three physical constraints of the fast independent component analysis algorithm in this decoupling calculation method for line loss in distribution substations are all targeted improvements for the line loss scenario and are not part of the standard configuration of general independent component analysis. This makes the decoupling calculation method for line loss in distribution substations suitable for the professional needs of low-voltage substation loss decomposition. Specifically: The non-negativity constraint of the Fast Independent Component Analysis (Fast IComponent Analysis) algorithm is that the output components of general IComponent Analysis can be positive or negative, adapting to general signals such as voice and images with positive and negative values; however, the power loss of a power system is physically always positive, and negative loss violates basic physical axioms. This decoupled calculation method for distribution substation line loss influencing factors embeds the non-negativity constraint into the iterative process through a dynamic penalty function method, forcing all loss component estimates to be positive, thus solving the problem of negative loss generated by the general algorithm. It is a customized improvement for the domain characteristic of non-negativity of line loss power. The energy conservation constraint of the Fast IComponent Analysis algorithm is that general IComponent Analysis only guarantees the statistical independence of the decomposed components, without constraining the sum of the components to be equal to the total amount of the original signal; however, power system losses must strictly satisfy the law of energy conservation, and the sum of the estimated values ​​of each loss component must equal the actual total loss of the distribution substation. This decoupled calculation method for distribution substation line loss influencing factors, through the projection operation after each iteration in step S5, proportionally distributes the total loss residual to each component, forcing the satisfaction of total conservation, and is used to solve the problem of mismatch between the sum of components and the total loss in the general algorithm. It is a customized improvement for the axiom of energy conservation in power systems. The proportional rationality constraint of the fast independent component analysis algorithm is that the component sorting of the general independent component analysis is random, the amplitude scaling is arbitrary, and the output results have no fixed physical correspondence. However, in the field of line loss, the relative magnitudes of the three types of additional losses—imbalance, harmonics, and reactive power—have clear physical common sense ranges. This decoupled calculation method for the influencing factors of distribution substation line loss uses the estimated value of the single-factor loss component as the physical anchor point, constrains the proportional deviation between the decomposed component and the initial loss value to not exceed 50%, and matches the physical meaning of the components through correlation coefficients. It is used to solve the problem of the component meaning being disordered and unable to correspond to specific power quality factors in the general algorithm. It is a customized improvement for the physical common sense in the field of line loss.

[0060] In other embodiments, during the process of obtaining the loss components of each factor, a multi-output long short-term memory (LSTM) neural network can be constructed, with the observed signal matrix as the input of the multi-output long short-term memory (LSTM) neural network, and the loss components of the three factors, the fundamental reference loss, and the coupling interaction loss of the five factors can be directly output.

[0061] It should be noted that using a multi-output long short-term memory (LSTM) neural network to obtain the five loss components of the actual total loss of the distribution area requires a large amount of labeled data for supervised training. However, the true values ​​of the loss components are difficult to obtain, and the cost of acquiring training data is high. Typically, it is necessary to combine pre-training with simulation data with a small amount of fine-tuning based on actual measurements. After training, the inference speed is extremely fast, making it suitable for scenarios with sufficient historical data and simulation samples. This decoupled calculation method for the influencing factors of distribution area line loss is an unsupervised method, which does not require labeled data and is more universal.

[0062] In this embodiment, the decoupled calculation method for distribution substation line loss influencing factors determines the data configuration through the three-level measurement adaptation mechanism in step S1. It only requires standard fusion terminal data to run and is compatible with substations with different measurement configurations. This solves the problem that existing high-precision methods can only be applied in demonstration substations and can be promoted on a large scale. The comprehensive equivalent resistance and no-load loss of the substation are determined through the ideal operating conditions in step S2, completely eliminating the dependence on the topology parameters and line parameters of the low-voltage substation. In the process of calculating the loss components of each factor, the general independent component analysis algorithm is improved with targeted physical constraints for the distribution network line loss scenario. This solves the three inherent problems that will inevitably occur when the general algorithm is directly applied: negative loss, energy non-conservation, and confusion of physical meaning. It is not a simple transfer of the general algorithm, but a targeted improvement with clear technical focus. It decomposes the actual total loss of the substation into five interpretable components with clear physical meaning, providing a reliable quantitative basis for targeted loss reduction.

[0063] It should be noted that this decoupling calculation method for distribution transformer substation line loss influencing factors, through data configuration of real-time electrical quantity data, enables the distribution network to be compatible with substations of different configurations. This achieves accurate quantification, interpretable decomposition, and engineering implementation of coupled interactive losses, providing core technical support for precise loss reduction in low-voltage substations. The method quantifies the multi-factor interaction effect by constructing a third-order symmetric coupling coefficient matrix and iteratively calculates the objective function by applying multiple constraints using an independent component analysis algorithm. This yields the loss components of each factor that meet the convergence condition, decomposing the actual total loss of the substation into multiple attribution components with clear physical meaning. This provides data basis for quantitative decision-making in targeted loss reduction and ensures the interpretability of the calculation results. Specifically, the method uses a third-order symmetric coupling coefficient matrix to characterize the interaction effects of pairwise and triadic factors, significantly improving the calculation accuracy of additional losses and achieving refined quantification of coupled interactive losses.

[0064] This application provides a decoupling calculation method for the influencing factors of line loss in distribution transformer substations. The method includes: acquiring historical and real-time electrical quantity data of the distribution network substations according to sampling time; performing data identification, missing value completion, and multi-sampling rate time synchronization alignment processing on the real-time electrical quantity data to obtain data configuration and real-time operating data; filtering ideal operating condition data from the historical electrical quantity data; applying univariate linear regression fitting processing to the ideal operating condition data to obtain the substation's comprehensive equivalent resistance and no-load loss; calculating the actual total loss and fundamental frequency reference loss of the substation based on the real-time operating data, the substation's comprehensive equivalent resistance, and no-load loss; determining the equivalent frequency coefficient based on the data configuration and real-time operating data; determining the initial values ​​of unbalanced additional loss, harmonic additional loss, and reactive power additional loss based on the real-time operating data, the substation's comprehensive equivalent resistance, the equivalent frequency coefficient, and the fundamental frequency reference loss; and obtaining a third-order symmetric coupling coefficient matrix. Based on the third-order symmetric coupling coefficient matrix, real-time operating data, the comprehensive equivalent resistance of the transformer area, the fundamental reference loss, the initial value of unbalanced additional loss, the initial value of harmonic additional loss, and the initial value of reactive power additional loss, the initial values ​​of coupling interaction loss and total transformer area loss are determined. Based on the initial values ​​of total transformer area loss, fundamental reference loss, unbalanced additional loss, harmonic additional loss, and reactive power additional loss, the observation signal matrix is ​​determined. Based on the observation signal matrix, a dynamic penalty function method is used to construct the objective function. Based on the actual total loss of the transformer area, multiple constraints are used to iteratively calculate the objective function until the loss components of each factor satisfy the convergence condition are obtained. Among them, the factor loss components include the fundamental reference loss, coupling interaction loss, unbalanced additional loss, harmonic additional loss, and reactive power additional loss. The convergence condition includes the number of iterations reaching the iteration number threshold or the difference between the calculation results of two adjacent iterations being less than the difference threshold. This decoupled calculation method for the influencing factors of distribution transformer area line loss determines the initial value of the calculated loss through data configuration, then constructs a third-order symmetric coupling coefficient matrix to quantify the interaction effect of multiple factors, and iteratively calculates the objective function in combination with multiple constraints to obtain the loss components of each factor that meet the convergence condition, thereby improving the calculation accuracy of additional loss and realizing refined calculation of additional loss in the transformer area. This provides a data basis for quantitative decision-making for targeted loss reduction and solves the technical problems of insufficient calculation accuracy and poor adaptability of existing refined calculation methods for additional loss in low-voltage transformer areas.

[0065] In one embodiment of this application, determining the initial values ​​of unbalanced additional loss, harmonic additional loss, and reactive power additional loss based on real-time operating data, the comprehensive equivalent resistance of the transformer area, the equivalent frequency coefficient, and the fundamental reference loss includes: The negative sequence current is calculated based on the three-phase current from real-time operating data. The initial value of unbalanced additional loss is obtained by calculating the comprehensive equivalent resistance of the transformer area, the negative sequence current, and the neutral line current and neutral line equivalent resistance of the real-time operating data. Obtain the skin effect engineering correction coefficient, and calculate the initial value of the harmonic additional loss based on the skin effect engineering correction coefficient, the equivalent frequency coefficient and the fundamental reference loss. Obtain the target economic operating power factor, and calculate the initial value of reactive power additional loss based on the fundamental reference loss, the target economic operating power factor, and the actual operating power factor of real-time operating data.

[0066] It should be noted that the negative sequence current is calculated using the symmetrical component current formula based on the three-phase current from real-time operating data. The initial value of unbalanced additional loss is calculated using the first loss formula based on the comprehensive equivalent resistance of the transformer area, the negative sequence current, and the neutral line current and equivalent resistance from real-time operating data. The skin effect engineering correction coefficient is obtained, and the initial value of harmonic additional loss is calculated using the second loss formula based on the skin effect engineering correction coefficient, the equivalent frequency coefficient, and the fundamental reference loss. The target economic operating power factor is obtained, and the initial value of reactive power additional loss is calculated using the third loss formula based on the fundamental reference loss, the target economic operating power factor, and the actual operating power factor from real-time operating data. In this embodiment, the skin effect engineering correction coefficient is dimensionless, and this parameter comes from the measured results of the conductor skin effect, for example, the conductor cross-sectional area of ​​a copper core cable is 70 mm². 2 Therefore, the skin effect engineering correction factor can be selected as 0.85; the conductor cross-sectional area of ​​the copper core cable is 70mm². 2 ~240mm 2 Therefore, the skin effect engineering correction factor can be selected as 0.75; the conductor cross-sectional area of ​​the copper core cable is 240mm². 2 Therefore, the skin effect engineering correction factor can be selected as 0.65. This skin effect engineering correction factor can be adjusted according to the actual conductor cross-section of the copper core cable in the distribution area. The target economic operating power factor can be selected as 0.95, or adjusted according to the actual situation.

[0067] In this embodiment, the formula for the symmetrical component current is: ; ; The first loss formula is: ; The second loss formula is: ; The third loss formula is: ; In the formula, I2 is the negative sequence current, in amperes (A); I A I B I CThe currents for phases A, B, and C are respectively, in amperes (A). a It is a symmetric component operator and is dimensionless; I N The neutral current is expressed in amperes (A); R N R is the equivalent resistance of the neutral line. eq P is the equivalent resistance of the transformer area. unb0 The initial value for unbalanced additional losses, D is the skin effect engineering correction coefficient, and K is the initial value for unbalanced additional losses. har P is the equivalent frequency coefficient. har0 For the initial value of the additional harmonic loss, cos Φ The actual operating power factor for real-time data, cos Φ 0 represents the target economic operating power factor, P rea0 This is the initial value for reactive power additional losses.

[0068] It should be noted that the neutral current I N It is directly collected by the intelligent fusion terminal; in abnormal situations, it can be derived from the three-phase current. Neutral line equivalent resistance R N R can be used N =K×R eq This formula is used to calculate the resistance, where K is the ratio of the copper core cable resistance to the phase line resistance, which can be adjusted according to the actual conductor type in the transformer substation. In other embodiments, the neutral line equivalent resistance R... N For copper core cables, the phase resistance can be 0.6 times, and the neutral line equivalent resistance R can be... N The resistance of the phase line can be 0.8 times that of the overhead line.

[0069] In one embodiment of this application, determining the equivalent frequency coefficient based on data configuration and real-time operating data includes: If the data is configured as enhanced configuration level or the total harmonic distortion rate of the current in real-time running data is greater than the harmonic distortion rate threshold, then the equivalent frequency coefficient is calculated based on the harmonic order, secondary harmonic current and fundamental current of the real-time running data. If the data configuration is at the standard configuration level or the total harmonic distortion rate of the current in the real-time running data is not greater than the harmonic distortion rate threshold, then the skin effect correction coefficient is obtained. Based on the skin effect correction coefficient and the total harmonic distortion rate of the current in the real-time running data, the equivalent frequency coefficient is calculated.

[0070] It should be noted that the harmonic distortion rate threshold can be selected as 20%, or it can be set according to the actual situation. The equivalent frequency coefficient is obtained by calculating the harmonic order, second harmonic current, and fundamental current using the precise harmonic calculation formula based on real-time operating data. The equivalent frequency coefficient is also obtained by calculating the total harmonic distortion rate of the current using an engineering approximation formula based on the skin effect correction coefficient and the real-time operating data. In this embodiment, the precise harmonic calculation formula is as follows: ; The engineering approximation formula is: ; In the formula, K har I is the equivalent frequency coefficient of the harmonic, dimensionless; h is the harmonic order; I h Ih is the h-th harmonic current, in amperes (A); I1 is the fundamental current, in amperes (A); THDI is the total harmonic distortion of the current, dimensionless; K skin This is the skin effect correction coefficient, which is dimensionless.

[0071] In one embodiment of this application, the initial values ​​of coupling interaction loss and total loss of the transformer area are determined based on the third-order symmetric coupling coefficient matrix, real-time operating data, equivalent resistance of the transformer area, fundamental reference loss, initial value of unbalanced additional loss, initial value of harmonic additional loss, and initial value of reactive power additional loss, including: Based on the three-phase current calculations from real-time operating data, the negative sequence current and the three-phase average current are obtained, and the negative sequence current is taken as the three-phase unbalance characteristic current. Based on all harmonic currents from real-time operating data or the total harmonic distortion rate and fundamental current from real-time operating data, the harmonic characteristic current is obtained. Based on the three-phase average current and the actual operating power factor from real-time operating data, the reactive characteristic current is obtained. The three-factor high-order nonlinear amplification coefficient is obtained. Based on the initial values ​​of unbalance additional loss, harmonic additional loss, reactive power additional loss, three-phase unbalance characteristic current, harmonic characteristic current, reactive power characteristic current, three-factor high-order nonlinear amplification coefficient, and the coupling coefficient of the three factors in the third-order symmetric coupling coefficient matrix, the coupling interaction loss is calculated. The initial value of the total loss of the transformer area is obtained by calculating the initial values ​​of coupling interaction loss, fundamental frequency reference loss, unbalanced additional loss, harmonic additional loss, and reactive power additional loss.

[0072] It should be noted that the initial value of the total loss in the transformer area can be understood as the sum of the initial values ​​of coupling interaction loss, fundamental frequency reference loss, unbalanced additional loss, harmonic additional loss, and reactive power additional loss. The three-phase unbalanced characteristic current I... unb =I2. Three-phase average current I avg It can be obtained by calculating the average value of the three-phase current, i.e., I. avg =(I A +I B +I C ) / 3, I A I B I C These are the currents for phases A, B, and C, respectively.

[0073] In this embodiment, the harmonic characteristic current is calculated based on all subharmonic currents in the real-time operating data or the total harmonic distortion rate and fundamental current of the real-time operating data. This can be understood as follows: if the data is configured at the enhancement configuration level, the harmonic characteristic current can be the square root of all subharmonic currents, i.e.: ; If the data configuration is at the standard configuration level, the harmonic characteristic current is calculated using the harmonic characteristic current formula based on the total harmonic distortion (THD) and fundamental current of the real-time operating data. The harmonic characteristic current formula is as follows: ; In the formula, I har I2, I3, I5, and I7 are the harmonic characteristic currents, and I2, I3, I5, and I7 are the 2nd, 3rd, 5th, and 7th harmonic currents, respectively.

[0074] In this embodiment, the reactive characteristic current is calculated using the reactive characteristic current formula based on the three-phase average current and the actual operating power factor from real-time operating data. The reactive characteristic current formula is: ; In the formula, I rea The reactive characteristic current, I avg This represents the average current across the three phases.

[0075] In this embodiment, the coupling interaction loss is calculated using the coupling interaction loss formula based on the initial values ​​of unbalanced additional loss, harmonic additional loss, reactive power additional loss, three-phase unbalanced characteristic current, harmonic characteristic current, reactive power characteristic current, three-factor high-order nonlinear amplification coefficient, and the coupling coefficients of the pairwise coupling of the three factors in the third-order symmetrical coupling coefficient matrix. The coupling interaction loss formula is as follows: ; In the formula, P couple For coupling interaction loss, C uh I is the coupling coefficient between three-phase imbalance and harmonics, dimensionless; unb The characteristic current of three-phase unbalance is expressed in amperes (A); I har R represents the characteristic harmonic current, measured in amperes (A). eq The equivalent resistance of the transformer substation is expressed in ohms (Ω); C ur I is the coupling coefficient between three-phase imbalance and reactive power, dimensionless; rea The reactive characteristic current is expressed in amperes (A); C hr C is the coupling coefficient between harmonics and reactive power, dimensionless; uhr P is a three-factor high-order nonlinear amplification factor, dimensionless; unb0 Initial value for unbalanced additional losses, in kilowatts (kW); Phar0 P represents the initial value for harmonic additional losses, in kilowatts (kW). rea0 This is the initial value for reactive power additional losses, in kilowatts (kW).

[0076] It should be noted that the three-factor higher-order nonlinear amplification factor can be set according to requirements.

[0077] In one embodiment of this application, obtaining the third-order symmetric coupling coefficient matrix includes: Based on the type of distribution network area access, determine the typical values ​​of the coupling coefficients of the three pairwise coupling factors in the third-order symmetric coupling coefficient matrix; Based on the actual total loss of the transformer area, the fundamental reference loss, the initial value of the unbalanced additional loss, the initial value of the harmonic additional loss, the initial value of the reactive power additional loss, the real-time operation data, and the comprehensive equivalent resistance of the transformer area, the coupling coefficient calibration values ​​of the three factors in the third-order symmetrical coupling coefficient matrix are obtained. The coupling coefficient calibration values ​​include the first coupling calibration coefficient between the three-phase unbalance and harmonics, the second coupling calibration coefficient between the three-phase unbalance and reactive power, and the third coupling calibration coefficient between harmonics and reactive power. The coupling coefficient calibration value is determined based on the three-phase unbalance, total harmonic distortion rate of current, and actual operating power factor of real-time operating data; the coupling coefficient difference is determined by comparing the coupling coefficient calibration value with the corresponding typical value of coupling coefficient. If the difference in coupling coefficients is greater than the difference threshold, then the calibration value of the coupling coefficients of the pairwise coupling factors of the three factors in the third-order symmetric coupling coefficient matrix is ​​used as the coupling coefficient of the pairwise coupling of the three factors in the third-order symmetric coupling coefficient matrix. If the difference in coupling coefficients is not greater than the difference threshold, then the typical value of the coupling coefficient of the pairwise coupling factors of the three factors in the third-order symmetric coupling coefficient matrix is ​​taken as the coupling coefficient of the pairwise coupling of the three factors in the third-order symmetric coupling coefficient matrix.

[0078] It should be noted that the difference threshold can be selected as 0.3. The difference threshold can also be set according to actual needs.

[0079] In this embodiment, the types of distribution network access include residential distribution areas, centralized charging pile distribution areas, and industrial load distribution areas. Based on the type of distribution network access, a preset rule is used to determine the typical values ​​of the pairwise coupling factors of the three factors in the third-order symmetric coupling coefficient matrix. This preset rule can be understood as the typical value range for residential distribution areas being the typical value C of the coupling coefficient between three-phase imbalance and harmonics. uh =0.15~0.30, typical value of coupling coefficient C between three-phase imbalance and reactive power. ur =0.10~0.20, typical value of coupling coefficient C between harmonics and reactive power. hr =0.05~0.12.

[0080] It should be noted that the typical values ​​of the coupling coefficients for the pairwise coupling factors of the three factors in the third-order symmetric coupling coefficient matrix can be adjusted according to the type of distribution network area access. For example, if the type of distribution network area access is a centralized charging pile area, C uh =Take the upper limit, i.e., C uh =0.30; If the distribution network area is connected to an industrial load area, C hr Take the upper limit, i.e., C hr =0.12.

[0081] In this embodiment, the coupling coefficient calibration values ​​of the pairwise coupling factors of the three factors in the third-order symmetric coupling coefficient matrix are calculated using a coefficient calibration formula based on the actual total loss of the transformer area, the fundamental reference loss, the initial value of the unbalanced additional loss, the initial value of the harmonic additional loss, the initial value of the reactive power additional loss, real-time operating data, and the comprehensive equivalent resistance of the transformer area. The coefficient calibration formula is as follows: ; ; ; In the formula, C uh ' is the first coupling calibration coefficient between three-phase imbalance and harmonics, C hr ' is the third coupling calibration coefficient between harmonics and reactive power, C ur ' is the second coupling calibration coefficient between three-phase imbalance and reactive power.

[0082] In this embodiment, the coupling coefficient calibration value is determined based on the three-phase imbalance, total harmonic distortion of current, and actual operating power factor from real-time operating data using a preset judgment rule. The preset judgment rule includes the following: If the three-phase imbalance is greater than the first threshold, the total harmonic distortion of the current is greater than the second threshold, and the actual operating power factor is greater than the third threshold, then the first coupling calibration coefficient will be used as the coupling coefficient calibration value. If the three-phase imbalance is greater than the first threshold, the total harmonic distortion of the current is less than the fourth threshold, and the actual operating power factor is less than the fifth threshold, then the second coupling calibration coefficient will be used as the coupling coefficient calibration value. If the three-phase imbalance is less than the fourth threshold, the total harmonic distortion of the current is greater than the second threshold, and the actual operating power factor is less than the fifth threshold, then the third coupling calibration coefficient will be used as the coupling coefficient calibration value.

[0083] It should be noted that the first threshold can be selected as 10%, the second threshold as 8%, and the third threshold as 0.9. The fourth threshold can be selected as 5%, and the fifth threshold as 0.85. The specific values ​​of the first, second, third, fourth, and fifth thresholds can be adjusted according to the actual engineering of the distribution network. In this embodiment, the coupling coefficient difference can be understood as follows: if the first coupling calibration coefficient is used as the coupling coefficient calibration value, the typical value of the coupling coefficient corresponding to the calculation of the coupling coefficient difference is the typical value of the coupling coefficient between three-phase imbalance and harmonics; if the second coupling calibration coefficient is used as the coupling coefficient calibration value, the typical value of the coupling coefficient corresponding to the calculation of the coupling coefficient difference is the typical value of the coupling coefficient between three-phase imbalance and reactive power; if the third coupling calibration coefficient is used as the coupling coefficient calibration value, the typical value of the coupling coefficient corresponding to the calculation of the coupling coefficient difference is the typical value of the coupling coefficient between harmonics and reactive power.

[0084] In this embodiment, to improve the accuracy of the calculation results, the third-order symmetric coupling coefficient matrix can be obtained by employing an unsupervised self-calibration two-level calibration strategy to determine the coupling coefficients of the pairwise coupling of the three factors in the third-order symmetric coupling coefficient matrix. The two-level calibration strategy refers to the following: if the difference in coupling coefficients is greater than a threshold, the calibration value of the coupling coefficient of each pairwise coupling factor in the third-order symmetric coupling coefficient matrix is ​​used as the coupling coefficient of the pairwise coupling of the three factors in the third-order symmetric coupling coefficient matrix; if the difference in coupling coefficients is not greater than the threshold, the typical value of the coupling coefficient of each pairwise coupling factor in the third-order symmetric coupling coefficient matrix is ​​used as the coupling coefficient of the pairwise coupling of the three factors in the third-order symmetric coupling coefficient matrix.

[0085] It should be noted that the decoupling calculation method for the influencing factors of distribution substation line loss adopts a two-level calibration strategy of typical value initialization + unsupervised self-calibration of dual-factor active operating condition window (selecting three types of windows: unbalanced and harmonic dual-factor active, reactive power near ideal, etc., and taking the median after cross-section calculation, and replacing when the deviation between the calibration value and the typical value exceeds 0.3). This makes the coupling coefficients of the three factors in the third-order symmetric coupling coefficient matrix adaptive with the load characteristics of the substation area, overcoming the shortcomings of existing coupling coefficients which are fixed empirical values, do not consider the differences in load characteristics of the substation area, and do not model high-order couplings.

[0086] Example 2: Figure 2 This is a schematic diagram of the framework of the decoupling calculation device for the influencing factors of distribution substation line loss described in an embodiment of this application. Figure 3 This is a flowchart illustrating the decoupling calculation device for the influencing factors of distribution substation line loss described in an embodiment of this application.

[0087] like Figure 2As shown, this application embodiment provides a decoupling calculation device for the influencing factors of distribution transformer area line loss, including a field data layer 100, an edge computing layer 200 and a cloud platform layer 300 connected in sequence; The field data layer 100 is used to acquire historical and real-time electrical quantity data of the distribution network area according to the sampling time. Edge computing layer 200 is used to obtain the fundamental reference loss, coupling interaction loss and loss components of each factor of the distribution network area access based on historical electrical quantity data and real-time electrical quantity data according to the above-mentioned decoupling calculation method of distribution area line loss influencing factors. The cloud platform layer 300 is used to generate governance decision schemes based on the fundamental frequency reference loss, coupling interaction loss, and loss components of various factors.

[0088] It should be noted that the decoupling calculation method for distribution transformer area line loss in the device of Embodiment 2 has been described in Embodiment 1, and will not be repeated in this embodiment. This distribution transformer area line loss influencing factor decoupling calculation device constructs an edge-cloud collaborative deployment architecture through a field data layer, an edge computing layer, and a cloud platform layer. It uses data configuration to determine the initial value of the calculated loss, then constructs a third-order symmetric coupling coefficient matrix to quantify the multi-factor interaction effect. Combined with multiple constraints, iterative calculations of the objective function are performed to obtain the loss components of each factor that meet the convergence condition, improving the calculation accuracy of additional losses and realizing refined calculation of additional losses in the transformer area, providing data basis for targeted loss reduction decisions. Specifically, the cloud platform layer 300, through the loss tracing and governance decision module, traces each fundamental frequency reference loss, coupling interaction loss, and each factor loss component to branches and users under full-configuration level data configuration and generates a governance decision scheme with a governance priority sequence.

[0089] like Figure 2 and Figure 3 As shown in the embodiment of this application, the edge computing layer 200 includes a measurement hierarchical adaptation module 210, a multi-source data processing module 220, a reference loss calculation module 230, a factor loss initial calculation module 240, and a coupled component decoupling attribution module 250 connected in sequence. The measurement classification and adaptation module 210 is used to determine the operating level of the data source of real-time electrical quantity data to obtain data configuration; and to perform missing estimation processing on real-time electrical quantity data according to the data configuration to obtain standardized real-time data. The multi-source data processing module 220 is used to perform data identification, missing value completion and multi-sampling rate time synchronization alignment processing on standardized real-time data to obtain real-time running data. The reference loss calculation module 230 is used to filter ideal operating condition data from historical electrical quantity data, and to perform univariate linear regression fitting on the ideal operating condition data to obtain the comprehensive equivalent resistance and no-load loss of the transformer area; based on real-time operating data, comprehensive equivalent resistance and no-load loss of the transformer area, the actual total loss of the transformer area and the fundamental reference loss are calculated. The factor loss initial calculation module 240 is used to determine the equivalent frequency coefficient based on the data configuration and real-time operation data; and to determine the initial values ​​of unbalanced additional loss, harmonic additional loss and reactive power additional loss based on the real-time operation data, the comprehensive equivalent resistance of the transformer area, the equivalent frequency coefficient and the fundamental reference loss. The coupling component decoupling attribution module 250 is used to obtain the third-order symmetric coupling coefficient matrix. Based on the third-order symmetric coupling coefficient matrix, real-time operating data, the comprehensive equivalent resistance of the transformer area, the fundamental reference loss, the initial value of the unbalanced additional loss, the initial value of the harmonic additional loss, and the initial value of the reactive power additional loss, it determines the initial value of the coupling interaction loss and the initial value of the total loss of the transformer area. Based on the initial values ​​of the total loss of the transformer area, the fundamental reference loss, the initial value of the unbalanced additional loss, the initial value of the harmonic additional loss, and the initial value of the reactive power additional loss, it determines the observation signal matrix. Based on the observation signal matrix, a function is constructed using the dynamic penalty function method to obtain the objective function. Based on the actual total loss of the transformer area, the objective function is iteratively calculated using multiple constraints until the loss components of each factor that meet the convergence condition are obtained. Among them, the factor loss components include fundamental frequency reference loss, coupling interaction loss, unbalanced additional loss, harmonic additional loss and reactive power additional loss. The convergence condition includes the number of iterations reaching the iteration number threshold or the difference between the calculation results of two adjacent iterations being less than the difference threshold. The objective function is: ; ; ; In the formula, J(w) is the objective function, x is the observed signal matrix, T is the transpose of the matrix, and P... total0 Let P be the initial value of the total loss in the transformer area. base For the fundamental reference loss, P unb0 As the initial value for the unbalanced additional loss, P har0 P is the initial value for the additional harmonic loss. rea0 Let d be the initial value of reactive power loss, d be the eigenvalue diagonal matrix corresponding to the observed signal matrix, E be the eigenvector matrix corresponding to the observed signal matrix, z be the whitened matrix, V be the whitening matrix, G() be the contrast function of the original fast independent component analysis, λ' be the dynamic penalty coefficient of the dynamic penalty function method, and y be the initial value of reactive power loss. i Let E{·} be the factor loss component of the i-th factor, E{·} be the mathematical expectation operator, and w be the separation weight vector matrix of the dynamic penalty function method.

[0090] It should be noted that the contents of each module in the edge computing layer 200 correspond to the contents of each step in the decoupling calculation method of the distribution area line loss in Embodiment 1. In this embodiment, the specific contents of the measurement classification adaptation module 210, the multi-source data processing module 220, the reference loss calculation module 230, the factor loss initial calculation module 240, and the coupling component decoupling attribution module 250 will not be described. In this embodiment, the measurement classification and adaptation module 210 can automatically determine the operating level based on the data source accessed by the transformer area and support automatic upgrading and downgrading; the multi-source data processing module 220 completes the cleaning and time synchronization of multi-source data; the reference loss calculation module 230 calculates the actual total loss and fundamental reference loss of the transformer area based on ideal operating conditions regression; the factor loss initial calculation module 240 calculates the initial values ​​of three types of single-factor additional losses at different levels; the coupling component decoupling attribution module 250 is the core module, which is used to construct a third-order symmetric coupling coefficient matrix and use an improved independent component analysis algorithm with physical constraints (such as non-negativity constraints, energy conservation constraints, proportional rationality constraints, etc.) to perform attribution decomposition, and output five types of loss components with clear physical meanings at the transformer area level: coupling interaction loss, fundamental reference loss, initial value of unbalanced additional loss, harmonic additional loss, and reactive power additional loss.

[0091] In this embodiment, the edge computing layer 200 can be understood as follows: the measurement classification adaptation module 210 and the multi-source data processing module 220 can be deployed in the intelligent converged terminal (edge ​​gateway) of the distribution area, responsible for real-time data access and preprocessing; the benchmark loss calculation module 230, the factor loss preliminary calculation module 240, and the coupled component decoupling attribution module 250 can be deployed in the edge computing unit of the distribution area, responsible for real-time loss calculation and decoupling, and the calculation results are stored locally and uploaded to the cloud platform of the cloud platform layer 300 through the power Internet of Things MQTT protocol. The cloud platform layer 300 is deployed in the cloud platform line loss management system through the loss tracing and governance decision module, responsible for the generation of user-level tracing and governance decisions (such as governance decision schemes) under the full data, the time-series data is stored in the local database of the converged terminal, and the daily frozen data and calculation results are uploaded to the relational database of the cloud platform for storage.

[0092] It should be noted that, as Figure 2As shown, the field data layer 100 consists of four types of acquisition devices: intelligent fusion terminals for distribution areas, branch harmonic monitoring devices, branch monitoring units, and smart meters. It is responsible for acquiring raw electrical quantities (such as historical electrical quantity data and real-time electrical quantity data). The edge computing layer 200 is deployed on the distribution area side. The intelligent fusion terminal has a built-in measurement classification adaptation module 210 and a multi-source data processing module 220 responsible for data access, level determination, and preprocessing. The edge computing unit has three built-in core computing modules: a benchmark loss calculation module 230, a factor loss preliminary calculation module 240, and a coupled component decoupling attribution module 250. These modules are responsible for real-time loss calculation and decoupling. The edge side is equipped with a local time-series database to store real-time operating data. The cloud platform layer 300 deploys a loss tracing and governance decision module, which is responsible for generating branch user-level loss tracing and governance schemes under full data. The cloud platform relational database stores daily frozen data and calculation results. Each layer interacts with the power Internet of Things MQTT protocol to jointly complete the entire process from acquisition to decision-making.

[0093] In this embodiment, the cloud platform layer 300 is used to generate governance decision schemes by performing branch and user-level loss allocation based on the fundamental frequency reference loss, coupling interaction loss, and various factor loss components at the full configuration level. This enables comprehensive multi-dimensional calculation of governance priorities and matching with the governance measure library, generating a recommended sequence of loss reduction schemes, and outputting complete loss attribution results and governance decisions. The loss source tracing and governance decision module includes: If the data is configured at the full configuration level, the four independent components—fundamental reference loss, initial value of unbalanced additional loss, harmonic additional loss, and reactive power additional loss—are allocated to each branch and each user according to the square of the current. That is, the independent loss component borne by a branch is the ratio of its total loss multiplied by the square of the characteristic current corresponding to that branch to the square of the total characteristic current of the distribution area. Coupling and interaction losses are fairly allocated according to marginal contribution using a Shapley value cooperative game method, resolving the unfairness of simple proportional allocation. Three-phase unbalance, harmonics, and reactive power are considered as three participants in a cooperative game, and coupling and interaction losses are considered as cooperative gains. The Shapley value of each participant represents the proportion of coupling and interaction losses that the factor should bear. Then, the factor loss component corresponding to each factor is allocated to each branch and each user according to the proportion of the characteristic value of that branch to the total characteristic value of the distribution area. If the data is configured as standard or enhanced, i.e., the data obtained from the field data layer does not contain branch user table data, no branch user-level allocation is performed, and the generated governance decision scheme only outputs the results of each component at the substation level (such as coupling interaction loss, fundamental frequency reference loss, initial value of unbalanced additional loss, harmonic additional loss, and reactive power additional loss).

[0094] It should be noted that the decoupling calculation device for the influencing factors of the distribution transformer area line loss ensures that the governance decision-making scheme generated by the cloud platform layer 300 strictly adheres to the fair principle of "whoever pollutes is responsible," avoiding the unreasonable situation where one factor dominates while other factors share the losses. The Shapley value calculation formula is: ; In the formula, N ={imbalance, harmonics, reactive power} represents the set of participants, n=3 represents the number of participants, and S represents the subset of participants. i A subset of; v ( S ) is a subset S The coupling interaction loss generated by the combined effect of the factors is calculated using a third-order symmetric coupling coefficient matrix. i ( v (i.e., the participating party) i The proportion of coupling interaction losses to be borne; ! is the mathematical factorial symbol, for example, n! = n (n-1) (n-2) ... 2 1. v ( S ∪{ i}) represents a subset S With participating parties i The union of, containing S and i All numbers in. v ( S ∪{ i})-( S )] represents the participating parties i The marginal contribution caused.

[0095] In this embodiment, the object is the branch or user located by loss tracing, i.e., the end measurement unit (branch monitoring unit, smart meter, etc.) with data configured at the full configuration level. In the loss tracing and governance decision-making module, a governance priority score is calculated for each object based on three dimensions: comprehensive loss ratio, unit governance cost, and investment payback period. The scoring formula is as follows: ; In the formula, S corej For object j Governance priority score; w 1. w 2. w 3 represents the dimension weights of different values, the default value. w 1 = 0.5 w2 = 0.3 w 3=0.2, which can also be adjusted according to local investment preferences. The adjustment rules include: increasing the ratio in regions with a focus on reducing losses. w 1. Improve the efficiency of regions that prioritize investment. w 2 and w 3; P j For the first j Manageable losses for each object C unit The benchmark value for unit governance cost, C j For the first j The actual unit governance cost of each object; T bench This is the benchmark value for the investment payback period. T j For the first j The actual payback period for each object. Based on governance priority scores, corresponding governance measures are automatically matched, generating a recommended sequence of loss reduction solutions.

[0096] It should be noted that, C unit The benchmark value for unit governance cost is taken from the investment evaluation standard for distribution network loss reduction projects or the average unit cost of loss reduction projects in the region over the past three years (unit: yuan / kW or yuan / kWh). C j For the first j The actual unit cost of governance for each object, the estimated investment of matching measures / expected annual power loss reduction, the estimated investment is taken from the unit cost of governance measures (such as the typical design cost or market inquiry of phase-switching switches, active power filters, static var generators, capacitor banks), and the expected annual power loss reduction = P j Expected loss reduction rate of measures Annual operating hours of the transformer substation are calculated from the load curve. T bench The benchmark value for the investment payback period is taken from the prescribed value in the industry evaluation standard for energy-saving and loss-reduction projects, and is a policy parameter. T j For the first j The actual investment payback period for each object T j =Estimated investment in measures / Annual loss reduction benefit, Annual loss reduction benefit = Estimated annual loss reduction in electricity volume Line loss discount is a fixed value for the distribution network.

[0097] In this embodiment of the application, the corresponding governance measures are automatically matched according to the governance priority score, and a recommended sequence of loss reduction schemes is generated, including: The system compiles and organizes a database of loss reduction and mitigation measures. This database is pre-selected by the type of factor affecting the target, with each measure labeled with its applicable factors, applicable level (transformer area / branch / user), unit cost, expected loss reduction rate, and typical payback period. Three-phase imbalance measures correspond to load commutation adjustment, commutation switch modification, and phase-to-phase reactive power compensation devices. Harmonic measures correspond to active power filters, passive filters, and harmonic source mitigation on the user side. Reactive power measures correspond to parallel capacitor banks, static var generators, and local reactive power compensation on the user side. Coupled losses are addressed by combining multiple measures based on the dominant factors. For the governance targets j Compare the proportions of coupling and interaction losses of three-phase imbalance, harmonics, and reactive power (including their respective coupling shares) in the shared losses, select the one with the highest proportion of coupling and interaction losses as the dominant factor, and screen out the set of candidate measures applicable to this factor and applicable to this object level from the loss reduction and mitigation measures library as candidate measures. For each candidate measure, estimate the cost based on the expected reduction rate. C j , T j Together with P j Substitute the values ​​into the scoring formula to calculate the governance priority score for each candidate measure; All governance targets j Each candidate measure is ranked in descending order according to its governance priority score to form a recommended sequence of loss reduction schemes, thus obtaining a governance decision scheme.

[0098] It should be noted that the coupling and interaction losses are first allocated among the three factors of imbalance, harmonics, and reactive power according to the Shapley value. Then, the share of each factor is allocated to each branch and user according to the proportion of the branch characteristic value to the total characteristic value of the transformer area, thus achieving a two-level allocation of coupling and interaction losses. The recommended sequence of loss reduction schemes includes the ranking, the target of governance, the dominant factor, the recommended measures, the reducible loss, the estimated investment, and the expected payback period. In other embodiments, implementation batches can also be divided according to the governance priority score and the score threshold (e.g., governance priority scores higher than the score threshold are included in the recent governance plan), or selected sequentially from the top of the sequence according to the annual investment budget, etc.

[0099] In other embodiments, the governance object is determined. j The dominant factor can be determined using the partial least squares regression (PLS) method, with three-phase unbalance, total harmonic distortion of current, and actual operating power factor as independent variables and bus loss rate as dependent variable. The independent contribution of each factor is obtained through variable projection as the proportion of coupling interaction loss, and the factor with the highest proportion of coupling interaction loss is selected as the dominant factor.

[0100] In this embodiment, the distribution substation line loss influencing factor decoupling calculation device is configured to full capacity based on data through the loss tracing and governance decision-making module of the cloud platform layer. The four independent losses—fundamental wave, three-phase imbalance, harmonics, and reactive power—are allocated to each branch and user according to the square of the current. The coupling interaction loss is fairly allocated according to the marginal contribution of each factor using the Shapley value cooperative game method, achieving accurate loss tracing at the branch / user level. The governance priority score is calculated based on three dimensions: comprehensive loss ratio, unit governance cost, and investment payback period. The device automatically matches the governance measure library to generate a recommended sequence of loss reduction schemes as the governance decision scheme.

[0101] It should be noted that the cloud platform layer of this distribution transformer area line loss influencing factor decoupling calculation device introduces the sharing idea of ​​Shapley value cooperative game and combines it with a multi-dimensional governance priority scoring mechanism to achieve accurate source tracing of coupled interactive losses at the branch and user levels. At the same time, it automatically generates a recommended sequence of loss reduction schemes, which solves the fundamental limitations of traditional methods that can only output the line loss of the transformer area, cannot locate the responsibility for loss, and are prone to misallocation of loss reduction investment. It builds a full-chain technical support from loss calculation to attribution and source tracing to governance decision-making, effectively improving the return on investment for loss reduction.

[0102] Example 3: Figure 4 This is a schematic diagram of the terminal device described in an embodiment of this application.

[0103] like Figure 4 As shown, this application provides a terminal device, including a processor and a memory; Memory is used to store program code and transfer the program code to the processor; The processor is used to execute the above-mentioned decoupling calculation method for the influencing factors of distribution substation line loss according to the instructions in the program code.

[0104] It should be noted that the processor is used to execute the steps in the above-described embodiment of a method for decoupling the influencing factors of distribution substation line loss according to the instructions in the program code. Alternatively, when the processor executes the computer program, it implements the functions of each module / unit in the above-described system / device embodiments.

[0105] For example, a computer program can be divided into one or more modules / units, one or more of which are stored in memory and executed by a processor to complete this application. One or more modules / units can be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in a terminal device.

[0106] Terminal devices can be computing devices such as desktop computers, laptops, handheld computers, and cloud servers. Terminal devices may include, but are not limited to, processors and memory. Those skilled in the art will understand that this does not constitute a limitation on the terminal device, which may include more or fewer components than illustrated, or combinations of certain components, or different components. For example, a terminal device may also include input / output devices, network access devices, buses, etc.

[0107] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor.

[0108] Memory can be an internal storage unit of a terminal device, such as a hard drive or RAM. Memory can also be an external storage device, such as a plug-in hard drive, SmartMedia Card (SMC), Secure Digital (SD) card, or Flash Card. Furthermore, memory can include both internal and external storage units. Memory is used to store computer programs and other programs and data required by the terminal device. Memory can also be used to temporarily store data that has been output or will be output.

[0109] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0110] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection between apparatuses or units through some interfaces, and may be electrical, mechanical, or other forms.

[0111] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0112] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0113] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0114] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A decoupled calculation method for factors affecting line loss in distribution substations, characterized in that, Includes the following steps: Historical electrical quantity data and real-time electrical quantity data of the distribution network area are obtained according to the sampling time. Data identification, missing value completion and multi-sampling rate time synchronization alignment processing are performed on the real-time electrical quantity data to obtain data configuration and real-time operation data. Ideal operating condition data is selected from the historical electrical quantity data, and univariate linear regression is used to fit the ideal operating condition data to obtain the comprehensive equivalent resistance and no-load loss of the transformer area. Based on the real-time operating data, the comprehensive equivalent resistance of the transformer area, and the no-load loss, the actual total loss and fundamental reference loss of the transformer area are calculated. Based on the data configuration and the real-time operating data, the equivalent frequency coefficient is determined; based on the real-time operating data, the comprehensive equivalent resistance of the transformer area, the equivalent frequency coefficient, and the fundamental reference loss, the initial values ​​of unbalanced additional loss, harmonic additional loss, and reactive power additional loss are determined. Obtain the third-order symmetric coupling coefficient matrix. Based on the third-order symmetric coupling coefficient matrix, the real-time operating data, the comprehensive equivalent resistance of the transformer area, the fundamental reference loss, the initial value of the unbalanced additional loss, the initial value of the harmonic additional loss, and the initial value of the reactive power additional loss, determine the initial values ​​of the coupling interaction loss and the total loss of the transformer area. Based on the initial values ​​of the total loss of the transformer area, the fundamental reference loss, the initial values ​​of the unbalanced additional loss, the initial values ​​of the harmonic additional loss, and the initial values ​​of the reactive power additional loss, the observation signal matrix is ​​determined; based on the observation signal matrix, the objective function is constructed using the dynamic penalty function method; based on the actual total loss of the transformer area, the objective function is iteratively calculated using multiple constraints until the loss components of each factor that satisfy the convergence condition are obtained. The loss components include fundamental frequency reference loss, coupling interaction loss, unbalanced additional loss, harmonic additional loss, and reactive power additional loss. The convergence condition includes the number of iterations reaching the iteration number threshold or the difference between the calculation results of two adjacent iterations being less than the difference threshold.

2. The decoupling calculation method for the influencing factors of distribution transformer area line loss according to claim 1, characterized in that, Based on the real-time operating data, the equivalent resistance of the transformer area, the equivalent frequency coefficient, and the fundamental reference loss, the initial values ​​of the unbalanced additional loss, harmonic additional loss, and reactive power additional loss are determined as follows: The negative sequence current is calculated based on the three-phase current from the real-time operating data. The initial value of unbalanced additional loss is calculated based on the comprehensive equivalent resistance of the transformer area, the negative sequence current, and the neutral line current and neutral line equivalent resistance of the real-time operating data. Obtain the skin effect engineering correction coefficient, and calculate the initial value of the harmonic additional loss based on the skin effect engineering correction coefficient, the equivalent frequency coefficient and the fundamental reference loss; Obtain the target economic operating power factor, and calculate the initial value of reactive power additional loss based on the fundamental reference loss, the target economic operating power factor, and the actual operating power factor of the real-time operating data.

3. The decoupling calculation method for the influencing factors of distribution transformer area line loss according to claim 1, characterized in that, Based on the data configuration and the real-time operating data, the equivalent frequency coefficient is determined as follows: If the data is configured as an enhanced configuration level or the total harmonic distortion rate of the current in the real-time running data is greater than the harmonic distortion rate threshold, then the equivalent frequency coefficient is calculated based on the harmonic order, secondary harmonic current and fundamental current of the real-time running data. If the data is configured as a standard configuration or the total harmonic distortion rate of the current in the real-time running data is not greater than the harmonic distortion rate threshold, then the skin effect correction coefficient is obtained, and the equivalent frequency coefficient is calculated based on the skin effect correction coefficient and the total harmonic distortion rate of the current in the real-time running data.

4. The decoupling calculation method for the influencing factors of distribution transformer area line loss according to claim 1, characterized in that, Based on the third-order symmetric coupling coefficient matrix, the real-time operating data, the integrated equivalent resistance of the transformer area, the fundamental reference loss, the initial value of the unbalanced additional loss, the initial value of the harmonic additional loss, and the initial value of the reactive power additional loss, the initial values ​​of the coupling interaction loss and the total loss of the transformer area are determined as follows: Based on the three-phase current of the real-time operating data, the negative sequence current and the three-phase average current are calculated, and the negative sequence current is used as the three-phase unbalance characteristic current; based on all the subharmonic currents of the real-time operating data or the total harmonic distortion rate and fundamental current of the real-time operating data, the harmonic characteristic current is calculated; based on the three-phase average current and the actual operating power factor of the real-time operating data, the reactive characteristic current is calculated. The three-factor high-order nonlinear amplification coefficient is obtained, and the coupling interaction loss is calculated based on the initial values ​​of the unbalance additional loss, the harmonic additional loss, the reactive power additional loss, the three-phase unbalance characteristic current, the harmonic characteristic current, the reactive power characteristic current, the three-factor high-order nonlinear amplification coefficient, and the coupling coefficients of the three factors in the third-order symmetric coupling coefficient matrix. The initial value of the total loss of the transformer area is calculated based on the coupling interaction loss, the fundamental reference loss, the initial value of the unbalanced additional loss, the initial value of the harmonic additional loss, and the initial value of the reactive power additional loss.

5. The decoupling calculation method for the influencing factors of distribution transformer substation line loss according to any one of claims 1-4, characterized in that, Obtaining the third-order symmetric coupling coefficient matrix includes: Based on the type of distribution network area access, determine the typical values ​​of the coupling coefficients of the three pairwise coupling factors in the third-order symmetric coupling coefficient matrix; Based on the actual total loss of the transformer area, the fundamental reference loss, the initial value of the unbalanced additional loss, the initial value of the harmonic additional loss, the initial value of the reactive power additional loss, the real-time operating data, and the comprehensive equivalent resistance of the transformer area, the coupling coefficient calibration values ​​of the pairwise coupling factors of the three factors in the third-order symmetrical coupling coefficient matrix are obtained. The coupling coefficient calibration values ​​include the first coupling calibration coefficient between the three-phase unbalance and harmonics, the second coupling calibration coefficient between the three-phase unbalance and reactive power, and the third coupling calibration coefficient between the harmonics and reactive power. Based on the three-phase imbalance, total harmonic distortion of current, and actual operating power factor of the real-time operating data, determine the coupling coefficient calibration value; and compare the coupling coefficient calibration value with the corresponding typical value of the coupling coefficient to determine the coupling coefficient difference. If the difference in the coupling coefficients is greater than the difference threshold, then the calibration value of the coupling coefficients of the pairwise coupling factors of the three factors in the third-order symmetric coupling coefficient matrix is ​​used as the coupling coefficient of the pairwise coupling of the three factors in the third-order symmetric coupling coefficient matrix. If the difference in the coupling coefficients is not greater than the difference threshold, then the typical value of the coupling coefficient of the pairwise coupling factors of the three factors in the third-order symmetric coupling coefficient matrix is ​​taken as the coupling coefficient of the pairwise coupling of the three factors in the third-order symmetric coupling coefficient matrix.

6. The method for decoupling the calculation of factors affecting line loss in distribution substations according to any one of claims 1-4, characterized in that, The objective function is: ； ； ； In the formula, J(w) is the objective function, x is the observed signal matrix, T is the transpose of the matrix, and P... total0 Let P be the initial value of the total loss in the transformer area. base For the fundamental reference loss, P unb0 For the initial value of the unbalanced additional loss, P har0 P is the initial value for the additional harmonic loss. rea0 Let d be the initial value of reactive power loss, d be the eigenvalue diagonal matrix corresponding to the observed signal matrix, E be the eigenvector matrix corresponding to the observed signal matrix, z be the whitened matrix, V be the whitening matrix, G() be the contrast function of the original fast independent component analysis, λ' be the dynamic penalty coefficient of the dynamic penalty function method, and y be the initial value of reactive power loss. i Let E{·} be the factor loss component of the i-th factor, E{·} be the mathematical expectation operator, and w be the separation weight vector matrix of the dynamic penalty function method.

7. A decoupling calculation device for factors affecting line loss in a distribution substation, characterized in that, include: The field data layer, edge computing layer, and cloud platform layer are connected sequentially. The field data layer is used to acquire historical and real-time electrical quantity data of the distribution network area according to the sampling time. The edge computing layer is used to obtain the fundamental reference loss, coupling interaction loss and various factor loss components of the distribution network area access based on the historical electrical quantity data and the real-time electrical quantity data according to the decoupling calculation method of the distribution area line loss influencing factors as described in any one of claims 1-5. The cloud platform layer is used to generate a governance decision scheme based on the fundamental frequency reference loss, the coupling interaction loss, and the loss components of each factor.

8. The decoupling calculation device for the influencing factors of distribution transformer area line loss according to claim 7, characterized in that, The edge computing layer includes a measurement classification and adaptation module, a multi-source data processing module, a reference loss calculation module, a factor loss initial calculation module, and a coupled component decoupling and attribution module connected in sequence. The measurement classification and adaptation module is used to determine the operating level of the data source of the real-time electrical quantity data and obtain the data configuration. And based on the data configuration, perform missing estimation processing on the real-time electrical quantity data to obtain standardized real-time data; The multi-source data processing module is used to perform data identification, missing value completion, and multi-sampling rate time synchronization alignment processing on the standardized real-time data to obtain real-time running data; The reference loss calculation module is used to filter out ideal operating condition data from the historical electrical quantity data, and to perform univariate linear regression fitting on the ideal operating condition data to obtain the comprehensive equivalent resistance and no-load loss of the transformer area. Based on the real-time operating data, the comprehensive equivalent resistance of the transformer area, and the no-load loss, the actual total loss and fundamental reference loss of the transformer area are calculated. The factor loss initial calculation module is used to determine the equivalent frequency coefficient based on the data configuration and the real-time operating data; and to determine the initial values ​​of unbalanced additional loss, harmonic additional loss and reactive power additional loss based on the real-time operating data, the comprehensive equivalent resistance of the transformer area, the equivalent frequency coefficient and the fundamental reference loss. The coupling component decoupling attribution module is used to obtain a third-order symmetric coupling coefficient matrix. Based on the third-order symmetric coupling coefficient matrix, the real-time operating data, the comprehensive equivalent resistance of the transformer area, the fundamental reference loss, the initial value of the unbalanced additional loss, the initial value of the harmonic additional loss, and the initial value of the reactive power additional loss, it determines the coupling interaction loss and the initial value of the total loss of the transformer area. Based on the initial value of the total loss of the transformer area, the fundamental reference loss, the initial value of the unbalanced additional loss, the initial value of the harmonic additional loss, and the initial value of the reactive power additional loss, it determines the observation signal matrix. Based on the observation signal matrix, it constructs a function using a dynamic penalty function method to obtain the objective function. Based on the actual total loss of the transformer area, iteratively calculates the objective function using multiple constraints until the loss components of each factor satisfy the convergence condition are obtained. The loss components include fundamental frequency reference loss, coupling interaction loss, unbalanced additional loss, harmonic additional loss, and reactive power additional loss. The convergence condition includes the number of iterations reaching the iteration number threshold or the difference between the calculation results of two adjacent iterations being less than the difference threshold.

9. The decoupling calculation device for the influencing factors of distribution transformer area line loss according to claim 8, characterized in that, The objective function is: ； ； ； In the formula, J(w) is the objective function, x is the observed signal matrix, T is the transpose of the matrix, and P... total0 Let P be the initial value of the total loss in the transformer area. base For the fundamental reference loss, P unb0 For the initial value of the unbalanced additional loss, P har0 P is the initial value for the additional harmonic loss. rea0 Let d be the initial value of reactive power loss, d be the eigenvalue diagonal matrix corresponding to the observed signal matrix, E be the eigenvector matrix corresponding to the observed signal matrix, z be the whitened matrix, V be the whitening matrix, G() be the contrast function of the original fast independent component analysis, λ' be the dynamic penalty coefficient of the dynamic penalty function method, and y be the initial value of reactive power loss. i Let E{·} be the factor loss component of the i-th factor, E{·} be the mathematical expectation operator, and w be the separation weight vector matrix of the dynamic penalty function method.

10. A terminal device, characterized in that, Including the processor and memory; The memory is used to store program code and transmit the program code to the processor; The processor is configured to execute the decoupling calculation method for distribution substation line loss influencing factors as described in any one of claims 1-6, according to the instructions in the program code.

Images