Knowledge graph-based transformer line loss influence data processing method and system

Abstract

The application discloses a kind of based on knowledge graph's transformer line loss influence data processing method and system, it is related to data processing technical field, method includes: establishing transformer space coordinate system and dividing sub-region;Based on knowledge graph, dynamic correction impedance and power consumption characteristic coordinates are obtained by executing electrothermal coupling reasoning;According to coordinate offset distance, spatial drift dispersion and spatial drift penalty coefficient are calculated;Combined with power consumption typical time difference, initial line loss influence degree is obtained based on joule heating cross term model;The influence degree is corrected using spatial drift penalty coefficient, and according to betweenness centrality and dynamic correction impedance, topology weight is distributed, and finally weighted aggregation obtains transformer total line loss influence degree.The application solves the problem that existing evaluation does not consider electrothermal coupling impedance drift, time asynchronous cross heating and does not fuse network topology, which leads to deviation in line loss evaluation, and realizes accurate quantitative evaluation of distribution network dynamic physical loss.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, specifically to a method and system for processing data on the impact of transformer line loss based on knowledge graphs. Background Technology

[0002] In the management of line losses in distribution substations and the monitoring of the insulation status of medium and low voltage transformers and secondary circuits, the electricity consumption characteristics of substations exhibit spatial dispersion and temporal asynchrony with the connection of high-power nonlinear loads in industrial parks or residential areas. These abnormal electrothermal coupling losses not only affect the economic operation of the power grid, but the resulting localized abnormal heating is also a core factor accelerating the degradation of the insulation medium of medium and low voltage transformers and secondary circuits, leading to technical bottlenecks in the tracing and assessment of line losses and the accurate monitoring of insulation status.

[0003] First, regarding spatial load location, existing technologies typically bind user electricity consumption characteristics to static planning nodes or calculate based on fixed impedance under standard environments, failing to fully consider the electrothermal coupling effect during power transmission. This means that the Joule heat generated by the current raises the conductor temperature and increases its actual physical impedance. Under high-temperature or heavy-load conditions, the assumption based on static impedance can lead to deviations in the calculation of the electrical distance to the load center, resulting in drift in the spatial location of electricity consumption characteristics. Second, in terms of loss assessment in the time-coordination dimension, existing methods typically use linear time difference ratios or overall arithmetic averages to measure the asynchronicity of multi-user electricity consumption behavior. This fails to reflect the physical mechanism that line loss heating is proportional to the square of the current. Because the time-shifting span and transient current amplitude are not cross-term integrated, existing models struggle to accurately reflect the differences in heating caused by multiple currents under asynchronous or synchronous conditions on local lines, resulting in a lack of corresponding physical dimensional support for time-dimensional line loss assessment. Finally, when performing global weighted aggregation of the line loss impact in different areas of the entire distribution network, existing solutions mostly rely on two-dimensional geographical distance or static electrical hierarchy to allocate weights. They fail to incorporate graph theory to explore the betweenness centrality of specific areas in dynamic network power flow transmission, nor do they link it to the dynamic loss impedance medium of that area. This weight allocation method, which fails to comprehensively consider the criticality of the network structure and the physical loss medium, causes the data aggregation results of the line loss impact in the distribution network to deviate from the actual loss patterns of the power grid, making it difficult to meet the data support requirements of the power grid dispatching end to implement precise loss reduction strategies. Summary of the Invention

[0004] To address the technical problems described in the background section, this invention provides a knowledge graph-based method and system for processing data on the impact of transformer substation line losses. This solves the technical problems of existing methods for assessing line losses, which fail to consider spatial impedance drift caused by electrothermal coupling, lack a time-asynchronous cross-coupling quantization mechanism based on Joule's law, and fail to integrate network topology and dynamic impedance for weight allocation. These shortcomings lead to biased assessments of line loss impact and an inability to accurately reflect the dynamic physical loss patterns of the distribution network. This provides a precise basis for tracing the source of losses in transformer and secondary circuit insulation thermal breakdown early warning.

[0005] A knowledge graph-based method for processing data on line loss impact in transformer substations includes: acquiring the physical coordinates of each data acquisition node within the target substation, establishing a spatial coordinate system, and dividing the target substation into multiple sub-regions; performing electro-thermal coupling inference based on the substation equipment knowledge graph to obtain dynamically corrected impedance, and acquiring the power consumption characteristic coordinates of each power user unit based on the dynamically corrected impedance; calculating the spatial drift dispersion based on the offset distance between the power consumption characteristic coordinates and the physical coordinates of the associated data acquisition nodes, and acquiring the spatial drift penalty coefficient based on the spatial drift dispersion; acquiring the power consumption start time point of each power user unit, and based on the same sub-region... The typical time difference of a sub-region is obtained from multiple power consumption start times within the region; the historical load ramp-up time tolerance is obtained as the standard time difference; based on the typical time difference and the standard time difference, the initial line loss impact is obtained based on the Joule heating cross term model; the initial line loss impact of each sub-region is corrected according to the spatial drift penalty coefficient to obtain the target line loss impact; the betweenness centrality of each sub-region is obtained based on the transformer area equipment knowledge graph; the topology weight is allocated according to the coupling relationship between betweenness centrality and dynamic correction impedance; the target line loss impact of each sub-region is weighted and aggregated according to the topology weight to obtain the total line loss impact of the target transformer area.

[0006] Optionally, the dynamic correction impedance is obtained by performing electrothermal coupling reasoning based on the knowledge graph of the distribution equipment, and the power consumption characteristic coordinates of each power user unit are obtained according to the dynamic correction impedance, including: obtaining the reference voltage of the low-voltage side of the distribution transformer and the operating voltage and operating current of the power user unit; taking the difference between the reference voltage and the operating voltage as the voltage drop value, and dividing the voltage drop value by the operating current to obtain the equivalent electrical impedance.

[0007] Optionally, the process of obtaining dynamically corrected impedance by performing electrothermal coupling reasoning based on the knowledge graph of transformer area equipment, and obtaining the power consumption characteristic coordinates of each power user unit based on the dynamically corrected impedance, further includes: obtaining the current ambient temperature, and retrieving the unit length reference impedance, temperature resistivity, thermal resistivity, and historical average operating current of the corresponding power supply line from the knowledge graph of transformer area equipment; obtaining the temperature regulation multiplier based on the square of the temperature resistivity, current ambient temperature, thermal resistivity, and historical average operating current; and multiplying the unit length reference impedance by the temperature regulation multiplier to obtain the dynamically corrected impedance.

[0008] Optionally, the process of obtaining dynamically corrected impedance by performing electrothermal coupling reasoning based on the knowledge graph of transformer area equipment, and obtaining the power consumption characteristic coordinates of each power user unit based on the dynamically corrected impedance, further includes: dividing the equivalent electrical impedance by the dynamically corrected impedance to obtain the electrical distance to the load center; and extending the electrical distance to the load center along the line spatial direction vector parameters recorded in the knowledge graph of transformer area equipment, starting from the physical coordinates of the associated data acquisition node, to obtain the electrical consumption characteristic coordinates of the power user unit.

[0009] Optionally, the spatial drift dispersion is calculated based on the offset distance between the electricity consumption characteristic coordinates and the physical coordinates of the associated data acquisition node, and the spatial drift penalty coefficient is obtained based on the spatial drift dispersion. This includes: calculating the absolute distance between the electricity consumption characteristic coordinates and the physical coordinates of the nearest data acquisition node to obtain the unit offset distance; calculating the arithmetic mean of the unit offset distances of all power user units within the target distribution area to obtain the spatial drift dispersion; obtaining a preset standard offset distance; dividing the spatial drift dispersion by the standard offset distance to obtain a proportionality coefficient; if the proportionality coefficient is greater than 1, then the proportionality coefficient is defined as the spatial drift penalty coefficient; if the proportionality coefficient is not greater than 1, then the value 1 is defined as the spatial drift penalty coefficient.

[0010] Optionally, the starting time of electricity consumption for each power user unit is obtained, and the typical time difference of the sub-region is obtained based on multiple starting time points of electricity consumption within the same sub-region. This includes: obtaining the mutation rate of the load current of the power user unit, recording the timestamp when the mutation rate exceeds a preset threshold as the starting time point of electricity consumption; extracting the starting time points of electricity consumption for all power user units within the same sub-region, calculating the interval time between any two starting time points of electricity consumption, and extracting the maximum value of the interval time as the typical time difference of the sub-region.

[0011] Optionally, based on the typical time difference and the standard time difference, the initial line loss impact is obtained using the Joule heating cross-term model, including: extracting the real-time load current amplitudes of two power user units with typical time differences within the same sub-region; dividing the typical time difference by the standard time difference to obtain the quotient, subtracting the quotient from the value 1 to obtain the difference, comparing the difference with the value zero, and taking the maximum value as the time impact factor; adding the sum of the squares of the two real-time load current amplitudes to the product of twice the product of the two real-time load current amplitudes and the time impact factor to obtain the equivalent square value of the actual Joule heating current; squaring the sum of the two real-time load current amplitudes to obtain the square value of the extreme Joule heating current; and dividing the equivalent square value of the actual Joule heating current by the square value of the extreme Joule heating current to obtain the initial line loss impact.

[0012] Optionally, the initial line loss impact of each sub-region is corrected according to the spatial drift penalty coefficient to obtain the target line loss impact, including: multiplying the initial line loss impact of each sub-region by the previously obtained spatial drift penalty coefficient to obtain the target line loss impact of each sub-region.

[0013] Optionally, based on the knowledge graph of the distribution equipment, the betweenness centrality of each sub-region is obtained. Topology weights are assigned according to the coupling relationship between betweenness centrality and dynamic correction impedance. The target line loss impact of each sub-region is weighted and aggregated according to the topology weights to obtain the total line loss impact of the target distribution area. This includes: abstracting the distribution equipment knowledge graph into a graph network, treating each sub-region as a node in the graph network, and calculating the betweenness centrality of each sub-region in the topology path; obtaining the average dynamic correction impedance of all power supply lines within each sub-region; calculating the product of the betweenness centrality and the average dynamic correction impedance of each sub-region, dividing this product by the sum of the products of all sub-regions to obtain the topology weight of each sub-region; multiplying the target line loss impact of each sub-region by its corresponding topology weight, and performing a summation operation on the product results of all sub-regions to obtain the total line loss impact of the target distribution area.

[0014] A knowledge graph-based data processing system for transformer substation line loss impact is also provided. This system executes a knowledge graph-based method for processing transformer substation line loss impact data. The system includes: a coordinate mapping module for acquiring the physical coordinates of each data acquisition node within the target transformer substation, establishing a spatial coordinate system, and dividing the target substation into multiple sub-regions; performing electrothermal coupling reasoning based on the transformer substation equipment knowledge graph to obtain dynamically corrected impedance, and obtaining the electricity consumption characteristic coordinates of each power user unit based on the dynamically corrected impedance; a spatial convergence analysis module for calculating the spatial drift dispersion based on the offset distance between the electricity consumption characteristic coordinates and the associated physical coordinates of the data acquisition nodes, and obtaining the spatial drift penalty coefficient based on the spatial drift dispersion; and a spatiotemporal coordination module. This module is used to obtain the power consumption start time of each power user unit, and to obtain the typical time difference of the sub-region based on multiple power consumption start times within the same sub-region. It also obtains the historical load ramp-up time tolerance as the standard time difference, and, based on the typical time difference and the standard time difference, obtains the initial line loss impact based on the Joule heating cross-term model. A topology weighting module is used to correct the initial line loss impact of each sub-region based on the spatial drift penalty coefficient to obtain the target line loss impact. Finally, it obtains the betweenness centrality of each sub-region based on the transformer area equipment knowledge graph, allocates topology weights according to the coupling relationship between betweenness centrality and dynamic correction impedance, and weights and aggregates the target line loss impact of each sub-region based on the topology weights to obtain the total line loss impact of the target transformer area.

[0015] The beneficial effects of this invention are reflected in: In the knowledge graph-based data processing method for transformer substation line loss impact, firstly, by introducing an electrothermal coupling reasoning mechanism to obtain dynamically corrected impedance, this scheme overcomes the limitations of existing static impedance calculations under heavy load and high-temperature environments. This allows the mapping of electricity consumption characteristic coordinates to reflect impedance changes caused by Joule heating, thus objectively reflecting the dynamic drift trajectory of the load centroid in the spatial coordinate system and reducing spatial positioning errors. Secondly, this scheme constructs a Joule heating cross-term model, cross-coupling the time-shifting peak span reflecting macroscopic electricity consumption behavior with the microscopic load current amplitude. Combined with energy dissipation laws, it quantifies the actual impact of asynchronous electricity consumption behavior on line loss, providing a matching electrical and physical foundation for time-coordinated line loss assessment. Finally, in the global data aggregation stage, this scheme improves the weight allocation method based on geographical distance. By calculating the betweenness centrality of each sub-region in the graph network and coupling it with the average dynamically corrected impedance, it identifies nodes within the distribution network that bear power flow crossings and consume a lot of energy. This differentiated weighting mechanism, which combines graph theory topological criticality with dynamic physical impedance, ensures that the final output of the target transformer area's line loss impact conforms to the actual energy dispersion and loss patterns of the transformer area, providing an objective quantitative basis for maintenance personnel to carry out line loss management. Attached Figure Description

[0016] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the accompanying drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn to scale.

[0017] Figure 1 This is a schematic diagram illustrating the steps of the knowledge graph-based method for processing data on line loss impact in transformer substations according to the present invention. Figure 2 This is a schematic diagram of part of step S1 in the knowledge graph-based method for processing data on line loss impact in transformer substations according to the present invention. Figure 3 This is a schematic diagram of part of step S2 in the knowledge graph-based method for processing data on line loss impact in transformer substations according to the present invention; Figure 4 This is a schematic diagram of part of step S3 in the knowledge graph-based method for processing data on line loss impact in transformer substations according to the present invention. Figure 5 This is a schematic diagram of part of step S4 in the knowledge graph-based method for processing data on line loss impact in transformer substations according to the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0019] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0020] It should be noted that similar labels and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, the terms "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0021] This invention provides a knowledge graph-based method for processing data on the impact of line loss in transformer substations, such as... Figure 1 As shown, in one specific embodiment, the method includes: S1. Obtain the physical coordinates of each data acquisition node within the target distribution area, establish a spatial coordinate system, and divide the target distribution area into multiple sub-regions; perform electrothermal coupling reasoning based on the distribution area equipment knowledge graph to obtain the dynamically corrected impedance, and obtain the power consumption characteristic coordinates mapped by each power user unit based on the dynamically corrected impedance; S2. Calculate the spatial drift dispersion based on the unit offset distance between the electricity consumption characteristic coordinates and the physical coordinates of the associated data acquisition nodes, and obtain the spatial drift penalty coefficient based on the spatial drift dispersion. S3. Obtain the power consumption start time of each power user unit, and obtain the typical time difference of the sub-region based on multiple power consumption start times in the same sub-region; obtain the historical load ramp-up time tolerance as the standard time difference, and obtain the initial line loss impact based on the Joule heating cross term model based on the typical time difference and the standard time difference. S4. Correct the initial line loss impact of each sub-region according to the spatial drift penalty coefficient to obtain the target line loss impact; obtain the betweenness centrality of each sub-region based on the knowledge graph of the equipment in the distribution area, allocate topology weights according to the coupling relationship between betweenness centrality and dynamic correction impedance, and perform weighted aggregation of the target line loss impact of each sub-region according to the topology weights to obtain the bus line loss impact of the target distribution area.

[0022] In this embodiment, it should be noted that in S1, the spatial characteristic coordinates of the electrical load are objectively reconstructed mainly through electrothermal coupling reasoning. In specific execution, the target distribution area boundary is first determined and a spatial coordinate system is established, dividing the park into multiple sub-areas and associating them with 120 industrial power user units. Taking a large stamping workshop as an example, its operating current is collected as 120 amps, operating voltage as 210 volts, and the base voltage of the distribution transformer as 220 volts. From this, the voltage drop is calculated to be 10 volts, and the equivalent electrical impedance is approximately 0.0833 ohms. Subsequently, the current ambient temperature of 35 degrees Celsius is obtained, and the reference impedance of the line at 20 degrees Celsius (0.005 ohms per meter), temperature resistivity (0.004), thermal resistivity (0.0002), and historical average current (100 amps) are retrieved from the distribution area equipment knowledge graph. Based on these physical parameters, the dynamically corrected impedance after being affected by both temperature rise and current heating is deduced to be 0.00534 ohms per meter. Dividing the equivalent electrical impedance by the dynamically corrected impedance yields an electrical distance of 15.6 meters from the load center. Starting from the physical coordinates 100,100,0 of the data acquisition node associated with this workshop, extending 15.6 meters along the line vector direction, its electrical characteristic coordinates are mapped to 115.6,100,0. This step incorporates an electrothermal coupling mechanism into the spatial mapping logic, which effectively reduces spatial positioning errors caused by static impedance assumptions under heavy load conditions.

[0023] In S2, spatial offset and convergence analysis is performed based on the coordinate data obtained in the previous steps to quantify the drift state of the load centroid of the entire distribution area. Specifically, the three-dimensional absolute distance between the power consumption characteristic coordinates of each power user unit and its corresponding static physical node is calculated. Continuing with the data from the aforementioned stamping workshop, its unit offset distance is 15.6 meters. Then, the unit offset distances of all 120 units in the park are summarized and the arithmetic mean is calculated, resulting in a spatial drift dispersion of 18.5 meters for the entire park at that moment. To assess the substantial impact of this dispersion on the operating status of the distribution area, the standard offset distance of 15.0 meters, which was preset during the planning of the distribution area, is retrieved from the knowledge graph. The proportionality coefficient is obtained by calculating the ratio of the spatial drift dispersion to the standard offset distance, i.e., 18.5 meters divided by 15.0 meters, resulting in a proportionality coefficient of approximately 1.233. If the proportionality coefficient is greater than 1, it is defined as the spatial drift penalty coefficient; if the proportionality coefficient is not greater than 1, a value of 1 is defined as the spatial drift penalty coefficient. Since 1.233 is greater than 1, it indicates that the current drift has exceeded the ideal range, so the spatial drift penalty coefficient is set to 1.233. The technical approach used in this step is to construct a convergence model by comparing statistical averaging with a preset benchmark. Its beneficial effect is that it can objectively reflect the overall spatial aggregation degree of the load center deviating from the static planning node caused by the irregular operation of a large number of devices with a unified data dimension, providing a quantitative basis for subsequent line loss correction.

[0024] In S3, the focus is on analyzing the synergy of electricity consumption time dimension, quantifying the impact of asynchronous electricity consumption on line loss by combining Joule heating law. Current surges in equipment within sub-regions are monitored to obtain the starting time of electricity consumption, and the maximum time interval between the start-up of any two devices is calculated. Taking sub-region A as an example, an air compressor and an injection molding machine generate a maximum typical time difference of 12 seconds, with their real-time load currents being 150 amps and 100 amps, respectively. The historical load ramp-up time tolerance for this period is extracted from the knowledge graph as the standard time difference, with a value of 30 seconds. When calculating the initial line loss impact, a simple linear time ratio is no longer used; instead, a cross-term model based on the current square feature is constructed. The time staggering ratio is incorporated into the cross-term of the product of the current amplitudes of the two devices to calculate the equivalent square value of heating reflecting the actual staggering state, which is then divided and compared with the extreme heating square value when the two currents are completely synchronized. Based on the above data, the calculated initial line loss impact of sub-region A is approximately 0.808, while that of other sub-regions B and C are 0.650 and 0.900, respectively. The technical approach of this step is to cross-couple the macroscopic time peak shifting span with the microscopic transient current amplitude. Its beneficial effect is that the assessment in the time dimension conforms to the nonlinear physical law of energy dissipation, thereby objectively quantifying the actual heating impact of asynchronous power consumption on local circuits.

[0025] In S4, the previously acquired spatial and temporal data are coupled, and the weighted aggregation of the line loss impact of the distribution network is completed using graph theory topology attributes. First, the initial line loss impact of each sub-region is corrected using the spatial drift penalty coefficient of 1.233 obtained in S2, resulting in target line loss impacts of approximately 0.996, 0.801, and 1.110 for sub-regions A, B, and C, respectively. Then, the distribution network is abstracted as a graph network, and the betweenness centrality of each sub-region in the power flow transmission path is extracted. The topology weights are then assigned based on the average dynamic correction impedance of each region. The betweenness centralities of sub-regions A, B, and C are known to be 0.5, 0.3, and 0.2, respectively, and their average dynamic correction impedances are 0.005, 0.006, and 0.004 ohms per meter, respectively. Multiplying the centrality by the impedance yields the base weight for each region, which is then normalized to give a topology weight of approximately 0.490 for sub-region A, approximately 0.353 for sub-region B, and approximately 0.157 for sub-region C. Finally, the target line loss impact of each sub-region is multiplied and weighted by its topology weight, and the resulting aggregated value is approximately 0.945 for the total line loss impact of the industrial park. This global index can be directly mapped to the comprehensive insulation thermal stress level borne by the low-voltage transformers and secondary circuit systems in this region. The technical approach of this step is to establish a weight allocation mechanism based on betweenness centrality and physical loss impedance. Its beneficial effect is that the global assessment data can objectively consider both the criticality of the network structure and the actual distribution of energy-consuming media, providing reliable data model support for accurately monitoring and predicting the insulation health status of transformers and secondary circuits.

[0026] In summary, the knowledge graph-based method for processing line loss impact data in distribution areas firstly overcomes the limitations of existing static impedance calculations under heavy load and high-temperature environments by introducing an electrothermal coupling reasoning mechanism to obtain dynamically corrected impedance. This allows the mapping of electricity consumption characteristic coordinates to reflect impedance changes caused by Joule heating, thus objectively reflecting the dynamic drift trajectory of the load centroid in the spatial coordinate system and reducing spatial positioning errors. Secondly, this method constructs a Joule heating cross-term model, cross-coupling the time-shifting peak span reflecting macroscopic electricity consumption behavior with the microscopic load current amplitude. Combined with energy dissipation laws, it quantifies the actual impact of asynchronous electricity consumption behavior on line loss, providing a matching electrical and physical foundation for line loss assessment in the time-coordinated dimension. Finally, in the global data aggregation stage, this method improves the weight allocation method based on geographical distance. By calculating the betweenness centrality of each sub-region in the graph network and coupling it with the average dynamically corrected impedance, it identifies nodes within the distribution network that bear power flow crossings and consume a lot of energy. This differentiated weighting mechanism, which combines graph theory topological criticality with dynamic physical impedance, ensures that the final output of the target transformer area's line loss impact conforms to the actual energy dispersion and loss patterns of the transformer area, providing an objective quantitative basis for maintenance personnel to carry out line loss management.

[0027] like Figure 2 As shown, in one specific embodiment, S1 includes: S11, identifying the physical boundary of the target transformer area, obtaining the physical coordinates of each data acquisition node, and establishing a spatial coordinate system with its geometric center as the origin. The target transformer area is divided into multiple sub-regions of equal area along the horizontal direction, and an archive mapping relationship between power user units and sub-regions is established.

[0028] S12. Collect the operating voltage and operating current of the i-th power user unit at the current moment to obtain the reference voltage on the low-voltage side of the distribution transformer. Use the difference between the reference voltage and the operating voltage as the voltage drop value, and divide the voltage drop value by the operating current to obtain the equivalent electrical impedance.

[0029] S13. Obtain the current ambient temperature and retrieve the reference impedance per unit length, temperature resistivity, thermal resistivity, and historical average operating current of the corresponding power supply line from the equipment knowledge graph of the distribution area. Obtain the dynamically corrected impedance based on these parameters. The dynamically corrected impedance, as defined here, refers to the AC impedance per unit length of the conductor that has been corrected for real-time fluctuations in load current and ambient temperature. Specifically, its calculation logic is as follows:

[0030] in, Dynamic correction impedance for power supply lines (unit: ); For power supply lines in The reference impedance per unit length (unit: ); Temperature resistivity of power supply line material (unit: ); Current ambient temperature (unit: ); Thermal resistivity of power supply lines (unit: This characterizes the temperature rise effect produced by the square of a unit current. The historical average operating current of this power supply line within a preset time window (unit: ).

[0031] Among them, the reference impedance per unit length, temperature resistivity, and thermal resistivity retrieved from the knowledge graph of the equipment in the distribution area are static physical parameters. These parameters are theoretical initial values ​​obtained by the system during the initial commissioning phase of the distribution area by automatically parsing the electronic ledger of the cable nameplate of the power supply line. They are then combined with the on-site benchmark measured voltage drop data of the cable under no-load and constant preset ambient temperature for polynomial fitting calibration before being entered into the knowledge graph. For example, for a certain type of overhead aluminum wire, the system first reads the manufacturer's nominal reference impedance of 0.0048 ohms per meter, then uses the constant temperature benchmark test data before commissioning for fine-tuning, and finally solidifies the accurately calibrated 0.005 ohms per meter into the knowledge graph. This effectively reduces the physical reference drift error caused by cable manufacturing tolerances and long-term service aging.

[0032] Specifically, the selection of the preset time window in the above expression is determined by analyzing the historical production rhythm of the industrial load in the target area. The system fits the historical daily load curve of the electrical equipment on typical working days in the past six months and extracts the load main frequency fluctuation cycle. In order to fully cover a complete heat transfer process of heating and dissipation, one-third of the main frequency cycle is defined as the preset time window. For example, if the frequency domain analysis shows that the load of a heavy industry workshop exhibits a 45-minute main frequency fluctuation, the system will automatically use one-third of this time, i.e., 15 minutes, as the preset time window to extract the historical average operating current, ensuring that the assessment of the added value of temperature rise is neither delayed nor overly sensitive.

[0033] S14. Divide the equivalent electrical impedance by the dynamically corrected impedance to obtain the electrical distance to the load center. Starting from the physical coordinates of the associated data acquisition node, extend the electrical distance to the load center along the line spatial direction vector parameters recorded in the transformer area equipment knowledge graph to obtain the power consumption characteristic coordinates of the power user unit. These power consumption characteristic coordinates represent the dynamic electrical centroid position of the user's power load under the grid topology mapping.

[0034] In this embodiment, it should be noted that in S11, the physical boundary of the target distribution area is first identified, the physical coordinates of each data acquisition node are obtained, and a spatial coordinate system is established with its geometric center as the origin. The target distribution area is divided into multiple sub-regions of equal area along the horizontal direction, and a file mapping relationship between power user units and sub-regions is established. The specific number of sub-regions and the area grid size are preset. This parameter is determined by the system based on the historical average load density in the target distribution area and the spatial distribution topology density of the data acquisition nodes. The system uses a spatial clustering algorithm to ensure that each sub-region contains at least one key topological branch and minimizes the variance of the number of power users. Taking a distribution area in a heavy machinery manufacturing park as an example, the system counts 120 nodes distributed in the 50,000 square meter park. To balance the computing load in each grid, the preset grid size is automatically locked after density clustering deduction, and three sub-regions of equal area are objectively output, with a total of 120 industrial power user units. Through this division of physical boundaries and spatial grids, the power equipment is fixed in a unified three-dimensional reference frame, providing a stable spatial basis for subsequent calculations. The technical means of this step lies in constructing an objective geometric topological base. The beneficial effect is that it enables the abstract electrical parameters within the transformer area to have quantifiable physical coordinate references, thereby supporting the subsequent distance calculation algorithm.

[0035] In S12, the first... The operating voltage and current of each power user unit at the current moment are used to obtain the reference voltage on the low-voltage side of the distribution transformer. The difference between the reference voltage and the operating voltage is taken as the voltage drop value, and the voltage drop value is divided by the operating current to obtain the equivalent electrical impedance. In sub-area A of the aforementioned park, taking a large stamping workshop as an example, the smart meter collects its current operating current as 120 amps and operating voltage as 210 volts, while the reference voltage on the low-voltage side of the distribution transformer is 220 volts. By calculating the difference, the voltage drop value is found to be 10 volts, and further divided by the operating current, the current equivalent electrical impedance of the workshop is approximately 0.0833 ohms. This step converts voltage and current into impedance parameters using electrical laws. The beneficial effect is that it transforms the user's load status into an impedance dimension that reflects the grid transmission pressure, providing data for subsequent distance calculations.

[0036] In S13, the dynamic correction impedance is calculated to restore the true physical state of the energy-consuming medium. The core calculation expression for this step is: The original design intent of this operational logic was to address the location distortion problem caused by the long-standing static impedance assumption in existing line loss analysis. In the actual operating environment of power distribution networks, the physical resistivity of overhead lines or underground cables is not constant, but is affected by both external ambient temperature and Joule heating of internal current.

[0037] Furthermore, the basic subtraction part within the parentheses of the expression This is used to quantify the temperature difference between the current actual ambient temperature and the factory standard test temperature (usually 20 degrees Celsius). This temperature difference constitutes the baseline temperature offset of the conductor under no-load conditions; while the additive terms in the expression... This profoundly reflects the coupling relationship between thermodynamics and electromagnetism. It uses the square of the historical average operating current of the power supply line within a preset time window, multiplied by the thermal resistance coefficient, to objectively quantify the additional temperature rise caused by the heat converted and accumulated by the Joule effect when the current passes through the conductor. This temperature rise model based on the square of the current truly restores the self-heating equilibrium state of the conductor under heavy load.

[0038] Furthermore, by adding these two temperature rises together, we obtain the comprehensive absolute temperature rise of the conductor deviating from the standard temperature. This is then multiplied by the temperature resistivity of the power supply line material and added to the value of 1, thus constructing a dimensionless temperature adjustment multiplier. Finally, by multiplying the reference impedance per unit length under standard conditions by this adjustment multiplier, we obtain the dynamic correction impedance that fluctuates in real time with operating conditions.

[0039] Based on the specific data from the aforementioned heavy machinery manufacturing park application scenario, the current ambient temperature was collected. 35 The target line segment was retrieved from the equipment knowledge graph of the transformer area in 20... The reference impedance per unit length below It is 0.005 Temperature resistivity It is 0.004 Thermal resistance coefficient It is 0.0002 Furthermore, the historical average operating current of this section of the line... 100 Substitute the data from these specific physical units into the expression for calculation: First, calculate the ambient temperature rise offset as... Next, the additional temperature rise caused by the self-heating of the load current is calculated as follows: The sum of the two results in the actual total temperature rise of the conductor. Subsequently, the temperature regulation multiplier was calculated as follows: The final dynamically corrected impedance is obtained. .

[0040] This calculation process, by introducing bidirectional electrothermal physical coupling, objectively reflects that the actual impedance of the circuit under the current extreme operating conditions of high temperature and carrying hundreds of amperes of current has increased by 6.8% compared to the reference value. If this calculation logic is not used in subsequent steps and the 0.005 value is continued... When calculating the physical distance using the equivalent electrical impedance, the static impedance can lead to an overestimation of the electrical distance to the load center, resulting in a virtual forward shift in the spatial mapping of the electrical characteristic coordinates. Therefore, this expression corrects for the distortion of the microscopic parameters of the physical medium, providing a reliable data foundation for accurately determining the spatial discrete state of the load centroid.

[0041] In S14, the equivalent electrical impedance is divided by the dynamically corrected impedance to obtain the electrical distance to the load center. Starting from the physical coordinates of the associated data acquisition node, the electrical distance to the load center is extended along the line spatial direction vector parameters recorded in the knowledge graph of the distribution area equipment to obtain the power consumption characteristic coordinates of the power user unit. According to the aforementioned data, the equivalent electrical impedance of the stamping workshop is 0.0833 ohms, which is divided by the dynamically corrected impedance of 0.00534 ohms per meter, resulting in an electrical distance to the load center of 15.6 meters. The knowledge graph indicates that the coordinates of the data acquisition node of this workshop are (100, 100, 0) and extend along the positive X-axis, thus mapping the power consumption characteristic coordinates to (115.6, 100, 0). This step transforms the impedance ratio into spatial coordinates by combining the topological direction. The beneficial effect is that it gives the electrical characteristics an objective spatial mapping location, reducing the positioning deviation caused by relying on static planning data.

[0042] like Figure 3 As shown, in one specific implementation, S2 includes: S21, calculating the absolute distance between the electricity consumption characteristic coordinates of the i-th power user unit and the physical coordinates of the nearest data acquisition node, and obtaining the unit offset distance.

[0043] S22. Calculate the arithmetic mean of the unit offset distances of all power user units within the target distribution area to obtain the spatial drift dispersion. Spatial drift dispersion refers to the group average statistical deviation of each user's electricity consumption characteristic coordinates from their corresponding static physical reference point. This dispersion objectively quantifies the overall spatial dispersion of user electricity consumption behavior.

[0044] S23. Obtain the preset standard offset distance from the equipment knowledge graph of the distribution area. Calculate the difference between the standard offset distance and the spatial drift dispersion, and divide this difference by the standard offset distance to obtain the scaling factor. If the scaling factor is greater than 1, it is defined as the spatial drift penalty factor; if the scaling factor is not greater than 1, a value of 1 is defined as the spatial drift penalty factor. The spatial drift penalty factor here is used to describe the degree of spatial aggregation of user power distribution within the distribution area relative to the theoretical design center point.

[0045] The preset standard offset distance is derived from the historical ideal operating data of the target transformer area under the initial completion and without extreme heavy load conditions. Specifically, the system retrieves the electricity coordinate offset records of the transformer area for 30 days in the lightest month of the past three years, calculates the baseline dispersion by arithmetically averaging the spatial drift dispersion of all users in the network each day, and adds twice the historical standard deviation to it as the allowable reasonable offset boundary. For example, if the average offset dispersion of these 30 days is 11.0 meters and the standard deviation is 2.0 meters, then the preset standard offset distance is calculated to be 11.0 meters plus 4.0 meters, which equals 15.0 meters.

[0046] In this embodiment, it should be noted that in S21, the calculation of the first... The absolute distance between the electricity consumption characteristic coordinates of each power user unit and the physical coordinates of the nearest data acquisition node is used to obtain the unit offset distance. Based on the coordinate data calculated in the previous step, Euclidean geometric distance calculation is performed in a three-dimensional coordinate system. Taking the aforementioned stamping workshop as an example, its electricity consumption characteristic coordinates are (115.6, 100, 0), while the static physical coordinates of its nearest data acquisition node are (100, 100, 0). Using the distance calculation formula between the two points, the absolute straight-line distance between these two points, i.e., the unit offset distance, is 15.6 meters. The technical means used in this step is to perform spatial geometric difference calculation between the abstract electrical mapping coordinates and their original physical nodes. The beneficial effect is that it quantifies the degree of deviation of the actual electricity load of a single user from its physical design node, providing basic parameters for subsequent group analysis.

[0047] In step S22, the arithmetic mean of the unit offset distances of all power user units within the target distribution area is calculated to obtain the spatial drift dispersion. The unit offset distances of all 120 power user units within the area are summed, and the sum is divided by the total number of units, 120. This statistical calculation yields a spatial drift dispersion of 18.5 meters for the entire area at that moment. This value integrates the deviation states of high-load and low-load users, reflecting the overall spatial displacement of the load centroid within the distribution area. This step uses the arithmetic mean statistical method to perform central tendency analysis on the dispersed individual offset data. The beneficial effect is that it filters out the fluctuation interference from a single extreme user, objectively presenting the spatial dispersion level of the load distribution in the entire distribution area at that time section, providing a macroscopic indicator for measuring the energy convergence status of the power grid.

[0048] In step S23, a preset standard offset distance is obtained from the equipment knowledge graph of the distribution area, and the ratio of spatial drift dispersion to the standard offset distance is calculated to obtain a proportionality coefficient. The standard offset distance for this distribution area is retrieved from the graph as 15.0 meters, and the spatial drift dispersion is 18.5 meters. The proportionality coefficient is obtained by calculating the ratio of spatial drift dispersion to the standard offset distance, i.e., 18.5 meters divided by 15.0 meters, resulting in a proportionality coefficient of approximately 1.233. Since 1.233 is greater than 1, it indicates that the current drift has exceeded the ideal range; therefore, the spatial drift penalty coefficient is set to 1.233. This step, by setting a safety benchmark and combining logical judgment to construct a convergence model, achieves the beneficial effect of transforming the dispersion with physical length units into a dimensionless control coefficient, thus facilitating its subsequent participation in the weighted correction calculation of line loss impact.

[0049] like Figure 4 As shown, in one specific embodiment, S3 includes: S31, monitoring the mutation rate of the load current of the power user unit, and recording the timestamp when the mutation rate exceeds a preset threshold as the start time of electricity consumption.

[0050] S32. Extract the power consumption start time of all power user units in the same sub-region, calculate the interval time between any two power consumption start time points, and extract the maximum value of the interval time as the typical time difference of the sub-region.

[0051] S33. Extract the historical load ramp-up time tolerance of the sub-region in the corresponding time period from the equipment knowledge graph of the transformer area as the standard time difference. This standard time difference represents the asynchronous threshold that the transformer area can accept based on historical behavioral characteristics during operation.

[0052] S34. Extract the real-time load current amplitude of two power user units with typical time differences within the same sub-region, and calculate the initial line loss impact degree based on the time difference parameter. The initial line loss impact degree refers to the preliminary assessment component determined by the synchronicity of user electricity consumption timing within a specific time window. Its calculation logic is as follows:

[0053] in, This represents the initial line loss impact (dimensionless proportion) of this sub-region. and Real-time load current amplitudes (unit: ) for two power user units that generate a typical time difference. ); Typical time difference (unit: ); Standard time difference (unit: The numerator of this model represents the square of the equivalent heating current after considering time asynchrony, and the denominator represents the square of the extreme heating current when it is fully synchronized. Their ratio objectively reflects the impact of asynchronous power consumption on local line losses.

[0054] In this embodiment, it should be noted that in S31, the sudden change rate of the load current of the power user unit is monitored, and the timestamp when the sudden change rate exceeds a preset threshold is recorded as the start time of electricity consumption. In the actual operation of the industrial park, the current rises when different production equipment is turned on, and the smart meter monitors this change with a high-frequency sampling rate. When a large air compressor or injection molding machine is connected to the power grid, its load current rise rate exceeds the set fluctuation threshold, and the current timestamp data is recorded in real time. The technical means used in this step is to capture the timing of actions based on electrical sudden change characteristics. The beneficial effect is that it can locate the load jump start point of each device, eliminate data interference during equipment standby or stable operation, and provide a high-resolution timing reference point for subsequent evaluation of the time coordination of multi-user electricity consumption behavior.

[0055] The aforementioned preset threshold (i.e., the set fluctuation threshold) is obtained by the system through normal distribution analysis based on the massive high-frequency current sampling data of high-power equipment in the sub-region during the past three months of stable operation. The system extracts the upper boundary value of the 95% confidence interval of the current mutation rate as a basic reference, and then automatically calibrates it by multiplying it by a reliability margin coefficient of 1.2 to obtain the preset threshold. For example, historical statistics show that the maximum current mutation slope of the equipment in the region during stable operation is 5 amperes per millisecond. After multiplying by a margin of 1.2, the system accurately sets the preset threshold to 6 amperes per millisecond. Once the real-time mutation rate exceeds 6 amperes, it is immediately determined to be a start-up action, thereby effectively filtering out false triggering interference caused by conventional power grid current harmonic fluctuations.

[0056] In step S32, the power consumption start time points of all power user units within the same sub-region are extracted. The interval between any two power consumption start time points is calculated, and the maximum value of the interval is extracted as the typical time difference for that sub-region. Within sub-region A, the power consumption start time points of multiple production line devices are compared pairwise and subtracted. Among these, the start time interval between a large air compressor and an injection molding machine is the largest among all combinations. This largest time interval is extracted, and the typical time difference is 12 seconds. The technical approach of this step is to filter out the extreme values ​​of the time span by calculating the difference. The beneficial effect is that it can objectively capture the asynchronous phenomenon of user power consumption behavior within the same local power grid, thus providing boundary condition data support for subsequent evaluation of the time coordination dimension of line loss caused by peak-shifting startup in this region.

[0057] In S33, the historical load ramp-up time tolerance of the sub-region during the corresponding time period is extracted from the equipment knowledge graph of the distribution area as the standard time difference. Due to the objective differences in the production rhythm of the industrial park under different seasons and shifts, based on the accumulated historical operating data of the same type of time period in the graph, it was found that the historical load ramp-up time tolerance of the park during the summer afternoon period is 30 seconds. The specific logic for obtaining the historical load ramp-up time tolerance is as follows: the system iterates through and extracts all abnormal events that triggered the main transformer temperature rise exceeding the limit alarm in the target area over the past five years. For each exceeding limit event, it traces the peak-shifting start-up time interval of multiple high-power devices before its occurrence. After removing outliers, it calculates the arithmetic mean of the time intervals that triggered the exceeding limit and lowers it by 10% for safety margin. For example, historical data statistics show that the average historical peak-shifting interval of the devices that caused the overheating of the line is 33.3 seconds. The system deducts 10% margin, which is about 3.3 seconds, and finally sets the standard time difference to 30 seconds. This serves as a benchmark to measure whether the asynchronous power consumption behavior has worsened the actual heat impact on the local line to the dangerous edge.

[0058] This 30-second interval is directly used as the standard time difference for subsequent calculations and comparisons. The technical approach used in this step is to extract the time base using the historical statistical attributes of the knowledge graph. The beneficial effect is that it avoids using fixed constants that are detached from the actual business scenario, so that the time tolerance judgment standard can fit the actual load characteristics of the distribution area, improving the rationality and dynamic adaptability of the time parameter comparison.

[0059] In S34, based on the spatial coordinate foundation and feature extraction established in the previous steps, further computational expressions are applied. To quantify the time-domain impact of asynchronous power consumption on the transformer area, the design of this calculation logic aims to overcome the technical shortcomings of existing technologies that only use simple time difference ratios to evaluate synergy, so that the time-domain evaluation conforms to the energy dissipation nature of Joule's law.

[0060] Furthermore, the power loss of a line is proportional to the square of the current flowing through it. When two high-power devices start up completely synchronously, the heat loss generated by their superimposed current is proportional to the square of the sum of the currents of the two devices, that is... This forms the denominator of the expression, representing the baseline value of the square of the extreme heating current under extreme synchronous operating conditions. However, in actual operation, there is often a time difference during equipment startup. This time misalignment causes the total current to fail to reach the theoretical peak value. Mathematically, the actual equivalent square value of heating is equal to the sum of the squares of the individual currents. Add an intersection term .

[0061] Furthermore, this expression will represent a scaling factor for the degree of temporal asynchrony. A linear decay weight is applied to this cross term: when the actual time difference approaches 0, the decay weight approaches 1, taking into account the heating effect of all cross terms; while when the actual time difference reaches or exceeds the tolerable standard tolerance, the weight approaches 0, the cross term is eliminated, and only the basic sum-of-squares loss is borne. Meanwhile, This is a function that maximizes the time factor to ensure that the time influence factor is truncated to zero when the typical time difference exceeds the standard time difference, thus avoiding negative cross terms that violate energy conservation.

[0062] Applying the above logic to the real data of sub-region A of the park, the real-time load current amplitude of the air compressor that generates the maximum time interval is known. 150 Real-time load current amplitude of injection molding machine 100 The typical time difference between the two 12 The standard time difference is limited by the knowledge graph of the Taiwan region. 30 First, calculate the time impact factor: subtract 12 from 1. With 30 The quotient, that is This dimensionless value indicates that the cross-term thermal effect is reduced to 60% due to the 12-second peak shift. Next, the numerator of the expression, i.e., the equivalent square of the actual Joule heating current, is calculated: the square of the air compressor current is... The square of the injection molding machine current is The cross term after superposition and reduction is Adding the three together gives the numerator as follows: Finally, the denominator, i.e., the baseline value of the extreme synchronous heating, is calculated. The initial line loss impact of sub-region A is obtained by calculating the ratio of the numerator to the denominator. .

[0063] This calculation clearly demonstrates that although the 12-second peak shift accounts for 40% of the 30-second tolerance, the actual heat loss impact in this region still reaches 80.8% of the theoretical extreme value due to the nonlinear cross-amplification effect of the current amplitude. This computational architecture, which reduces the dimensionality of macroscopic time parameters and embeds them into the microscopic current square formula, solves the problem of the separation between time evaluation and energy dissipation in existing algorithms, giving the evaluation indicators real physical dimensions to support them. like Figure 5 As shown, in one specific implementation, S4 includes: S41, multiplying the initial line loss impact degree of each sub-region by the previously obtained spatial drift penalty coefficient to obtain the target line loss impact degree of each sub-region. This step reflects the adjustment of the local line loss weight by the spatial aggregation state.

[0064] S42. Abstract the knowledge graph of transformer area equipment into a graph network, and treat each sub-region as a node in the graph network. Calculate the betweenness centrality of each sub-region in the topology path. Betweenness centrality refers to the frequency at which a node in a sub-region acts as a critical link in all possible inter-node transmission paths in a power topology network. Assign topology weights based on the coupling between betweenness centrality and dynamically corrected impedance. The calculation logic is as follows:

[0065] in, To be assigned to the Topological weights of each sub-region (dimensionless); For the first Betweenness centrality of each subregion; For the first Average dynamic correction impedance of all power supply lines in each sub-region (unit: ); For summation auxiliary index, For the first Betweenness centrality of each subregion; The average dynamic correction impedance of all power supply lines in the m-th sub-region (unit: ); This represents the total number of sub-regions. The topological weight is based on the combined contribution of network structural criticality and physical impedance loss.

[0066] S43. Multiply the target line loss impact of each sub-region by its corresponding topology weight, and perform a summation operation on the product results of all sub-regions to obtain the bus line loss impact of the target area, which is the comprehensive evaluation value of the overall line loss level.

[0067] In this embodiment, it should be noted that in S41, the initial line loss impact degree of each sub-region is multiplied by the spatial drift penalty coefficient to obtain the target line loss impact degree of each sub-region. Based on the data from the previous step, the initial line loss impact degree of each sub-region is corrected using the spatial drift penalty coefficient of 1.233 obtained in S2. The initial line loss impact degree of sub-region A is retrieved as 0.808, and the target line loss impact degree of sub-region A is calculated to be approximately 0.996 (0.808 × 1.233). Similarly, the target line loss impact degree of sub-region B is calculated to be approximately 0.801 (0.650 × 1.233), and that of sub-region C is approximately 1.110 (0.900 × 1.233). The technical means of this step is to couple the result of the time dimension assessment with the degree of dispersion of the spatial dimension through division. The beneficial effect obtained is to transform the spatial dispersion state caused by the centroid drift of the transformer load into an amplification factor for the local line loss level, realizing the data fusion of the dual influencing factors of time and space.

[0068] In S42, in order to aggregate the spatiotemporal evaluation results of each sub-region into a global index, a calculation expression is used. The topology weights are assigned using this calculation process to address the inaccuracy in existing clustering algorithms that rely solely on two-dimensional geographical distance or equal partitioning. In complex distribution substations, the contribution of different regions to bus losses depends not only on the power consumption status of their users but also on their macroscopic topological location within the entire power transmission network and their microscopic dielectric conditions.

[0069] Furthermore, the betweenness centrality parameter in the numerator of the expression is extracted from a graph theory algorithm. It objectively records the frequency at which a specific sub-region acts as a bridge on the shortest power flow transmission path between all node pairs in the target transformer area. The higher the frequency, the more network penetration current is gathered in that region, and the more critical its topological position. However, high-frequency power flow penetration alone does not necessarily lead to high losses. Only when these currents flow through high-impedance media will they be converted into actual line losses. Therefore, the expression multiplies the betweenness centrality with the average dynamic correction impedance of all power lines in the region, thus constructing a joint characteristic index that simultaneously considers the criticality of the network structure and the severity of physical losses. To ensure the mathematical closure and conservation of the subsequent weighted aggregation, the denominator is calculated for all... The joint feature indices of each sub-region are summed, and this summation is used as a global benchmark to normalize each sub-region.

[0070] Based on the specific data of the application scenario, the entire station area was divided into 3 sub-regions. The knowledge graph analysis revealed that the betweenness centrality of sub-regions A, B, and C were respectively... , , (Dimensionless), and the average dynamic correction impedances for these three regions, calculated from previous steps, are respectively... , , First, calculate the numerator product of each sub-region: the characteristic index of sub-region A is... The characteristic index of subregion B is The characteristic index of subregion C is Then, the denominator is calculated as the sum of the global joint characteristic indices: Finally, a division operation is performed to obtain the respective topological weights: the weights assigned to sub-region A. The weight of subregion B The weight of subregion C .

[0071] This calculation result intuitively shows that although the average heating impedance of the line in sub-region B is the highest (0.006), However, because sub-region A occupies a more central hub position in the power grid (with a betweenness centrality as high as 0.5), more current is forced to flow through the lines in region A. Therefore, nearly half (49.0%) of the aggregation weight is ultimately assigned to sub-region A. This calculation logic, by combining graph theory centrality and Ohm's law, elevates a single electrical parameter to a network physical composite weight, effectively preventing data misleading caused by high-energy-consuming areas at the edges or low-energy-consuming areas in the center, and providing a crucial data aggregation path for the accurate calculation of the overall line loss impact.

[0072] In step S43, the target line loss impact of each sub-region is multiplied by its corresponding topology weight, and the product results of all sub-regions are summed to obtain the bus line loss impact of the target transformer area. Based on previous calculations, the target line loss impact of sub-region A is 0.996, and its topology weight is 0.490; for sub-region B, it is 0.801 and 0.353; and for sub-region C, it is 1.110 and 0.157. Calculating the product of these three sets of data, the product for sub-region A is approximately 0.488, for sub-region B approximately 0.283, and for sub-region C approximately 0.174. Finally, these three products are weighted and summed to obtain a bus line loss impact of approximately 0.945 for the transformer area in this heavy industrial park. This step uses a weighted summation mechanism to summarize the impact indicators of each region, achieving the beneficial effect of outputting a comprehensive evaluation value that considers both spatiotemporal fluctuations and topology impedance, providing a quantitative reference for operation and maintenance.

[0073] This invention also provides a knowledge graph-based system for processing data on the impact of transformer substation line losses. The system is used to implement a knowledge graph-based method for processing data on the impact of transformer substation line losses. The system includes: The coordinate mapping module is used to obtain the physical coordinates of each data acquisition node in the target distribution area, establish a spatial coordinate system and divide the target distribution area into multiple sub-regions; it performs electrothermal coupling reasoning based on the distribution area equipment knowledge graph to obtain the dynamic correction impedance, and obtains the power consumption characteristic coordinates of each power user unit based on the dynamic correction impedance; The spatial convergence analysis module is used to calculate the spatial drift dispersion based on the offset distance between the electricity consumption characteristic coordinates and the physical coordinates of the associated data acquisition nodes, and to obtain the spatial drift penalty coefficient based on the spatial drift dispersion. The spatiotemporal coordination module is used to obtain the power consumption start time of each power user unit, obtain the typical time difference of the sub-region based on multiple power consumption start time points in the same sub-region, obtain the historical load ramp-up time tolerance as the standard time difference, and obtain the initial line loss impact based on the Joule heating cross term model based on the typical time difference and the standard time difference. The topology weighting module is used to correct the initial line loss impact of each sub-region based on the spatial drift penalty coefficient to obtain the target line loss impact. Based on the knowledge graph of the equipment in the distribution area, the betweenness centrality of each sub-region is obtained, and the topology weight is allocated according to the coupling relationship between the betweenness centrality and the dynamic correction impedance. The target line loss impact of each sub-region is weighted and aggregated according to the topology weight to obtain the bus line loss impact of the target distribution area.

[0074] To further clarify the operating mechanism and physical quantification process of the technical solution of this invention, the following analysis will be conducted in detail on the underlying derivation logic of the knowledge graph-based method for processing data on line loss impact in transformer substations, using a scenario containing specific parameters and data.

[0075] Taking a certain heavy machinery manufacturing park as an example The distribution station area is described in detail for a specific application scenario. This target area has a standard rectangular layout, and the system divides it into sections based on the distribution of data acquisition nodes. There are three sub-regions (labeled as sub-region A, sub-region B, and sub-region C respectively), with a total of [number] mounts. Individual industrial power user units. During peak summer load periods (when the ambient temperature is...). The system first enters S1 for electrothermal coupling inference. Taking the large stamping workshop within sub-region A (the...) as an example... Taking an individual electricity user unit as an example, the smart meter collects its current operating current. A, Operating voltage V, while the low-voltage side reference voltage of the distribution transformer V, from which the voltage drop value is derived. V, equivalent electrical impedance is The system then extracts the physical parameters of the overhead insulated aluminum conductor from the knowledge graph: The reference impedance per unit length below Temperature resistivity thermal resistance coefficient The historical average operating current of this line A. Substitute into the dynamic correction impedance formula Calculate the additional value of temperature rise ,but This indicates that under the combined effects of high current and high temperature, the actual impedance of the circuit has increased compared to the factory reference value. .

[0076] Following this, at the end of S1, the system divides the obtained equivalent electrical impedance by the dynamically corrected impedance to obtain the electrical distance to the load center. m. The knowledge graph indicates that the physical coordinates of the data acquisition nodes associated with this stamping workshop are... Furthermore, the route vector extends strictly along the positive X-axis. Therefore, the system accurately maps the current power consumption characteristics of the stamping workshop to... These coordinates do not represent the location of the workshop's main entrance, but rather objectively reflect the equivalent characteristic location of electrical energy in three-dimensional space under its current power consumption and line heat loss. The system covers all [items / entities / areas] within the park. Each power user unit performs the above physical mapping.

[0077] The system then enters S2 for spatial convergence analysis. Through three-dimensional Euclidean distance calculation, the electrical characteristic coordinates of the stamping workshop and its static physical nodes are determined. The unit offset distance between them is m. The system monitors the area within the park. The unit offset distances of each unit are summed and their arithmetic mean is calculated, yielding a spatial drift dispersion of 18.5m for the entire park at that moment. The system retrieves the standard offset distance designed for this area from the knowledge graph. The ratio of spatial drift dispersion to standard offset distance is calculated as 18.5 / 15.0 ≈ 1.233. Since 1.233 > 1, the system determines 1.233 as the spatial drift penalty coefficient at the current moment. This value is greater than 1, indicating that the irregular start-up and shutdown of a large number of heavy equipment in the park has caused a serious drift in the overall load center of gravity of the distribution area, and the energy distribution is in a highly discrete state.

[0078] After completing the spatial dimension analysis, the system enters S3 for Joule heating cross-term collaborative analysis. Taking sub-region A as an example, multiple production lines in this region frequently start and stop. The smart meter detects two extreme users (e.g., a large air compressor and an injection molding machine) that trigger the largest time intervals. The real-time load current amplitude of the air compressor... A, Real-time load current amplitude of the injection molding machine A. Extract the starting time of their electricity consumption and calculate the typical time difference between the two. The system retrieves the historical load ramp-up time tolerance for the afternoon period of that season from the knowledge graph, using it as the standard time difference. s. This data means that the two high-power devices are in The devices were connected to the grid within seconds, while the grid's original design buffer time for shock absorption was... Second.

[0079] To accurately quantify the impact of this asynchronous startup on losses, the system substitutes the above data into the Joule heating cross-term model formula. First, calculate the time-related factors. The numerator (the square of the actual Joule heating current) is calculated as follows: The denominator (the square of the extreme synchronous Joule heating current) is calculated as follows: The initial line loss impact of sub-region A is obtained by dividing by the ratio. Calculation results show that, despite peak shifting... However, due to the squared-term surge effect caused by the superposition of two huge currents, the local power grid still suffered a heating effect equivalent to 80.8% of the theoretical limit loss. Similarly, the system calculated the initial line loss impact of sub-regions B and C in parallel. and .

[0080] Finally, the system enters the S4 execution graph topology coupling weighted aggregation. First, the system uses the spatial drift penalty coefficient obtained in S2. The initial line loss of each sub-region is amplified and corrected. Taking sub-region A as an example, the target line loss impact is 0.808×1.233≈0.996, sub-region B is 0.650×1.233≈0.801, and sub-region C is 0.900×1.233≈1.110. Subsequently, the system extracts the topological features of the graph. Sub-region A is located at a key topological node of the main road leading from the main distribution room, and its betweenness centrality is... B and C are on branches, respectively , The average dynamic correction impedances of the three are respectively , , .

[0081] According to the formula Perform topological weight allocation. The numerator products of each sub-region are as follows: Region A Area B Area C The sum of the products is Therefore, the topological weights of each sub-region are as follows: , , Ultimately, the system weighted and summed the target line loss impact with the topology weights, obtaining the bus line loss impact of the heavy industrial park as 0.490×0.996+0.353×0.801+0.157×1.110≈0.488+0.283+0.174≈0.945. This value fully integrates the additional impedance of high-temperature environments, spatial drift of giant loads, cross-heating of asynchronous currents, and the vulnerability of the backbone network. In the scenario of monitoring the insulation status of medium and low voltage transformers and secondary circuits, this value can help the system identify potential insulation degradation hazards caused by a surge in local heat loss in advance, providing objective and quantitative physical data support for the power grid dispatching end to implement precise reactive power compensation, peak-shifting and power outage commands, and insulation hazard investigation.

[0082] The preferred embodiments of this disclosure have been described in detail above with reference to the accompanying drawings. However, this disclosure is not limited to the specific details of the above embodiments. Within the scope of the technical concept of this disclosure, various simple modifications can be made to the technical solutions of this disclosure, and these simple modifications all fall within the protection scope of this disclosure.

[0083] It should also be noted that the various specific technical features described in the above embodiments can be combined in any suitable manner without contradiction. To avoid unnecessary repetition, this disclosure will not describe the various possible combinations separately.

[0084] Furthermore, various different embodiments of this disclosure can be combined in any way, as long as they do not violate the spirit of this disclosure, they should also be regarded as the content disclosed in this disclosure.

[0085] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them. Although the present invention 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 or all of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be covered within the scope of the claims and specification of the present invention.

Claims

1. A method for processing data on the impact of transformer substation line loss based on knowledge graphs, characterized in that the method... include: Obtain the physical coordinates of each data acquisition node within the target area, establish a spatial coordinate system, and divide the target area into multiple sub-regions; Based on the knowledge graph of distribution transformer equipment, electrothermal coupling reasoning is performed to obtain dynamically corrected impedance. Then, based on the dynamically corrected impedance, the power consumption characteristic coordinates of each power user unit are obtained, including: obtaining the reference voltage on the low-voltage side of the distribution transformer and the operating voltage and current of the power user unit; using the difference between the reference voltage and the operating voltage as the voltage drop value, and dividing the voltage drop value by the operating current to obtain the equivalent electrical impedance; obtaining the current ambient temperature, and retrieving the unit length reference impedance, temperature resistivity, thermal resistivity, and historical average operating current of the corresponding power supply line from the distribution transformer equipment knowledge graph; and obtaining the dynamically corrected impedance based on these parameters. Specifically... , For dynamic correction impedance of power supply lines. For power supply lines in The reference impedance per unit length is below. The temperature resistivity of the power supply line material. The current ambient temperature. The thermal resistivity of a power supply line characterizes the temperature rise effect caused by a unit current squared. The historical average operating current of the power supply line within a preset time window is used; the equivalent electrical impedance is divided by the dynamically corrected impedance to obtain the electrical distance to the load center; starting from the physical coordinates of the associated data acquisition node, the electrical distance to the load center is extended along the line spatial direction vector parameters recorded in the knowledge graph of the transformer area equipment to obtain the power consumption characteristic coordinates of the power user unit. Based on the offset distance between the electricity consumption characteristic coordinates and the physical coordinates of the associated data acquisition nodes, the spatial drift dispersion is calculated, and the spatial drift penalty coefficient is obtained based on the spatial drift dispersion. Obtain the power consumption start time of each power user unit, and obtain the typical time difference of the sub-region based on multiple power consumption start times in the same sub-region; obtain the historical load ramp-up time tolerance as the standard time difference, and obtain the initial line loss impact based on the Joule heating cross term model based on the typical time difference and the standard time difference. Extract the real-time load current amplitude of two power user units with typical time differences within the same sub-region, and calculate the initial line loss impact by combining the time difference parameter. Specifically, , This represents the initial line loss impact level for this sub-region. and These represent the real-time load current amplitudes of two power user units that generate a typical time difference. This is a typical time difference. Standard time difference; The initial line loss impact of each sub-region is corrected based on the spatial drift penalty coefficient to obtain the target line loss impact. The betweenness centrality of each sub-region is obtained based on the knowledge graph of the equipment in the distribution area. The topology weight is allocated according to the coupling relationship between the betweenness centrality and the dynamic correction impedance. The target line loss impact of each sub-region is weighted and aggregated according to the topology weight to obtain the bus line loss impact of the target distribution area.

2. The method for processing data on the impact of transformer substation line loss based on knowledge graphs according to claim 1, characterized in that, The process of calculating the spatial drift dispersion based on the offset distance between the electricity consumption characteristic coordinates and the physical coordinates of the associated data acquisition nodes, and obtaining the spatial drift penalty coefficient based on the spatial drift dispersion, includes: Calculate the absolute distance between the electricity consumption characteristic coordinates and the physical coordinates of the nearest data acquisition node, and obtain the unit offset distance; The spatial drift dispersion is obtained by calculating the arithmetic mean of the unit offset distance of all power user units within the target distribution area. Obtain the preset standard offset distance, and divide the spatial drift dispersion by the standard offset distance to obtain the scaling factor; If the scaling factor is greater than 1, then the scaling factor is defined as the space drift penalty factor; if the scaling factor is not greater than 1, then the value 1 is defined as the space drift penalty factor.

3. The method for processing data on the impact of transformer substation line loss based on knowledge graphs according to claim 1, characterized in that, The process of obtaining the electricity consumption start time of each power user unit and obtaining the typical time difference of the sub-region based on multiple electricity consumption start time points within the same sub-region includes: Obtain the mutation rate of the load current of the power user unit, and record the timestamp when the mutation rate exceeds the preset threshold as the start time of electricity consumption; Extract the start time of electricity consumption for all power user units within the same sub-region, calculate the interval between any two start times of electricity consumption, and extract the maximum value of the interval as the typical time difference for that sub-region.

4. The method for processing data on the impact of transformer substation line loss based on knowledge graphs according to claim 1, characterized in that, The step of correcting the initial line loss impact of each sub-region based on the spatial drift penalty coefficient to obtain the target line loss impact includes: Multiply the initial line loss impact of each sub-region by the previously obtained spatial drift penalty coefficient to obtain the target line loss impact of each sub-region.

5. The method for processing data on the impact of transformer substation line loss based on knowledge graphs according to claim 1, characterized in that, The method involves obtaining the betweenness centrality of each sub-region based on the knowledge graph of the transformer area equipment, allocating topology weights according to the coupling relationship between betweenness centrality and dynamic correction impedance, and weighting and aggregating the target line loss impact of each sub-region according to the topology weights to obtain the total line loss impact of the target transformer area, including: The knowledge graph of equipment in the distribution area is abstracted into a graph network, and each sub-region is treated as a node in the graph network. The betweenness centrality of each sub-region in the topological path is calculated. Obtain the average dynamic correction impedance of all power supply lines in each sub-region; Calculate the product of the betweenness centrality and the average dynamic correction impedance of each sub-region, and divide the product of each sub-region by the sum of the products of all sub-regions to obtain the topological weight of each sub-region; Multiply the target line loss impact of each sub-region by its corresponding topology weight, and sum the product results of all sub-regions to obtain the bus line loss impact of the target station area.

6. A knowledge graph-based data processing system for the impact of transformer substation line loss, characterized in that, The system is used to execute the knowledge graph-based data processing method for transformer substation line loss impact as described in any one of claims 1 to 5, and the system includes: The coordinate mapping module is used to obtain the physical coordinates of each data acquisition node in the target distribution area, establish a spatial coordinate system and divide the target distribution area into multiple sub-regions; it performs electrothermal coupling reasoning based on the distribution area equipment knowledge graph to obtain the dynamic correction impedance, and obtains the power consumption characteristic coordinates of each power user unit based on the dynamic correction impedance; The spatial convergence analysis module is used to calculate the spatial drift dispersion based on the offset distance between the electricity consumption characteristic coordinates and the physical coordinates of the associated data acquisition nodes, and to obtain the spatial drift penalty coefficient based on the spatial drift dispersion. The spatiotemporal coordination module is used to obtain the power consumption start time of each power user unit, obtain the typical time difference of the sub-region based on multiple power consumption start time points in the same sub-region, obtain the historical load ramp-up time tolerance as the standard time difference, and obtain the initial line loss impact based on the Joule heating cross term model based on the typical time difference and the standard time difference. The topology weighting module is used to correct the initial line loss impact of each sub-region based on the spatial drift penalty coefficient to obtain the target line loss impact. Based on the knowledge graph of the equipment in the distribution area, the betweenness centrality of each sub-region is obtained, and the topology weight is allocated according to the coupling relationship between the betweenness centrality and the dynamic correction impedance. The target line loss impact of each sub-region is weighted and aggregated according to the topology weight to obtain the bus line loss impact of the target distribution area.

Images