Power distribution network line loss optimization method and system

Abstract

The invention provides a power distribution network line loss optimization method and system, and relates to the technical field of power distribution network line loss optimization, and the method comprises the steps: receiving local operation data collected by a measurement unit in real time, and generating the approximate resistance characteristics of a line segment based on the data and the operation information of an adjacent measurement unit, according to the method, approximate resistance characteristics and local operation data are fused to generate more reliable measurement data, an instantaneous power loss integral method is adopted to calculate instantaneous power loss aiming at bidirectional power flow and harmonic problems caused by distributed power supply access, and energy loss is generated through sliding time window integration. And according to the fused measurement data and energy loss, determining the global line loss of the power distribution network and generating an optimization control instruction, so that challenges such as complex operation environment, dynamic change of line parameters and bidirectional power flow of the power distribution network can be effectively handled, and the accuracy of line loss evaluation and the effectiveness of optimization control are remarkably improved. Accurate evaluation and efficient optimization of the line loss of the power distribution network are realized.

Description

Technical Field

[0001] This application relates to the field of power distribution network line loss optimization technology, and more specifically, to a method and system for optimizing power distribution network line losses. Background Technology

[0002] As a crucial component of the power system, the operating efficiency of the distribution network directly impacts the economy and reliability of electricity. In practice, line losses are one of the main forms of energy waste; therefore, effectively reducing line losses in distribution networks has always been a focus of industry attention. Traditional line loss optimization methods typically rely on accurate line parameters and a stable operating environment. However, with the increasing complexity of distribution network operating environments—such as line aging, segmented repairs, the integration of new loads, and the widespread application of distributed power sources—deviations may arise between the actual electrical characteristics of the lines and the system model. This presents a significant challenge to the accurate assessment and effective optimization of line losses.

[0003] Specifically, in power distribution networks, the actual electrical characteristics of lines (such as equivalent resistance and its temperature dependence) may become disconnected from the system's internal parameter database due to historical segmented repairs. For example, during emergency repairs, on-site maintenance teams may use conductor segments from different production batches or manufacturers to quickly restore power. These conductor segments may have slight differences in alloy composition or manufacturing processes, resulting in inconsistent electrical characteristics. However, this detailed material change information is not updated in a timely or comprehensive manner to the central digital "Line Basic Parameter Database," preventing the database from continuously reflecting the line's original, homogeneous design parameters. Long-term segmented repairs cause the line to physically become a "hybrid line" composed of multiple conductors, resulting in discrepancies between its overall equivalent resistance value and its temperature dependence and the predictions of the system model.

[0004] Furthermore, the complexity of the distribution network operating environment exacerbates this problem. Sustained high load demand causes lines to carry high currents for extended periods, amplifying subtle, temperature-dependent differences in resistance characteristics within "hybrid lines." Frequent adjustments to the on-load tap positions of distribution transformers alter the voltage distribution and power flow paths of local grids, making hidden errors in line loss calculations more difficult to detect. Simultaneously, the widespread integration of intermittent power sources such as distributed photovoltaic systems leads to drastic and frequent changes in the magnitude and direction of power flow. Traditional line loss calculation and optimization methods, based on assumptions of relatively stable, unidirectional power flow and static network parameters, are ill-suited to this dynamic and bidirectional power flow environment, resulting in line loss estimates that deviate significantly from actual physical reality.

[0005] Ultimately, the distribution network's line loss optimization system was found to heavily rely on a "line basic parameter database" that had a long-standing and significant contradiction with the actual electrical characteristics of the physical branches. This inconsistency in basic data was significantly amplified and masked under complex operating environments characterized by high temperatures, high loads, frequent transformer tap changes, and drastic power flow fluctuations caused by distributed generation, rendering the system's generated line loss optimization commands ineffective and potentially even exacerbating power losses within the grid. Maintenance personnel struggled to pinpoint the root cause of these persistently abnormal line loss figures because the system model data they relied on was disconnected from the physical reality of the grid from the outset, creating a profound and intractable operational challenge. Summary of the Invention

[0006] This application provides a method and system for optimizing line losses in distribution networks, aiming to solve the problem that the complex operating environment of distribution networks and the discrepancy between the actual electrical characteristics of the lines and the system model pose challenges to the accurate assessment and effective optimization of line losses.

[0007] On the one hand, this application provides a method for optimizing line losses in a distribution network, including: Receive local operation data collected in real time by measurement units deployed at key nodes of the distribution network, wherein the key nodes are distributed photovoltaic grid connection points, high-load user access points, or on-load tap adjustment points of distribution transformers, and wherein the key nodes have edge computing capabilities. Based on the local operating data of the measurement unit and the operating information obtained by the measurement unit from adjacent measurement units, an approximate resistance characteristic of the line segment under the responsibility of the measurement unit is generated. Based on preset multi-dimensional credibility indicators, the approximate resistance characteristics and the local operating data are evaluated for differential weight credibility, and measurement data with differential weight fusion is generated. Based on the measurement data fused by the differentiated weights, the instantaneous power loss of the line segment is calculated using the instantaneous power loss integral method, and then the energy loss is generated by integration through a sliding time window. Based on the measurement data fused by the differentiated weights and the energy loss, the global line loss of the distribution network is determined, and an optimized control command is generated based on the global line loss.

[0008] Optionally, the step of performing differentiated weighted reliability assessment on the approximate resistance characteristics and the local operating data based on preset multi-dimensional reliability indicators, and generating differentiated weighted fused measurement data, includes: Define a local closed-loop region in the distribution network; Obtain the energy imbalance in the local closed-loop region; Based on the energy imbalance, locate the measurement unit that causes the energy imbalance trend and obtain the target measurement unit. Based on the operating data of the target measurement unit, generate micro impedance self-calibration instructions; By executing the micro impedance self-calibration command, the local segment equivalent resistance estimation algorithm of the target measurement unit is adjusted to generate differentially weighted fused measurement data.

[0009] Optionally, the step of generating an approximate resistance characteristic of the line segment under the responsibility of the measurement unit based on the local operating data of the measurement unit and the operating information obtained by the measurement unit from adjacent measurement units includes: Obtain the topology probe signal response sent by the measurement unit and construct a local topology connection table; Detect changes in the local topology connection table to determine whether the topology structure has changed; When a change in the topology is detected, the system receives the topology change event reported by the measurement unit and obtains the latest topology connection relationship. Based on the latest topology connection relationship, adjust the adjacent measurement unit information acquisition strategy of the measurement unit to ensure that the measurement unit only obtains operation information from the currently actually connected adjacent measurement units; Based on the operational information obtained by the measurement unit from the adjusted adjacent nodes, and combined with the local operational data, an approximate resistance characteristic of the line segment under the responsibility of the measurement unit is generated.

[0010] Optionally, the step of calculating the instantaneous power loss of the line segment using the instantaneous power loss integration method based on the measurement data fused with the differentiated weights, and then generating the energy loss through integration via a sliding time window, includes: Receive instantaneous voltage and instantaneous current data at both ends of the line segment; Harmonic separation is performed on the instantaneous voltage and instantaneous current data to obtain the fundamental component and each harmonic component; Based on the fundamental component and each harmonic component, the fundamental power loss and each harmonic power loss are calculated respectively. When calculating the fundamental power loss, the fundamental current component and the fundamental voltage component are used, combined with the measurement data fused by the differentiated weights. When calculating the harmonic power loss, the harmonic current component and each harmonic voltage component are used, combined with the measurement data fused by the differentiated weights. The fundamental power loss and the power loss of each harmonic are superimposed to obtain the total instantaneous power loss of the line segment. The energy loss is obtained by numerically integrating all the total instantaneous power loss within the sliding time window.

[0011] Optionally, the step of determining the global line loss of the distribution network based on the measurement data fused by the differentiated weights and the energy loss, and generating optimized control instructions based on the global line loss, includes: Identify the current operating mode of the power distribution network; Based on the operating mode, determine the priority of the line loss optimization objective and other operating objectives; Based on the aforementioned priority, the line loss optimization objective is dynamically weighted with other operational objectives; Based on the dynamically weighted line loss optimization target and other operational targets, optimization control instructions are generated.

[0012] Optionally, the step of identifying the current operating mode of the distribution network includes: Continuously monitor the voltage, current, power, load curves of each measurement unit in the distribution network and the output data of distributed power sources, and extract features to obtain multi-dimensional operating feature vectors; The multidimensional running feature vector is matched in real time with the feature spaces of multiple preset running modes, and the similarity between the multidimensional running feature vector and each feature space is calculated. When the similarity of multiple modes is within a preset threshold range, the mode conflict resolution mechanism is activated to obtain the evaluation results. The mode conflict resolution mechanism dynamically evaluates the activity level of each mode by analyzing the historical duration, current power flow direction and amplitude change rate of each mode, as well as the voltage fluctuation trend of key nodes. Based on the evaluation results, the mode with the highest activity level is identified as the current operating mode.

[0013] Optionally, the step of determining the priority of the line loss optimization objective and other operational objectives based on the operating mode includes: Based on the operating mode, a preset mode-priority mapping table is queried to obtain the initial priority of the line loss optimization target corresponding to the operating mode and other operating targets. The initial priority is dynamically adjusted based on the real-time operating status of the distribution network to obtain the adjusted priority. Based on the adjusted priorities, the priority of the line loss optimization objective is determined to be higher than that of other operational objectives.

[0014] Optionally, the step of determining the global line loss of the distribution network based on the measurement data fused by the differentiated weights and the energy loss includes: Monitor the data integrity and real-time performance of the measurement unit; Identify measurement units with missing or degraded data; The missing data is calculated based on the real-time operating data of the adjacent measurement units of the measurement unit with missing or degraded data, the historical operating trend of the measurement unit with missing or degraded data, and the approximate resistance characteristics of the line segment to which the measurement unit with missing or degraded data is responsible. The missing data is labeled with a confidence level to obtain confidence level inference data; Global line loss is assessed based on confidence-based data and other complete and reliable measurement data. The confidence-based data is assigned a lower weight, while the other complete and reliable measurement data is assigned a higher weight. This ensures that the global line loss of the distribution network can still be accurately assessed even when the data is incomplete or unreliable.

[0015] Optionally, the multi-dimensional reliability index includes at least one of the following: data freshness, consistency with data from adjacent measurement units, historical error performance of the measurement unit, and equipment health status of the measurement unit.

[0016] On the other hand, this application provides a distribution network line loss optimization system, which includes: The data receiving module is used to receive local operating data collected in real time by measurement units deployed at key nodes of the distribution network. The key nodes are distributed photovoltaic grid connection points, high-load user access points, or on-load tap adjustment points of distribution transformers. The key nodes have edge computing capabilities. Based on the local operating data of the measurement units and the operating information obtained by the measurement units from adjacent measurement units, the module generates approximate resistance characteristics of the line segment under the responsibility of the measurement unit. The credibility assessment module is used to perform differentiated weight credibility assessment on the approximate resistance characteristics and the local operating data based on preset multi-dimensional credibility indicators, and generate differentiated weight fusion measurement data. The energy loss calculation module is used to calculate the instantaneous power loss of the line segment using the instantaneous power loss integration method based on the measurement data fused by the differentiated weights, and then generate the energy loss by integrating through a sliding time window. The optimization control module is used to determine the global line loss of the distribution network based on the measurement data fused by the differentiated weights and the energy loss, and to generate optimization control instructions based on the global line loss.

[0017] The distribution network line loss optimization method and system disclosed in this application effectively solves the problem of long-term disconnect between the actual electrical characteristics of the line and the system model in traditional methods by receiving local operating data collected in real time by measurement units deployed at key nodes of the distribution network, and generating approximate resistance characteristics of line segments based on this data and the operating information of adjacent measurement units. Furthermore, this method introduces a differentiated weighted reliability evaluation mechanism based on multi-dimensional reliability indicators to fuse approximate resistance characteristics and local operating data, thereby generating more reliable measurement data and overcoming the problem of inaccurate parameters caused by line aging and segmented repairs. Addressing the bidirectional power flow and harmonic issues brought about by distributed power source integration, this method uses the instantaneous power loss integral method to calculate instantaneous power loss and generates energy loss through sliding time window integration, significantly improving the accuracy of line loss calculation in dynamic and complex environments. Finally, based on the fused measurement data and energy loss, the global line loss of the distribution network is determined and optimized control commands are generated. This can effectively address challenges such as the complex operating environment of the distribution network, dynamic changes in line parameters, and bidirectional power flow, significantly improving the accuracy of line loss assessment and the effectiveness of optimized control. This solves the problems of inaccurate line loss estimation and invalid optimization commands in complex distribution network environments using traditional methods, avoids invalid optimization commands caused by data inconsistency, and achieves accurate assessment and efficient optimization of distribution network line losses. Attached Figure Description

[0018] To illustrate this application more clearly, the accompanying drawings used in the embodiments will be briefly described below. Obviously, those skilled in the art can obtain other drawings based on these drawings without any creative effort.

[0019] Figure 1 The diagram above illustrates a flowchart of a distribution network line loss optimization method in an embodiment. Figure 2 The diagram above illustrates a schematic representation of a power distribution network line loss optimization system in an embodiment.

[0020] Figure reference numerals: 100, Distribution network line loss optimization system; 10, Data receiving module; 20, Reliability assessment module; 30, Energy loss calculation module; 40, Optimization control module. Detailed Implementation

[0021] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments. The components of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0022] It should be noted that similar reference numerals 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, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0023] Traditional methods for optimizing distribution network line losses typically rely on accurate line parameters and a stable operating environment. However, with the increasing complexity of distribution network operating environments, such as line aging, segmented repairs, the integration of new loads, and the widespread application of distributed power sources, discrepancies may arise between the actual electrical characteristics of the lines and the system model. This poses a significant challenge to the accurate assessment and effective optimization of line losses. Specifically, the actual electrical characteristics of the lines may become disconnected from the system's internal parameter database due to historical segmented repairs, leading to deviations between the overall equivalent resistance value and its temperature-dependent variations and the predictions of the system model. Furthermore, sustained high load demand, frequent adjustments to the on-load tap positions of distribution transformers, and the widespread integration of intermittent power sources such as distributed photovoltaics cause drastic and frequent changes in the magnitude and direction of power flow. Traditional line loss calculation and optimization methods struggle to cope with this dynamic and bidirectional power flow environment, resulting in line loss estimates that significantly deviate from actual physical realities. Ultimately, the distribution network's line loss optimization system discovered that its core operational decisions and optimization strategies heavily relied on a "line basic parameter library" that had a long-standing and significant contradiction with the actual electrical characteristics of physical branches. This resulted in the invalidation of the line loss optimization instructions generated by the system, and may even have exacerbated power losses within the power grid.

[0024] like Figure 1 The diagram illustrates a flowchart of a distribution network line loss optimization method in an embodiment. This application proposes a distribution network line loss optimization method, comprising: S10, receiving local operation data collected in real time by measurement units deployed at key nodes of the distribution network, wherein the key nodes are distributed photovoltaic grid connection points, high-load user access points or on-load tap adjustment points of distribution transformers, and wherein the key nodes have edge computing capabilities. Among them, key node measurement units in the distribution network refer to measurement devices deployed on nodes in the distribution network that have important electrical characteristics or operational control significance. These nodes can be distributed photovoltaic grid-connected points, high-load user access points, or on-load tap adjustment points of distribution transformers. The main function of the measurement units is to collect local operating data in real time, such as voltage, current, power, and frequency. In addition, these measurement units also have edge computing capabilities, which means that they can perform preliminary data processing, analysis, and decision-making near the data source, thereby reducing data transmission latency, improving response speed, and alleviating the computational burden on the central server.

[0025] Local operating data refers to electrical quantity data directly collected by the measurement unit at its node, reflecting the real-time operating status of that node.

[0026] S20, Based on the local operating data of the measurement unit and the operating information obtained by the measurement unit from the adjacent measurement units, an approximate resistance characteristic of the line segment under the responsibility of the measurement unit is generated. The approximate resistance characteristic refers to the estimated equivalent resistance parameter of the line segment under the responsibility of a measurement unit, obtained by analyzing the local operating data of the measurement unit and the operating information obtained from adjacent measurement units. Since the actual line parameters may vary due to various factors, this approximate resistance characteristic can more accurately reflect the actual electrical behavior of the line.

[0027] S30, Based on the preset multi-dimensional credibility index, perform differentiated weight credibility evaluation on the approximate resistance characteristics and the local operating data, and generate differentiated weight fusion measurement data. Among them, multi-dimensional reliability indicators are multiple metrics used to evaluate the reliability of approximate resistance characteristics and local operating data. These include data freshness, consistency with data from adjacent measurement units, historical error performance of the measurement unit, and the equipment health status of the measurement unit. These indicators allow for differentiated evaluation of data from different sources and of varying quality.

[0028] Differential weighted fusion measurement data refers to measurement data obtained by fusion after assigning different weights to different data based on credibility assessment. Data with high credibility will be given higher weights, thereby ensuring the accuracy and reliability of the final fused data.

[0029] S40, based on the measurement data fused by the differentiated weights, the instantaneous power loss of the line segment is calculated using the instantaneous power loss integral method, and then the energy loss is generated by integration through a sliding time window. The instantaneous power loss integration method is a method for calculating the power loss of a line segment, and it is particularly suitable for line segments. This method processes the instantaneous voltage and instantaneous current data at both ends of the line segment to calculate the instantaneous power loss, and then obtains the energy loss through integration.

[0030] Sliding time window integration refers to generating continuous energy loss data by continuously moving the time window and integrating the data within the window over a certain time range.

[0031] S50, based on the measurement data fused by the differentiated weights and the energy loss, determine the global line loss of the distribution network, and generate an optimized control command based on the global line loss.

[0032] Global line loss refers to the total energy loss of the entire distribution network within a certain period of time.

[0033] Optimization control commands are control commands generated based on the global line loss assessment results to adjust distribution network equipment (such as on-load tap changers of distribution transformers, reactive power compensation devices, etc.) in order to reduce line losses.

[0034] This application aims to address the problems of inaccurate data, model mismatch, and difficulty in handling dynamic power flow in complex distribution network environments by introducing edge computing capabilities, approximate resistance characteristic generation, differentiated weight credibility assessment, and instantaneous power loss integration method. This will enable accurate assessment and effective optimization of distribution network line losses, thereby improving the economy and reliability of power grid operation.

[0035] Firstly, in the step of receiving local operational data collected in real time by measurement units deployed at key nodes of the distribution network, these measurement units are located at critical nodes, such as distributed photovoltaic grid-connected points, high-load user access points, or on-load tap adjustment points of distribution transformers. These measurement units possess edge computing capabilities, enabling them to collect local operational data in real time. For example, a measurement unit can be a smart meter, whose built-in microprocessor can monitor and record electrical parameters such as voltage, current, active power, and reactive power in real time. This data can be directly uploaded to a local data storage device or edge computing server via wired or wireless communication modules. As one implementation, the measurement unit can use a cellular network-based wireless communication module to send the collected local operational data to the edge computing node in real time. Another implementation is that the measurement unit can use a wired communication method based on fiber optics or Ethernet to transmit the data to a local data concentrator, which then uploads it to the edge computing node.

[0036] Secondly, in the step of generating the approximate resistance characteristics of the line segment under the responsibility of the measurement unit based on its local operating data and the operating information obtained from neighboring measurement units, the measurement unit not only utilizes its own collected local operating data but also obtains operating information from neighboring measurement units. For example, the measurement unit can periodically send requests to its neighboring measurement units to obtain their voltage, current, and other operating data. By combining this data, the measurement unit can estimate the approximate resistance characteristics of the line segment it is responsible for. As one implementation method, the measurement unit can adopt a collaborative computing approach based on distributed algorithms to exchange data with neighboring measurement units and jointly calculate the approximate resistance characteristics of the line segment they are responsible for. For example, each measurement unit can send its locally measured voltage and current data, as well as its topological connection information with neighboring measurement units, to neighboring nodes. After receiving the data, the neighboring nodes use this information to iteratively calculate the equivalent resistance of the line segment connecting them to their neighboring nodes using power system analysis methods such as the node voltage method or the branch current method. Another approach is that the measurement unit can send its local operating data and operating information obtained from neighboring measurement units to a regional edge computing server, which will centrally calculate the approximate resistance characteristics of each line segment and feed the results back to the corresponding measurement unit.

[0037] Furthermore, in the step of evaluating the approximate resistance characteristics and the local operating data based on preset multi-dimensional reliability indicators to generate differentiated weighted fused measurement data, this application introduces multi-dimensional reliability indicators to ensure data accuracy. These indicators may include data freshness, consistency with adjacent measurement unit data, historical error performance of the measurement unit, and equipment health status of the measurement unit. For example, if the data of a certain measurement unit has not been updated for a long time, or if its data deviates significantly from the data of adjacent measurement units, its reliability will decrease. As one implementation method, a scoring mechanism can be set for each reliability indicator. For example, data freshness can be scored based on the interval between the data update time and the current time, with a higher score for a shorter interval; consistency with adjacent measurement unit data can be scored by comparing the absolute or relative value of the data difference, with a higher score for a smaller difference. Then, these scores are weighted and summed to obtain a comprehensive reliability score. Based on this score, different weights are assigned to the approximate resistance characteristics and the local operating data, and then fused. Another implementation method is to use fuzzy logic or machine learning models to evaluate reliability. For example, a classifier can be trained to classify data into "high confidence", "medium confidence" and "low confidence" levels based on multi-dimensional indicators, and different fusion weights can be preset for each level.

[0038] Next, in the step of calculating instantaneous power loss for the line segment using the instantaneous power loss integration method based on the measurement data fused with the differentiated weights, and then generating energy loss through integration via a sliding time window, this application employs the instantaneous power loss integration method to accurately calculate line loss, especially for the increasingly common bidirectional power flow segments in distribution networks. For example, for a line segment connecting distributed photovoltaic and loads, the power flow direction may frequently reverse with changes in irradiance and load. Traditional calculation methods based on average power are difficult to accurately reflect this dynamic loss. As an implementation method, instantaneous voltage and current data at both ends of the line segment can be collected in real time, and instantaneous power loss can be calculated based on these instantaneous values. Specifically, instantaneous power loss can be expressed as the product of the voltage difference between the two ends of the line segment and the instantaneous current flowing through the line segment. Then, by integrating the instantaneous power loss within a certain time window, the energy loss within that time window is obtained. To obtain continuous energy loss data, a sliding time window integration method can be used, that is, the instantaneous power loss within the latest time window (e.g., 5 minutes) is integrated at regular time steps (e.g., 1 second). Another approach is to use signal processing techniques such as Fourier transform to decompose instantaneous voltage and instantaneous current into fundamental components and harmonic components, then calculate the fundamental power loss and harmonic power loss separately, and finally sum them to obtain the total instantaneous power loss.

[0039] Finally, in the step of determining the global line loss of the distribution network based on the measured data fused with the differentiated weights and the energy loss, and generating optimized control commands based on the global line loss, after obtaining the energy loss of each line segment, these losses are summarized to determine the global line loss of the entire distribution network. For example, the global line loss within the same time window can be obtained by accumulating the energy losses of all line segments within that time window. Based on the evaluation results of the global line loss, corresponding optimized control commands will be generated. As one implementation method, the optimal control strategy can be calculated using optimization algorithms (such as genetic algorithms, particle swarm optimization, etc.) based on the magnitude and trend of the global line loss, combined with preset optimization objectives (such as reducing line loss, improving voltage quality, and ensuring power supply reliability). For example, when the global line loss is too high, commands may be generated to adjust the position of the on-load tap changer of the distribution transformer to optimize the voltage distribution; or the reactive power compensation device may be controlled to be put on or cut off to improve the power factor, thereby reducing the line loss. Another approach is to dynamically adjust the priority of line loss optimization targets and other operational targets based on different operating modes (such as peak load, off-peak load, peak output of distributed power sources, etc.) and generate corresponding optimization control commands.

[0040] The distribution network line loss optimization method proposed in this application forms a complete closed loop for line loss assessment and optimization through the close coordination of the aforementioned technical features. First, measurement units deployed at key nodes of the distribution network collect local operational data in real time and perform preliminary processing using their edge computing capabilities. These measurement units not only utilize their own data but also collaborate with adjacent measurement units to acquire operational information, thereby more accurately generating approximate resistance characteristics for the line segments they are responsible for. This step solves the problem of the disconnect between line parameters and actual physical characteristics in traditional methods, providing more reliable basic data for subsequent line loss calculations.

[0041] Subsequently, to address the challenge of inconsistent data quality, this application introduces a differentiated weighted reliability assessment mechanism based on multi-dimensional reliability indicators. By assessing the reliability of approximate resistance characteristics and local operating data, and assigning different weights based on the assessment results, differentiated weighted fused measurement data is ultimately generated. This mechanism effectively improves the overall reliability of the input data and avoids line loss assessment bias caused by a single data source or low-quality data.

[0042] To address the increasingly complex power flow characteristics of distribution networks, this application innovatively employs an instantaneous power loss integration method for line segments to calculate instantaneous power loss, and generates energy loss through integration via a sliding time window. This method can accurately capture dynamic losses under bidirectional power flow, overcoming the limitations of traditional methods in handling frequent power flow changes after the integration of distributed generation sources, making the line loss calculation results closer to reality.

[0043] Finally, based on highly reliable fused measurement data and accurately calculated energy losses, the global line loss of the distribution network can be accurately determined. Based on this global line loss assessment, optimized control commands are generated, such as adjusting the on-load tap changers of distribution transformers and controlling reactive power compensation devices, to effectively reduce line losses. The entire process forms an intelligent closed loop from data acquisition, parameter estimation, data fusion, line loss calculation to optimized control, ensuring the accuracy of line loss assessment and the effectiveness of optimized control in complex and ever-changing distribution network operating environments.

[0044] This application introduces the technical feature of "generating approximate resistance characteristics of the line segment under the responsibility of a measurement unit based on local operating data of the measurement unit and operating information obtained by the measurement unit from adjacent measurement units." This enables dynamic and real-time estimation of the actual electrical parameters of the line, effectively solving the problem of line parameters being out of sync with physical reality in traditional methods. Furthermore, existing technologies often lack effective evaluation and fusion mechanisms when dealing with data quality issues, leading to low-quality data potentially contaminating the entire line loss assessment result. This application, by "evaluating the reliability of approximate resistance characteristics and local operating data with differentiated weights based on preset multi-dimensional reliability indicators, generating differentiated weighted fused measurement data," can finely process data from different sources and of different qualities, ensuring the reliability of the input data and thus improving the accuracy of line loss assessment.

[0045] More importantly, with the widespread integration of distributed power sources, bidirectional power flow is becoming increasingly common in distribution networks, rendering traditional line loss calculation methods based on unidirectional power flow and static parameters inadequate. This application employs an "instantaneous power loss integration method to calculate instantaneous power loss, and then generates energy loss through integration via a sliding time window," which accurately captures dynamic losses under bidirectional power flow, significantly improving the accuracy of line loss calculation. These innovations enable the method in this application to achieve more accurate line loss assessment and more effective optimization control in complex and ever-changing distribution network operating environments, thereby significantly reducing power loss and improving the operating efficiency and economy of the distribution network.

[0046] In some embodiments, the step of performing differentiated weighted reliability assessment on the approximate resistance characteristics and the local operating data based on a preset multi-dimensional reliability index, and generating differentiated weighted fused measurement data, includes: Define a local closed-loop region in the distribution network; Obtain the energy imbalance in the local closed-loop region; Based on the energy imbalance, locate the measurement unit that causes the energy imbalance trend and obtain the target measurement unit. Based on the operating data of the target measurement unit, generate micro impedance self-calibration instructions; By executing the micro impedance self-calibration command, the local segment equivalent resistance estimation algorithm of the target measurement unit is adjusted to generate differentially weighted fused measurement data.

[0047] Specifically, in a distribution network, through topology analysis or pre-configured settings, several electrically closed loop regions are identified, defining local closed-loop regions within the distribution network. These regions typically consist of multiple measurement units, characterized by energy flow within them satisfying Kirchhoff's current law and voltage law, meaning energy input and output should be essentially balanced. This provides clear boundaries and ranges for subsequent calculations of energy imbalances.

[0048] The acquisition of energy imbalance in the local closed-loop region can be understood as a summary calculation of the real-time power (including active and reactive power) of all measurement units within the defined local closed-loop region. Ideally, the total input power within a closed-loop region should equal the total output power plus the losses within the region. When actual measurement data leads to input-output mismatch, i.e., when energy imbalance exists, it indicates that there may be deviations in the measurement data or equivalent resistance estimation within that region. This is used to preliminarily determine whether there are data anomalies within a local region based on the macroscopic principle of energy conservation.

[0049] In practical applications, based on the energy imbalance, the measurement unit causing the energy imbalance trend is located, and the target measurement unit is obtained. Specifically, when a significant energy imbalance is detected in a local closed-loop region, the operating data, historical error performance, and consistency with adjacent measurement units within that region are further analyzed. Through data mining, statistical analysis, or rule-based reasoning, the measurement unit most likely to cause the energy imbalance is identified. For example, each measurement unit can be ranked according to its contribution to the energy imbalance, data volatility, or deviation from expected values ​​to determine one or more target measurement units. The purpose is to accurately pinpoint the source of data anomalies and avoid unnecessary adjustments.

[0050] Furthermore, once the target measurement unit is located, its real-time operating data, such as voltage, current, and power, combined with its historical operating data and known line segment topology information, will be used to calculate instructions for calibrating its local line segment equivalent resistance estimation algorithm. These instructions may include suggestions for adjusting the resistance estimation model parameters or directly provide new equivalent resistance values. The purpose is to provide precise guidance for subsequent adjustments to the local line segment equivalent resistance estimation algorithm.

[0051] Therefore, the generated micro-impedance self-calibration command is applied to the equivalent resistance estimation algorithm of the line segment handled by the target measurement unit. This adjustment makes the measurement unit's estimation of the line segment resistance more accurate, thereby improving the reliability of its local operating data. In the subsequent differential weight reliability evaluation process, due to the improved data quality of the measurement unit, the weights assigned to it during the fusion process will be more reasonable, ultimately generating more accurate and reliable differential weight fusion measurement data. The aim is to improve the accuracy of measurement data from the source, thereby optimizing the overall line loss assessment and control effect.

[0052] The technical solution of this application effectively solves the problem in the basic technical solution that the preset multi-dimensional reliability indicators may not be able to fully capture local data deviations under dynamic or abnormal operating conditions by introducing a local closed-loop region energy imbalance detection and micro-impedance self-calibration mechanism. Specifically, when the data of a certain measurement unit in the distribution network is abnormal (for example, due to sensor drift or local fault causing inaccurate estimation of the equivalent resistance of the line segment), this will manifest as an energy imbalance in its local closed-loop region. By monitoring and analyzing this energy imbalance, the target measurement unit causing the imbalance trend can be accurately located. Once the target measurement unit is identified, its operating data will be used to generate a micro-impedance self-calibration command, which can specifically adjust the equivalent resistance estimation algorithm of the line segment under the responsibility of the measurement unit. This self-calibration mechanism enables the measurement unit to dynamically correct its perception of the line segment characteristics, thereby improving the accuracy and reliability of its local operating data. Ultimately, when conducting the differential weight credibility assessment, the calibrated measurement data will obtain more reasonable weights, ensuring that the fused measurement data can more accurately reflect the actual operating status of the distribution network and avoid global line loss assessment errors caused by local data deviations.

[0053] In some alternative embodiments, it is assumed that a local closed-loop area within a power distribution network includes measurement units A, B, and C. Under normal circumstances, the energy input and output of this area should be balanced. However, due to a slight drift in a sensor of measurement unit B, its measured current value for the line segment it is responsible for is slightly higher than the actual value, thus affecting its estimation of the equivalent resistance of the local line segment. At this time, an energy imbalance is detected in the local closed-loop area; for example, the total input power is significantly greater than the total output power. Based on the energy imbalance, by analyzing the data contribution and historical performance of each measurement unit, measurement unit B is identified as the target measurement unit causing the energy imbalance trend. Subsequently, using the real-time voltage and current data and historical operating trends of measurement unit B, a micro-impedance self-calibration command is generated. This command is sent to measurement unit B to adjust its internal local line segment equivalent resistance estimation algorithm, for example, by fine-tuning a certain resistance parameter by 0.5%. After adjustment, measurement unit B's estimation of the line segment equivalent resistance becomes more accurate, and its output local operating data is also corrected accordingly. In the subsequent differential weight credibility assessment, due to the improved data quality of measurement unit B, the weight assigned to it during the data fusion process will be more reasonable, thereby generating more accurate differential weight fusion measurement data and ensuring the accuracy of line loss calculation.

[0054] In some embodiments, the step of generating an approximate resistance characteristic of the line segment under the responsibility of the measurement unit based on the local operating data of the measurement unit and the operating information obtained by the measurement unit from adjacent measurement units includes: Obtain the topology probe signal response sent by the measurement unit and construct a local topology connection table; Detect changes in the local topology connection table to determine whether the topology structure has changed; When a change in the topology is detected, the system receives the topology change event reported by the measurement unit and obtains the latest topology connection relationship. Based on the latest topology connection relationship, adjust the adjacent measurement unit information acquisition strategy of the measurement unit to ensure that the measurement unit only obtains operation information from the currently actually connected adjacent measurement units; Based on the operational information obtained by the measurement unit from the adjusted adjacent nodes, and combined with the local operational data, an approximate resistance characteristic of the line segment under the responsibility of the measurement unit is generated.

[0055] Specifically, a measurement unit can periodically or under the trigger of specific events send probe signals to its physically connected neighboring nodes. These probe signals can be handshake messages, heartbeat packets, or specially designed topology probe messages based on communication protocols. After receiving a response from a neighboring node, the measurement unit constructs or updates a local topology connection table based on the response information (e.g., the responding node's unique identifier, communication quality, etc.). This table records the connection status and basic information between the current measurement unit and each of its directly connected neighboring measurement units.

[0056] The detection of changes in the local topology connection table to determine whether the topology has changed can be understood as the measurement unit continuously monitoring its local topology connection table. By comparing the current local topology connection table with the previous time step or a preset baseline topology connection table, additions or deletions of connection relationships, changes in connection status, or anomalies in connection parameters can be identified. When these changes exceed preset thresholds or rules, it is determined that the topology of the distribution network has changed.

[0057] In practical applications, when a topology change is detected, the system receives topology change events reported by measurement units and obtains the latest topology connectivity. Specifically, once a topology change is confirmed, the relevant measurement unit generates a topology change event and reports it to the distribution network management system or regional coordination controller. This event typically includes the identifier of the measurement unit that changed, the type of change (e.g., connection disconnected, new connection established), information on the adjacent nodes involved, and a timestamp of the change. After receiving these events, the system comprehensively analyzes reports from multiple measurement units and, in conjunction with the overall distribution network topology model, calculates and confirms the latest, globally consistent topology connectivity.

[0058] Furthermore, after obtaining the latest topology connections, instructions are issued to the affected measurement units to update their internal lists of adjacent measurement units and data acquisition configurations. For example, if a measurement unit disconnects from its original adjacent node, its data acquisition strategy will be adjusted to no longer request data from that node; if it establishes a connection with a new node, a new task will be added to obtain operational information from that new node. This adjustment ensures that the data source of the measurement unit strictly matches the actual physical connection state of the distribution network.

[0059] Therefore, based on the operational information obtained by the measurement unit from the adjusted adjacent nodes, combined with the local operational data, an approximate resistance characteristic of the line segment under the responsibility of the measurement unit is generated. This means that, after ensuring the accuracy of the information from adjacent measurement units, the measurement unit can use its own real-time collected local operational data (such as voltage, current, power, etc.) and the operational information obtained from the currently actually connected adjacent measurement units to accurately estimate the approximate resistance characteristic of the line segment under its responsibility through a preset algorithm (e.g., based on the principle of voltage drop or power balance).

[0060] The technical solution of this application effectively addresses the impact of distribution network topology changes on the accuracy of generating approximate resistance characteristics of line segments by introducing a dynamic topology sensing and adaptation mechanism. First, through periodic or event-triggered topology detection signal responses, the measurement unit can construct and maintain its local topology connection table in real time, thus laying the foundation for topology change detection. When a change in the local topology connection table is detected, the unit can promptly determine that the topology structure has changed and receive the topology change event reported by the measurement unit to obtain the latest topology connection relationships. Based on this latest relationship, the measurement unit's adjacent information acquisition strategy is precisely adjusted to ensure that it only obtains operational data from adjacent measurement units with current actual physical connections. Therefore, even in complex environments with frequent changes in distribution network topology, the measurement unit can still obtain accurate adjacent operational information. Combined with local operational data, this generates more accurate and reliable approximate resistance characteristics of line segments, providing a solid data foundation for subsequent line loss calculations.

[0061] Through the above technical solution, this application can significantly improve the adaptability and accuracy of the distribution network line loss optimization method in dynamic topology environments. Compared with traditional methods that do not consider topology changes, this technical solution ensures that the measurement unit always obtains operational information from the correct adjacent nodes by sensing, judging, and adapting to changes in the distribution network topology in real time. This effectively avoids data source errors caused by topology changes, thereby ensuring that the generated approximate resistance characteristics of the line segments have higher accuracy and reliability.

[0062] In some optional embodiments, it is assumed that within a certain distribution network area, a distributed photovoltaic grid-connected point (as a key node measurement unit) is operating normally, with its adjacent measurement units being the upstream substation outlet measurement unit and the downstream load access point measurement unit. At a certain moment, due to a line fault, the distribution network performs fault isolation and load transfer operations, causing the connection between the distributed photovoltaic grid-connected point and the original downstream load access point to be broken, and a new connection to a measurement unit on another backup line to be established.

[0063] At this time, the distributed photovoltaic grid-connected measurement unit periodically sends topology probe signals and receives responses from its physically connected nodes, thereby updating its internal local topology connection table. When a change in the local topology connection table is detected (for example, a previously connected node stops responding while a new node begins to respond), it is determined that the topology has changed. The measurement unit then reports a topology change event and obtains the latest topology connection relationship, i.e., it is now connected to a measurement unit on a backup line.

[0064] Based on this updated topology, the information acquisition strategy of the adjacent measurement units of this measurement unit is immediately adjusted, causing it to stop acquiring information from the original downstream load connection point measurement unit and instead acquire operational information from the newly connected backup line measurement unit. Based on the operational information acquired from this adjusted adjacent node, combined with its own local operational data, this distributed photovoltaic grid-connected point measurement unit can accurately generate the approximate resistance characteristics of the line segment it is responsible for (i.e., the line segment connected to the backup line). In this way, even when the distribution network topology changes dynamically, the generation of the approximate resistance characteristics of the line segment can still maintain high accuracy, avoiding line loss calculation deviations caused by lagging or incorrect topology information.

[0065] In some embodiments, the step of calculating the instantaneous power loss of a line segment using the instantaneous power loss integral method based on the measurement data fused by the differentiated weights, and then generating the energy loss through integration via a sliding time window, includes: Receive instantaneous voltage and instantaneous current data at both ends of the line segment; Harmonic separation is performed on the instantaneous voltage and instantaneous current data to obtain the fundamental component and each harmonic component; Based on the fundamental component and each harmonic component, the fundamental power loss and each harmonic power loss are calculated respectively. When calculating the fundamental power loss, the fundamental current component and the fundamental voltage component are used, combined with the measurement data fused by the differentiated weights. When calculating the harmonic power loss, the harmonic current component and each harmonic voltage component are used, combined with the measurement data fused by the differentiated weights. The fundamental power loss and the power loss of each harmonic are superimposed to obtain the total instantaneous power loss of the line segment. The energy loss is obtained by numerically integrating all the total instantaneous power loss within the sliding time window.

[0066] Specifically, high-sampling-rate instantaneous voltage and current values ​​are acquired in real time from measurement units deployed at both ends of the distribution network segment. This data forms the basis for accurate power loss calculations, especially in bidirectional power flow and dynamically changing environments.

[0067] Harmonic separation is performed on instantaneous voltage and current data to obtain the fundamental component and various harmonic components. This can be understood as using Fourier transform (e.g., Fast Fourier Transform, FFT) or other signal processing techniques to decompose the original non-sinusoidal voltage and current waveforms into the fundamental frequency (i.e., the power frequency component) and a series of higher harmonic components. The purpose is to distinguish the energy components of different frequencies so that their contribution to losses can be calculated separately.

[0068] In practical applications, fundamental power loss and harmonic power loss are calculated separately based on the fundamental component and each harmonic component. When calculating the fundamental power loss, the fundamental current and voltage components are used, combined with differentially weighted fused measurement data to ensure accuracy. Similarly, when calculating each harmonic power loss, the harmonic current and voltage components are used, combined with differentially weighted fused measurement data. This separate calculation aims to more precisely assess the contribution of different frequency components to line loss, especially in distribution networks with significant harmonic pollution. The differentially weighted fused measurement data is used here to calibrate or weight the voltage and current data to improve the reliability of the calculation.

[0069] Furthermore, the fundamental power loss and the power losses of each harmonic are superimposed to obtain the total instantaneous power loss of the line segment. This means summing the instantaneous power losses generated by all frequency components to obtain the true total power loss of the line segment at a certain moment.

[0070] Finally, the total instantaneous power loss within the sliding time window is numerically integrated to obtain the energy loss. The sliding time window is a dynamic integration method that accumulates instantaneous power loss within a window of a certain length and slides over time, thus generating the energy loss of a line segment in real time and continuously. This method effectively reflects the dynamic changes in line loss and provides accurate energy loss data for subsequent line loss assessment and optimization.

[0071] The technical solution of this application provides a foundation for accurate calculation by introducing the reception of instantaneous voltage and current data. By performing harmonic separation on these data, complex non-sinusoidal waveforms can be decomposed into fundamental and harmonic components, allowing for the separate calculation of fundamental and harmonic power losses. This separation calculation mechanism, combined with differentially weighted fusion of measurement data, ensures that the contribution of different frequency components to line loss can be accurately assessed even in the presence of harmonic distortion, avoiding calculation errors that may exist in traditional methods. Subsequently, by superimposing the fundamental and harmonic power losses, the total instantaneous power loss of the line segment can be obtained, comprehensively reflecting the actual loss situation of the line segment at a certain moment. Finally, a sliding time window is used to numerically integrate the total instantaneous power loss, enabling dynamic and continuous energy loss calculation, effectively capturing real-time changes in line loss, and thus overcoming the limitations that traditional methods may encounter when dealing with complex power grid environments and high-precision line loss calculations.

[0072] The above technical solution enables more accurate and comprehensive calculation of instantaneous power and energy losses in distribution network segments. Specifically, by performing harmonic separation on instantaneous voltage and current data and calculating the power losses of the fundamental wave and each harmonic separately, this application can effectively address the increasingly serious harmonic pollution problem in modern distribution networks, avoiding underestimation of line losses due to neglecting harmonic losses. Simultaneously, combining measurement data with differentiated weighting further improves the reliability of the calculation results. The use of sliding time window integration ensures the dynamic and real-time nature of energy loss calculation, more accurately reflecting changes in the operating status of the distribution network. This provides a more solid and reliable data foundation for subsequent global line loss assessment and the generation of optimized control commands, thereby improving the overall performance and effectiveness of the distribution network line loss optimization method.

[0073] In some alternative embodiments, it is assumed that in a distribution network, a line segment connecting a distributed photovoltaic grid connection point and a high-load user has measurement units deployed at both ends. These measurement units collect instantaneous voltage and current data at both ends of the line segment in real time. Due to the intermittent output of distributed photovoltaic power and the nonlinear characteristics of high-load users (e.g., including a large number of frequency converters), the voltage and current waveforms on the line segment exhibit obvious harmonic distortion and bidirectional power flow characteristics.

[0074] To accurately calculate the energy loss of the line segment, the received instantaneous voltage and current data are first sent to the processing unit. The processing unit uses the Fast Fourier Transform (FFT) algorithm to perform harmonic separation on these instantaneous data, thereby accurately extracting the fundamental component (50Hz or 60Hz) and the third, fifth, seventh, and other harmonic components.

[0075] Subsequently, based on these separated fundamental voltage and current components and harmonic voltage and current components, the fundamental power loss and the power loss of each harmonic are calculated respectively. For example, the fundamental power loss can be calculated using the effective value of the fundamental voltage, the effective value of the fundamental current, and the fundamental power factor, and corrected by combining the measurement data fused with differentiated weights. Similarly, the power loss of each harmonic is calculated based on the corresponding harmonic voltage and current components, and also combined with the measurement data fused with differentiated weights.

[0076] Next, the calculated fundamental power loss and harmonic power loss are superimposed to obtain the total instantaneous power loss of the line segment at each sampling time.

[0077] Finally, a preset sliding time window (e.g., 15 minutes) is used to numerically integrate these continuous instantaneous power loss data. At shorter time intervals (e.g., 1 minute), the window slides forward, and the integral value of all instantaneous power losses within the window is recalculated, thus generating the energy loss of the line segment in real time. In this way, high-precision line loss data can be obtained even in complex bidirectional power flow and harmonic environments, providing accurate data for the refined management and optimized control of the distribution network.

[0078] In some embodiments, the step of determining the global line loss of the distribution network based on the measurement data fused by the differentiated weights and the energy loss, and generating optimized control instructions based on the global line loss, includes: Identify the current operating mode of the power distribution network; Based on the operating mode, determine the priority of the line loss optimization objective and other operating objectives; Based on the aforementioned priority, the line loss optimization objective is dynamically weighted with other operational objectives; Based on the dynamically weighted line loss optimization target and other operational targets, optimization control instructions are generated.

[0079] Specifically, identifying the current operating mode of the distribution network refers to analyzing and judging the current operating status of the distribution network to determine its typical operating scenario. These operating modes may include, but are not limited to, normal operation mode, high penetration mode of distributed generation, heavy load mode, light load mode, and fault recovery mode. By identifying the operating mode, important contextual information can be provided for subsequent optimization decisions.

[0080] The operational strategies and priorities of the distribution network will differ under different operating modes. For example, under heavy load mode, voltage stability and power supply reliability may have higher priorities; while under high distributed generation penetration mode, maximizing renewable energy absorption and suppressing power flow fluctuations may be more critical. Line loss optimization objectives, along with other operational objectives, can include minimizing voltage deviation, stabilizing power flow, minimizing equipment losses, and maximizing power supply reliability. By prioritizing these objectives, it can be ensured that key operational indicators are given priority during the optimization process.

[0081] In practical applications, different operational objectives are assigned corresponding weights based on predetermined priorities. Dynamic weighting means that these weights are not fixed but are adjusted according to real-time operational patterns and priorities. For example, higher-priority objectives are given greater weights, making them more dominant in the optimization function and guiding the optimization algorithm to converge in a direction that better suits the current operational needs.

[0082] Therefore, multiple operational objectives, after dynamic weighting, are used as part of a comprehensive optimization function. Solving this function using an optimization algorithm ultimately generates a series of specific control commands. These commands can include adjusting the output of distributed power sources, controlling the charging and discharging of energy storage devices, regulating the tap changers of on-load tap changers, and switching reactive power compensation equipment, thereby achieving comprehensive optimized operation of the distribution network.

[0083] The technical solution of this application introduces the identification of distribution network operation modes, making line loss optimization no longer an isolated, single-objective optimization, but closely integrated with the overall operation status of the distribution network. Specifically, by identifying the current operation mode of the distribution network, the main challenges and operational characteristics faced by the current grid can be obtained. Based on this, the priority of line loss optimization objectives and other operational objectives is dynamically determined according to different operation modes. For example, in some modes, voltage stability may be more important than line loss optimization, while in other modes, line loss optimization can be given higher weight. Furthermore, by dynamically weighting these objectives, it is ensured that high-priority objectives are given more full consideration and satisfaction during the optimization process. As a result, the generated optimization control commands can not only effectively reduce line losses, but also take into account the operational needs of the distribution network in terms of voltage quality, power supply reliability, equipment lifespan, and other aspects, thereby maximizing the comprehensive benefits of the distribution network.

[0084] Through the above technical solution, this application overcomes the limitations of traditional line loss optimization methods in complex and variable distribution network environments. Specifically, by identifying the distribution network's operating mode, the line loss optimization strategy can adaptively adjust according to actual operating conditions, avoiding suboptimal solutions that may result from a "one-size-fits-all" optimization approach. Simultaneously, by dynamically determining and weighting the line loss optimization objective with other operating objectives, it ensures that key operating indicators such as voltage stability and power supply reliability are not sacrificed while pursuing line loss reduction, thereby significantly improving the overall efficiency and robustness of distribution network operation. This multi-objective, adaptive optimization control command generation mechanism enables the distribution network to better cope with complex operating scenarios such as distributed power source integration and load fluctuations, improving the intelligence and refined management level of the distribution network.

[0085] In some embodiments, the step of identifying the current operating mode of the distribution network includes: Continuously monitor the voltage, current, power, load curves of each measurement unit in the distribution network and the output data of distributed power sources, and extract features to obtain multi-dimensional operating feature vectors; The multidimensional running feature vector is matched in real time with the feature spaces of multiple preset running modes, and the similarity between the multidimensional running feature vector and each feature space is calculated. When the similarity of multiple modes is within a preset threshold range, the mode conflict resolution mechanism is activated to obtain the evaluation results. The mode conflict resolution mechanism dynamically evaluates the activity level of each mode by analyzing the historical duration, current power flow direction and amplitude change rate of each mode, as well as the voltage fluctuation trend of key nodes. Based on the evaluation results, the mode with the highest activity level is identified as the current operating mode.

[0086] Specifically, various operating parameters are collected and reported in real time by measurement units deployed at key nodes in the distribution network, continuously monitoring the voltage, current, power, load curves, and distributed generation output data of each measurement unit in the distribution network. This data reflects the real-time operating status of the distribution network and forms the basis for pattern recognition. Feature extraction is performed on this data to obtain a multi-dimensional operating feature vector. The purpose is to transform the raw, high-dimensional operating data into refined features that can effectively characterize the operating mode of the distribution network. These features may include statistical or time-series characteristics such as voltage amplitude, current amplitude, active power, reactive power, load factor, and distributed generation output level.

[0087] This process involves real-time matching of the multidimensional operational feature vector with the feature spaces of various preset operational modes, calculating the similarity between the multidimensional operational feature vector and each feature space. This can be understood as comparing the currently extracted operational features with feature templates of historical or predefined typical operational modes (e.g., heavy load mode, light load mode, distributed power supply high output mode, islanded mode, etc.). Similarity calculation can employ machine learning algorithms such as Euclidean distance, cosine similarity, or support vector machine (SVM), with the aim of initially determining which preset mode(s) the current operational state might belong to.

[0088] In practical applications, when the distribution network is operating at mode boundaries, in a transitional phase, or when multiple mode characteristics overlap, traditional single similarity matching may not provide a clear mode identification result. In such cases, a mode conflict resolution mechanism is activated. Its purpose is to further distinguish these ambiguous or conflicting modes by introducing deeper dynamic evaluation indicators. This mechanism dynamically assesses the activity level of each mode by analyzing its historical duration, current power flow direction and amplitude change rate, and voltage fluctuation trends at key nodes. For example, modes with longer historical durations may have higher stability; power flow direction and amplitude change rate can reflect the dynamic evolution trend of the mode; and voltage fluctuation trends at key nodes can indicate the stability of the power grid. These indicators, taken together, can more comprehensively characterize the essence of the current operating state.

[0089] Therefore, based on the evaluation results, the mode with the highest activity level is identified as the current operating mode. The purpose is to select the mode that best represents the actual operating state of the current distribution network through a dynamic evaluation mechanism in the event of mode conflicts, thereby providing an accurate mode basis for subsequent line loss optimization decisions.

[0090] The technical solution of this application can comprehensively capture the real-time operating status of the distribution network by continuously monitoring various operational data and extracting multi-dimensional operational feature vectors. Based on this, by matching these feature vectors with the feature spaces of multiple preset operating modes in real time, the possible modes corresponding to the current operating status can be preliminarily identified. Especially when the distribution network operating modes are complex and changeable, or in the transitional phase of mode switching, multiple modes may have similarity within a preset threshold range, indicating ambiguity or conflict in mode recognition. It is precisely because of the introduction of a mode conflict resolution mechanism that this mechanism can further analyze dynamic indicators such as the historical duration of each mode, the current power flow direction and amplitude change rate, and the voltage fluctuation trend of key nodes, thereby dynamically evaluating these potential modes and quantifying their activity level. Through this multi-dimensional and dynamic evaluation method, even in scenarios where the operating mode is unclear or rapidly changing, the most active and realistic operating mode can be effectively distinguished and determined, avoiding deviations in line loss optimization strategies caused by mode misjudgment.

[0091] Through the above technical solutions, this application can significantly improve the accuracy and robustness of distribution network operation mode identification. Especially in scenarios where the distribution network operation is complex, mode boundaries are ambiguous, or rapid switching occurs, traditional single-matching methods are prone to misjudgment. The mode conflict resolution mechanism proposed in this application, by comprehensively considering multiple dimensions such as the historical persistence of the mode, dynamic changes in power flow, and voltage stability, can effectively solve the ambiguity and conflict problems in mode identification, ensuring that the identified operation mode more accurately reflects the actual operating conditions of the distribution network. This provides a more reliable foundation for subsequent determination of line loss optimization targets, dynamic adjustment of priorities, and generation of optimization control commands, thereby further improving the overall effect and operational stability of distribution network line loss optimization.

[0092] In some embodiments, the step of determining the priority of the line loss optimization objective over other operational objectives based on the operating mode includes: Based on the operating mode, a preset mode-priority mapping table is queried to obtain the initial priority of the line loss optimization target corresponding to the operating mode and other operating targets. The initial priority is dynamically adjusted based on the real-time operating status of the distribution network to obtain the adjusted priority. Based on the adjusted priorities, the priority of the line loss optimization objective is determined to be higher than that of other operational objectives.

[0093] Specifically, the mode-priority mapping table can be understood as a pre-configured data structure that stores the initial priority relationships between different power distribution network operation modes (e.g., normal operation mode, heavy load mode, distributed power high penetration mode, fault recovery mode, etc.) and various operation objectives (e.g., line loss optimization, voltage stability, power factor optimization, equipment life management, etc.). This mapping table aims to provide a reasonable priority benchmark for different operation scenarios, ensuring that under a specific operation mode, there is a preliminary optimization direction that conforms to the preset strategy.

[0094] The real-time operating status of the distribution network may include, but is not limited to, real-time voltage, current, power, load curves, distributed generation output, equipment health status, changes in grid topology, and abnormal event alarms from various measurement units. Dynamically adjusting the initial priority refers to modifying or weighting the preset initial priority based on this real-time operating data to reflect the actual needs and urgency of the current grid. For example, when a local voltage is detected to deviate significantly from the allowable range, the priority of voltage stability may be significantly increased, even if its initial priority is not high under the current operating mode. The aim is to enable the optimization strategy to flexibly respond to the rapidly changing grid, avoiding poor optimization results or new operational problems caused by fixed priority settings.

[0095] Therefore, by comprehensively considering the initial direction provided by the operating mode and the dynamic correction of the real-time operating status, the priority of the line loss optimization target and other operating targets is finally determined, providing an accurate basis for the generation of subsequent dynamic weighting and optimization control commands.

[0096] The technical solution of this application first obtains an initial priority configuration by querying a preset mode-priority mapping table based on the identified operating mode. This sets a benchmark for the optimization target that conforms to the current macroscopic operating state. Based on this, the real-time operating state of the distribution network is further introduced as a dynamic adjustment factor to correct the initial priority. It is precisely this mechanism, combining static preset and dynamic adjustment, that enables not only adherence to predetermined operating strategies but also flexible responses to sudden situations or subtle changes in the power grid. For example, under normal operating mode, line loss optimization may have a higher priority; however, if abnormal voltage fluctuations are detected in a local area in real time, the priority of voltage stability will be dynamically increased to prioritize the reliable operation of the power grid. This dynamic adjustment ensures that the priority setting of the optimization target is more realistic, thereby avoiding limited optimization effects or potential operational risks caused by improper priority settings.

[0097] Through the above technical solution, this application enables intelligent and dynamic management of the priority of distribution network line loss optimization targets and other operational targets. Compared with technical solutions that rely solely on static priority settings based on operating modes, this application can more accurately reflect the real-time operational needs and urgency of the distribution network, effectively avoiding the rigidity or suboptimal optimization strategy caused by fixed priorities in complex and ever-changing operating environments. Therefore, it significantly improves the adaptability and robustness of the distribution network line loss optimization strategy, ensuring that line losses are reduced to the maximum extent possible while guaranteeing the safe and stable operation of the power grid, thereby improving the overall operating efficiency and economy of the distribution network.

[0098] In some optional embodiments, assuming the distribution network is currently identified as operating in "heavy load mode," a preset mode-priority mapping table is queried. In this mapping table, "heavy load mode" may correspond to a medium initial priority for line loss optimization, a high initial priority for voltage stability, and a high initial priority for equipment overload protection. However, during real-time operation, if the voltage of a critical node is consistently below the safety threshold and the load curve shows that the load in that area is still growing rapidly, this indicates that the real-time operating state of the power grid is becoming unstable. At this time, the initial priorities are dynamically adjusted based on this real-time operating data. Specifically, the priority of the voltage stability target is further increased to the highest level, while the priority of the line loss optimization target may be temporarily slightly reduced to prioritize ensuring the voltage stability and operational safety of the power grid. Through this dynamic adjustment, the priority of the line loss optimization target and other operating targets at the current moment is finally determined, enabling the subsequently generated optimization control commands to prioritize solving the most pressing voltage stability problem while also considering line loss optimization, thereby ensuring the safe, stable, and efficient operation of the distribution network in heavy load mode.

[0099] In some embodiments, the step of determining the global line loss of the distribution network based on the measurement data fused by the differentiated weights and the energy loss includes: Monitor the data integrity and real-time performance of the measurement unit; Identify measurement units with missing or degraded data; The missing data is calculated based on the real-time operating data of the adjacent measurement units of the measurement unit with missing or degraded data, the historical operating trend of the measurement unit with missing or degraded data, and the approximate resistance characteristics of the line segment to which the measurement unit with missing or degraded data is responsible. The missing data is labeled with a confidence level to obtain confidence level inference data; Global line loss is assessed based on confidence-based data and other complete and reliable measurement data. The confidence-based data is assigned a lower weight, while the other complete and reliable measurement data is assigned a higher weight. This ensures that the global line loss of the distribution network can still be accurately assessed even when the data is incomplete or unreliable.

[0100] Specifically, continuous monitoring of the data flow of all measurement units in the distribution network is conducted. This can include checking indicators such as packet loss rate, data transmission delay, and data field integrity. When the data integrity or real-time performance of a measurement unit fails to meet preset standards, it is marked as a measurement unit with missing or degraded data. For example, if a measurement unit fails to report data for a continuous period, or if the reported data contains a large number of outliers exceeding reasonable ranges, that measurement unit will be identified. For measurement units identified as having missing or degraded data, data completion is performed using multiple information sources. Real-time operating data from adjacent measurement units can provide spatial correlation information; for example, the trends in voltage and current changes at adjacent nodes are often correlated. Historical operating trends of missing or degraded measurement units can provide temporal regularity; for example, the periodic changes in load curves. The approximate resistance characteristics of the line segment managed by the missing or degraded measurement unit provide constraints on the physical model, helping to more accurately estimate missing voltage, current, or power values. This information is comprehensively utilized through methods such as interpolation, prediction, or estimation based on physical models to generate reasonable estimates for the missing data. Crediting the missing data involves assessing its reliability after the estimated missing data is obtained and assigning it a corresponding credit. This credit can reflect the accuracy and reliability of the estimated data, as well as the degree of deviation it may have from the actual value. For example, the estimated data can be labeled as high, medium, or low reliability based on factors such as the richness of the data source on which the estimation is based, the accuracy of the estimation algorithm, and historical verification results. Global line loss assessment involves integrating the credited estimated data and other complete and reliable measurement data to calculate the global line loss of the entire distribution network. In this process, the credited estimated data is given a lower weight, while other complete and reliable measurement data is given a higher weight. This means that when calculating global line loss, complete and reliable actual measurement data will play a dominant role, while the estimated data serves as an auxiliary and supplementary factor, and its influence on the final result is appropriately reduced to avoid introducing excessive deviation due to estimation errors.

[0101] The technical solution of this application introduces a monitoring mechanism for the integrity and real-time performance of measurement unit data, enabling timely detection and identification of measurement units with data problems. Given the complexity of distribution network operation and the susceptibility of data acquisition to interference, this mechanism ensures proactive intervention when data quality is poor. It is precisely because problematic measurement units are identified that subsequent estimation of missing data becomes possible. By comprehensively utilizing real-time operating data, historical operating trends, and approximate resistance characteristics of line segments from adjacent measurement units, reasonable and physically meaningful estimations of missing data can be made from multiple dimensions, thereby filling data gaps. Based on this, the estimated data is labeled with credibility and fused with differentiated weights to complete and reliable measurement data. This allows for full utilization of all available information during global line loss assessment, while effectively suppressing the uncertainties that may arise from the estimated data, ensuring the robustness and accuracy of the assessment results.

[0102] Through the above technical solution, this application effectively addresses the challenge of missing or degraded measurement data in distribution networks, avoiding the problems of interrupted or reduced accuracy in line loss assessment due to incomplete data. Compared to traditional methods that rely solely on complete and reliable data, this application can still accurately assess the overall line loss of the distribution network even when data is incomplete or unreliable, significantly improving the continuity and reliability of line loss assessment. Furthermore, by assigning lower weights to extrapolated data and higher weights to complete and reliable data, it ensures that the accuracy of the assessment results is not excessively affected by extrapolation errors, thus providing a more solid data foundation for the refined management and optimized control of distribution networks.

[0103] In some embodiments, the multi-dimensional reliability index includes at least one of the following: data freshness, consistency with data from adjacent measurement units, historical error performance of the measurement unit, and device health status of the measurement unit.

[0104] Among these, data freshness refers to the time interval between the acquisition and processing of measurement data; the shorter the time interval, the higher the data freshness and reliability. The degree of mutual verification between the data from the current measurement unit and the data collected by adjacent measurement units, thus assessing the consistency with the data from adjacent measurement units, is also considered. For example, consistency is evaluated by comparing the power balance relationship between adjacent nodes; higher consistency leads to higher reliability. The historical error performance of a measurement unit refers to the statistical characteristics of the data error of that measurement unit over a past period. For example, it is evaluated through indicators such as the deviation and volatility between historical data and the actual value; smaller errors and lower volatility lead to higher reliability. The equipment health status of a measurement unit refers to the operating status of the measurement unit's hardware and software. For example, it is evaluated by monitoring indicators such as the equipment's operating temperature, battery life, communication quality, and software version; good equipment health status leads to higher reliability.

[0105] The technical solution of this application, by introducing the aforementioned multi-dimensional credibility indicators, enables a more comprehensive and refined evaluation of measurement data. Data freshness ensures the evaluation is based on the latest operational status; consistency with data from adjacent measurement units utilizes the topology and physical laws of the distribution network to cross-validate the data; the historical error performance of the measurement units considers the inherent measurement accuracy and stability of the equipment; and the equipment health status of the measurement units reflects the reliability of data acquisition from the source. The comprehensive consideration of these indicators allows for a more accurate assessment of the authenticity and reliability of the data, providing a solid foundation for subsequent differentiated weighting.

[0106] By employing the aforementioned technical solution, when evaluating the reliability of differentiated weights for approximate resistance characteristics and local operating data, the quality and reliability of the data can be comprehensively considered from multiple dimensions. This not only improves the accuracy and robustness of the evaluation, effectively avoiding the bias that may arise from evaluating a single indicator, but also enables a reasonable judgment of data reliability based on multi-dimensional information even in complex situations such as data anomalies, equipment failures, or communication interruptions. This results in the generation of more accurate and reliable differentiated weighted fusion measurement data, laying a solid data foundation for subsequent line loss calculation and optimized control.

[0107] This application also proposes a distribution network line loss optimization system, such as... Figure 2 As shown, a power distribution network line loss optimization system 100 includes: The data receiving module 10 is used to receive local operating data collected in real time by the measurement unit deployed at the key node of the distribution network, wherein the key node is a distributed photovoltaic grid connection point, a high-load user access point, or an on-load tap adjustment point of the distribution transformer, and wherein the key node has edge computing capability; and, based on the local operating data of the measurement unit and the operating information obtained by the measurement unit from adjacent measurement units, to generate an approximate resistance characteristic of the line segment under the responsibility of the measurement unit. The credibility assessment module 20 is used to perform differentiated weight credibility assessment on the approximate resistance characteristics and the local operating data based on preset multi-dimensional credibility indicators, and generate differentiated weight fusion measurement data. The energy loss calculation module 30 is used to calculate the instantaneous power loss of the line segment using the instantaneous power loss integration method based on the measurement data fused by the differentiated weights, and then generate the energy loss by integrating through a sliding time window. The optimization control module 40 is used to determine the global line loss of the distribution network based on the measurement data fused by the differentiated weights and the energy loss, and to generate optimization control instructions based on the global line loss.

[0108] The system of this application, through its data receiving module, can dynamically and in real-time receive and process operational data from local and adjacent measurement units, thereby generating approximate resistance characteristics of line segments. This effectively solves the problem of line parameters being disconnected from physical reality in traditional systems. Furthermore, existing systems often lack effective evaluation and fusion mechanisms when dealing with data quality issues, leading to low-quality data potentially contaminating the entire line loss assessment result. The system of this application, through its reliability assessment module, can perform differentiated weighted reliability assessments of approximate resistance characteristics and local operational data based on multi-dimensional reliability indicators, and generate differentiated weighted fused measurement data. This allows for refined processing of data from different sources and of different qualities, ensuring the reliability of input data and significantly improving the accuracy of line loss assessment.

[0109] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for optimizing line losses in a distribution network, characterized in that, include: Receive local operation data collected in real time by measurement units deployed at key nodes of the distribution network, wherein the key nodes are distributed photovoltaic grid connection points, high-load user access points, or on-load tap adjustment points of distribution transformers, and wherein the key nodes have edge computing capabilities. Based on the local operating data of the measurement unit and the operating information obtained by the measurement unit from adjacent measurement units, an approximate resistance characteristic of the line segment under the responsibility of the measurement unit is generated. Based on preset multi-dimensional credibility indicators, the approximate resistance characteristics and the local operating data are evaluated for differential weight credibility, and measurement data with differential weight fusion is generated. Based on the measurement data fused by the differentiated weights, the instantaneous power loss of the line segment is calculated using the instantaneous power loss integral method, and then the energy loss is generated by integration through a sliding time window. Based on the measurement data fused by the differentiated weights and the energy loss, the global line loss of the distribution network is determined, and an optimized control command is generated based on the global line loss.

2. The method for optimizing distribution network line losses according to claim 1, characterized in that, The step of performing differentiated weighted reliability assessment on the approximate resistance characteristics and the local operating data based on preset multi-dimensional reliability indicators, and generating differentiated weighted fused measurement data includes: Define a local closed-loop region in the distribution network; Obtain the energy imbalance in the local closed-loop region; Based on the energy imbalance, locate the measurement unit that causes the energy imbalance trend and obtain the target measurement unit. Based on the operating data of the target measurement unit, generate micro impedance self-calibration instructions; By executing the micro impedance self-calibration command, the local segment equivalent resistance estimation algorithm of the target measurement unit is adjusted to generate differentially weighted fused measurement data.

3. The method for optimizing distribution network line losses according to claim 1, characterized in that, The step of generating the approximate resistance characteristics of the line segment under the responsibility of the measurement unit based on the local operating data of the measurement unit and the operating information obtained by the measurement unit from adjacent measurement units includes: Obtain the topology probe signal response sent by the measurement unit and construct a local topology connection table; Detect changes in the local topology connection table to determine whether the topology structure has changed; When a change in the topology is detected, the system receives the topology change event reported by the measurement unit and obtains the latest topology connection relationship. Based on the latest topology connection relationship, adjust the adjacent measurement unit information acquisition strategy of the measurement unit to ensure that the measurement unit only obtains operation information from the currently actually connected adjacent measurement units; Based on the operational information obtained by the measurement unit from the adjusted adjacent nodes, and combined with the local operational data, an approximate resistance characteristic of the line segment under the responsibility of the measurement unit is generated.

4. The method for optimizing distribution network line losses according to claim 1, characterized in that, The step of calculating the instantaneous power loss of the line segment using the instantaneous power loss integration method based on the measurement data fused by the differentiated weights, and then generating the energy loss through integration via a sliding time window, includes: Receive instantaneous voltage and instantaneous current data at both ends of the line segment; Harmonic separation is performed on the instantaneous voltage and instantaneous current data to obtain the fundamental component and each harmonic component; Based on the fundamental component and each harmonic component, the fundamental power loss and each harmonic power loss are calculated respectively. When calculating the fundamental power loss, the fundamental current component and the fundamental voltage component are used, combined with the measurement data fused by the differentiated weights. When calculating the harmonic power loss, the harmonic current component and each harmonic voltage component are used, combined with the measurement data fused by the differentiated weights. The fundamental power loss and the power loss of each harmonic are superimposed to obtain the total instantaneous power loss of the line segment. The energy loss is obtained by numerically integrating all the total instantaneous power loss within the sliding time window.

5. The method for optimizing distribution network line losses according to claim 1, characterized in that, The step of determining the global line loss of the distribution network based on the measurement data fused by the differentiated weights and the energy loss, and generating optimized control commands based on the global line loss, includes: Identify the current operating mode of the power distribution network; Based on the operating mode, determine the priority of the line loss optimization objective and other operating objectives; Based on the aforementioned priority, the line loss optimization objective is dynamically weighted with other operational objectives; Based on the dynamically weighted line loss optimization target and other operational targets, optimization control instructions are generated.

6. The method for optimizing distribution network line losses according to claim 5, characterized in that, The steps for identifying the current operating mode of the power distribution network include: Continuously monitor the voltage, current, power, load curves of each measurement unit in the distribution network and the output data of distributed power sources, and extract features to obtain multi-dimensional operating feature vectors; The multidimensional running feature vector is matched in real time with the feature spaces of multiple preset running modes, and the similarity between the multidimensional running feature vector and each feature space is calculated. When the similarity of multiple modes is within a preset threshold range, the mode conflict resolution mechanism is activated to obtain the evaluation results. The mode conflict resolution mechanism dynamically evaluates the activity level of each mode by analyzing the historical duration, current power flow direction and amplitude change rate of each mode, as well as the voltage fluctuation trend of key nodes. Based on the evaluation results, the mode with the highest activity level is identified as the current operating mode.

7. The method for optimizing distribution network line losses according to claim 5, characterized in that, The step of determining the priority of the line loss optimization objective and other operational objectives based on the operating mode includes: Based on the operating mode, a preset mode-priority mapping table is queried to obtain the initial priority of the line loss optimization target corresponding to the operating mode and other operating targets. The initial priority is dynamically adjusted based on the real-time operating status of the distribution network to obtain the adjusted priority. Based on the adjusted priorities, the priority of the line loss optimization objective is determined to be higher than that of other operational objectives.

8. The method for optimizing distribution network line losses according to claim 1, characterized in that, The step of determining the global line loss of the distribution network based on the measurement data fused by the differentiated weights and the energy loss includes: Monitor the data integrity and real-time performance of the measurement unit; Identify measurement units with missing or degraded data; The missing data is calculated based on the real-time operating data of the adjacent measurement units of the measurement unit with missing or degraded data, the historical operating trend of the measurement unit with missing or degraded data, and the approximate resistance characteristics of the line segment to which the measurement unit with missing or degraded data is responsible. The missing data is labeled with a confidence level to obtain confidence level inference data; Global line loss is assessed based on confidence-based data and other complete and reliable measurement data. The confidence-based data is assigned a lower weight, while the other complete and reliable measurement data is assigned a higher weight. This ensures that the global line loss of the distribution network can still be accurately assessed even when the data is incomplete or unreliable.

9. The method for optimizing distribution network line losses according to claim 1, characterized in that, The multi-dimensional reliability indicators include at least one of the following: data freshness, consistency with data from adjacent measurement units, historical error performance of the measurement unit, and equipment health status of the measurement unit.

10. A power distribution network line loss optimization system, characterized in that, The system includes: The data receiving module is used to receive local operating data collected in real time by measurement units deployed at key nodes of the distribution network. The key nodes are distributed photovoltaic grid connection points, high-load user access points, or on-load tap adjustment points of distribution transformers. The key nodes have edge computing capabilities. Based on the local operating data of the measurement units and the operating information obtained by the measurement units from adjacent measurement units, the module generates approximate resistance characteristics of the line segment under the responsibility of the measurement unit. The credibility assessment module is used to perform differentiated weight credibility assessment on the approximate resistance characteristics and the local operating data based on preset multi-dimensional credibility indicators, and generate differentiated weight fusion measurement data. The energy loss calculation module is used to calculate the instantaneous power loss of the line segment using the instantaneous power loss integration method based on the measurement data fused by the differentiated weights, and then generate the energy loss by integrating through a sliding time window. The optimization control module is used to determine the global line loss of the distribution network based on the measurement data fused by the differentiated weights and the energy loss, and to generate optimization control instructions based on the global line loss.

Images