Method for dynamically adjusting three-phase imbalance of low-voltage distribution area

Abstract

The invention discloses a three-phase imbalance dynamic adjustment method for a low-voltage distribution area, and the method comprises the steps: synchronously collecting real-time three-phase current data, environment temperature data and current time data of the distribution area, carrying out the timestamp alignment of the collected data, and generating a multi-source data set; analyzing the multi-source data set, calculating a current three-phase unbalance degree, and forming a current scene parameter; performing scene association matching on the current scene parameter and the historical load data, and generating a load prediction result based on a matching result and a typical load mode in the historical load data; the current three-phase unbalance degree and the future unbalance trend are fused, adjustment tasks are distributed in combination with the real-time operation state of an adjustment device in a power distribution area, and a cooperative adjustment strategy is generated; and executing the coordinated regulation strategy, collecting regulated power grid state feedback data, recalculating the three-phase unbalance degree, and iteratively optimizing the coordinated regulation strategy based on a recalculated result until the three-phase current value is converged to a preset target range.

Description

Technical Field

[0001] This invention relates to the field of electrical variable regulation and control technology, specifically a method for dynamic adjustment of three-phase imbalance in a low-voltage distribution transformer area. Background Technology

[0002] Low-voltage distribution substations are the final link in the power system supplying electricity to end users, and their operational stability and power quality directly impact users' electricity experience. In low-voltage distribution networks, the connection of numerous single-phase electrical devices and the randomness and time-varying nature of their power consumption easily leads to uneven three-phase load distribution, i.e., three-phase imbalance. This phenomenon increases line losses, reduces transformer efficiency, and may even burn out electrical equipment, seriously threatening the safe and economical operation of the power grid. Therefore, dynamic adjustment of three-phase imbalance in low-voltage distribution substations is crucial.

[0003] Existing three-phase imbalance regulation technologies typically employ automatic regulating devices. These devices monitor the three-phase current in real time, and when the detected imbalance exceeds a preset fixed threshold, they trigger internal switching elements to automatically switch some single-phase loads from the heavily loaded phase to the lightly loaded phase, thereby achieving load redistribution. Some regulating systems also use different regulation settings based on simple time periods to adapt to coarse load peak and off-peak variations.

[0004] However, the aforementioned existing technologies have significant shortcomings in practical applications. First, these adjustment methods are essentially a delayed, passive response mechanism, intervening only after an imbalance has occurred and reached a certain severity, lacking the ability to predict load changes. Second, their control logic is relatively simple, typically neglecting complex factors affecting user electricity consumption such as ambient temperature and holidays, resulting in poor adjustment performance and low adaptability in variable operating environments. Finally, their adjustment actions are relatively isolated, lacking consideration of future load trends; seemingly reasonable adjustments in the present may lead to new or more severe imbalances in the next period due to natural load evolution. Summary of the Invention

[0005] To address the shortcomings and deficiencies of existing technologies, this invention provides a method and system for dynamic adjustment of three-phase imbalance in low-voltage distribution transformer areas. It aims to solve the problems of passive lag, neglect of complex scenario factors, and isolated adjustment actions in traditional three-phase imbalance adjustment methods, and to achieve accurate and adaptive balance of three-phase current in low-voltage distribution transformer areas.

[0006] This method first synchronously collects real-time three-phase current data, ambient temperature data, and current time data (including date, time period, and holiday information) from the distribution substation. A multi-source dataset is generated through timestamp alignment to ensure data consistency and validity across time dimensions. This multi-source dataset is then analyzed. On one hand, the current three-phase imbalance is calculated (first, the arithmetic mean of the effective values ​​of the three-phase currents is obtained; then, the absolute deviation of each phase current from the mean is determined and the maximum value is selected; finally, the imbalance degree is quantified through normalization). On the other hand, the ambient temperature data and current time data are extracted and combined to form scenario parameters that characterize the current operating scenario. Next, these scenario parameters are correlated and matched with historical load data to filter out historical load data corresponding to similar historical scenarios. This is then combined with clustering from the historical load data. The system analyzes and identifies typical load patterns to generate load forecasts that can predict future imbalance trends. It then integrates the current three-phase imbalance with future imbalance trends, and combines this with the real-time operating status (including switch status, remaining adjustable capacity, most recent action time, and fault alarm information) of regulating devices within the distribution substation (such as automatic three-phase imbalance regulators and smart circuit breakers) to allocate regulating tasks. Based on future imbalance trends, the tasks are time-sequentially scheduled to generate a coordinated regulating strategy that balances immediate regulating needs with future risk predictions. Finally, the coordinated regulating strategy is executed, and the three-phase imbalance is recalculated using the adjusted grid status feedback data. The regulating strategy parameters are iteratively optimized based on the deviation between the new imbalance and the preset target range, and the load forecast results are corrected until the three-phase current values ​​converge to the preset target range.

[0007] Correspondingly, this invention also provides a dynamic adjustment system adapted to this method, including a data acquisition and processing module, a state and scenario analysis module, a load prediction module, a collaborative strategy generation module, and an execution and optimization module. The functions of each module correspond one-to-one with the steps of the aforementioned method, jointly achieving dynamic management of three-phase imbalance. This solution effectively avoids the limitations of traditional methods by combining multi-source data fusion, scenario-related prediction, collaborative strategy generation, and closed-loop optimization, thereby improving the stability of power grid operation and power quality, and reducing line losses and equipment maintenance costs.

[0008] The present invention specifically adopts the following technical solution:

[0009] A method for dynamic adjustment of three-phase imbalance in a low-voltage distribution substation includes:

[0010] Simultaneously collect real-time three-phase current data, ambient temperature data, and current time data containing date, time period, and holiday information from the distribution substation. Perform timestamp alignment processing on the collected data to generate a multi-source dataset.

[0011] The multi-source dataset is analyzed to calculate the current three-phase imbalance, and ambient temperature data and current time data are extracted from the multi-source dataset to form the current scenario parameters.

[0012] The current scenario parameters are matched with historical load data, and load prediction results are generated based on the matching results and typical load patterns in the historical load data to infer future imbalance trends.

[0013] By integrating the current three-phase imbalance with the future imbalance trend, and combining the real-time operating status of the regulating devices in the distribution substation area to allocate regulating tasks, a coordinated regulating strategy is generated.

[0014] The coordinated adjustment strategy is executed, the adjusted power grid state feedback data is collected and the three-phase imbalance is recalculated, and the coordinated adjustment strategy is iteratively optimized based on the recalculated results until the three-phase current values ​​converge to the preset target range.

[0015] Furthermore, the method for performing timestamp alignment processing to generate a multi-source dataset specifically includes: according to a preset time interval, associating the most recently collected real-time three-phase current data and ambient temperature data before each time interval node with the current time data containing date, time period and holiday information corresponding to that time interval node to form a single structured data record, and continuously generating structured data records according to the time sequence to constitute a multi-source dataset.

[0016] Furthermore, the method for calculating the current three-phase imbalance specifically includes: extracting the effective values ​​of the currents of phases A, B, and C at the same timestamp from a multi-source dataset and calculating their arithmetic mean; calculating the absolute deviation between the effective value of the current of each phase and the arithmetic mean, and selecting the largest absolute deviation as the current deviation value; and expressing the ratio of the current deviation value to the arithmetic mean as a percentage to obtain the current three-phase imbalance.

[0017] Furthermore, the method for achieving the scenario association matching specifically includes: parsing hourly features, weekday features, and holiday identifiers from the current time data, and combining them with ambient temperature data to form current scenario parameters; pre-setting weight coefficients for temperature and time features, with the sum of the two being 1; calculating the similarity between the current ambient temperature and historical ambient temperatures, and the matching degree between the current time feature and historical time features, respectively; obtaining the similarity score between the current scenario and each historical scenario by weighted summation of the similarity and matching degree through the weight coefficients; and filtering historical load data corresponding to historical scenarios with similarity scores higher than a preset threshold to complete the scenario association matching.

[0018] Furthermore, the method for generating load prediction results specifically includes: using a clustering analysis algorithm to cluster historical load data to obtain at least one typical load pattern; assigning dynamic weights to the historical load data after association and matching based on the fitting degree between the current scenario parameters and each typical load pattern, wherein the higher the fitting degree, the greater the dynamic weight; and performing a weighted moving average calculation on the historical load data after assigning dynamic weights to obtain the load prediction results for a future preset time period.

[0019] Furthermore, the method of allocating regulation tasks and generating a coordinated regulation strategy based on the real-time operating status of the regulating devices within the distribution substation area specifically includes: collecting the real-time operating status of the regulating devices, which includes the current switch status, remaining adjustable capacity, the time of the most recent action, and fault alarm information; the regulating devices include at least one of a three-phase imbalance automatic regulating device, an intelligent circuit breaker, and a controllable load controller; determining the load level to be transferred based on the current three-phase imbalance, filtering available regulating devices and allocating load transfer tasks based on the remaining adjustable capacity of the regulating devices and fault alarm information; and arranging the allocated regulation tasks in a time sequence based on future imbalance trends: if the future imbalance increases, pre-setting preventive regulation actions; if the future imbalance decreases, delaying non-emergency regulation actions; and combining the time-sequenced regulation tasks with the action commands of the regulating devices to form a coordinated regulation strategy.

[0020] Furthermore, the iterative optimization of the coordinated regulation strategy based on the adjusted grid state feedback data specifically includes: collecting grid state feedback data to recalculate the three-phase imbalance, and calculating the deviation between the recalculated three-phase imbalance and the preset target range; adjusting the calculation parameters of the coordinated regulation strategy according to the deviation value and a preset incremental rule, while using grid state feedback data to correct the load prediction results to obtain updated load prediction results; and regenerating the coordinated regulation strategy by combining the adjusted calculation parameters and the updated load prediction results to complete the iterative optimization.

[0021] And, a three-phase imbalance dynamic adjustment system for a low-voltage distribution substation, comprising:

[0022] The data acquisition and processing module is used to synchronously acquire real-time three-phase current data, ambient temperature data, and current time data containing date, time period, and holiday information of the distribution substation, and perform timestamp alignment processing to generate multi-source datasets;

[0023] The status and scenario analysis module is used to parse the multi-source dataset to calculate the current three-phase imbalance, and extract the ambient temperature data and current time data to form the current scenario parameters.

[0024] The load forecasting module is used to perform scenario association matching with the current scenario parameters and historical load data, and generate load forecasting results based on the matching results and typical load patterns in the historical load data, so as to infer future imbalance trends.

[0025] The collaborative strategy generation module is used to integrate the current three-phase imbalance with the future imbalance trend, and combine the real-time operating status of the regulating device in the distribution area to allocate regulating tasks and generate a collaborative regulating strategy.

[0026] The execution and optimization module is used to execute the coordinated adjustment strategy, collect the adjusted power grid state feedback data and recalculate the three-phase imbalance, and iteratively optimize the coordinated adjustment strategy based on the recalculated results until the three-phase current values ​​converge to the preset target range.

[0027] And a computer device, characterized in that it includes a processor and a memory, the memory storing a computer program, wherein when the processor executes the computer program, it implements the method described above.

[0028] A non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described above.

[0029] Compared with the prior art, the present invention and its preferred embodiments have at least the following beneficial effects:

[0030] First, this invention generates a multi-source dataset by synchronously collecting real-time three-phase current data, ambient temperature data, and current time data containing date, time period, and holiday information, and aligning them with timestamps. This breaks through the limitation of traditional regulation methods that rely solely on single current data, enabling a more comprehensive and accurate perception of the operating status and external conditions of low-voltage distribution substations. This provides a more reliable data foundation for subsequent imbalance calculations and load forecasting, effectively improving the completeness and accuracy of power grid status assessment.

[0031] Secondly, this invention innovatively introduces a scenario-related prediction mechanism, which uses current scenario parameters to match historical load data and combines typical load patterns to generate load prediction results to infer future imbalance trends. This changes the traditional method's passive response only after imbalance occurs, making the regulation strategy forward-looking and able to predict the direction of load changes in advance. It avoids secondary imbalances after regulation due to ignoring future trends, and significantly enhances the adaptability of the regulation scheme to dynamic changes in the power grid.

[0032] Furthermore, by integrating the current three-phase imbalance with the future imbalance trend, and combining the real-time operating status of the regulating device to allocate tasks and perform timing arrangement, this invention generates a coordinated regulating strategy. This ensures both immediate response to the current imbalance and proactive response to future imbalance risks. At the same time, it avoids isolated operation or frequent start-stop of the regulating device, optimizes the utilization efficiency of regulating resources, reduces equipment wear and tear, and extends the service life of the regulating device.

[0033] Finally, this invention constructs a closed-loop mechanism of execution-feedback-iterative optimization, which continuously corrects the load prediction results and strategy calculation parameters based on the adjusted grid state feedback data until the three-phase current converges to the target range. Unlike the limitation of traditional methods that only adjust once without optimization, this invention can dynamically adapt to the fluctuations and uncertainties in the grid operation process, continuously improve the adjustment accuracy, ensure the long-term stable operation of the grid, reduce line losses and overall operation and maintenance costs, and improve the power experience and power quality of end users. Attached Figure Description

[0034] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments:

[0035] Figure 1 This is a schematic diagram of the three-phase imbalance dynamic adjustment method for low-voltage distribution substations according to an embodiment of the present invention;

[0036] Figure 2 This is a diagram illustrating the dynamic weight allocation mechanism in an embodiment of the present invention.

[0037] Figure 3 This is a diagram showing the collaborative strategy generation in an embodiment of the present invention;

[0038] Figure 4 This is a schematic diagram of the three-phase unbalanced dynamic adjustment system for low-voltage distribution substations according to an embodiment of the present invention. Detailed Implementation

[0039] In the following, specific embodiments of this application will be described in detail with reference to the accompanying drawings. Based on these detailed descriptions, those skilled in the art will be able to clearly understand and implement this application. Without departing from the principles of this application, features from various embodiments can be combined to obtain new implementations, or certain features from some embodiments can be substituted to obtain other preferred implementations.

[0040] To make the features and advantages of the present invention more apparent and understandable, specific embodiments are described below in conjunction with the accompanying drawings:

[0041] To address the problems existing in the current technology, referring to Figure 1In the three-phase imbalance dynamic adjustment method of the low-voltage distribution substation in this embodiment of the invention, multi-source data fusion and scenario association technology are adopted, combined with load prediction and feedback optimization mechanism, which can achieve accurate and adaptive three-phase current balance in the low-voltage distribution substation.

[0042] The implementation of this embodiment includes:

[0043] Acquire real-time three-phase current data, ambient temperature data, and current time data of the distribution substation, and combine them to generate a multi-source dataset;

[0044] Analyze the multi-source dataset to calculate the current three-phase imbalance.

[0045] Acquire historical load data, use multi-source datasets to perform scenario correlation on historical load data, and generate load prediction results;

[0046] By integrating the current three-phase imbalance with load forecast results, a coordinated regulation strategy is calculated and generated.

[0047] The coordinated adjustment strategy is implemented, and the adjusted grid state is monitored to generate feedback data. The coordinated adjustment strategy is then iteratively optimized using the feedback data to obtain the target convergence current value.

[0048] Specifically, in this embodiment, a multi-source dataset that comprehensively reflects the current power grid operation scenario is first constructed by integrating real-time three-phase current, ambient temperature, and time information. Based on this dataset, two core tasks are executed in parallel: firstly, real-time analysis to accurately calculate the current three-phase imbalance degree of the power grid, serving as the basis for immediate adjustment; secondly, forward-looking analysis to intelligently correlate the current scenario with historical load data, thereby generating predictions of future load changes. The core innovation of this method lies in the deep integration of real-time imbalance state assessment and future imbalance trend prediction, forming a coordinated adjustment strategy that takes into account both the present and the future. Finally, through a closed-loop process of executing the strategy, monitoring the effect, feeding back data, and iterative optimization, the adjustment strategy can self-correct and iterate based on actual results, ultimately guiding the power grid state to converge to the ideal balance target.

[0049] As a preferred embodiment, the combination of generating multi-source datasets includes:

[0050] Acquire real-time three-phase current data, ambient temperature data, and current time data;

[0051] Real-time three-phase current data, ambient temperature data, and current time data are timestamped to generate a multi-source dataset.

[0052] Specifically, in this embodiment, the first step is to deploy corresponding data acquisition devices in the low-voltage distribution transformer area. Data is collected in real time via intelligent terminals or metering devices installed at the low-voltage side outgoing lines of the distribution transformer. The three-phase line current is measured to obtain real-time three-phase current data. Simultaneously, temperature sensors are installed at representative locations in the surrounding environment of the distribution substation, or ambient temperature data is obtained by connecting to a local meteorological service data interface. Current time data is accurately acquired using a system clock synchronized with a standard time source to ensure a unified time reference for all data. After acquiring these three types of heterogeneous data, the core operation is timestamp alignment. Because different sensors may have different sampling frequencies and data reporting mechanisms, direct combination can lead to time misalignment of data, affecting the accuracy of the analysis. Timestamp alignment aims to integrate data from different sources to the same time point based on a unified time series. For example, a fixed time interval, such as every minute, can be set to associate the most recently acquired real-time three-phase current data and ambient temperature data before that time point with the current time data at that time point, forming a structured data record containing timestamps, three-phase current values, and ambient temperature values. By continuously performing this operation, a strictly synchronized multi-source dataset is ultimately generated, where each record accurately reflects the electrical state and external environmental conditions of the distribution substation at a specific moment.

[0053] As a preferred embodiment, the current three-phase imbalance is calculated by parsing the multi-source dataset, including:

[0054] The average value of the three-phase current is calculated based on the three-phase current data;

[0055] Determine the current deviation between the average three-phase current and the corresponding current of each phase;

[0056] The current deviation value is normalized to generate the current three-phase imbalance.

[0057] Specifically, in this embodiment, the datasets with the same timestamp are first extracted from the multi-source datasets. The effective three-phase current value is denoted as... , , Based on these three current data points, the average three-phase current was calculated. This average value reflects the overall load level of the current transformer area. The calculation formula is as follows:

[0058] ,

[0059] In the formula, , , They were obtained through real-time monitoring. Mutually, Harmony The effective value of the phase current, The calculated average value of the three-phase current, This represents the number of three-phase currents. Subsequently, to assess the deviation of each phase load from the average level, it is necessary to determine the current deviation between the average three-phase current and the corresponding current in each phase. This value represents the current difference with the most severe deviation under the current imbalance condition. It is calculated by determining the absolute difference between the current in each phase and the average current, using the following formula:

[0060] ,

[0061] In the formula, This represents the current deviation value in the three-phase current. The function selects the current difference among the three-phase currents that deviates most severely from the average level at the current moment. Finally, to eliminate the influence of the overall load level on the imbalance evaluation and to make the evaluation results comparable, the obtained current deviation values ​​are normalized to generate the current three-phase imbalance degree. This normalization process typically involves expressing the current deviation value as a percentage, relative to the average three-phase current. The calculation formula is as follows:

[0062] ,

[0063] In the formula, This is the final generated current three-phase imbalance, which can intuitively reflect the degree of imbalance of the current three-phase current in the low-voltage distribution substation.

[0064] As a preferred embodiment, the load prediction results generated by using multi-source datasets to perform scenario correlation on historical load data include:

[0065] Ambient temperature data and current time data are extracted from multi-source datasets to form current scenario parameters;

[0066] Match historical scenarios with parameters similar to the current scenario in historical load data, and filter to obtain related historical load data;

[0067] Perform trend analysis on historical load data to generate load forecast results.

[0068] Specifically, in this embodiment, the first step is to extract key information characterizing the current operating scenario from the real-time collected and aligned multi-source dataset. This information includes ambient temperature data and current time data, which are then combined to form the current scenario parameters. The current time data here is not simply a timestamp, but is parsed into features with specific physical meaning, such as hour, day of the week, and whether it is a holiday. These features collectively characterize the current social activities and electricity consumption patterns. Subsequently, the stored historical load database is accessed. In this database, each historical load data entry is associated with the scenario parameters at the time of its occurrence. By setting a scenario similarity evaluation function, the current scenario parameters are compared with each scenario parameter in the historical database. This evaluation function can comprehensively consider the matching degree of time features and the proximity of temperature values, for example:

[0069] ,

[0070] In the formula, For the current situation and the first Similarity score of each historical scenario and These are the weighting coefficients for temperature and time characteristics, which sum to 1 and are preset according to the characteristics of the transformer area. It is a function used to calculate the current temperature. With historical temperature The similarity is such that the smaller the temperature difference, the larger the function value. It is a function used to calculate the current time feature. Characteristics of historical time The similarity score is set to 1 if the match is correct, and 0 or a smaller value otherwise. By setting a similarity threshold, all matches are filtered out. Historical scenarios exceeding this threshold are identified, and the corresponding three-phase load data are extracted to form a set of correlated historical load data. Finally, trend analysis is performed on this filtered set of correlated historical load data. In simple terms, a weighted average can be calculated over a future time window for the historical load curves under similar scenarios, with the weights proportional to their scenario similarity scores, thereby generating a three-phase load prediction result for a future period.

[0071] As a preferred embodiment, the coordinated regulation strategy calculated by integrating the current three-phase imbalance and load prediction results includes:

[0072] Based on the load forecast results, the future imbalance trend is inferred and a future imbalance prediction is generated.

[0073] Based on the current three-phase imbalance, determine the urgency of the current adjustment and generate immediate adjustment requirements;

[0074] By combining predictions of future imbalances with immediate adjustment needs, a coordinated adjustment strategy is generated.

[0075] Specifically, in this embodiment, the predicted three-phase current values ​​for each future time segment in the load prediction results are first substituted into the same calculation model used to calculate the current three-phase imbalance, thereby obtaining the predicted imbalance at a series of future time points. The calculation process is as follows:

[0076]

[0077] In the formula, Represents a point in the future. It is obtained from the load forecast results. The predicted three-phase current values ​​at time t. It is the average value of the three-phase predicted current calculated based on these three predicted values. It is calculated The three-phase imbalance is predicted at any given time. By performing this calculation on all key time points within a future time window, a time series describing the trend of imbalance changes can be formed, i.e., future imbalance prediction. Simultaneously, based on the calculated current three-phase imbalance, the urgency of current regulation is determined, generating immediate regulation demand. This step is achieved by comparing the current three-phase imbalance value with preset multi-level thresholds. For example, warning thresholds and severity thresholds are set. When the imbalance exceeds the warning threshold but is below the severity threshold, the immediate regulation demand is defined as "regulation needed"; when it exceeds the severity threshold, the demand is defined as "urgent regulation." This demand not only includes the priority of regulation but also implies the magnitude of the load to be transferred, and its magnitude is positively correlated with the severity of the current imbalance. Finally, by comprehensively considering the future imbalance prediction and the immediate regulation demand, the final coordinated regulation strategy is generated. This decision-making process is a trade-off process. For example, if the immediate need for adjustment is "urgent adjustment", the strategy will prioritize large-scale and rapid adjustment actions, even if the future imbalance prediction shows a positive trend. Conversely, if the current imbalance is low, but the future imbalance prediction shows that it will deteriorate sharply in a short period of time, the strategy will initiate preventive and gradual adjustment actions in advance to avoid possible severe imbalances in the future by smoothing out peaks and valleys.

[0078] Preferably, the above-mentioned coordinated adjustment strategy can be generated by those skilled in the art based on the above-mentioned scheme in this embodiment and in combination with existing known technologies, by establishing and solving a multi-objective optimization model. This model can consider minimizing the adjustment cost (penalizing frequent and large-scale switching actions) and the future imbalance risk (penalizing the imbalance of future cycle predictions based on load prediction results) as objectives. The coordinated adjustment strategy generated by this multi-objective optimization model will serve as the initial strategy input of the closed-loop optimization system. When subsequent feedback data shows that the adjustment effect has not reached the target, the closed-loop system will adjust the calculation parameters of the model (such as weight coefficients and penalty factors) based on the measured deviation, and re-solve the optimization model to generate the iterative strategy, thereby achieving deep coupling between the optimization model and the closed-loop feedback.

[0079] The constraints to be set may include:

[0080] Load phase uniqueness constraint: Each load can only be connected to one phase at any given time.

[0081] Regulation device capacity constraint: The current of each phase must not exceed its safe upper limit after regulation.

[0082] Action count limit: There is a limit to the number of times a single load can switch over in a short period of time (e.g., within 1 hour).

[0083] Irreversible operation cycle: After load switching, at least K control cycles must be maintained before switching again to prevent equipment damage; where K is a positive integer preset according to the technical specifications of the regulating device.

[0084] Solving the optimization model is generally a mixed-integer programming problem. For small to medium-sized distribution areas, standard optimization solvers (such as Gurobi and CPLEX) can be used. For large-scale or real-time-critical scenarios, heuristic algorithms (such as genetic algorithms and particle swarm optimization) or model predictive control (MPC) frameworks can be used for rolling optimization, solving the optimization problem in the future finite time domain at each decision point and executing the actions of the first cycle. The boundary value of the number of load nodes can be adjusted according to specific hardware performance and real-time requirements.

[0085] The selection of the solution method in this reference is determined based on the scale of the distribution area and the real-time requirements: For small-to-medium-scale scenarios with fewer than 50 load nodes in the distribution area, standard optimized solvers such as Gurobi and CPLEX are preferred to ensure the global optimality of the strategy; for large-scale scenarios with more than 100 load nodes, or high real-time scenarios with a response latency requirement of less than 10 seconds for adjustment commands, heuristic algorithms such as genetic algorithms and particle swarm optimization are used to improve the solution speed within an acceptable accuracy range; for scenarios that require continuous dynamic adjustment of the strategy, the MPC framework is used for rolling optimization, updating the prediction time domain and solving every 5 minutes to achieve dynamic adaptation of the strategy.

[0086] As a preferred embodiment, a coordinated regulation strategy is implemented, and feedback data is generated by monitoring the regulated grid state. The coordinated regulation strategy is then iteratively optimized using the feedback data, including:

[0087] Issue and execute coordinated control strategies to multiple control devices;

[0088] After execution, new three-phase current data is collected to generate feedback data;

[0089] Based on the feedback data, a new three-phase imbalance is calculated;

[0090] Determine whether the new three-phase imbalance meets the preset adjustment target. If not, use the new three-phase imbalance to trigger a recalculation of the coordinated adjustment strategy.

[0091] Specifically, in this embodiment, the execution phase of the regulation strategy is initiated first. The control system, based on specific instructions included in the coordinated regulation strategy, such as phase sequence switching, load transfer amount, and execution time, issues and executes control commands to multiple regulating devices deployed in the low-voltage distribution substation. These regulating devices can be three-phase imbalance automatic regulating devices, intelligent circuit breakers, controllable load controllers, etc. For reference, the three-phase imbalance automatic regulating device, as the core regulating unit, is responsible for the precise transfer of loads between phases, directly improving the three-phase current distribution; the intelligent circuit breaker is used to control the on / off state of the loads on important branches within the substation, cooperating to achieve phase switching of large loads; the controllable load controller performs power regulation for dispatchable loads (such as energy storage devices and adjustable power appliances), assisting in smoothing out three-phase imbalance, and the actions of each device are uniformly scheduled by the coordinated regulation strategy.

[0092] After the command is issued, these devices will automatically complete the switching or adjustment of phase-to-phase loads according to the strategy requirements. After a short and stable time period following the execution of the adjustment action, the monitoring and feedback process is immediately initiated. New three-phase current data after the adjustment action takes effect is collected through three-phase current data acquisition terminals deployed in the distribution substation. This real-time data forms the direct basis for evaluating the adjustment effect, i.e., generating feedback data. Subsequently, based on this feedback data, the aforementioned three-phase imbalance calculation method is called again to calculate a new three-phase imbalance. This new imbalance value directly reflects the implementation effect of this adjustment strategy. Next is the judgment and iteration phase, which compares this new three-phase imbalance with the preset adjustment target. This adjustment target is usually an acceptable imbalance threshold, such as the value specified by national standards. If the judgment result indicates that the new three-phase imbalance has met or exceeded the preset adjustment target, the adjustment is considered successful, and the system will enter continuous monitoring mode, awaiting the next imbalance exceeding the limit event. If the judgment result is not met, it indicates that the effect of a single adjustment did not meet expectations, or that an unforeseen change in load occurred during the adjustment process. In this case, the new three-phase imbalance that did not meet the standard will be used as the new current state input to trigger a recalculation of the coordinated adjustment strategy. This recalculation process will repeat the aforementioned steps of data acquisition, imbalance calculation, load prediction, and strategy generation, thereby forming a dynamic and continuously optimized closed-loop adjustment system.

[0093] As a preferred approach, the closed-loop optimization convergence determination of this scheme adopts a dual-indicator principle: 1) Static indicator: the three-phase imbalance is lower than the preset adjustment target (e.g., 15%) for three consecutive data acquisition cycles (1 minute per cycle); 2) Dynamic indicator: the absolute value of the imbalance change between two adjacent cycles is less than 1%, and the adjustment device has no pending instructions. When both indicators are met, the system is determined to have converged, the closed-loop optimization stops, and the system enters continuous monitoring mode. If only the static indicator is met but the dynamic indicator is not met, the fine-tuning strategy needs to be executed again before re-determining the convergence.

[0094] Preferably, the above-mentioned regulation strategies of this invention can be implemented using existing or specially deployed regulation devices within the distribution area, and coordinated control can be achieved through a unified communication protocol (such as ModbusTCP, IEC61850, MQTT). The application of communication protocols in this scheme follows a layered adaptation principle: the ModbusTCP protocol is used between the regulation devices and the control system within the distribution area to achieve real-time command issuance and status acquisition, ensuring the timeliness of regulation actions; the IEC61850 protocol is used between the control system and the power grid dispatching layer for data interaction, adapting to the standardized communication requirements of the power system; and the MQTT protocol is used between the data acquisition module and the cloud-based historical load database for non-real-time historical data retrieval, balancing communication efficiency and resource consumption.

[0095] As a preferred embodiment, trend analysis is performed on the associated historical load data to generate load prediction results, including:

[0096] Identify typical load patterns in correlated historical load data;

[0097] Dynamic weights are assigned to associated historical load data based on the similarity between current scenario parameters and typical load patterns.

[0098] A weighted moving average is calculated on the associated historical load data with the assigned dynamic weights to generate load forecast results.

[0099] The load prediction results generated based on typical load patterns and dynamic weights can accurately predict the three-phase load change trend in future periods, which can then be converted into the future three-phase imbalance trend. This trend, together with the current three-phase imbalance, will serve as the core input, providing a forward-looking decision-making basis for the generation of subsequent coordinated adjustment strategies.

[0100] Specifically, the first step is to preprocess the entire historical load database to identify typical load patterns. This step can employ clustering algorithms, such as K-Means or fuzzy C-means clustering, to categorize the massive daily load curves according to their peak time, peak-to-valley difference, volatility, and other morphological characteristics. The center or representative curve of each category is defined as a typical load pattern, such as "weekday residential pattern," "weekend commercial pattern," and "holiday special pattern." After obtaining the associated historical load data, the current scenario parameters are correlated with these identified typical load patterns to determine which pattern(s) are most likely to be followed. Next, a dynamic weight is assigned to each historical data point in the associated historical load data set. The determination of this weight integrates two dimensions: the similarity between its original scenario and the current scenario, and the fit between its own load curve and the currently dominant typical load pattern. The dynamic weight allocation mechanism is as follows: Figure 2 As shown. This dual consideration makes the weight allocation more accurate.

[0101] The preferred dynamic weight allocation method of this scheme is as follows: the dynamic weight assigned to each associated historical load data is obtained by proportionally weighting and summing the corresponding scenario similarity score and the fit degree of the load curve and typical load pattern; the scenario similarity score and the fit degree both range from 0 to 1, and the fit degree can be calculated by the cosine similarity algorithm; the weight allocation ratio of the two is controlled by an allocation coefficient, which can be set according to the load characteristics of the transformer area. For example, the scenario similarity ratio can be set to 0.4 and the pattern fit degree ratio can be set to 0.6 for residential transformer areas, and vice versa for commercial transformer areas, so as to represent the different importance ratios of the two dimensions in the weight allocation.

[0102] Finally, a weighted moving average is calculated on the associated historical load data with assigned dynamic weights to generate the final load forecast result. The calculation formula can be expressed as:

[0103] ,

[0104] In the formula, It is the predicted future. Load values ​​at each point in time. Current time. Summation symbol. Iterate through all the filtered historical load data. Is assigned to the first The dynamic weights associated with historical load data are determined by both scenario similarity and pattern fit; the larger the value, the higher its reference value for prediction. It is the first The associated historical load data on its occurrence date The actual load value at a given point in time.

[0105] As the preferred solution for this implementation, the coordinated adjustment strategy, which combines future imbalance predictions with immediate adjustment needs, includes:

[0106] Acquire the real-time operating status of multiple regulating devices and generate the regulating resource status;

[0107] Based on the immediate adjustment needs, allocate adjustment tasks according to the adjustment of resource status;

[0108] Based on predictions of future imbalances, the allocated adjustment tasks are time-series arranged to generate a coordinated adjustment strategy.

[0109] Specifically, in this embodiment, the collaborative strategy is generated as follows: Figure 3As shown, the first step is to acquire the real-time operating status of multiple regulating devices and generate a regulating resource status list. The control system needs to communicate with all three-phase imbalance regulating devices in the distribution area to query and collect their key status information, including but not limited to the current switching status of the device, remaining adjustable capacity, the time of the most recent action, and whether there are any fault alarms. This information is integrated into a dynamically updated regulating resource status table, providing a realistic basis for subsequent task allocation. Next, based on the immediate regulating needs generated in the previous steps, regulating tasks are allocated on the basis of the regulating resource status. This step is an optimization solution process, aiming to achieve the lowest regulating cost or the highest regulating efficiency while meeting the regulating needs. For example, if the immediate regulating need is to transfer a certain amount of load from the heavily loaded phase to the lightly loaded phase, the regulating resource status table is traversed to find the combination of regulating devices that can complete this task with the fewest actions or the least equipment loss, and a specific load transfer amount is assigned to each selected device to form a preliminary regulating task set. The final step, which is also the key to reflecting coordination and foresight, is to schedule the allocated regulating tasks according to the prediction of future imbalances to generate the final coordinated regulating strategy. By examining the future imbalance forecast curve, if the forecast indicates that the imbalance will naturally improve or worsen at some point in the future, the timing of the currently assigned regulation tasks will be adjusted. For example, a non-urgent regulation task might be delayed to avoid conflict with natural load fluctuations, or a regulation task might be broken down into multiple smaller steps and executed gradually at different points in the future to smooth the impact of the regulation process on the power grid.

[0110] As a preferred implementation for reference, the timing arrangement of this scheme follows the rule of "load level adaptation + grid tolerance constraint": 1) Task decomposition rule: When the load to be transferred in a single instance exceeds 30% of the rated capacity of the regulating device, it is decomposed into 2 to 3 sub-tasks, and the load transfer amount of each sub-task does not exceed 20% of the rated capacity of the device; 2) Time interval determination method: The execution interval of the sub-task is determined according to the voltage fluctuation tolerance of the power grid in the distribution area. It is 5 minutes for conventional distribution areas and 10 minutes for distribution areas with high voltage stability requirements. The interval time must be greater than the duration corresponding to the irreversible cycle K of the regulating device.

[0111] The final output of the coordinated regulation strategy is a detailed action plan that specifies when, what operation, and the magnitude of each regulation device should perform.

[0112] As a preferred embodiment, the recalculation of the coordinated regulation strategy triggered by the new three-phase imbalance includes:

[0113] Use the new three-phase imbalance as the current state input;

[0114] The load forecast results are updated based on the feedback data to generate updated load forecast results.

[0115] Based on the deviation between the new three-phase imbalance and the regulation target, adjust the calculation parameters of the coordinated regulation strategy;

[0116] The updated collaborative adjustment strategy is generated by using the adjusted strategy calculation parameters and the updated load prediction results.

[0117] Specifically, firstly, the new three-phase imbalance that failed to meet the target after this adjustment is used as the initial state input for the next decision cycle. This allows the formulation of the next strategy to be directly based on the actual effect of the previous operation, forming the basis of feedback control. Secondly, the feedback data containing the new three-phase current data is used to update the load prediction model. Specifically, this set of latest measured data is used as a new anchor point and input into the prediction algorithm to correct the original load prediction results in real time, especially to correct the short-term prediction curve, thereby generating an updated load prediction result that is closer to the current actual evolution trend of the power grid. Next, the strategy calculation model itself will be self-tuned. This is achieved by quantifying the deviation between the new three-phase imbalance and the adjustment target. This deviation value, as an error signal, is used to adjust the calculation parameters of the coordinated adjustment strategy. A key adjustment gain parameter is adjusted in the following way:

[0118] ,

[0119] In the formula, The calculation parameters for the adjusted coordinated regulation strategy, These are the parameter values ​​before adjustment. It is a preset learning rate or adjustment coefficient used to control the step size of parameter adjustment. It is the new three-phase imbalance measured after implementing the adjustment strategy. This is the preset adjustment target. This formula means that if the actual effect differs significantly from the target, the adjustment range of the parameters will also increase accordingly, thus making the next adjustment action more forceful. Finally, using the calculation parameters of this dynamically adjusted collaborative adjustment strategy, and the load prediction results updated based on the latest feedback data, the collaborative adjustment strategy generation algorithm will be re-executed to calculate a corrected and optimized updated collaborative adjustment strategy, which will then be issued and executed.

[0120] Based on the same inventive concept, such as Figure 4 As shown, this embodiment of the invention also provides a three-phase imbalance dynamic adjustment system for a low-voltage distribution substation corresponding to the above method, comprising:

[0121] The data acquisition module is used to acquire real-time three-phase current data, ambient temperature data and current time data of the distribution substation, and combine them to generate a multi-source dataset;

[0122] The data analysis module is used to parse multi-source datasets and calculate the current three-phase imbalance.

[0123] The load forecasting generation module is used to acquire historical load data, perform scenario correlation on the historical load data using multi-source datasets, and generate load forecasting results.

[0124] The collaborative strategy generation module is used to integrate the current three-phase imbalance with the load forecast results to calculate and generate a collaborative adjustment strategy.

[0125] The execution and feedback module is used to execute the coordinated regulation strategy, monitor the regulated grid state to generate feedback data, and use the feedback data to iteratively optimize the coordinated regulation strategy to obtain the target convergence current value.

[0126] In the above embodiments, parameters that are not explicitly defined (such as similarity functions) can be implemented based on techniques known in the art (such as linear interpolation or threshold comparison).

[0127] To verify the feasibility of this invention in practice, the proposed solution was applied to a typical low-voltage distribution substation in a city with mixed residential and commercial use. This substation has a complex power load, with the tidal nature of residential electricity consumption and the concentrated daytime demand from commercial electricity consumption overlapping, leading to frequent and unpredictable three-phase imbalance problems. Traditional regulation methods often rely on passive responses after the imbalance exceeds limits, resulting in problems such as regulation lag, frequent actions, and inability to adapt to rapid load changes, making it difficult to achieve ideal mitigation results. This substation aims to use the method of this invention to achieve dynamic prediction, proactive regulation, and closed-loop optimization of three-phase imbalance.

[0128] In this application example, the dynamic adjustment method of this invention is deployed in the transformer substation. The system first collects real-time three-phase current data (A, B, and C) from a smart terminal installed on the low-voltage side of the transformer, and obtains ambient temperature data from the local meteorological service, along with the current time data from the system clock. These data are timestamped and combined to generate a multi-source dataset that provides the basis for subsequent analysis. Based on this dataset, the system calculates the current three-phase imbalance in real time. Simultaneously, it extracts the current time and ambient temperature as current scenario parameters, matches similar scenarios in historical load data, and performs dynamic weighted calculations based on pre-identified typical load patterns such as the "weekday peak business pattern" through cluster analysis, generating a load forecast and imbalance trend prediction for the next two hours.

[0129] During the specific adjustment process, the system integrates the current imbalance level with future imbalance predictions to generate a coordinated adjustment strategy. For example, at a certain time on a certain day, the system detects that the current three-phase imbalance level is 18.5%, exceeding the preset 15% warning threshold, and generates an immediate adjustment request that "adjustment is needed." Simultaneously, load forecasting results show that due to the continuous influx of commercial loads, the imbalance level will further deteriorate to approximately 25% within the next hour. Based on this, the system immediately queries the real-time operating status of the automatic three-phase imbalance adjustment devices in the distribution area. After confirming that their adjustable capacity is sufficient, a coordinated adjustment strategy is generated. This strategy not only issues an immediately executed load transfer command to resolve the current 18.5% imbalance but also, based on future imbalance predictions, presets a preventative fine-tuning action 30 minutes later to smoothly cope with the upcoming imbalance peak and avoid secondary disturbances to the power grid.

[0130] After the strategy was executed, the system entered the closed-loop feedback optimization phase. Five minutes after the adjustment was completed, the system collected new three-phase current data and calculated a new imbalance of 11.8%, which was lower than the adjustment target of 15%, indicating that the adjustment was successful. In another load surge test, the imbalance after the first adjustment was 16.1%, still below the target. The system then triggered a recalculation mechanism, using 16.1% as the new current state input, and adaptively adjusted the strategy calculation parameters based on the 1.1% adjustment deviation. Simultaneously, the load prediction results were updated using the latest measured current feedback data, generating a more targeted updated collaborative adjustment strategy, which was then re-executed, ultimately successfully reducing the imbalance to 13.5%, achieving the preset target.

[0131] Data comparison shows that the dynamic adjustment method of this invention has advantages in adjustment effect, response speed, and intelligence level. Taking the adjustment event of the aforementioned time as an example, this invention, through a collaborative adjustment strategy, achieves a response time of approximately 5 minutes and reduces the imbalance from 18.5% to 11.8%. In contrast, the traditional threshold-triggered adjustment method typically has a response time of over 15 minutes, and due to a lack of predictive ability, the imbalance after adjustment still hovers around 16%, resulting in poor performance. The closed-loop optimization mechanism enables this system to adapt to complex operating conditions, ensuring the ultimate achievement of the adjustment target.

[0132] Therefore, compared with the prior art, the advantages of the above solutions in the embodiments of the present invention include:

[0133] This method, by collecting and fusing real-time information on three-phase current, ambient temperature, and time, can comprehensively perceive changes in the power grid's operating status. This collaborative analysis of multi-source data enables the system to accurately calculate the current three-phase imbalance, providing a reliable data foundation for subsequent precise adjustments and improving the accuracy and comprehensiveness of status perception.

[0134] This invention innovatively integrates the current operating state with load forecasting results based on scenario association. By combining immediate imbalance with future load trends, the system can generate a coordinated adjustment strategy that takes into account both current needs and future changes. This avoids the drawback of traditional methods that only focus on the current state and ignore dynamic development trends, thus enhancing the foresight and adaptability of adjustment measures.

[0135] After executing the regulation strategy, this invention continuously monitors the power grid status and generates feedback data, forming a closed-loop optimization mechanism. By continuously recalculating and optimizing the strategy using the feedback data, the regulation process can dynamically converge to the target current value. This not only improves the accuracy of regulation but also ensures the system's continuous and stable operation in the face of power grid fluctuations.

[0136] By comprehensively considering immediate adjustment needs and future imbalance predictions, this method can intelligently allocate and schedule tasks for multiple regulating devices. This collaborative scheduling mechanism optimizes the utilization efficiency of regulating resources, avoids frequent or conflicting actions of regulating devices, thereby improving three-phase imbalance, extending equipment lifespan, and reducing overall operating costs.

[0137] Based on the same inventive concept, this invention also provides a computer device, comprising: one or more processors, and a memory for storing one or more computer programs; the programs include program instructions, and the processor executes the program instructions stored in the memory. The processor may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing and control core of the terminal, used to implement one or more instructions, specifically for loading and executing one or more instructions stored in a computer storage medium to implement the above-described method.

[0138] It should be further explained that, based on the same inventive concept, the present invention also provides a computer storage medium storing a computer program, which, when executed by a processor, performs the above-described method. This storage medium can be any combination of one or more computer-readable media. A computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. A computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In the present invention, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0139] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this disclosure. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0140] The foregoing has shown and described the basic principles, main features, and advantages of this disclosure. Those skilled in the art should understand that this disclosure is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of this disclosure. Various changes and modifications can be made to this disclosure without departing from its spirit and scope, and all such changes and modifications fall within the scope of this disclosure as claimed.

[0141] This invention is not limited to the above-described preferred embodiment. Anyone inspired by this invention can derive other forms of dynamic adjustment method for three-phase imbalance in low-voltage distribution substations. All equivalent changes and modifications made within the scope of the claims of this invention shall fall within the scope of this invention.

Claims

1. A method for dynamic adjustment of three-phase imbalance in a low-voltage distribution transformer area, characterized in that, include: Simultaneously collect real-time three-phase current data, ambient temperature data, and current time data containing date, time period, and holiday information from the distribution substation. Perform timestamp alignment processing on the collected data to generate a multi-source dataset. The multi-source dataset is analyzed to calculate the current three-phase imbalance, and ambient temperature data and current time data are extracted from the multi-source dataset to form the current scenario parameters. The current scenario parameters are matched with historical load data, and load prediction results are generated based on the matching results and typical load patterns in the historical load data to infer future imbalance trends. By integrating the current three-phase imbalance with the future imbalance trend, and combining the real-time operating status of the regulating devices in the distribution substation area to allocate regulating tasks, a coordinated regulating strategy is generated. The coordinated adjustment strategy is executed, the adjusted power grid state feedback data is collected and the three-phase imbalance is recalculated, and the coordinated adjustment strategy is iteratively optimized based on the recalculated results until the three-phase current values ​​converge to the preset target range.

2. The method for dynamic adjustment of three-phase imbalance in a low-voltage distribution substation according to claim 1, characterized in that: The method for performing timestamp alignment processing to generate a multi-source dataset specifically includes: according to a preset time interval, associating the most recently collected real-time three-phase current data and ambient temperature data before each time interval node with the current time data containing date, time period and holiday information corresponding to that time interval node to form a single structured data record, and continuously generating structured data records according to the time sequence to constitute a multi-source dataset.

3. The method for dynamic adjustment of three-phase imbalance in a low-voltage distribution substation according to claim 1, characterized in that: The method for calculating the current three-phase imbalance specifically includes: extracting the effective values ​​of the currents of phases A, B, and C at the same timestamp from a multi-source dataset and calculating their arithmetic mean; calculating the absolute deviation between the effective value of the current of each phase and the arithmetic mean, and selecting the largest absolute deviation as the current deviation value; and expressing the ratio of the current deviation value to the arithmetic mean as a percentage to obtain the current three-phase imbalance.

4. The method for dynamic adjustment of three-phase imbalance in a low-voltage distribution substation according to claim 1, characterized in that: The method for achieving the scenario association matching specifically includes: parsing hourly features, weekday features, and holiday identifiers from the current time data, and combining them with ambient temperature data to form current scenario parameters; pre-setting weight coefficients for temperature and time features, with the sum of the two being 1; calculating the similarity between the current ambient temperature and historical ambient temperatures, and the matching degree between the current time feature and historical time features; obtaining the similarity score between the current scenario and each historical scenario by weighted summation of the similarity and matching degree using the weight coefficients; and filtering historical load data corresponding to historical scenarios with similarity scores higher than a preset threshold to complete the scenario association matching.

5. The method for dynamic adjustment of three-phase imbalance in a low-voltage distribution substation according to claim 1, characterized in that: The method for generating load prediction results specifically includes: using a clustering analysis algorithm to cluster historical load data to obtain at least one typical load pattern; assigning dynamic weights to the historical load data after association and matching based on the fitting degree between the current scenario parameters and each typical load pattern, wherein the higher the fitting degree, the greater the dynamic weight; and performing a weighted moving average calculation on the historical load data after assigning dynamic weights to obtain the load prediction results for a future preset time period.

6. The method for dynamic adjustment of three-phase imbalance in a low-voltage distribution substation according to claim 1, characterized in that: The method of allocating regulation tasks and generating a coordinated regulation strategy by combining the real-time operating status of the regulating devices within the distribution substation area specifically includes: collecting the real-time operating status of the regulating devices, which includes the current switch status, remaining adjustable capacity, the time of the most recent action, and fault alarm information; the regulating devices include at least one of a three-phase imbalance automatic regulating device, an intelligent circuit breaker, and a controllable load controller; determining the load level to be transferred based on the current three-phase imbalance, selecting available regulating devices and allocating load transfer tasks based on the remaining adjustable capacity of the regulating devices and fault alarm information; and arranging the allocated regulating tasks in a time sequence based on future imbalance trends: if the future imbalance increases, pre-setting preventive regulation actions; if the future imbalance decreases, delaying non-emergency regulation actions; and combining the time-sequentially arranged regulating tasks with the action commands of the regulating devices to form a coordinated regulation strategy.

7. The method for dynamic adjustment of three-phase imbalance in a low-voltage distribution substation according to claim 1, characterized in that: The iterative optimization of the coordinated regulation strategy based on the adjusted grid state feedback data specifically includes: collecting grid state feedback data to recalculate the three-phase imbalance, and calculating the deviation between the recalculated three-phase imbalance and the preset target range; adjusting the calculation parameters of the coordinated regulation strategy according to the preset incremental rules based on the deviation, and simultaneously using grid state feedback data to correct the load prediction results to obtain updated load prediction results; and regenerating the coordinated regulation strategy by combining the adjusted calculation parameters and the updated load prediction results to complete the iterative optimization.

8. A three-phase imbalance dynamic adjustment system for a low-voltage distribution substation, characterized in that, include: The data acquisition and processing module is used to synchronously acquire real-time three-phase current data, ambient temperature data, and current time data containing date, time period, and holiday information of the distribution substation, and perform timestamp alignment processing to generate multi-source datasets; The status and scenario analysis module is used to parse the multi-source dataset to calculate the current three-phase imbalance, and extract the ambient temperature data and current time data to form the current scenario parameters. The load forecasting module is used to perform scenario association matching with the current scenario parameters and historical load data, and generate load forecasting results based on the matching results and typical load patterns in the historical load data, so as to infer future imbalance trends. The collaborative strategy generation module is used to integrate the current three-phase imbalance with the future imbalance trend, and combine the real-time operating status of the regulating device in the distribution area to allocate regulating tasks and generate a collaborative regulating strategy. The execution and optimization module is used to execute the coordinated adjustment strategy, collect the adjusted power grid state feedback data and recalculate the three-phase imbalance, and iteratively optimize the coordinated adjustment strategy based on the recalculated results until the three-phase current values ​​converge to the preset target range.

9. A computer device, characterized in that, It includes a processor and a memory, the memory storing a computer program, which, when executed by the processor, implements the three-phase imbalance dynamic adjustment method for low-voltage distribution substations according to any one of claims 1-7.

10. A non-transitory computer-readable storage medium, characterized in that, The storage medium stores a computer program, which, when executed by a computer processor, implements the three-phase imbalance dynamic adjustment method for low-voltage distribution substations as described in any one of claims 1-7.

Images