Power management method and device based on intelligent fusion terminal, equipment and medium

Abstract

The application provides a power management method and device based on an intelligent fusion terminal, equipment and a medium, and belongs to the technical field of power management. The method comprises the following steps: obtaining a total output prediction curve of all distributed power sources in a preset future time period based on total load time sequence data of a transformer area, output time sequence data of each distributed power source of the transformer area and environmental data; obtaining a plurality of control instructions based on the total output prediction curve; simulating the working condition of the transformer area in the preset future time period based on each control instruction, and obtaining an adaptability value corresponding to each control instruction based on the simulated working condition data; extracting a candidate control instruction with an adaptability value greater than a preset adaptability threshold; determining a target control instruction based on the candidate control instruction, and managing the distributed power sources corresponding to the transformer area based on the target control instruction. The power management method and device based on the intelligent fusion terminal, the equipment and the medium provided by the application can adapt to the collaborative management requirements of multiple types of distributed power sources under a high penetration rate.

Description

Technical Field

[0001] This application belongs to the field of power management technology, and more specifically, relates to a power management method, device, equipment, and medium based on an intelligent converged terminal. Background Technology

[0002] With the development of distributed energy technologies, distributed power sources such as photovoltaics, energy storage, and wind power have been widely integrated into low-voltage distribution networks, becoming an important support for improving energy utilization efficiency. Currently, the management of distributed power sources in distribution areas largely relies on traditional monitoring systems. These systems combine historical load data with simple statistical models to generate short-term output forecasts, and then, based on the real-time collected total power difference in the distribution area, directly allocate target output values ​​to each distributed power source to form control commands. This type of management adjusts the power supply operating status through real-time feedback, which can meet basic power supply continuity requirements and is widely used in scenarios with low distributed power source penetration.

[0003] However, the output estimation of existing management methods relies on only a single data dimension and a simplified model, without fully integrating multi-source information such as environmental parameters and load timing patterns, resulting in insufficient estimation accuracy. Control commands generated based on real-time power differences lack prediction of future operating conditions and are easily disconnected from the actual regulation capabilities of distributed power sources. This may not only cause safety issues such as voltage exceeding limits and power imbalance in the distribution area, but also reduce the absorption efficiency of distributed power sources due to insufficient command targeting, making it difficult to adapt to the collaborative management needs of multiple types of distributed power sources under high penetration rates. Summary of the Invention

[0004] The purpose of this application is to provide a power management method, device, equipment, and medium based on an intelligent converged terminal, so as to adapt to the collaborative management needs of various types of distributed power sources under high penetration.

[0005] A first aspect of this application provides a power management method based on a smart converged terminal, comprising:

[0006] Based on the total load time series data of the distribution area, the output time series data of each distributed power source in the distribution area, and environmental data, the total output prediction curve of all distributed power sources in the preset future time period is obtained through a long short-term memory network. Each distributed power source includes photovoltaic power sources, energy storage power sources, and wind power sources. The output time series data is continuously recorded at fixed time intervals, and it is an ordered set of power generation data of each distributed power source at different times.

[0007] Multiple sets of control commands are obtained based on the total output prediction curve; each set of control commands contains the target output value of each distributed power source within a preset future time period.

[0008] The operating conditions of the transformer area within a preset future time period are simulated based on each set of control commands to obtain simulated operating condition data, and the fitness value corresponding to each set of control commands is obtained based on the simulated operating condition data.

[0009] Extract multiple sets of control instructions with fitness values ​​greater than a preset fitness threshold as candidate control instructions;

[0010] The target control command is determined based on the candidate control command, and the distributed power supply corresponding to the transformer area is managed based on the target control command.

[0011] A second aspect of this application provides a power management device based on a smart converged terminal, comprising:

[0012] The power output prediction module is used to obtain the total power output prediction curve of all distributed power sources within a preset future time period based on the total load time series data of the transformer area, the power output time series data of each distributed power source in the transformer area, and environmental data, and through a long short-term memory network. Each distributed power source includes photovoltaic power sources, energy storage power sources, and wind power sources. The power output time series data is continuously recorded at fixed time intervals, and it is an ordered set of power generation data of each distributed power source at different times.

[0013] The control command generation module is used to obtain multiple sets of control commands based on the total output prediction curve; each set of control commands contains the target output value of each distributed power source within a preset future time period.

[0014] The fitness calculation module is used to simulate the operating conditions of the transformer area within a preset future time period based on each set of control commands, obtain simulated operating condition data, and obtain the fitness value corresponding to each set of control commands based on the simulated operating condition data.

[0015] The control instruction filtering module is used to extract multiple sets of control instructions with fitness values ​​greater than a preset fitness threshold as candidate control instructions;

[0016] The power management module is used to determine the target control command based on the candidate control commands, and to manage the distributed power supply corresponding to the transformer area based on the target control command.

[0017] A third aspect of this application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement the steps of the power management method based on a smart fusion terminal described above.

[0018] A fourth aspect of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the power management method based on a smart fusion terminal described above.

[0019] The beneficial effects of the power management method, device, equipment, and medium based on intelligent fusion terminals provided in this application are as follows: This application, by integrating the characteristics of distribution area load, power output, and environmental time-series data, and combining the ability of long short-term memory networks to capture time-series dependencies, achieves accurate prediction of total power output, significantly improving prediction accuracy and providing a reliable basis for control command generation, avoiding power imbalance caused by blind adjustments. By generating multiple sets of control commands and quantifying their merits through operating condition simulation and fitness calculation, and then selecting high-quality candidate commands to determine the target command, it ensures that the command adapts to the operating needs of the distribution area, reducing the risk of inefficient and illegal adjustments. Finally, the target command achieves closed-loop management of distributed power sources, forming a complete process of prediction, command, simulation, selection, and execution. This application effectively smooths out intermittent fluctuations in distributed power output, reduces safety risks such as voltage exceeding limits and line overload, improves the efficiency of distributed power absorption, and ultimately achieves the dual goals of safe and stable operation of the distribution area and efficient utilization of distributed power sources. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 A flowchart illustrating a power management method based on a smart converged terminal provided in an embodiment of this application;

[0022] Figure 2 A structural block diagram of a power management device based on a smart fusion terminal provided in an embodiment of this application;

[0023] Figure 3 This is a schematic block diagram of an electronic device provided in an embodiment of this application. Detailed Implementation

[0024] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0025] It is understood that in the embodiments of this application, data such as user information are involved. When the embodiments of this application are applied to specific products or technologies, user permission or consent is required, and the collection, use and processing of related data must comply with relevant laws, regulations and standards.

[0026] It should be noted that the terms "first," "second," etc., used in the specification, claims, and drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such use of data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in sequences other than those illustrated or described herein.

[0027] Before providing a further detailed description of the embodiments of this application, the nouns and terms involved in the embodiments of this application will be explained, and the nouns and terms involved in the embodiments of this application shall be interpreted as follows.

[0028] The intelligent integrated terminal is the main body for the control and management of distributed power sources (photovoltaics, energy storage, wind power). It guides the power source to adjust its operating status by issuing target control commands and receives the actual operating data of the power source to form a closed-loop management, thereby realizing precise control and efficient coordination of the power source.

[0029] In this embodiment, an intelligent fusion terminal with a built-in edge computing module (based on an ARM Cortex-A72 architecture processor, supporting local data processing and model running) can be installed on the low-voltage side of the distribution transformer or key branch nodes in the distribution area to ensure coverage of all distributed power sources and load monitoring points in the distribution area and avoid remote transmission delays.

[0030] A distribution area refers to a region in a power system that is supplied by a single distribution transformer. It encompasses the distribution lines, distributed power sources, electrical loads, and related distribution equipment within that region. It is the basic unit for distribution network management. In this embodiment, the distribution area is a low-voltage distribution area that includes distributed power sources such as photovoltaic, energy storage, and wind power.

[0031] To make the objectives, technical solutions, and advantages of this application clearer, the following description will be provided in conjunction with the accompanying drawings and specific embodiments.

[0032] Please refer to Figure 1 , Figure 1 This is a flowchart illustrating a power management method based on a smart converged terminal according to an embodiment of this application. The power management method based on a smart converged terminal provided in this embodiment can be executed by an electronic device, and the method may include:

[0033] S101: Based on the total load time series data of the distribution area, the output time series data of each distributed power source in the distribution area, and environmental data, the total output prediction curve of all distributed power sources in the preset future time period is obtained through a long short-term memory network; each distributed power source includes photovoltaic power source, energy storage power source and wind power source, and the output time series data is continuously recorded at fixed time intervals, which is an ordered set of power generation data of each distributed power source at different times.

[0034] In this embodiment, the total load time series data of the transformer area refers to the set of electricity consumption data of all users (residential, commercial and industrial) in the transformer area within a historical time period. The collection frequency is 1 minute / time. The collected content may include active power and reactive power. The total load time series data of the transformer area can be stored in the local database of the smart fusion terminal for a storage period of 1 year, which is used for subsequent training and prediction of the long short-term memory network.

[0035] The output time-series data of each distributed power source in the distribution area includes: photovoltaic power generation output time-series data can be collected through the monitoring module of the photovoltaic inverter at a frequency of 1 second / time, and the collected data can include output active power and reactive power; energy storage power generation output time-series data can be collected through the energy storage converter at a frequency of 1 second / time, and the collected data can include charging and discharging power and charging and discharging status; wind power generation output time-series data can be collected through the wind turbine controller at a frequency of 2 seconds / time, and the collected data can include output active power and wind turbine speed. All of the above data are stored in order according to timestamps, forming the output time-series dataset for each distributed power source.

[0036] Environmental data refers to the impact of distributed power sources in the distribution area. For photovoltaic power sources, data such as solar irradiance, ambient temperature, and cloud cover can be collected. For wind power sources, data such as wind speed and direction can be collected. Environmental data can be collected using environmental sensors installed near the distributed power sources at a frequency of once per minute.

[0037] In this embodiment, a three-layer Long Short-Term Memory (LSTM) network structure can be adopted. The input layer has 30 neurons (corresponding to 30 feature dimensions, including the temporal features of historical output data, environmental data, and load data), the hidden layer has 64 neurons, and the output layer has 1 neuron (corresponding to the total output prediction value). The Adam optimizer can be used for network training, with an initial learning rate of 0.001, decaying by 10% every 50 iterations. The root mean square error loss function is used, and the training iterations are 200. The training data uses historical data from the transformer area over the past year, divided into training, validation, and test sets in a 7:2:1 ratio to ensure the model's generalization ability.

[0038] The preset future time period is set to 15-60 minutes in the future, and the prediction time granularity is 1 minute / point. That is, the total output prediction curve contains 15-60 prediction data points, and each data point corresponds to the total output prediction value at a certain moment.

[0039] By using a trained Long Short-Term Memory (LSTM) network, the system takes into account the total load time-series data of the distribution area, the output time-series data of each distributed power source, and environmental data from the past 1-2 hours. It then outputs a sequence of predicted total output values ​​for a preset future time period. This sequence is arranged in chronological order to form a total output prediction curve, which also includes the output fluctuation range (determined based on model prediction error, with fluctuation amplitude ≤ 5%).

[0040] S102: Based on the total output prediction curve, multiple sets of control commands are obtained; each set of control commands contains the target output value of each distributed power source within a preset future time period.

[0041] In this embodiment, multiple sets of control commands are a collection of instructions used to guide the operation of distributed power sources. Each set of control commands sets a target output value for photovoltaic power sources, energy storage power sources, and wind power sources. The setting of the target output value needs to be combined with the adjustment capability of each distributed power source and the adjustment requirements of the distribution area. For example, in a certain set of control commands, the target output value of photovoltaic power source A is 85kW, the target output value of energy storage power source B is 35kW (charging state), and the target output value of wind power source C is 120kW. All target output values ​​are within the adjustable margin range of the corresponding distributed power source. The number of control commands is set to 100 sets to ensure sufficient selection space for subsequent optimization.

[0042] In this embodiment, the total output prediction curve reflects the future power generation potential of distributed power sources. By comparing the matching relationship between this curve and the load demand of the distribution area, the adjustment direction of power surplus that needs to be absorbed or power shortage that needs to be supplemented can be clarified. Based on this direction, target output values ​​are allocated to each power source according to the different adjustment characteristics of photovoltaic, energy storage or wind power (such as flexible charging and discharging of energy storage and rapid reactive power adjustment of photovoltaic). Multiple sets of control commands are generated. The design of multiple sets of commands is to avoid the limitations of a single scheme and to provide sufficient selection space for subsequent selection of the optimal solution.

[0043] S103: Simulate the operating conditions of the transformer area within a preset future time period based on each set of control commands to obtain simulated operating condition data, and obtain the fitness value corresponding to each set of control commands based on the simulated operating condition data.

[0044] In this embodiment, the transformer substation operating condition simulation is the output operating data after simulating the execution effect of control commands through a transformer substation simulation model. The model includes components such as transformers, lines, distributed power sources, and loads. The transformer parameters are set according to the actual transformer models in the transformer substation (e.g., 500kVA, 10kV / 0.4kV), and the line parameters are calculated based on the conductor type (e.g., JKLYJ-120) and length. The target output value in each set of control commands is input into the simulation model to simulate the operating state of the transformer substation at each moment within a preset future time period, and the simulated operating condition data is output, including the voltage of key nodes in the transformer substation, the main line current, the actual grid-connected power of the distributed power source, and the regulation cost. The fitness value is a comprehensive performance index reflecting the quality of control commands, with a value range of 0-100 points. It is calculated by weighting the safety target score (voltage and current compliance) and the efficiency target score (output absorption rate and regulation cost). The higher the score, the better the safety and efficiency performance of the command.

[0045] In this embodiment, a transformer substation simulation model including transformers, lines, power sources, and loads is built. The target output value in each set of control commands is substituted into the model to simulate the operating status of the substation's voltage, current, and power in a future time period, thus obtaining simulated operating condition data. Based on this data, a fitness value is calculated using evaluation criteria of safety and efficiency dimensions. This value transforms the actual operating effect of the commands into a comparable numerical value, providing an objective basis for judging the merits of the commands.

[0046] S104: Extract multiple sets of control instructions with fitness values ​​greater than the preset fitness threshold as candidate control instructions.

[0047] In this embodiment, the preset fitness threshold is the minimum score criterion for determining whether a control instruction has further optimization value. It can be set in conjunction with the safety operation requirements and efficiency targets of the distribution area to screen high-quality candidate control instructions. Candidate control instructions are the set of control instructions with fitness values ​​greater than the preset fitness threshold.

[0048] In this embodiment, the preset fitness threshold is the minimum passing line (e.g., 80 points) set by combining the safety operation standards and efficiency targets of the transformer area. By comparing the fitness value of each control instruction with this threshold, instructions that meet the basic safety and efficiency requirements are selected as candidate control instructions. This process eliminates obviously non-compliant or inefficient instructions, reduces the redundancy of subsequent optimization calculations, and ensures that all instructions entering the next round have the value for further optimization.

[0049] S105: Determine the target control command based on the candidate control command, and manage the distributed power supply corresponding to the transformer area based on the target control command.

[0050] In this embodiment, the target control instruction is the instruction with the best overall performance selected from the candidate control instructions through optimization algorithms such as crossover and mutation. It is the execution instruction that is finally issued to the distributed power source to ensure the safety and economy of the transformer area operation.

[0051] In this embodiment, candidate control commands can be iteratively optimized using optimization algorithms such as crossover mutation to select the target control command with the best overall performance from multiple sets of high-quality commands. Subsequently, the intelligent fusion terminal sends the command to the controllers of each distributed power source through an adapted communication protocol. The controllers adjust the power source operating status according to the command (such as photovoltaic power output adjustment and energy storage switching charging and discharging modes). At the same time, the terminal monitors the actual power output and the operating conditions of the distribution area in real time to ensure a closed loop in the management process, ultimately achieving precise and efficient operation of the distributed power source and balancing the power supply and demand of the distribution area.

[0052] As can be seen from the above, this embodiment achieves accurate prediction of total output by integrating the characteristics of distribution area load, power output, and environmental time-series data, combined with the ability of long short-term memory networks to capture time-series dependencies. This significantly improves prediction accuracy, provides a reliable basis for generating control commands, and avoids power imbalance caused by blind adjustments. By generating multiple sets of control commands and quantifying their merits through operating condition simulation and fitness calculations, and then selecting high-quality candidate commands to determine the target command, it ensures that the command adapts to the operating needs of the distribution area and reduces the risk of inefficient and illegal adjustments. Finally, the target command achieves closed-loop management of distributed power sources, forming a complete process of prediction, command, simulation, selection, and execution. This embodiment effectively smooths out intermittent fluctuations in distributed power output, reduces safety risks such as voltage exceeding limits and line overload, improves the absorption efficiency of distributed power sources, and ultimately achieves the dual goals of safe and stable operation of the distribution area and efficient utilization of distributed power sources.

[0053] In one embodiment of this application, multiple sets of control commands are obtained based on the total output prediction curve, including:

[0054] Based on the total output prediction curve, the total load prediction curve of the transformer area for a preset future time period, and the predicted value of the line loss of the transformer area, the total power difference that needs to be adjusted is obtained.

[0055] Based on the total power difference and the real-time adjustable margin of each distributed power source, an initial population containing multiple sets of initial control commands is generated through random initialization; in each set of control commands, the target output value of each distributed power source must fall within its real-time adjustable margin.

[0056] Using the regulation direction of each distributed power source, the regulation ratio of each distributed power source, and the voltage support contribution as clustering features, and based on the K-means algorithm, the initial population is clustered to obtain multiple feature clusters;

[0057] For each set of initial control commands in each feature cluster, fitness prediction is performed to obtain the fitness prediction value corresponding to each set of initial control commands. The set of initial control commands with the highest fitness prediction value in each feature cluster is selected as the representative control command.

[0058] Based on each set of representative control commands within each feature cluster, Gaussian perturbations are applied to the target output values ​​of each distributed power source to obtain multiple sets of control commands.

[0059] In existing technologies, the generation of control commands based on the total output prediction curve lacks a clear generation logic and constraint mechanism. This leads to issues such as the generated control commands potentially violating the distributed power supply's regulation capabilities or mismatching regulation requirements with the operating conditions of the transformer area. Subsequent extensive screening is required, reducing algorithm efficiency. To address these problems, this embodiment refines the steps for obtaining multiple sets of control commands based on the total output prediction curve. Through a process of total power difference calculation, initial population generation, cluster screening, and Gaussian perturbation amplification, control commands that meet the constraints and possess diversity are generated.

[0060] In this embodiment, the total load forecast curve for the transformer area in the preset future time period is generated by the same long short-term memory network as the total output forecast curve. The input data is the time series data of the total load of the transformer area in the most recent 1-2 hours and environmental data (such as season, time period, etc.). The prediction time granularity and the preset future time period are consistent with the total output forecast curve to ensure that the time dimensions of the two are matched.

[0061] The predicted line loss value for the transformer substation is calculated based on the line parameters (resistance, reactance) and the predicted total output and total load, using the formula... Where I is the predicted line current (obtained by dividing the difference between the predicted total output and the predicted total load by the line voltage), and R is the total line resistance. The predicted line loss is calculated separately for each moment within a preset future time period, forming a predicted line loss sequence.

[0062] The calculation logic for the total power difference that needs to be adjusted is: Total power difference = (Total output forecast - Total load forecast) - Line loss forecast. When the total power difference is positive, it indicates that the power in the distribution area is excessive, and the excess power needs to be absorbed by reducing the output of distributed power sources or charging energy storage power sources. When the total power difference is negative, it indicates that the power in the distribution area is insufficient, and the power gap needs to be supplemented by increasing the output of distributed power sources or discharging energy storage power sources.

[0063] The real-time adjustable margin of each distributed power source is the range of power output that can be safely adjusted under the current operating conditions. For photovoltaic (PV) power sources, the real-time adjustable margin is the difference between the current output value and the maximum power point (MPPT). For example, if the current PV output is 95kW and the MPPT maximum power point is 100kW, the adjustable margin is 80-100kW (adjustable downwards to a minimum of 80kW and upwards to a maximum of 100kW). The real-time adjustable margin of energy storage power sources is determined based on the current state of charge (SOC). When SOC = 60%, the adjustable margin for discharging is 0-100kW, and for charging it is 0-80kW. The real-time adjustable margin of wind power sources is determined based on the current wind speed. When the wind speed is stable, the adjustable margin is ±15% of the rated power. For example, when the rated power is 200kW, the adjustable margin is 170-230kW.

[0064] The initial population is a set of multiple initial control commands generated through random initialization. In this embodiment, the initial population size can be set to 100 initial control commands. The target output value of each command is generated by random number generation, ensuring that the target output value of each distributed power source falls within its real-time adjustable margin. For example, if the adjustable margin of a photovoltaic power source is 80-100kW, then a value between 80-100kW is randomly generated as its target output value; if the adjustable margin of a storage power source's charging is 0-80kW, then a value between 0-80kW is randomly generated as its target charging power.

[0065] Clustering features are indicators used to distinguish differences in control command regulation strategies. The regulation direction of each distributed power source includes three states: increasing output, decreasing output, and maintaining output, represented by 1, -1, and 0, respectively. The regulation ratio of each distributed power source = the regulation of a certain distributed power source / the total regulation (the total regulation is the sum of the absolute values ​​of the regulation of each distributed power source). Voltage support contribution = reactive power regulation of photovoltaic power source / voltage deviation threshold of the distribution area, where the reactive power regulation of photovoltaic power source is the reactive power output value set in the control command, and the voltage deviation threshold of the distribution area is the voltage fluctuation range specified by the national standard (e.g., ±3% × 220V = ±6.6V for a 0.4kV distribution area).

[0066] The K-means algorithm is used to group control commands with similar regulation strategies in the initial population into multiple feature clusters, thus classifying and filtering the commands. In this embodiment, the number of clusters can be set to 1 / 5 of the initial population size, i.e., 20 feature clusters. The number of iterations of the K-means algorithm is set to 50, and the convergence condition is that the change in cluster centers ≤ 0.001. By clustering, initial control commands with similar regulation strategies are grouped into the same cluster, avoiding redundancy of control commands in the initial population. Then, simplified safety and efficiency objectives are used for scoring. The safety objective only assesses whether the voltage is within the preset voltage threshold range, and the efficiency objective only assesses whether the total absorption rate is ≥ 85%. The scoring rule is: 60 points for meeting the voltage target, 0 points for not meeting the target; 40 points for a total absorption rate ≥ 85%, otherwise points are deducted proportionally. The fitness estimate = safety score + efficiency score.

[0067] Within each feature cluster, the initial control command with the highest fitness prediction is selected as the representative control command, resulting in a total of 20 representative control commands to ensure coverage of different regulation strategy types. Using the target output value of the representative control commands as the mean, the perturbation standard deviation is set to 5% of the corresponding real-time adjustable margin of the distributed power source. For example, if the adjustable margin of a photovoltaic power source is 20kW (80-100kW), then the perturbation standard deviation is 1kW. Random perturbation values ​​conforming to a Gaussian distribution are generated and superimposed on the target output value of the representative control commands to obtain new control commands. During the perturbation process, it is ensured that the new target output value still falls within the real-time adjustable margin; if it exceeds it, the adjustable margin boundary value is used.

[0068] As can be seen from the above, this embodiment clarifies the adjustment requirements through the calculation of total power difference, making the generation of control commands more targeted; the constraint mechanism in the initial population generation process ensures the compliance of control commands and avoids the generation of invalid commands; clustering and screening reduce the redundancy of the initial population and improve the efficiency of subsequent optimization; Gaussian perturbation increases the diversity of control commands, ensuring that the algorithm can explore more potential optimal solutions. The final generated control commands have significantly improved safety and efficiency performance, while reducing the computational complexity of the algorithm.

[0069] In one embodiment of this application, obtaining the fitness value corresponding to each set of control commands based on simulated operating condition data includes:

[0070] The simulated operating data is scored based on the objective function to obtain the safety target score and the efficiency target score. The safety target score is the safety operation score of the distribution area when each distributed power source in the distribution area is running based on the target output value within a preset future time period. The efficiency target score is the overall utilization efficiency score of the distributed power source when each distributed power source in the distribution area is running based on the target output value within a preset future time period.

[0071] The fitness value corresponding to each set of control instructions is obtained by weighting the safety target score and the efficiency target score.

[0072] In this embodiment, the objective function is a mathematical model used to reflect the simulated operating data corresponding to the evaluation control commands. It includes two types: a safety objective function and an efficiency objective function. These functions score the safe operation performance of the transformer substation and the utilization efficiency of distributed power sources, respectively, and are the core basis for calculating the safety and efficiency target scores. The safety target score is obtained by scoring the simulated operating data using the safety objective function. It focuses on the safety of the transformer substation operation within a preset future time period. Evaluation dimensions include voltage compliance at key nodes and current compliance of main lines. The score directly reflects the control commands' ability to ensure the safe operation of the transformer substation. The efficiency target score is obtained by scoring the simulated operating data using the efficiency objective function. It focuses on the utilization efficiency of distributed power sources within a preset future time period. Evaluation dimensions include the distributed power source output absorption rate and the degree of optimization of regulation costs. The score directly reflects the improvement effect of the control commands on resource utilization efficiency. Weighted calculation combines the operational principle of prioritizing safety while also considering efficiency in the distribution area. It assigns preset weights (e.g., 60% for safety target score and 40% for efficiency target score) to safety target score and efficiency target score, and calculates the comprehensive score using the formula "fitness value = safety target score × weight + efficiency target score × weight". This comprehensive score is used to measure the overall performance of control commands.

[0073] In this embodiment, simulated operating condition data is input into a first objective function to calculate a safety target score, quantifying the control command's ability to guarantee voltage and current in the distribution area. Simulated operating condition data is input into a second objective function to calculate an efficiency target score, quantifying the optimization effect of the control command on the distributed power absorption rate and regulation cost. The safety target score and efficiency target score are weighted and summed using preset weights to obtain a fitness value, which comprehensively reflects the overall performance of the control command and provides a quantitative basis for subsequent selection and optimization of control commands.

[0074] As can be seen from the above, this embodiment achieves quantitative scoring of safety and efficiency objectives through the objective function, avoiding the bias of subjective evaluation; the weighted calculation method comprehensively considers the safety and efficiency requirements of the transformer area operation, enabling the fitness value to objectively and comprehensively reflect the quality of control commands; the unified scoring standard ensures the comparability between different control commands, provides a reliable basis for the selection of candidate control commands and the determination of target control commands, and improves the stability and reliability of the algorithm.

[0075] In one embodiment of this application, the simulated operating condition data is scored based on an objective function to obtain a safety target score and an efficiency target score, including:

[0076] The safety target score is obtained by scoring the simulated working condition data based on the first objective function;

[0077] The first objective function is:

[0078]

[0079] in, Indicates the score for safety objectives. This indicates the weight corresponding to voltage compliance. This indicates the weight corresponding to current compliance. This represents the ratio of the duration during which the voltage at key nodes in the transformer substation remains within a preset voltage threshold range to the total duration of that time period. It represents the ratio of the duration during which the main line current of the transformer area is within the preset current threshold range to the total duration of that time period within a preset future time period;

[0080] The efficiency target score is obtained by scoring the simulated operating data based on the second objective function;

[0081] The second objective function is:

[0082]

[0083] in, This represents the score for the efficiency target. This indicates the weight corresponding to the absorption efficiency. This indicates the weights corresponding to cost optimization. This represents the ratio of the total actual grid-connected power of all distributed power sources in the distribution area to the total predicted output of each distributed power source within a preset future time period. It represents the ratio of the difference between the benchmark adjustment cost and the actual adjustment cost to the benchmark adjustment cost within a preset future time period.

[0084] In this embodiment, Set to 0.6. Set to 0.4. 1. This reflects the importance of voltage compliance in the safe operation of distribution transformer areas, as voltage exceeding limits has a more direct impact on electrical equipment and grid stability. The preset voltage threshold range must comply with national standards for distribution network voltage. The preset voltage threshold range for 0.4kV distribution transformer areas is 209-239V, and for 10kV distribution transformer areas it is 9.3-10.7kV, determined specifically based on the voltage level of the distribution transformer area. For example, if the preset future time period is 30 minutes (1800 seconds), and the duration for which the voltage at critical nodes in the distribution transformer area is within the 209-239V range is 1782 seconds, then... =1782 / 1800=0.99, which is 99%. The preset current threshold range is the safe operating range for the rated current carrying capacity of the main line in the transformer area, set to 80%-100% of the rated current carrying capacity. For example, if the rated current carrying capacity of the main line is 400A, then the preset current threshold range is 320-400A. For example, if the preset future time period is 30 minutes, and the duration for which the current in the main line of the transformer area is within the 320-400A range is 1800 seconds, then... =1800 / 1800=1.0, which is 100%.

[0085] In this embodiment, Set to 0.6. Set to 0.4. This reflects the core role of grid integration efficiency in the utilization of distributed power sources, prioritizing the full utilization of power generation resources. For example, if the predicted total output of all distributed power sources is 8100 kWh (270 kW × 30 minutes) within a future time period, and the actual total grid-connected power is 7614 kWh, then... =7614 / 8100=0.94, or 94%. The baseline adjustment cost can be the distribution area adjustment cost under conventional adjustment strategies (such as passively disconnecting distributed power sources or fixed power limits), including equipment loss costs and energy consumption costs. For example, under a conventional adjustment strategy, the baseline adjustment cost within 30 minutes is 100 yuan. The actual adjustment cost can be the distribution area adjustment cost corresponding to the current control command, using the same accounting standard as the baseline adjustment cost. For example, under the current control command, the actual adjustment cost within 30 minutes is 80 yuan. The calculation logic is as follows = (Benchmark adjustment cost - actual adjustment cost) / benchmark adjustment cost. For example, (100-80) / 100=0.2, which is 20%, indicating that the adjustment cost of the current control command is reduced by 20% compared with the conventional adjustment strategy.

[0086] In this embodiment, the voltage compliance duration and current compliance duration from the simulated operating condition data are substituted into the first objective function to calculate the safety target score. This score quantifies the control command's ability to guarantee the voltage and current of the transformer area. The actual grid-connected power, the predicted output, the baseline regulation cost, and the actual regulation cost from the simulated operating condition data are substituted into the second objective function to calculate the efficiency target score. This score quantifies the optimization effect of the control command on the distributed power absorption rate and regulation cost. Both objective functions are dimensionless, ensuring that the score range is uniformly 0-100 points, which facilitates subsequent weighted calculations.

[0087] As can be seen from the above, the formulas for the first and second objective functions in this embodiment are scientifically and rationally designed, covering the core evaluation indicators of safe operation of the transformer substation and the utilization efficiency of distributed power sources; dimensionless calculation ensures the comparability and uniformity of scores, avoiding scoring deviations caused by differences in dimensions; specific weight values ​​and indicator definitions provide clear operational standards for the scoring process, allowing those skilled in the art to directly perform calculations based on the formulas and definitions, thus improving the feasibility of the technical solution; at the same time, the formulas are highly flexible, allowing the weights and thresholds to be adjusted according to the actual needs of different transformer substations, adapting to different application scenarios.

[0088] In one embodiment of this application, determining a target control command based on candidate control commands includes:

[0089] Step 1: Perform single-point crossover on the candidate control commands to obtain multiple new sets of candidate control commands;

[0090] Step 2: Randomly mutate multiple sets of new candidate control instructions to obtain multiple sets of adjusted candidate control instructions;

[0091] Step 3: Simulate the operating conditions of the transformer area within a preset future time period based on the target output value corresponding to each set of adjusted candidate control commands to obtain simulated operating condition data; and obtain the fitness value corresponding to each set of adjusted candidate control commands based on the simulated operating condition data, and use the fitness value corresponding to each set of adjusted candidate control commands as candidate control commands.

[0092] Step 4: Repeat steps 1-3 until the preset conditions are met, and determine the target candidate control command based on the fitness obtained when the preset conditions are met;

[0093] The preset conditions include at least one of the following:

[0094] The number of iterations has reached the preset number of iterations;

[0095] The deviation of the optimal fitness value between two consecutive cycles is less than the preset fitness deviation threshold.

[0096] In this embodiment, the instruction optimization operation in the genetic algorithm during single-point crossover refers to selecting a unique crossover point in the distributed power supply regulation term sequence of candidate control instructions, exchanging the regulation terms (such as the target output values ​​of each power supply) before and after the crossover point of two different sets of candidate control instructions, and merging the high-quality regulation logic of the two sets of instructions to generate multiple new sets of candidate control instructions. The purpose is to retain the gene combination of efficient regulation strategies.

[0097] For example: The crossover probability is set to 0.8, meaning there is an 80% probability of crossover for each pair of candidate control commands. The crossover point is selected within the distributed power source regulation item sequence of the control command. For instance, if the regulation item sequence is [target output value of PV A, target output value of PV B, target output value of energy storage C, target output value of wind power D], the crossover point is selected between the target output values ​​of PV B and energy storage C. Regulation items before the crossover point retain the value of the first candidate control command, and regulation items after the crossover point retain the value of the second candidate control command, forming a new candidate control command. For example:

[0098] Candidate control command 1: [85kW, 85kW, 35kW, 120kW]

[0099] Candidate control command 2: [87kW, 83kW, 32kW, 115kW]

[0100] Intersection point: between the second and third items

[0101] New candidate control commands: [85kW, 85kW, 32kW, 115kW]

[0102] Random mutation is an optimization operation that makes small-probability, small-amplitude adjustments to new candidate control commands. It randomly selects some adjustment terms of some commands with a preset mutation probability (e.g., 5%) and applies small perturbations within the real-time adjustable margin of the corresponding distributed power source to avoid the algorithm getting trapped in local optima and maintain the diversity of the command population.

[0103] For example, the mutation probability is set to 0.05, meaning that each adjustment item of each new candidate control command has a 5% probability of mutation. The mutation method involves applying a random disturbance to the target output value, with the disturbance amplitude ≤ 10% of the corresponding distributed power source's real-time adjustable margin. For example, if the adjustable margin of photovoltaic A is 20kW and the target output value is 85kW, then the mutated target output value will be between 83-87kW. During the mutation process, it is ensured that the target output value still falls within the real-time adjustable margin.

[0104] The adjusted candidate control instructions are formed after two optimization operations: single-point crossover and random mutation. They retain the superior adjustment logic of the parent instructions while introducing new adjustment possibilities through mutation, making them the targets for the next round of iterative optimization. Preset conditions are the criteria for terminating the iterative optimization process, including two cases: reaching the target number of iterations and convergence of the optimal fitness value. This ensures that the optimization process avoids blind iteration and waste of resources while fully exploring the optimal instructions. The preset number of iterations is an upper limit (e.g., 50 times) based on the algorithm's optimization efficiency and the adjustment needs of the transformer area. It is a hard condition for terminating the iteration, avoiding excessive iteration that leads to increased computational resource consumption and response latency. The optimal fitness value deviation is the difference between the highest fitness values ​​of all adjusted candidate control instructions in two adjacent iterations. It reflects the degree of convergence of the algorithm; the smaller the deviation, the closer the instruction performance is to the global optimum.

[0105] The preset fitness deviation threshold is a critical value (e.g., 1%) for determining algorithm convergence. When the deviation between the optimal fitness values ​​of two adjacent iterations is less than this threshold, it indicates that the instruction performance has stabilized, and further iteration cannot improve the optimization effect, triggering iteration termination. The target candidate control instruction is the control instruction with the highest fitness value selected from the final adjusted candidate control instruction set after iterative optimization meets the preset conditions. This is the target control instruction used to manage the distributed power supply and represents the optimal execution scheme after multiple rounds of optimization.

[0106] In this embodiment, firstly, single-point crossover is performed on the candidate control commands to integrate the superior adjustment logic of different candidate control commands and generate new candidate control commands; then, the new candidate control commands are randomly mutated to introduce a small amount of random perturbation to avoid the algorithm getting trapped in local optima; the adjusted candidate control commands are subjected to operating condition simulation and fitness value calculation to update the candidate control command set; the above process is repeated until the number of iterations reaches the preset number of iterations or the algorithm converges, and finally, the control command with the highest fitness value is selected as the target control command to ensure the global optimality of the target control command.

[0107] As can be seen from the above, the single-point crossover operation in this embodiment achieves the fusion of high-quality adjustment logic and improves the overall performance of the new candidate control instructions; the random mutation operation introduces population diversity and effectively avoids the algorithm getting trapped in local optima; the explicit iteration termination condition ensures the efficiency of the algorithm and avoids unnecessary waste of computing resources; through iterative optimization, the fitness value of the target control instruction is improved compared with the initial candidate control instruction, ultimately improving the safety of the transformer area operation and the utilization efficiency of distributed power sources.

[0108] In one embodiment of this application, the target control instruction includes multiple instructions;

[0109] Management of distributed power sources corresponding to the transformer area based on target control commands includes:

[0110] The priority of each instruction is determined based on the type of multiple instructions;

[0111] Multiple instructions are sent to the controllers corresponding to each distributed power source according to priority, so as to realize the management of the distributed power sources corresponding to the transformer area.

[0112] In this embodiment, the types of instructions are classified according to the adjustment target, the adjustment object, and the degree of urgency. Specifically, these include, but are not limited to, photovoltaic power active power adjustment instructions, photovoltaic power reactive power adjustment instructions, energy storage power charge / discharge adjustment instructions, wind power active power adjustment instructions, emergency output reduction instructions, and emergency reactive power compensation instructions. Different types of instructions correspond to different operational needs of the distribution area. Priority is an execution order level set based on the importance and response timeliness of the instruction type. It is used to distinguish between the urgency and criticality of instructions (e.g., emergency operation handling instructions have higher priority than regular adjustment instructions), ensuring that critical instructions are executed by distributed power sources first, and avoiding impact on the safe operation of the distribution area due to improper instruction execution order.

[0113] For example: Priority is divided into three levels from high to low. The first priority is emergency operation handling instructions (such as emergency output reduction instructions and emergency reactive power compensation instructions). These instructions are used to deal with emergencies such as voltage surges / dips and line overloads, and need to be executed immediately. The second priority is reactive power regulation instructions (such as photovoltaic power reactive power regulation instructions). These instructions are used to maintain voltage stability in the distribution area and require a high response speed. The third priority is active power regulation instructions (such as photovoltaic power active power regulation instructions, energy storage power charge and discharge regulation instructions, and wind power active power regulation instructions). These instructions are used to balance power in the distribution area and have relatively low response speed requirements.

[0114] The controller corresponding to the distributed power source is a local control unit (such as photovoltaic inverter controller, energy storage PCS controller, wind turbine controller) that is matched one by one with each distributed power source. It has the function of receiving control commands issued by the intelligent fusion terminal and parsing the command content, and driving the distributed power source to adjust its operating status (such as adjusting output and switching working modes). It is the core hardware unit for realizing the execution of commands.

[0115] For example: The intelligent converged terminal first issues a first-priority command to ensure rapid handling of emergency situations; after the first-priority command is executed, a second-priority command is issued; and finally, a third-priority command is issued. The communication protocol for issuing commands is determined based on priority and the distance of the distributed power source. First- and second-priority commands preferentially use low-latency communication protocols (such as RS485 / Modbus wired protocol and LoRa wireless protocol), while third-priority commands can use conventional communication protocols (such as 4G / 5G). For example, for short-distance distributed power sources (distance ≤ 50 meters), the RS485 / Modbus wired protocol is used to issue first- and second-priority commands with a transmission latency ≤ 100ms; for long-distance distributed power sources (distance > 50 meters), the LoRa wireless protocol is used to issue first- and second-priority commands with a transmission latency ≤ 500ms; all third-priority commands for distributed power sources are issued via the 4G / 5G protocol with a transmission latency ≤ 1s.

[0116] In this embodiment, the multiple instructions included in the target control command are first classified by type, and their priorities are determined according to the importance of the instructions and the response speed requirements. The instructions are then issued in the order of "first-level priority → second-level priority → third-level priority" and the appropriate communication protocol is selected based on the distance and communication conditions of the distributed power supply. After receiving the instructions, the controllers of each distributed power supply execute the instructions in priority order to ensure that emergency instructions and critical instructions are executed first, thereby improving the timeliness and effectiveness of instruction execution and ensuring the safe and stable operation of the distribution area.

[0117] As can be seen from the above, this embodiment ensures the priority execution of emergency handling commands and voltage regulation commands by prioritizing them, which significantly shortens the processing delay of emergencies such as voltage exceeding limits in the distribution area and improves the safe and stable operation capability of the distribution area; selecting the communication protocol according to priority ensures the timeliness of command transmission while reducing communication costs; executing commands according to priority avoids conflicts caused by the simultaneous execution of multiple commands, and improves the coordination and accuracy of distributed power supply regulation.

[0118] In one embodiment of this application, after managing the distributed power supply corresponding to the transformer area based on the target control command, the method further includes:

[0119] Collect the actual total output value of each distributed power source in the distribution area after responding to the target control command and the real-time operating condition data of the distribution area, and calculate the output deviation between the actual total output value and the target output value;

[0120] If the output deviation exceeds the preset deviation threshold, a correction control command is generated based on the output deviation and the real-time operating condition data of the distribution area, and then sent to the controller of the corresponding distributed power source.

[0121] In this embodiment, the actual total output value is the total power generation / charging and discharging power value collected and summarized in real time after each distributed power source in the distribution area performs adjustment operations in response to the target control command. It is the core quantitative indicator of the actual operating status of the distributed power source and is compared with the target total output value to judge the adjustment effect.

[0122] The real-time operating status data of the transformer substation is the operating status data of the substation collected in real time at fixed intervals (such as 2 seconds / time) after the target control command is executed. It includes, but is not limited to, the voltage of key nodes in the substation, the current of the main line, the real-time value of the total load, the actual output of the distributed power supply unit, and environmental parameters, providing real-time basis for deviation analysis and the generation of correction commands.

[0123] Output deviation is an indicator used to quantify the difference between the actual operating effect of a distributed power source and the command requirements. The calculation logic is: Output deviation = |Actual total output value - Target total output value| / Target total output value × 100%, which reflects the execution accuracy of the target control command.

[0124] The preset deviation threshold is a critical value set based on the operating accuracy requirements of the transformer area and the regulation capability of the distributed power supply. It is the standard for determining whether deviation correction is needed (e.g., 3%). When the output deviation exceeds this threshold, the correction control command generation process is triggered.

[0125] Correction control commands are generated to address situations where output deviation exceeds the standard. They are specific adjustment commands generated by combining the magnitude of the output deviation, the real-time operating conditions of the distribution area (such as whether the voltage is compliant and whether the load fluctuates), and the real-time adjustable margin of the distributed power source. The purpose is to correct the actual output value of the distributed power source to near the target output value, thereby ensuring power balance and stable operation of the distribution area.

[0126] In this embodiment, after the target control command is executed, the actual total output value of each distributed power source and the real-time operating condition data of the transformer area are collected in real time. The output deviation between the actual total output value and the target total output value is calculated. The output deviation is compared with a preset deviation threshold. If the threshold is exceeded, the cause of the deviation is analyzed in conjunction with the real-time operating condition of the transformer area (such as voltage, current, load, etc.), and a targeted correction control command is generated and issued. After the distributed power source executes the correction control command, it continues to monitor the actual output value and operating condition data to form a closed-loop feedback correction mechanism to ensure that the output deviation is always controlled within the allowable range.

[0127] As can be seen from the above, the closed-loop feedback correction mechanism in this embodiment realizes real-time monitoring and dynamic adjustment of the operating status of distributed power sources, keeping the output deviation within the specified range and improving the regulation accuracy of distributed power sources; timely correction of output deviation avoids problems such as voltage exceeding limits and power imbalance caused by the accumulation of deviation, ensuring the safe and stable operation of the distribution area; the correction control commands are highly targeted and can be flexibly adjusted according to the real-time operating conditions of the distribution area, improving the adaptability and flexibility of the technical solution.

[0128] Based on the same inventive concept, this application also provides a power management device for a smart converged terminal to implement the power management method for a smart converged terminal as described above. The solution provided by this device is similar to the solution described in the above method. Therefore, the specific limitations of one or more power management device embodiments for a smart converged terminal provided below can be found in the limitations of the power management method for a smart converged terminal described above, and will not be repeated here.

[0129] This application provides a power management device based on a smart converged terminal, such as... Figure 2 As shown, the power management device 20 based on the intelligent fusion terminal includes: an output prediction module 21, a control command generation module 22, a fitness calculation module 23, a control command filtering module 24, and a power management module 25.

[0130] The output prediction module 21 is used to obtain the total output prediction curve of all distributed power sources within a preset future time period based on the total load time series data of the transformer area, the output time series data of each distributed power source in the transformer area, and environmental data, and through a long short-term memory network. Each distributed power source includes photovoltaic power sources, energy storage power sources, and wind power sources. The output time series data is continuously recorded at fixed time intervals and is an ordered set of power generation data of each distributed power source at different times.

[0131] The control command generation module 22 is used to obtain multiple sets of control commands based on the total output prediction curve; each set of control commands contains the target output value of each distributed power source within a preset future time period.

[0132] The fitness calculation module 23 is used to simulate the operating conditions of the transformer area within a preset future time period based on each set of control commands, obtain simulated operating condition data, and obtain the fitness value corresponding to each set of control commands based on the simulated operating condition data.

[0133] The control instruction filtering module 24 is used to extract multiple sets of control instructions with fitness values ​​greater than a preset fitness threshold as candidate control instructions;

[0134] The power management module 25 is used to determine the target control command based on the candidate control commands, and to manage the distributed power supply corresponding to the transformer area based on the target control command.

[0135] In one embodiment of this application, the control instruction generation module 22 is specifically used for:

[0136] Based on the total output prediction curve, the total load prediction curve of the transformer area for a preset future time period, and the predicted value of the line loss of the transformer area, the total power difference that needs to be adjusted is obtained.

[0137] Based on the total power difference and the real-time adjustable margin of each distributed power source, an initial population containing multiple sets of initial control commands is generated through random initialization; in each set of control commands, the target output value of each distributed power source must fall within its real-time adjustable margin.

[0138] Using the regulation direction of each distributed power source, the regulation ratio of each distributed power source, and the voltage support contribution as clustering features, and based on the K-means algorithm, the initial population is clustered to obtain multiple feature clusters;

[0139] For each set of initial control commands in each feature cluster, fitness prediction is performed to obtain the fitness prediction value corresponding to each set of initial control commands. The set of initial control commands with the highest fitness prediction value in each feature cluster is selected as the representative control command.

[0140] Based on each set of representative control commands within each feature cluster, Gaussian perturbations are applied to the target output values ​​of each distributed power source to obtain multiple sets of control commands.

[0141] In one embodiment of this application, the fitness calculation module 23 is specifically used for:

[0142] The simulated operating data is scored based on the objective function to obtain the safety target score and the efficiency target score. The safety target score is the safety operation score of the distribution area when each distributed power source in the distribution area is running based on the target output value within a preset future time period. The efficiency target score is the overall utilization efficiency score of the distributed power source when each distributed power source in the distribution area is running based on the target output value within a preset future time period.

[0143] The fitness value corresponding to each set of control instructions is obtained by weighting the safety target score and the efficiency target score.

[0144] In one embodiment of this application, the fitness calculation module 23 is further configured to:

[0145] The safety target score is obtained by scoring the simulated working condition data based on the first objective function;

[0146] The first objective function is:

[0147]

[0148] in, Indicates the score for safety objectives. This indicates the weight corresponding to voltage compliance. This indicates the weight corresponding to current compliance. This represents the ratio of the duration during which the voltage at key nodes in the transformer substation remains within a preset voltage threshold range to the total duration of that time period. It represents the ratio of the duration during which the main line current of the transformer area is within the preset current threshold range to the total duration of that time period within a preset future time period;

[0149] The efficiency target score is obtained by scoring the simulated operating data based on the second objective function;

[0150] The second objective function is:

[0151]

[0152] in, This represents the score for the efficiency target. This indicates the weight corresponding to the absorption efficiency. This indicates the weights corresponding to cost optimization. This represents the ratio of the total actual grid-connected power of all distributed power sources in the distribution area to the total predicted output of each distributed power source within a preset future time period. It represents the ratio of the difference between the benchmark adjustment cost and the actual adjustment cost to the benchmark adjustment cost within a preset future time period.

[0153] In one embodiment of this application, the power management module 25 is specifically used for:

[0154] Step 1: Perform single-point crossover on the candidate control commands to obtain multiple new sets of candidate control commands;

[0155] Step 2: Randomly mutate multiple sets of new candidate control instructions to obtain multiple sets of adjusted candidate control instructions;

[0156] Step 3: Simulate the operating conditions of the transformer area within a preset future time period based on the target output value corresponding to each set of adjusted candidate control commands to obtain simulated operating condition data; and obtain the fitness value corresponding to each set of adjusted candidate control commands based on the simulated operating condition data, and use the fitness value corresponding to each set of adjusted candidate control commands as candidate control commands.

[0157] Step 4: Repeat steps 1-3 until the preset conditions are met, and determine the target candidate control command based on the fitness obtained when the preset conditions are met;

[0158] The preset conditions include at least one of the following:

[0159] The number of iterations has reached the preset number of iterations;

[0160] The deviation of the optimal fitness value between two consecutive cycles is less than the preset fitness deviation threshold.

[0161] In one embodiment of this application, the target control command includes multiple commands; the power management module 25 is further configured to:

[0162] The priority of each instruction is determined based on the type of multiple instructions;

[0163] Multiple instructions are sent to the controllers corresponding to each distributed power source according to priority, so as to realize the management of the distributed power sources corresponding to the transformer area.

[0164] In one embodiment of this application, the power management device 20 based on the intelligent fusion terminal further includes: a correction module, specifically used for:

[0165] Collect the actual total output value of each distributed power source in the distribution area after responding to the target control command and the real-time operating condition data of the distribution area, and calculate the output deviation between the actual total output value and the target output value;

[0166] If the output deviation exceeds the preset deviation threshold, a correction control command is generated based on the output deviation and the real-time operating condition data of the distribution area, and then sent to the controller of the corresponding distributed power source.

[0167] See Figure 3 , Figure 3 This is a schematic block diagram of an electronic device provided according to an embodiment of this application. Figure 3 The electronic device 300 in this embodiment may include one or more processors 301, one or more input devices 302, one or more output devices 303, and one or more memories 304. The processors 301, input devices 302, output devices 303, and memories 304 communicate with each other via a communication bus 305. The memories 304 store computer programs, including program instructions. The processors 301 execute the program instructions stored in the memories 304. Specifically, the processors 301 are configured to invoke the program instructions to perform the functions of the modules in the aforementioned device embodiments, for example... Figure 2 The functions of the output prediction module 21, control command generation module 22, fitness calculation module 23, control command filtering module 24, and power management module 25 are shown.

[0168] It should be understood that, in the embodiments of this application, the processor 301 may be a central processing unit (CPU), or it may be 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. The general-purpose processor may be a microprocessor or any conventional processor.

[0169] Input device 302 may include a touchpad, a fingerprint sensor (for collecting the user's fingerprint information and fingerprint orientation information), a microphone, etc., and output device 303 may include a display (LCD, etc.), a speaker, etc.

[0170] The memory 304 may include read-only memory and random access memory, and provides instructions and data to the processor 301. A portion of the memory 304 may also include non-volatile random access memory. For example, the memory 304 may also store multiple sets of control instructions and other information.

[0171] In specific implementations, the processor 301, input device 302, and output device 303 described in the embodiments of this application can execute the implementation methods described in the power management method based on the intelligent fusion terminal provided in the embodiments of this application, or they can execute the implementation methods of the electronic devices described in the embodiments of this application, which will not be repeated here.

[0172] In another embodiment of this application, a computer-readable storage medium is provided. This computer-readable storage medium stores a computer program, which includes program instructions. When executed by a processor, the program instructions implement all or part of the processes in the methods described above. Alternatively, the computer program can instruct related hardware to complete the process. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include any entity or device capable of carrying computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc.

[0173] The computer-readable storage medium can be an internal storage unit of the electronic device in any of the foregoing embodiments, such as a hard disk or memory of the electronic device. The computer-readable storage medium can also be an external storage device of the electronic device, such as a plug-in hard disk, smart media card (SMC), secure digital card (SD), flash card, etc., provided on the electronic device. Furthermore, the computer-readable storage medium can include both internal and external storage units of the electronic device. The computer-readable storage medium is used to store computer programs and other programs and data required by the electronic device. The computer-readable storage medium can also be used to temporarily store data that has been output or will be output.

[0174] Those skilled in the art will recognize that the modules / units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this application.

[0175] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the electronic devices and units described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0176] In the several embodiments provided in this application, it should be understood that the disclosed electronic devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For instance, the division of modules / units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules may be combined or integrated into another system, or some features may be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces or modules / units, or it may be an electrical, mechanical, or other form of connection.

[0177] The modules / units described as separate components may or may not be physically separate. Similarly, the components shown as modules / units may or may not be physical modules / units; they may be located in one place or distributed across multiple network modules / units. Some or all of the modules / units can be selected to achieve the purpose of the embodiments of this application, depending on actual needs.

[0178] Furthermore, the functional modules in the various embodiments of this application can be integrated into one processing unit, or each module can exist physically separately, or two or more modules can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional module / unit.

[0179] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A power management method based on an intelligent converged terminal, characterized in that, include: Based on the total load time series data of the transformer area, the output time series data of each distributed power source in the transformer area, and environmental data, a total output prediction curve of all distributed power sources within a preset future time period is obtained through a long short-term memory network; the distributed power sources include photovoltaic power sources, energy storage power sources, and wind power sources, and the output time series data is continuously recorded at fixed time intervals, which is an ordered set of power generation data of each distributed power source at different times. Multiple sets of control commands are obtained based on the total output prediction curve; each set of control commands contains the target output value of each distributed power source within a preset future time period. The multiple sets of control commands obtained based on the total output prediction curve include: Based on the total output prediction curve, the total load prediction curve of the transformer area for the preset future time period, and the predicted value of the line loss of the transformer area, the total power difference that needs to be adjusted is obtained. Based on the total power difference and the real-time adjustable margin of each distributed power source, an initial population containing multiple sets of initial control commands is generated through random initialization; in each set of control commands, the target output value of each distributed power source must fall within its real-time adjustable margin. Using the regulation direction of each distributed power source, the regulation ratio of each distributed power source, and the voltage support contribution as clustering features, the initial population is clustered based on the K-means algorithm to obtain multiple feature clusters; For each set of initial control commands in each feature cluster, fitness prediction is performed to obtain the fitness prediction value corresponding to each set of initial control commands. The set of initial control commands with the highest fitness prediction value in each feature cluster is selected as the representative control command. Based on each set of representative control commands within each feature cluster, Gaussian perturbation is applied to the target output value of each distributed power source to obtain multiple sets of control commands. The operating conditions of the transformer area within a preset future time period are simulated based on each set of control commands to obtain simulated operating condition data, and the fitness value corresponding to each set of control commands is obtained based on the simulated operating condition data. Extract multiple sets of control instructions whose fitness values ​​are greater than a preset fitness threshold, and use them as candidate control instructions; The target control command is determined based on the candidate control command, and the distributed power supply corresponding to the transformer area is managed based on the target control command.

2. The power management method based on a smart converged terminal as described in claim 1, characterized in that, The process of obtaining the fitness value corresponding to each set of control commands based on the simulated operating condition data includes: The simulated operating data is scored based on the objective function to obtain a safety target score and an efficiency target score. The safety target score is the safety operation score of the distribution area when each distributed power source in the distribution area is running based on the target output value within a preset future time period. The efficiency target score is the overall utilization efficiency score of the distributed power source when each distributed power source in the distribution area is running based on the target output value within a preset future time period. The fitness value corresponding to each set of control commands is obtained by weighting the safety target score and the efficiency target score.

3. The power management method based on an intelligent converged terminal as described in claim 2, characterized in that, The scoring of the simulated operating condition data based on the objective function to obtain a safety target score and an efficiency target score includes: The simulated operating condition data is scored based on the first objective function to obtain the safety target score; The first objective function is: in, Indicates the score for safety objectives. This indicates the weight corresponding to voltage compliance. This indicates the weight corresponding to current compliance. This represents the ratio of the duration during which the voltage at key nodes in the transformer substation remains within a preset voltage threshold range to the total duration of that time period. It represents the ratio of the duration during which the main line current of the transformer area is within the preset current threshold range to the total duration of that time period within a preset future time period; The simulated operating data is scored based on the second objective function to obtain the efficiency target score; The second objective function is: in, This represents the score for the efficiency target. This indicates the weight corresponding to the absorption efficiency. This indicates the weights corresponding to cost optimization. This represents the ratio of the total actual grid-connected power of all distributed power sources in the distribution area to the total predicted output of each distributed power source within a preset future time period. This represents the ratio of the difference between the benchmark adjustment cost and the actual adjustment cost to the benchmark adjustment cost within a preset future time period.

4. The power management method based on a smart converged terminal as described in claim 1, characterized in that, The step of determining the target control command based on the candidate control commands includes: Step 1: Perform single-point crossover on the candidate control commands to obtain multiple new sets of candidate control commands; Step 2: Randomly mutate the multiple sets of new candidate control instructions to obtain multiple sets of adjusted candidate control instructions; Step 3: Simulate the operating conditions of the transformer area within a preset future time period based on the target output value corresponding to each set of adjusted candidate control commands to obtain simulated operating condition data; and obtain the fitness value corresponding to each set of adjusted candidate control commands based on the simulated operating condition data, and use the fitness value corresponding to each set of adjusted candidate control commands as the candidate control commands. Step 4: Repeat steps 1-3 until the preset conditions are met, and determine the target candidate control command based on the fitness obtained when the preset conditions are met; The preset conditions include at least one of the following: The number of iterations has reached the preset number of iterations; The deviation of the optimal fitness value between two consecutive cycles is less than the preset fitness deviation threshold.

5. The power management method based on a smart converged terminal as described in claim 1, characterized in that, The target control command includes multiple commands; The management of distributed power sources corresponding to the transformer area based on the target control command includes: The priority of each instruction is determined based on its type. The multiple instructions are sent to the controllers corresponding to each distributed power source according to the priority, so as to realize the management of the distributed power sources corresponding to the transformer area.

6. The power management method based on a smart converged terminal as described in claim 1, characterized in that, After managing the distributed power supply corresponding to the transformer area based on the target control command, the method further includes: Collect the actual total output value of each distributed power source in the distribution area after responding to the target control command and the real-time operating condition data of the distribution area, and calculate the output deviation between the actual total output value and the target output value; If the output deviation is greater than a preset deviation threshold, a correction control command is generated based on the output deviation and the real-time operating condition data of the transformer area, and sent to the controller of the corresponding distributed power source.

7. A power management device based on an intelligent converged terminal, characterized in that, include: The power output prediction module is used to obtain the total power output prediction curve of all distributed power sources within a preset future time period based on the total load time series data of the transformer area, the power output time series data of each distributed power source in the transformer area, and environmental data, and through a long short-term memory network. The distributed power sources include photovoltaic power sources, energy storage power sources, and wind power sources. The power output time series data is continuously recorded at fixed time intervals and is an ordered set of power generation data of each distributed power source at different times. The control command generation module is used to obtain multiple sets of control commands based on the total output prediction curve; each set of control commands contains the target output value of each distributed power source within a preset future time period. The control instruction generation module is specifically used for: Based on the total output prediction curve, the total load prediction curve of the transformer area for the preset future time period, and the predicted value of the line loss of the transformer area, the total power difference that needs to be adjusted is obtained. Based on the total power difference and the real-time adjustable margin of each distributed power source, an initial population containing multiple sets of initial control commands is generated through random initialization; in each set of control commands, the target output value of each distributed power source must fall within its real-time adjustable margin. Using the regulation direction of each distributed power source, the regulation ratio of each distributed power source, and the voltage support contribution as clustering features, the initial population is clustered based on the K-means algorithm to obtain multiple feature clusters; For each set of initial control commands in each feature cluster, fitness prediction is performed to obtain the fitness prediction value corresponding to each set of initial control commands. The set of initial control commands with the highest fitness prediction value in each feature cluster is selected as the representative control command. Based on each set of representative control commands within each feature cluster, Gaussian perturbation is applied to the target output value of each distributed power source to obtain multiple sets of control commands. The fitness calculation module is used to simulate the operating conditions of the transformer area within a preset future time period based on each set of control commands, obtain simulated operating condition data, and obtain the fitness value corresponding to each set of control commands based on the simulated operating condition data. The control instruction filtering module is used to extract multiple sets of control instructions whose fitness values ​​are greater than a preset fitness threshold as candidate control instructions; The power management module is used to determine the target control command based on the candidate control commands, and to manage the distributed power supply corresponding to the transformer area based on the target control command.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method as described in any one of claims 1 to 6.

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

Images