Edge decision power distribution method based on intelligent electric meter

Abstract

The invention provides an edge decision power distribution method based on an intelligent electric meter, and relates to the technical field of power distribution, and the method comprises the steps: monitoring the data of an access circuit, and transmitting an edge power distribution decision signal to a plurality of edge power utilization ends; continuously acquiring power utilization characteristic data, and performing edge power distribution decision; the intelligent electric meter monitors and receives an edge power distribution decision result sent by any edge power utilization end, or selects historical power utilization characteristic data sent and recorded by any edge power utilization end to carry out power utilization characteristic data prediction and central power distribution decision; carrying out balanced power distribution analysis, and carrying out power distribution on the plurality of edge power utilization ends; and balance power distribution analysis is continued. Through the method and the device, the technical problem of time delay caused by the fact that a plurality of edge power utilization ends sequentially send power distribution requirements and decision information in the prior art can be solved, the technical target of parallel sending of the edge power utilization ends is achieved, and the technical effect of improving the real-time performance of power distribution is achieved.

Description

Technical Field

[0001] The present application relates to the field of power distribution technology, and in particular to an edge decision-making power distribution method based on a smart meter. Background Art

[0002] In modern power systems, with the rapid growth of electricity demand and the diversification of power-consuming devices, the traditional power distribution model faces more and more challenges. Especially in the context of the gradual popularization of edge computing and smart meters, how to achieve efficient and balanced power distribution in a distributed environment has become an urgent problem to be solved.

[0003] At present, the existing multiple edge power users often need to send power distribution requirements and decision information in sequence. Although this sequential transmission mode can meet the needs in some simple scenarios, in complex multi-edge environments, the delay of sequential transmission will seriously affect the overall performance of the system. When each edge power user generates and sends its power distribution decision, it must wait for the response of the previous power user, resulting in delays in the overall decision-making process. In scenarios with high real-time requirements, such as responding to sudden power fluctuations or rapidly changing power demand, the reaction speed of the existing system is not enough to ensure the rationality and timeliness of power distribution, resulting in power outages or power waste.

[0004] To sum up, there is a technical problem in the prior art that due to multiple edge power users sending power distribution demands and decision information in sequence, the system will produce significant time delays when facing a complex multi-edge environment, further affecting the real-time and overall performance of power distribution. Summary of the invention

[0005] The purpose of this application is to provide an edge decision-making power distribution method based on smart meters, so as to solve the technical problem in the prior art that due to multiple edge power users sending power distribution demands and decision information in sequence, the system produces significant time delays when facing a complex multi-edge environment, further affecting the real-time and overall performance of power distribution.

[0006] In view of the above problems, the present application provides an edge decision-making power distribution method based on smart meters.

[0007] The present application provides an edge decision-making power distribution method based on a smart meter, comprising: monitoring access circuit data through an access monitor in the smart meter, sending an edge power distribution decision signal to multiple edge power terminals after a preset fluctuation event occurs, and obtaining multiple signal sending times, wherein the multiple edge power terminals periodically transmit power consumption feature data to the smart meter; the multiple edge power terminals receive the edge power distribution decision signal, continuously collect power consumption feature data within a preset number of time frames, and make edge power distribution decisions, and after any edge power terminal completes the edge power distribution decision, send a first edge power distribution decision result to the smart meter, and obtain a first decision sending time; the smart meter continuously monitors and receives the edge power distribution decision result sent by any edge power terminal according to the preset monitoring time, and when the first decision sending time and the corresponding first signal sending time are less than or equal to the preset When the monitoring time is set, the first edge power distribution decision result is received and the monitoring reception continues; otherwise, the smart meter randomly selects any historical power consumption feature data sent and recorded by the edge power user to perform power consumption feature data prediction and central power distribution decision, and obtains the first central power distribution decision result; the smart meter performs balanced power distribution analysis according to the first edge power distribution decision result or the first central power distribution decision result, obtains the first balanced power distribution analysis result including multiple first power distribution parameters, and distributes power to the multiple edge power users; continues to monitor and obtain the second edge power distribution decision result or process and obtain the second central power distribution decision result according to twice the preset monitoring time, and performs balanced power distribution analysis in combination with the first edge power distribution decision result or the first central power distribution decision result, until the balanced power distribution analysis is performed based on the power distribution decision results of the multiple edge power users, or the preset fluctuation event ends.

[0008] One or more technical solutions provided in this application have at least the following technical effects or advantages: The access circuit data is monitored by the access monitor in the smart meter, and after a preset fluctuation event occurs, an edge power distribution decision signal is sent to multiple edge power terminals, and multiple signal sending times are obtained, wherein the multiple edge power terminals periodically transmit power consumption feature data to the smart meter; the multiple edge power terminals receive the edge power distribution decision signal, continuously collect power consumption feature data within a preset number of time frames, and make edge power distribution decisions. After any edge power terminal completes the edge power distribution decision, a first edge power distribution decision result is sent to the smart meter, and a first decision sending time is obtained; the smart meter continuously monitors and receives the edge power distribution decision result sent by any edge power terminal according to the preset monitoring time, and receives the first edge power distribution decision result when the first decision sending time and the corresponding first signal sending time are less than or equal to the preset monitoring time, and continues to monitor and receive; otherwise, the smart meter The electric meter randomly selects any historical power consumption characteristic data sent and recorded by an edge power user to perform power consumption characteristic data prediction and central power distribution decision, and obtains a first central power distribution decision result; the smart electric meter performs balanced power distribution analysis according to the first edge power distribution decision result or the first central power distribution decision result, obtains a first balanced power distribution analysis result including multiple first power distribution parameters, and distributes power to the multiple edge power users; continues to monitor and obtain a second edge power distribution decision result or process and obtain a second central power distribution decision result according to twice the preset monitoring time, and performs balanced power distribution analysis in combination with the first edge power distribution decision result or the first central power distribution decision result, until a balanced power distribution analysis is performed based on the power distribution decision results of the multiple edge power users, or a preset fluctuation event ends, thereby achieving the technical goal of multiple edge power users sending power distribution demands and decision information in parallel, and achieving the technical effect of improving the real-time performance of power distribution and the overall system performance.

[0009] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented according to the contents of the specification, and in order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are specifically cited below. It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easy to understand through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In order to more clearly illustrate the technical solutions in the present application or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only exemplary, and for ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0011] Figure 1 A flowchart of an edge decision-making power distribution method based on a smart meter is provided for this application; Figure 2 This is a schematic diagram of the process of edge power distribution decision-making in an edge decision-making power distribution method based on a smart meter in the present application. DETAILED DESCRIPTION

[0012] This application provides an edge decision-making power distribution method based on smart meters, which solves the technical problem in the prior art that multiple edge power users send power distribution requirements and decision information in sequence, resulting in significant delays in the system when facing a complex multi-edge environment, further affecting the real-time performance and overall performance of power distribution. The technical goal of multiple edge power users sending power distribution requirements and decision information in parallel is achieved, achieving the technical effect of improving the real-time performance of power distribution and the overall system performance.

[0013] Below, the technical solutions in the present application will be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments of the present application. It should be understood that the present application is not limited to the example embodiments described herein. Based on the embodiments of the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present application. It should also be noted that, for the convenience of description, only the parts related to the present application are shown in the accompanying drawings, rather than all of them.

[0014] Please see attached Figure 1 , the present application provides an edge decision-making power distribution method based on a smart meter, which specifically includes the following steps: Step 1: Monitor the access circuit data through the access monitor in the smart meter. After a preset fluctuation event occurs, send an edge power distribution decision signal to multiple edge power users and obtain multiple signal sending times, wherein the multiple edge power users periodically transmit power consumption characteristic data to the smart meter.

[0015] Specifically, the access monitor in the smart meter can monitor the voltage and current data in the access circuit in real time, that is, the access circuit data. When the access circuit data fluctuates beyond the preset range, the preset fluctuation event is identified and action is taken immediately. At this time, the smart meter will send edge distribution decision signals to multiple edge power users, and then adjust the power consumption according to the current power situation to relieve the pressure on the power grid. Edge power users may include residential areas, industrial facilities, etc. The time when the edge distribution decision signal is sent is recorded for subsequent analysis and optimization. At the same time, the edge power user will regularly feed back the power consumption feature data to the smart meter to ensure continuous monitoring and timely adjustment of power distribution to maintain the stability and efficient operation of the power grid, thereby effectively responding to power fluctuation events, and optimizing the use of power resources to ensure the reliability of power supply. For example, in a distribution station, by monitoring the voltage and current to be distributed, when the voltage and current fluctuate by preset values ​​such as reaching the fluctuation threshold, it means that the current power distribution cannot meet the needs of multiple edge power users, and edge power distribution is required to make decisions to reduce the power consumption of edge power users. For example, edge power users are multiple households or multiple factory workshops. At the same time, send a signal and obtain multiple signal sending times.

[0016] Step 2: The multiple edge power consumption terminals receive the edge power distribution decision signal, continuously collect power consumption characteristic data within a preset number of time frames, and make edge power distribution decisions. After any edge power consumption terminal completes the edge power distribution decision, the first edge power distribution decision result is sent to the smart meter, and the first decision sending time is obtained.

[0017] Specifically, when multiple edge power consumption terminals receive edge power distribution decision signals, they continuously collect power consumption characteristic data within a preset number of time frames to analyze the amplitude at which power can be attenuated and maintain normal operation as an acceptable distribution attenuation coefficient. For example, power consumption characteristic data includes parameters such as voltage and current that can reflect power consumption, such as household power consumption and workshop production power consumption. Each edge power consumption terminal independently makes edge power distribution decisions to ensure the balance between power load and grid supply. When any edge power consumption terminal completes the edge power distribution decision, the distribution attenuation coefficient is used as the edge power distribution decision result. Among them, the first one to complete the edge power distribution decision sends the edge power distribution decision result to the smart meter, ensuring that the power grid can obtain the latest status of each edge power consumption terminal in real time. Then the smart meter will record the sending time of the first edge power distribution decision result so as to analyze and optimize it in subsequent processing, so as to achieve efficient allocation of power resources and stable operation of the power grid. The preset number of time frames refers to the number of pre-set time intervals, which is used to define the duration or frequency of data collection or processing. In the case of edge power consumption, the number of preset time frames represents the collection cycle or time period of power consumption characteristic data, that is, the power consumption characteristic data is collected continuously or periodically within the preset time frame. The setting of the number of preset time frames ensures that the system can obtain enough representative data to support accurate power distribution decisions.

[0018] Step three: The smart meter continuously monitors and receives the edge power distribution decision result sent by any edge power user according to the preset monitoring time. When the first decision sending time and the corresponding first signal sending time are less than or equal to the preset monitoring time, the smart meter receives the first edge power distribution decision result and continues to monitor and receive. Otherwise, the smart meter randomly selects the historical power consumption characteristic data sent and recorded by any edge power user to perform power consumption characteristic data prediction and central power distribution decision, and obtains the first central power distribution decision result.

[0019] Specifically, the smart meter continuously monitors the edge power distribution decision results sent from any edge power user within the preset monitoring time. When the first decision sending time and the corresponding first signal sending time are within the preset monitoring time, the smart meter will receive the first edge power distribution decision result and continue to monitor to ensure the timeliness and accuracy of real-time data. However, if the first edge power distribution decision result is not received within the preset time, the smart meter will randomly select and use the recorded historical power consumption feature data from any edge power user to predict the power consumption feature data, and make a central power distribution decision based on the selection result, thereby generating the first central power distribution decision result to ensure the continuity and stability of power distribution. Among them, the computing power of the edge power user is weak, the communication may be unstable, and the time from receiving the signal to completing the decision processing is long, but the smart meter needs to quickly perform power distribution processing, so a preset monitoring time is set for detection. When the first decision sending time and the corresponding first signal sending time are less than or equal to the preset monitoring time, it is considered that the edge decision processing time meets the requirements, and subsequent power distribution is performed according to the first edge power distribution decision result sent. If it is greater than the preset monitoring time, it is considered untimely. The smart meter randomly selects historical power consumption characteristic data sent to the smart meter and recorded within the historical time at an edge power user to perform power consumption characteristic data prediction and central power distribution decision-making, thereby improving the power distribution response speed and avoiding unreasonable power distribution at some edge power users and stopping work due to insufficient power.

[0020] Step 4: The smart meter performs a balanced power distribution analysis based on the first edge power distribution decision result or the first central power distribution decision result, obtains a first balanced power distribution analysis result including multiple first power distribution parameters, and distributes power to the multiple edge power consumption terminals.

[0021] Specifically, after obtaining the distribution attenuation coefficient in the first edge distribution decision result or the first central distribution decision result, the smart meter performs a balanced distribution analysis, that is, according to the current power demand and supply situation, it determines the reasonable allocation of power resources among multiple edge power terminals, and generates multiple first distribution parameters, which reflect the amount of power that should be allocated to each power terminal under the current situation. For example, based on the generative adversarial network, multiple distribution attenuation coefficients of multiple edge power terminals are generated, the distribution of distribution decision results is obtained, and then power distribution is performed. Then, the actual distribution operation is performed according to the first distribution parameter to ensure that each edge power terminal can obtain the required power while maintaining the stability and efficiency of the entire power grid, so that power resources can be optimally configured, avoiding power overload or waste of resources, and ensuring reliable and continuous power supply.

[0022] Step 5: Continue to monitor and obtain the second edge power distribution decision result or process and obtain the second central power distribution decision result at twice the preset monitoring time, and perform balanced power distribution analysis in combination with the first edge power distribution decision result or the first central power distribution decision result, until a balanced power distribution analysis is performed based on the power distribution decision results of the multiple edge power consumption ends, or the preset fluctuation event ends.

[0023] Specifically, after the power distribution analysis is completed for the first time, monitoring is continued at twice the preset monitoring time to obtain the second edge power distribution decision result or process to obtain the second central power distribution decision result. The initial power distribution analysis is based on multiple distribution attenuation coefficients generated by generative adversarial generation, which may not be accurately adapted enough. The edge power distribution decision result sent by the second completed edge power user is monitored during the next preset monitoring time. By combining the newly obtained decision result with the previous first edge or central power distribution decision result, a balanced power distribution analysis is performed again. The strategy of gradually extending the monitoring time allows for flexible adjustment of the power distribution plan in a continuously changing power consumption environment, thereby more accurately responding to actual power demand. If no information is received, any edge power terminal other than the first edge power terminal is randomly selected, and a central power distribution decision is made based on the recorded historical data. Based on the distribution attenuation coefficients in the two distribution decision results, an adversarial network is generated, and the steps are repeated until the final balanced distribution analysis can be completed based on the distribution decision results of multiple edge power terminals. At this time, the accuracy and adaptability are the highest, or until the preset fluctuation event ends, ensuring that power resources are always reasonably allocated during the entire fluctuation period, avoiding power supply interruptions or overloads, and ensuring the stable operation of the power grid and the power safety of users.

[0024] The described edge decision-making power distribution method based on smart meters can achieve the technical goal of sending power distribution demands and decision information in parallel at multiple edge power users, thereby achieving the technical effect of improving the real-time performance of power distribution and the overall system performance.

[0025] Furthermore, the present application also includes: Access circuit data is collected through an access monitor in the smart meter, wherein the access circuit data includes access circuit voltage and access circuit current; it is determined whether the access circuit data meets a fluctuation event threshold, if so, a preset fluctuation event occurs, and if not, a preset fluctuation event does not occur; after a preset fluctuation event occurs, an edge power distribution decision signal is sent to multiple edge power consumption terminals, and multiple signal sending times of successful sending are obtained, wherein the multiple edge power consumption terminals periodically transmit power consumption characteristic data to the smart meter.

[0026] Specifically, the access circuit data is collected through the access monitor in the smart meter to ensure real-time monitoring and data collection of the power grid. The access monitor in the smart meter can continuously record and analyze the access circuit data to form a dynamic database. For example, the access monitor is a sensor module built into the smart meter, which is used to monitor and collect key data in the access circuit, such as voltage and current. Among them, the access circuit data mainly includes the access circuit voltage and the access circuit current, and the access circuit voltage and the access circuit current reflect the working status of the circuit. In addition, changes in voltage and current can provide information about power demand, power supply stability, equipment health, etc. Therefore, real-time monitoring of the access circuit voltage and access circuit current can foresee potential failure risks.

[0027] Next, the access circuit data is analyzed to determine whether the access circuit data meets the fluctuation event threshold. The fluctuation event threshold is a preset standard used to determine whether the fluctuation of the access circuit voltage and the access circuit current exceeds the normal range. The fluctuation event threshold is set based on long-term historical data analysis and experience in power grid operation to ensure that the fluctuation event threshold can effectively identify abnormal conditions that may threaten the stability of the power grid. If the access circuit data does not meet the fluctuation event threshold, it means that the preset fluctuation event has not occurred, and normal monitoring continues; but if the access circuit data meets the fluctuation event threshold, it means that a fluctuation event has occurred, and the response measures are triggered to ensure the continuity and stability of the power supply and avoid equipment damage or power outages caused by abnormal fluctuations in voltage or current.

[0028] Then, after a preset fluctuation event occurs, an edge power distribution decision signal is immediately sent to multiple edge power users, thereby adjusting the power distribution in time to cope with sudden load changes in the power grid. Among them, the edge power distribution decision signal is a signal used to quickly respond to and adjust the power distribution strategy when abnormal fluctuations occur in the power grid. Edge power distribution refers to the power distribution management for power-consuming equipment or users located at the end or edge of the power grid, such as edge power users including residential areas, small industrial parks, commercial buildings, etc. The edge power distribution decision signal is a control signal issued when a smart meter or central control system detects potential power supply problems or load imbalances in the power grid, which is used to guide the edge power user to adjust the power consumption behavior, thereby maintaining the stability and reliability of the power grid. When a preset fluctuation event occurs, the edge power user adjusts the power demand according to the edge power distribution decision signal to reduce the pressure on the power grid, thereby effectively alleviating the load on the power grid without affecting the overall power supply quality, and avoiding insufficient power supply or equipment failure caused by excessive power consumption.

[0029] Finally, the edge power user periodically transmits the power consumption characteristic data to the smart meter to continuously monitor and update the operation status of the power grid. The power consumption characteristic data includes information such as the real-time power consumption, power consumption mode, and equipment operation status of each edge power user. By regularly transmitting power consumption characteristic data, maintaining a comprehensive understanding of the entire power grid and updating the power grid operation model in real time, it can not only promptly discover and solve potential power supply problems, but also make more accurate power distribution decisions based on real-time data.

[0030] The access circuit data is collected through the access monitor in the smart meter, and the access circuit data is judged to identify the preset fluctuation events. Then, when the fluctuation event occurs, the edge distribution decision signal is sent to multiple edge power consumption ends, and the sending time of the edge distribution decision signal is obtained. Finally, the entire process of transmitting the power consumption characteristic data to the smart meter through the edge power consumption end is completed at a regular interval. Through real-time monitoring, data analysis and decision feedback, the efficient use of electricity and the stable operation of the power grid are ensured, and reliable power services are provided to users.

[0031] Further, if Figure 2 As shown, this application also includes: After receiving the edge power distribution decision signal, the multiple edge power consumption terminals continuously collect power consumption characteristic data within a preset number of time frames to obtain multiple edge power consumption characteristic data; within the multiple edge power consumption terminals, power distribution attenuation coefficient decisions are made according to the multiple edge power consumption characteristic data respectively; after any edge power consumption terminal completes the power distribution attenuation coefficient decision, a first power distribution attenuation coefficient of the first edge power consumption terminal is obtained as a first edge power distribution decision result and sent to the smart meter, and the first decision sending time of successful sending is obtained.

[0032] Specifically, after receiving the edge power distribution decision signal, multiple edge power users will immediately start a process, that is, continuously collecting power consumption characteristic data within a preset time frame, and quickly understand the power consumption of each edge power user when a preset fluctuation event occurs in the power grid, so as to allocate power resources more accurately. Power consumption characteristic data includes multiple parameters such as voltage, current, power factor, etc., which reflect the real-time load, demand changes, and power quality of the power user. By collecting data from multiple time frames within a preset time, a more comprehensive and accurate set of power consumption characteristic data can be obtained, providing a reliable basis for subsequent power distribution decisions.

[0033] Next, at each edge power consumer, a decision on the distribution attenuation coefficient is made based on the corresponding power consumption characteristic data. The distribution attenuation coefficient is a parameter that reflects the degree of match between the current power consumption status of the power consumer and the power supply capacity of the power grid. By calculating the distribution attenuation coefficient, it is determined whether it is necessary to reduce or adjust the power supply to the edge power consumer under the current circumstances, and then accurately match the power supply capacity of the power grid with the power demand of the user, thereby avoiding power waste or insufficient power supply due to imbalance between supply and demand. By making independent distribution attenuation coefficient decisions based on its own edge power consumption characteristic data, each edge power consumer improves the response speed and flexibility of the entire system.

[0034] Next, after any edge power user completes the decision on the distribution attenuation coefficient, a specific distribution attenuation coefficient is generated at the corresponding edge power user as the distribution decision result, that is, the first distribution attenuation coefficient of the first edge power user, as the first edge distribution decision result. The first edge distribution decision result is sent to the smart meter, which can obtain the latest decision information of each edge power user in real time, ensuring that the overall regulation of the power grid can timely reflect the status changes of each edge power user. In addition, the time of sending the first edge distribution decision result will also be recorded, that is, the first decision sending time will be used as important data for subsequent analysis and optimization.

[0035] Through distributed decision-making and real-time feedback, the power system can respond quickly to fluctuations and achieve the optimal allocation of power resources, which not only improves the stability of the power grid, but also ensures the reliability and efficiency of users' electricity use.

[0036] Furthermore, the present application also includes: In the multiple edge power consumption terminals, based on the power consumption characteristic data in the historical time, multiple sample power consumption characteristic data sets are collected according to the preset time frame number, and based on the proportion of power distribution attenuation that can be performed under each sample power consumption characteristic data, multiple sample power distribution attenuation coefficient sets are labeled to obtain; using the multiple sample power consumption characteristic data sets and the multiple sample power distribution attenuation coefficient sets, multiple edge power distribution attenuation decision makers are trained respectively, and configured in the multiple edge power consumption terminals; based on the multiple edge power distribution attenuation decision makers, power distribution attenuation coefficient decisions are made for the multiple edge power consumption characteristic data.

[0037] Specifically, in multiple edge power consumption terminals, based on the power consumption characteristic data in the historical time, multiple sample power consumption characteristic data sets are collected according to the preset number of time frames. By analyzing the power consumption characteristic data in different time periods, common power consumption patterns and abnormal situations are identified to obtain a sample power consumption characteristic data set. The sample power consumption characteristic data set includes not only daily power consumption data, but also covers the changes in power consumption under different environmental conditions, such as seasonal fluctuations or power consumption peaks caused by special events. By collecting sample power consumption characteristic data sets according to the preset number of time frames, the power consumption characteristics can be analyzed more accurately to ensure that the collected data can truly reflect the power consumption status of each edge power consumption terminal under different circumstances, providing strong support for subsequent decision-making.

[0038] Next, according to the proportion of power distribution attenuation that can be performed under each sample power consumption characteristic data, multiple sets of sample power distribution attenuation coefficients were labeled and obtained, and the power consumption characteristic data in the historical time period was converted into reference coefficients that can be used for actual decision-making. Power distribution attenuation reflects the degree of power regulation of the system under a specific power consumption state, and thus more clearly understands the effective adjustment of power distribution to maintain the stability of the power grid and the efficiency of power consumption under different power consumption conditions.

[0039] Subsequently, multiple sample power consumption feature data sets and multiple sample power distribution attenuation coefficient sets are used to train multiple edge power distribution attenuation decision makers respectively. Through machine learning, different power consumption modes can be automatically identified and adapted, so that power distribution decisions can be made more flexibly and efficiently in actual operation. The edge power distribution attenuation decision maker at each edge power consumption end continuously optimizes the decision model in the edge power distribution attenuation decision maker during the training process, so that the edge power distribution attenuation decision maker can respond quickly and make reasonable power adjustments under different power consumption environments, and thus can cope with various complex power consumption situations, whether it is a sudden load increase or a long-term change in power consumption, it can ensure the reasonable allocation of power resources. Then, multiple edge power distribution attenuation decision makers are configured in each edge power consumption end.

[0040] Next, based on multiple edge distribution attenuation decision makers, distribution attenuation coefficient decisions are made for multiple edge power consumption feature data. This means that in actual operation, when new power consumption feature data is detected, the edge distribution attenuation decision maker calculates the distribution attenuation coefficient that best suits the current situation, and adjusts power consumption based on the distribution attenuation coefficient. This not only improves the resilience of the power grid, but also ensures that power supply can meet actual demand under any circumstances, avoiding resource waste or insufficient power supply due to improper power allocation, and achieving optimal allocation of power resources and efficient operation of the power grid.

[0041] The combination of machine learning and real-time decision-making ensures that the power grid can operate efficiently and stably under various complex situations and achieves the rational allocation of power resources.

[0042] Furthermore, the present application also includes: The smart meter continuously monitors and receives the edge power distribution decision result sent by any edge power user according to the preset monitoring time. When the first decision sending time and the corresponding first signal sending time are less than or equal to the preset monitoring time, the first edge power distribution decision result is received within the preset monitoring time; when the first decision sending time and the corresponding first signal sending time are greater than the preset monitoring time, the first edge power distribution decision result is not received within the preset monitoring time. The smart meter randomly selects the historical power consumption characteristic data sent and recorded by any edge power user, obtains the first historical power consumption characteristic data, performs power consumption characteristic data prediction and central power distribution decision, and obtains the first central power distribution decision result.

[0043] Specifically, the smart meter continuously monitors and receives the edge power distribution decision results sent from any edge power user at the preset monitoring time. The preset monitoring time refers to the time to wait for the edge power user to transmit the decision data within the set time window, ensuring timely acquisition of the latest status information of the edge end and providing data support for the stable operation of the power grid. Through continuous monitoring, the smart meter can dynamically update the distribution strategy of power supply to ensure rapid response when the power grid fluctuates or the power demand suddenly changes.

[0044] Then, when the first decision sending time and the corresponding first signal sending time are less than or equal to the preset monitoring time, the smart meter will receive the first edge power distribution decision result within the preset monitoring time, that is, if the edge power consumption end can respond and send its decision result in time within the preset monitoring time, the smart meter can obtain and process the decision result within the preset monitoring time, thereby ensuring the timeliness of the decision result and providing a basis for subsequent power distribution.

[0045] However, when the first decision sending time and the corresponding first signal sending time are greater than the preset monitoring time, the smart meter fails to receive the first edge power distribution decision result within the preset monitoring time, and thus cannot rely on real-time data to make decisions. Therefore, the smart meter will randomly select any historical power consumption characteristic data sent and recorded by the edge power consumption end, and obtain the first historical power consumption characteristic data. Even if the real-time data is missing, it can still be judged and predicted through the historical power consumption characteristic data to ensure that power distribution is not seriously affected by data delays. By analyzing the historical power consumption characteristic data, current demand can be predicted based on past power consumption patterns, thereby making reasonable central power distribution decisions.

[0046] Finally, the smart meter predicts the power consumption characteristic data based on the selected first historical power consumption characteristic data, and makes a central power distribution decision based on the power consumption characteristic data prediction, and obtains the first central power distribution decision result. Then, even in the absence of real-time data, a relatively reasonable power distribution decision can be made through the prediction algorithm and historical data. The result of the first central power distribution decision will serve as the basis for the current power distribution, ensuring that the stability and efficiency of the power grid can be maintained in the case of data delay or loss, and avoiding improper power distribution due to data loss.

[0047] By ensuring that the power grid can operate efficiently under all circumstances, the timely satisfaction of electricity demand and the rational allocation of power resources are guaranteed.

[0048] Furthermore, the present application also includes: According to the power consumption characteristic data of the multiple edge power consumption terminals recorded in the smart meter, multiple sample historical power consumption characteristic data sets are collected, and the power consumption characteristic data after each sample historical power consumption characteristic data is obtained as multiple sample predicted power consumption characteristic data sets; the multiple sample historical power consumption characteristic data sets and the multiple sample predicted power consumption characteristic data sets are used to train multiple power consumption characteristic data prediction branches to obtain a power consumption characteristic data predictor, which is configured in the smart meter; the first historical power consumption characteristic data is input into the power consumption characteristic data prediction branch corresponding to the edge power consumption terminal in the power consumption characteristic data predictor to predict and obtain first predicted power consumption characteristic data; according to the power consumption characteristic data of the multiple edge power consumption terminals recorded in the smart meter, the distribution attenuation coefficient is marked, multiple central distribution attenuation decision makers are trained and configured in the smart meter, and central distribution prediction is performed on the first predicted power consumption characteristic data to obtain a first central distribution decision result.

[0049] Specifically, based on the power consumption characteristic data of multiple edge power consumption terminals recorded in the smart meter, multiple sample historical power consumption characteristic data sets are collected, and the power consumption characteristic data after each sample historical power consumption characteristic data is further obtained to form multiple sample predicted power consumption characteristic data sets, that is, by analyzing historical data and corresponding future data, a sample predicted power consumption characteristic data set that can reflect the power consumption trend is established. The sample predicted power consumption characteristic data set contains the changes in power consumption in different time periods and different power consumption environments, and can provide a wide range of reference information to help understand and predict future power consumption patterns, thereby further expanding and optimizing the prediction capabilities and providing more accurate support for subsequent decision-making processes.

[0050] Next, multiple sample historical electricity consumption feature data sets and corresponding multiple sample predicted electricity consumption feature data sets are used to train the corresponding multiple electricity consumption feature data prediction branches, and finally multiple electricity consumption feature data prediction branches are integrated to obtain a complete electricity consumption feature data predictor, and the electricity consumption feature data predictor is configured in the smart meter. Training multiple electricity consumption feature data prediction branches can grasp the changing rules of electricity consumption characteristics by learning the relationship between historical data and future data. When future electricity consumption occurs, accurate predictions are made through the electricity consumption feature data predictor to ensure that the electricity consumption feature data predictor can provide high-precision prediction results under different edge electricity consumption terminals. After the electricity consumption feature data predictor is embedded in the smart meter, electricity consumption predictions can be made at any time according to actual conditions.

[0051] Then, when a prediction is needed, the first historical electricity consumption characteristic data is input into the electricity consumption characteristic data predictor, and specifically enters the corresponding edge electricity consumption characteristic data prediction branch, so as to obtain the first predicted electricity consumption characteristic data, which can judge the future electricity demand in advance and make control decisions in advance when necessary, so as to avoid power supply imbalance caused by sudden increase or decrease in electricity consumption.

[0052] Finally, according to the power consumption characteristic data of multiple edge power consumption terminals recorded in the smart meter, the distribution attenuation coefficient is marked, and multiple central distribution attenuation decision makers are trained using the marked distribution attenuation coefficient, and finally multiple central distribution attenuation decision makers are configured in the smart meter. Multiple central distribution attenuation decision makers are used to perform central distribution prediction on the first predicted power consumption characteristic data, so as to obtain the first central distribution decision result, ensuring that power resources can be reasonably allocated according to the predicted demand. By marking and training multiple central distribution attenuation decision makers, it is possible to effectively cope with complex and changeable power consumption environments and ensure the efficient operation of the power grid and the stability of power supply.

[0053] Through the learning of historical data and real-time prediction, it ensures that the power grid can respond flexibly to various power usage scenarios and achieve optimal power resource management.

[0054] Furthermore, the present application also includes: According to the edge power distribution decision data of the multiple edge power terminals when a preset fluctuation event occurs within a historical time, a sample first power distribution decision result set of the first edge power terminal and a sample power distribution decision result distribution set are collected, wherein each sample power distribution decision result distribution includes the power distribution decision results of the multiple edge power terminals; the sample first power distribution decision result set and the sample power distribution decision result distribution set are used to train a first power distribution decision distribution restorer; the first power distribution decision distribution restorer is used to perform power distribution decision distribution restoration on the first edge power distribution decision result or the first central power distribution decision result to obtain a power distribution decision result distribution, wherein the power distribution decision result distribution includes the power distribution decision results of multiple edge power terminals, wherein each power distribution decision result includes a distribution attenuation coefficient; according to the smart meter, the current access circuit data is collected and obtained, and combined with the standard access circuit data, a calculation is performed. The access attenuation is obtained by calculating, and the multiple maximum distribution attenuations of the multiple edge power terminals are obtained according to the multiple distribution attenuation coefficients and the rated distribution data of the multiple edge power terminals; when the sum of the multiple maximum distribution attenuations is greater than or equal to the access attenuation, balanced distribution is performed according to the multiple distribution attenuation coefficients to obtain a first balanced distribution analysis result including multiple first distribution parameters, wherein the ratio of the actual distribution attenuation coefficient in each first distribution parameter is the same as the distribution attenuation coefficient, and the actual distribution attenuation coefficient is greater than the corresponding distribution attenuation coefficient; when the sum of the multiple distribution attenuations is less than the access attenuation, the edge power terminals with small distribution attenuation coefficients are preferentially distributed according to the multiple distribution attenuation coefficients to obtain a first balanced distribution analysis result including multiple first distribution parameters, wherein the actual distribution attenuation coefficient in each first distribution parameter is greater than or equal to the corresponding distribution attenuation coefficient.

[0055] Specifically, when a preset fluctuation event occurs within a historical period, edge power distribution decision data of multiple edge power terminals are collected. The edge power distribution decision data reflects the changes in the power demand of each edge power terminal in response to the preset fluctuation event. Among them, the sample first distribution decision result set of the first edge power terminal extracted from the edge power distribution decision data contains all the distribution decision results made by the edge power terminal when the preset fluctuation event occurs at different historical times, providing a reference for subsequent analysis. At the same time, a sample distribution decision result distribution set is collected, and the sample distribution decision result distribution set includes the distribution decision results of multiple edge power terminals, that is, each sample distribution decision result distribution is the overall performance of the distribution decision of multiple edge power terminals under the preset fluctuation event, and thus can identify the behavior patterns of different edge power terminals in response to the preset fluctuation event, providing a rich historical basis for the subsequent optimization of the distribution strategy, which is helpful to improve the system's responsiveness and decision-making accuracy in similar situations in the future.

[0056] Next, the first power distribution decision distribution restorer is trained using the collected sample first power distribution decision result set and the sample power distribution decision result distribution set. The function of the first power distribution decision distribution restorer is to reconstruct and simulate these decision distributions in practical applications by learning the distribution patterns of the first power distribution decision result set and the sample power distribution decision result distribution set. By learning from historical decision data, it is possible to infer how the power distribution decisions between different edge power users are distributed under similar conditions, so that the first power distribution decision distribution restorer can quickly and accurately predict the possible future power distribution decision distribution when faced with new data, thereby providing support for making more accurate power distribution decisions.

[0057] Then, the first power distribution decision distribution restorer performs power distribution decision distribution restoration on the first edge power distribution decision result or the first central power distribution decision result, and can perform more detailed analysis and simulation on the power distribution decision result of each edge power consumption end to ensure that the power distribution can accurately reflect the actual demand and power supply capacity of each edge power consumption end, thereby obtaining the distribution of power distribution decision results. The distribution of power distribution decision results includes the power distribution decision results of multiple edge power consumption ends, and each power distribution decision result contains the corresponding power distribution attenuation coefficient, which can better manage power resources and ensure that power distribution can remain balanced and stable under any fluctuation conditions.

[0058] Next, based on the current access circuit data collected by the smart meter and combined with the standard access circuit data, the access attenuation is calculated to evaluate the attenuation impact of the current circuit in actual operation. Then, using multiple distribution attenuation coefficients and the rated distribution data of multiple edge power terminals, multiple maximum distribution attenuations of the edge power terminals are calculated. That is, by comparing the actual data with the standard data, the most reasonable power distribution method for each edge power terminal under the current environment is determined to ensure that power resources can be effectively utilized, reduce waste and maintain the stability of the power grid.

[0059] Then, when the sum of multiple maximum distribution attenuations is greater than or equal to the access attenuation, balanced distribution is performed according to multiple distribution attenuation coefficients to obtain a first balanced distribution analysis result including multiple first distribution parameters. By comparing the ratio of the actual distribution attenuation coefficient and the distribution attenuation coefficient, that is, the ratio of the actual distribution attenuation coefficient in each edge power consumption terminal to the distribution attenuation coefficient generated by the decision is the same, for example, both are 0.5, for example, the distribution attenuation coefficients generated by the decision of the A and B edge power consumption terminals are 20% and 30%, and the actual distribution attenuation coefficients are 10% and 15%. Ensuring that the actual distribution attenuation coefficient is always greater than the corresponding distribution attenuation coefficient helps to reasonably allocate electricity in the case of abundant resources, ensure that each edge power consumption terminal can obtain sufficient power supply, and maintain the overall balance of the system. For example, the total ratio of the sum of multiple maximum distribution attenuations to the access attenuation is calculated, and then the total ratio is multiplied by the maximum distribution attenuation of each edge power consumption terminal to obtain the actual distribution attenuation, and divided by the rated distribution data of each edge power consumption terminal to obtain the actual distribution attenuation coefficient.

[0060] However, if the sum of multiple distribution attenuations is less than the access attenuation, it means that the distribution of some edge distribution terminals may not be met. Then, according to multiple distribution attenuation coefficients, priority is given to edge power consumption terminals with smaller distribution attenuation coefficients, and power is directly distributed according to the distribution attenuation coefficients to obtain a first balanced distribution analysis result including multiple first distribution parameters. For example, edge power consumption terminals with smaller distribution attenuation coefficients are edge power consumption terminals with tight production or power demand. Then, in the case of relatively limited power resources, priority is given to powering edge power consumption terminals with higher or more critical power demand, thereby ensuring the effective use of power resources and avoiding power shortages caused by uneven resource allocation. The actual distribution attenuation coefficient within each first distribution parameter is greater than or equal to the corresponding distribution attenuation coefficient, thereby ensuring that power distribution can still meet key needs when resources are tight.

[0061] By calculating the attenuation and the maximum distribution attenuation, the power distribution strategy can be flexibly adjusted to ensure that the power grid can maintain balanced and stable operation regardless of whether resources are abundant or scarce, ultimately achieving efficient power resource management.

[0062] Furthermore, the present application also includes: Based on the generative adversarial network, the first power distribution decision distribution restorer is constructed, and the first power distribution decision distribution restorer includes a generation unit and an adversarial unit; the sample first power distribution decision result set and the sample power distribution decision result distribution set are used to perform supervised training on the generation unit and the adversarial unit until the training converges, so as to obtain the trained first power distribution decision distribution restorer.

[0063] Specifically, the generative adversarial network is a machine learning framework. Based on the generative adversarial network, a model of the first distribution decision distribution reducer is constructed, including a generation unit and an adversarial unit. The task of the generation unit is to generate a distribution decision distribution similar to the real data, while the adversarial unit attempts to distinguish the generated data from the real data. Through adversarial training, the generation unit will continuously improve its output and eventually produce results that are very close to the real data. In the first distribution decision distribution reducer, the collaborative work of the generation unit and the adversarial unit enables the system to simulate and reconstruct complex distribution decision distributions, thereby providing support for subsequent power distribution.

[0064] Next, the generation unit and the adversarial unit are supervised and trained using the sample first distribution decision result set and the sample distribution decision result distribution set. The generation unit and the adversarial unit can continuously learn and improve. The generation unit attempts to generate a new distribution decision distribution, while the adversarial unit evaluates the authenticity of the new distribution decision distribution. In the continuous iteration, the output of the generation unit is closer and closer to the real distribution decision distribution. As the training progresses, the confrontation between the generation unit and the adversarial unit gradually tends to balance, and the training process finally reaches convergence. Training convergence means that the generation unit can generate data that is highly similar to the real distribution decision distribution, while the adversarial unit can no longer distinguish between the generated data and the real data. At this time, the first distribution decision distribution restorer that has been trained has the ability to restore and simulate the distribution of distribution decisions in practical applications. When facing new data, the first distribution decision distribution restorer can generate a distribution decision distribution that conforms to the actual situation based on the patterns learned during the training process, providing accurate prediction and support for the dispatch and allocation of the power system.

[0065] By fully training the model, it is possible to provide powerful forecasting and allocation support for the power system, ensuring efficient use of power resources and stable operation of the system.

[0066] In summary, the edge decision-making power distribution method based on smart meters provided by this application has the following technical effects: The access circuit data is monitored by the access monitor in the smart meter, and after a preset fluctuation event occurs, an edge power distribution decision signal is sent to multiple edge power terminals, and multiple signal sending times are obtained, wherein the multiple edge power terminals periodically transmit power consumption feature data to the smart meter; the multiple edge power terminals receive the edge power distribution decision signal, continuously collect power consumption feature data within a preset number of time frames, and make edge power distribution decisions. After any edge power terminal completes the edge power distribution decision, a first edge power distribution decision result is sent to the smart meter, and a first decision sending time is obtained; the smart meter continuously monitors and receives the edge power distribution decision result sent by any edge power terminal according to the preset monitoring time, and receives the first edge power distribution decision result when the first decision sending time and the corresponding first signal sending time are less than or equal to the preset monitoring time, and continues to monitor and receive; otherwise, the smart meter The electric meter randomly selects any historical power consumption characteristic data sent and recorded by an edge power user to perform power consumption characteristic data prediction and central power distribution decision, and obtains a first central power distribution decision result; the smart electric meter performs balanced power distribution analysis according to the first edge power distribution decision result or the first central power distribution decision result, obtains a first balanced power distribution analysis result including multiple first power distribution parameters, and distributes power to the multiple edge power users; continues to monitor and obtain a second edge power distribution decision result or process and obtain a second central power distribution decision result according to twice the preset monitoring time, and performs balanced power distribution analysis in combination with the first edge power distribution decision result or the first central power distribution decision result, until a balanced power distribution analysis is performed based on the power distribution decision results of the multiple edge power users, or a preset fluctuation event ends, thereby achieving the technical goal of multiple edge power users sending power distribution demands and decision information in parallel, and achieving the technical effect of improving the real-time performance of power distribution and the overall system performance.

[0067] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

[0068] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application belong to the scope of the present application and its equivalent technology, the present application is also intended to include these modifications and variations.

Claims

1. An edge decision-making power distribution method based on smart meters, characterized in that: The method comprises: The access circuit data is monitored by an access monitor in the smart meter. After a preset fluctuation event occurs, an edge power distribution decision signal is sent to multiple edge power consumption terminals, and multiple signal sending times are obtained, wherein the multiple edge power consumption terminals transmit power consumption characteristic data to the smart meter at regular intervals. The multiple edge power consumption terminals receive the edge power distribution decision signal, continuously collect power consumption feature data within a preset number of time frames, make edge power distribution decisions, and after any edge power consumption terminal completes the edge power distribution decision, send a first edge power distribution decision result to the smart meter, and obtain a first decision sending time; The smart meter continuously monitors and receives the edge power distribution decision result sent by any edge power consumption terminal according to the preset monitoring time, and receives the first edge power distribution decision result when the first decision sending time and the corresponding first signal sending time are less than or equal to the preset monitoring time, and continues to monitor and receive, otherwise, the smart meter randomly selects any historical power consumption feature data sent and recorded by any edge power consumption terminal to perform power consumption feature data prediction and central power distribution decision, and obtains the first central power distribution decision result; The smart electric meter performs a balanced power distribution analysis according to the first edge power distribution decision result or the first central power distribution decision result, obtains a first balanced power distribution analysis result including a plurality of first power distribution parameters, and distributes power to the plurality of edge power consumption terminals; Continue to monitor and obtain the second edge power distribution decision result or process and obtain the second central power distribution decision result at twice the preset monitoring time, and perform balanced power distribution analysis in combination with the first edge power distribution decision result or the first central power distribution decision result, until a balanced power distribution analysis is performed based on the power distribution decision results of the multiple edge power consumption ends, or the preset fluctuation event ends.

2. The edge decision-making power distribution method based on smart meters according to claim 1 is characterized in that: The access monitor in the smart meter monitors the access circuit data. After a preset fluctuation event occurs, an edge power distribution decision signal is sent to multiple edge power consumption terminals, and multiple signal sending times are obtained, including: Collecting access circuit data through an access monitor in the smart meter, wherein the access circuit data includes an access circuit voltage and an access circuit current; Determine whether the access circuit data meets the fluctuation event threshold, if yes, a preset fluctuation event occurs, if no, a preset fluctuation event does not occur; After a preset fluctuation event occurs, an edge power distribution decision signal is sent to multiple edge power consumption terminals, and multiple signal sending times of successful sending are obtained, wherein the multiple edge power consumption terminals periodically transmit power consumption characteristic data to the smart meter.

3. The edge decision-making power distribution method based on smart meters according to claim 1 is characterized in that: The plurality of edge power consumption terminals receive the edge power distribution decision signal, continuously collect power consumption characteristic data within a preset number of time frames, and make edge power distribution decisions, including: After receiving the edge power distribution decision signal, the plurality of edge power consumption terminals continuously collect power consumption characteristic data within a preset number of time frames to obtain a plurality of edge power consumption characteristic data; In the plurality of edge power consumption terminals, respectively, a power distribution attenuation coefficient decision is made according to the plurality of edge power consumption characteristic data; After any edge power user completes the power distribution attenuation coefficient decision, the first power distribution attenuation coefficient of the first edge power user is obtained, which is sent to the smart meter as the first edge power distribution decision result, and the first decision sending time of successful sending is obtained.

4. The edge decision-making power distribution method based on smart meters according to claim 3 is characterized in that: In the plurality of edge power consumption terminals, respectively, according to the plurality of edge power consumption characteristic data, a power distribution attenuation coefficient decision is made, including: In the plurality of edge power consumption terminals, according to the power consumption characteristic data in the historical time, a plurality of sample power consumption characteristic data sets are collected according to the preset time frame number, and according to the proportion of power distribution attenuation that can be performed under each sample power consumption characteristic data, a plurality of sample power distribution attenuation coefficient sets are obtained by marking; Using the plurality of sample power consumption feature data sets and the plurality of sample power distribution attenuation coefficient sets, respectively training a plurality of edge power distribution attenuation decision makers, which are configured in the plurality of edge power consumption terminals; Based on the multiple edge power distribution attenuation decision makers, power distribution attenuation coefficient decisions are made on the multiple edge power consumption feature data.

5. The edge decision-making power distribution method based on smart meters according to claim 1 is characterized in that: include: The smart meter continuously monitors and receives the edge power distribution decision result sent by any edge power consumption terminal according to the preset monitoring time, and receives the first edge power distribution decision result within the preset monitoring time when the first decision sending time and the corresponding first signal sending time are less than or equal to the preset monitoring time; When the first decision sending time and the corresponding first signal sending time are greater than the preset monitoring time, and the first edge power distribution decision result is not received within the preset monitoring time, the smart meter randomly selects any historical power consumption characteristic data sent and recorded by the edge power consumption terminal, obtains the first historical power consumption characteristic data, performs power consumption characteristic data prediction and central power distribution decision, and obtains the first central power distribution decision result.

6. The edge decision-making power distribution method based on smart meters according to claim 5 is characterized in that: Conduct power consumption characteristic data prediction and central power distribution decision-making to obtain the first central power distribution decision result, including: According to the power consumption characteristic data of the plurality of edge power consumption terminals recorded in the smart meter, a plurality of sample historical power consumption characteristic data sets are collected, and the power consumption characteristic data after each sample historical power consumption characteristic data is obtained as a plurality of sample predicted power consumption characteristic data sets; Using the plurality of sample historical power consumption feature data sets and the plurality of sample predicted power consumption feature data sets, training a plurality of power consumption feature data prediction branches, obtaining a power consumption feature data predictor, and configuring the predictor in the smart meter; Inputting the first historical power consumption characteristic data into the power consumption characteristic data prediction branch corresponding to the edge power consumption terminal in the power consumption characteristic data predictor to predict and obtain first predicted power consumption characteristic data; According to the power consumption characteristic data of the multiple edge power consumption terminals recorded in the smart meter, the distribution attenuation coefficient is marked, multiple central distribution attenuation decision makers are trained and configured in the smart meter, and central distribution prediction is performed on the first predicted power consumption characteristic data to obtain a first central distribution decision result.

7. The edge decision-making power distribution method based on smart meters according to claim 1 is characterized in that: The smart meter performs balanced power distribution analysis according to the first edge power distribution decision result or the first central power distribution decision result to obtain a first balanced power distribution analysis result including a plurality of first power distribution parameters, including: According to the edge power distribution decision data of the plurality of edge power terminals when a preset fluctuation event occurs within a historical time, a sample first power distribution decision result set of the first edge power terminal and a sample power distribution decision result distribution set are collected, wherein each sample power distribution decision result distribution includes the power distribution decision results of the plurality of edge power terminals; Using the sample first power distribution decision result set and the sample power distribution decision result distribution set to train a first power distribution decision distribution restorer; The first power distribution decision distribution restorer is used to restore the first edge power distribution decision result or the first central power distribution decision result to obtain a power distribution decision result distribution, wherein the power distribution decision result distribution includes power distribution decision results of multiple edge power consumption ends, wherein each power distribution decision result includes a distribution attenuation coefficient; According to the smart meter, current access circuit data is collected and obtained, and access attenuation is calculated in combination with standard access circuit data, and multiple maximum distribution attenuation amounts of the multiple edge power terminals are calculated according to multiple distribution attenuation coefficients and rated distribution data of multiple edge power terminals; When the sum of the multiple maximum power distribution attenuations is greater than or equal to the access attenuation, balanced power distribution is performed according to the multiple power distribution attenuation coefficients to obtain a first balanced power distribution analysis result including multiple first power distribution parameters, wherein the ratio of the actual power distribution attenuation coefficient in each first power distribution parameter is the same as the power distribution attenuation coefficient, and the actual power distribution attenuation coefficient is greater than the corresponding power distribution attenuation coefficient; When the sum of the multiple distribution attenuations is less than the access attenuation, priority is given to edge power consumption terminals with small distribution attenuation coefficients according to the multiple distribution attenuation coefficients to obtain a first balanced distribution analysis result including multiple first distribution parameters, wherein the actual distribution attenuation coefficient within each first distribution parameter is greater than or equal to the corresponding distribution attenuation coefficient.

8. The edge decision-making power distribution method based on smart meters according to claim 7 is characterized in that: The sample first power distribution decision result set and the sample power distribution decision result distribution set are used to train a first power distribution decision distribution restorer, including: Based on the generative adversarial network, construct the first power distribution decision distribution reducer, wherein the first power distribution decision distribution reducer includes a generation unit and an adversarial unit; The sample first power distribution decision result set and the sample power distribution decision result distribution set are used to perform supervised training on the generation unit and the adversarial unit until the training converges, thereby obtaining the trained first power distribution decision distribution restorer.

Images