Multi-dimensional power distribution optimization method based on ammeter big data and edge calculation

Abstract

The invention provides a multi-dimensional power distribution optimization method based on ammeter big data and edge calculation, and relates to the technical field of intelligent power distribution, and the method comprises the steps: receiving user distribution topological information of a transformer area; extracting a first ammeter user type and a first ammeter user scale until an Nth ammeter user type and an Nth ammeter user scale; obtaining the predicted running capacity of the first ammeter in a preset time zone; obtaining the predicted running capacity of the Nth ammeter in a preset time zone; predicting the power consumption of the court to obtain the power consumption of the court in a preset time zone; and according to the preset time zone district power consumption, carrying out district power distribution management. According to the application, the technical problem that the power consumption prediction is not fine enough and the power distribution refinement degree is further influenced due to the lack of individual difference analysis of the power consumption of the transformer area users in the prior art can be solved, and the purposes of improving the power consumption prediction and power distribution refinement degree of the transformer area and improving the power distribution refinement degree are achieved. And the power distribution management efficiency and accuracy of the whole transformer area are improved.

Description

Technical Field

[0001] The present application relates to the field of intelligent power distribution technology, and in particular to a multi-dimensional power distribution optimization method based on electric meter big data and edge computing. Background Art

[0002] In the current power system, efficient management and optimized dispatching of distribution networks are important tasks to ensure safe and reliable power supply. With the development of social economy and the continuous growth of electricity demand, the complexity of distribution networks is increasing. Users of different types and sizes have diverse demands for electricity and there are significant time zone differences. Traditional distribution management methods mainly rely on historical experience and static models, which are difficult to adapt to rapidly changing power load conditions, resulting in low distribution efficiency and large distribution losses. Especially in large-scale substations, due to the complexity of user types and electricity consumption behaviors, it has become increasingly challenging to accurately predict and manage users' electricity demand.

[0003] With the development of smart grid technology, meter big data analysis and edge computing technology have gradually been applied in power distribution management. Through real-time monitoring of user electricity consumption data by smart meters and distributed processing with edge computing technology, the distribution system can greatly improve its ability to perceive and predict user electricity demand. However, the existing meter data analysis methods often start from the top-level data, that is, directly analyzing the overall electricity consumption data of the substation to distribute electricity to the substation. However, each user in the substation has individual differences, and individual changes may cause instability in the overall electricity consumption data analysis of the substation, which in turn affects the degree of refinement of power distribution.

[0004] In summary, the existing technology has a technical problem that the power consumption forecast is not precise enough due to the lack of individualized difference analysis of power consumption of users in the substation area, which further affects the degree of refinement of power distribution. Summary of the invention

[0005] The purpose of this application is to provide a multi-dimensional power distribution optimization method based on electric meter big data and edge computing, so as to solve the technical problem in the prior art that the power consumption forecast is not precise enough due to the lack of individualized difference analysis of the power consumption of users in the substation area, which further affects the degree of refinement of power distribution.

[0006] In view of the above problems, the present application provides a multi-dimensional power distribution optimization method based on electric meter big data and edge computing, including: receiving user distribution topology information of a substation, wherein the user distribution topology information includes a first electric meter distribution node to an Nth electric meter distribution node; traversing the first electric meter distribution node to the Nth electric meter distribution node, extracting a first electric meter user type and a first electric meter user scale to the Nth electric meter user type and the Nth electric meter user scale; performing electric meter big data analysis on the first electric meter user type and the first electric meter user scale through a first edge computing power center, Obtain the predicted power consumption of the first electricity meter in a preset time zone; perform electricity meter big data analysis on the user type and user scale of the N electricity meter through the Nth edge computing center to obtain the predicted power consumption of the Nth electricity meter in the preset time zone, wherein the first edge computing center and the Nth edge computing center operate simultaneously; perform a substation power consumption forecast based on the user distribution topology information according to the predicted power consumption of the first electricity meter in the preset time zone until the predicted power consumption of the N electricity meter in the preset time zone, and obtain the substation power consumption in the preset time zone; perform substation power distribution management according to the substation power consumption in the preset time zone.

[0007] One or more technical solutions provided in this application have at least the following technical effects or advantages: Receive user distribution topology information of a substation, wherein the user distribution topology information includes a first meter distribution node up to an Nth meter distribution node; traverse the first meter distribution node up to the Nth meter distribution node, extract a first meter user type and a first meter user scale, up to an Nth meter user type and an Nth meter user scale; perform meter big data analysis on the first meter user type and the first meter user scale through a first edge computing center, and obtain a predicted power outage amount of the first meter in a preset time zone; perform meter big data analysis on the Nth meter user type and the Nth meter user scale through an Nth edge computing center, and obtain a predicted power outage amount of the Nth meter in a preset time zone, wherein the first edge computing center and the Nth edge computing center operate simultaneously; perform substation power consumption prediction based on the user distribution topology information according to the predicted power outage amount of the first meter in a preset time zone up to the predicted power outage amount of the Nth meter in a preset time zone, and obtain the substation power consumption in a preset time zone; perform power distribution management in the substation according to the substation power consumption in the preset time zone. By receiving the topological information of user distribution in the substation, conducting big data analysis based on user type and scale, and combining the edge computing center to parallelly process the power consumption forecast of different meter distribution nodes, while taking distribution losses into consideration, the forecast results are corrected through reverse calibration, thereby achieving the technical effect of improving the power consumption forecast and distribution refinement of the substation, thereby improving the efficiency and accuracy of distribution management in the entire substation.

[0008] 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

[0009] 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.

[0010] Figure 1 This is a flow chart of a multi-dimensional power distribution optimization method based on electric meter big data and edge computing in this application; Figure 2 This is a flow chart of obtaining the power consumption of a preset time zone in a multi-dimensional power distribution optimization method based on electric meter big data and edge computing in this application. DETAILED DESCRIPTION

[0011] This application solves the technical problem in the prior art that the power consumption forecast is not refined enough due to the lack of individualized difference analysis of the power consumption of users in the substation area, which further affects the refinement of power distribution by providing a multi-dimensional power distribution optimization method based on meter big data and edge computing. By receiving the topological information of the user distribution in the substation area, performing big data analysis based on user type and scale, combining the edge computing center to parallelly process the power consumption forecast of different meter distribution nodes, and considering the power distribution loss at the same time, the prediction results are corrected by reverse calibration, so as to achieve the technical effect of improving the power consumption forecast and the refinement of power distribution in the substation area, thereby improving the efficiency and accuracy of power distribution management in the entire substation area.

[0012] 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.

[0013] Please see attached Figure 1, this application provides a multi-dimensional power distribution optimization method based on electric meter big data and edge computing, which specifically includes the following steps: Step 1: Receive user distribution topology information of the substation, wherein the user distribution topology information includes the first electricity meter distribution node to the Nth electricity meter distribution node.

[0014] Specifically, user distribution topology information refers to the spatial distribution structure of each meter distribution node and its interconnection relationship within the power substation, which covers all node data from the first meter distribution node to the Nth meter distribution node, where N is an integer greater than 1. Meter distribution nodes refer to the locations of meters installed at each user end in the power network, and these nodes constitute the basic network structure of the entire power substation.

[0015] For example, in a small community, there are several residential buildings, each of which is connected to the power grid of the substation. Each household's electric meter represents an electric meter distribution node, from which the first electric meter distribution node to the Nth electric meter distribution node are obtained. All electric meter distribution nodes are interconnected through power lines to form a complete power distribution network, which is the user distribution topology information. By receiving the user distribution topology information, basic data support is provided for subsequent power distribution and optimization operations.

[0016] Step 2: traverse the first electricity meter distribution node until the Nth electricity meter distribution node, extract the first electricity meter user type and the first electricity meter user scale until the Nth electricity meter user type and the Nth electricity meter user scale.

[0017] Specifically, after receiving the user distribution topology information of the substation, each meter distribution node is traversed, that is, from the first meter distribution node to the Nth meter distribution node, and the user type and user scale of each node are extracted one by one. User type refers to the type of user served by the meter, such as residential users, commercial users or industrial users. Different types of users usually have different electricity consumption characteristics. For example, the peak electricity consumption of residential users may be concentrated at night, while the peak of commercial users may be during the day. The user scale represents the number of users, that is, the different number of users will also have different electricity consumption. Thus, the first meter user type and the first meter user scale are obtained, until the Nth meter user type and the Nth meter user scale. Users of different types and scales will have different load impacts on the power system. Therefore, when distributing electricity, these differences need to be comprehensively considered to achieve accurate distribution optimization.

[0018] Step three: Through the first edge computing center, the meter big data analysis is performed on the user type and user scale of the first meter to obtain the predicted running power of the first meter in the preset time zone.

[0019] Specifically, the edge computing center refers to the computing nodes located at the edge of the power network. These nodes have certain data processing capabilities and can perform real-time data analysis and processing at locations close to the data source (such as meter distribution nodes). By performing calculations at the edge, the delay in data transmission can be reduced and the processing efficiency can be improved. First, a big data analysis request is initiated to the first edge computing center based on the user type of the first meter and the user scale of the first meter. For example, suppose that the first meter is connected to a residential household, its user type is a residential user, and the user scale is a single household. The first edge computing center will perform data analysis based on the household's historical electricity consumption data and the electricity consumption patterns of other similar users.

[0020] Specifically, the first edge computing center will collect the user's electricity consumption data in different time periods and divide these data into different time zones, such as peak time zones, flat time zones, and valley time zones. Based on the data in these time zones, it is possible to predict the user's power consumption in a specific period in the future, that is, the expected value of power consumption. For example, assuming that past data shows that the family's peak electricity consumption is between 7 and 10 p.m., and the lowest electricity consumption is between 2 and 5 a.m., then the first edge computing center can be used to predict the family's electricity consumption in the same time period in the future, and the predicted power consumption in the preset time zone of the first meter can be obtained, so as to realize the individualized difference analysis of the power consumption of users in the area, which can provide key references for subsequent power distribution and load management, and more accurately predict the power demand of each meter distribution node, thereby avoiding over-allocation or under-allocation during power dispatching, and improving the refinement of distribution management.

[0021] Step 4: Performing an electricity meter big data analysis on the user type and user scale of the Nth electricity meter through the Nth edge computing center to obtain the predicted running power of the Nth electricity meter in a preset time zone, wherein the first edge computing center and the Nth edge computing center operate simultaneously.

[0022] Specifically, in order to comprehensively and accurately predict the electricity demand of the entire substation, it is necessary not only to analyze the first meter distribution node, but also to perform a similar analysis on the Nth meter distribution node. That is, the Nth edge computing center conducts meter big data analysis on the Nth meter user type and the Nth meter user scale to obtain the predicted running power of the Nth meter in the preset time zone. The Nth edge computing center refers to the computing node responsible for processing the Nth meter distribution node, which is located at the edge of the power network like the first edge computing center. Its main function is to perform real-time analysis and processing of the data of the meter distribution node, reduce data transmission delays, and provide fast and accurate power consumption forecasts.

[0023] Among them, the method of calculating the predicted power consumption of the preset time zone of the Nth electric meter by the Nth edge computing center is exactly the same as the method of calculating and obtaining the predicted power consumption of the preset time zone of the first electric meter by the first edge computing center, which will not be repeated here. The first edge computing center and the Nth edge computing center will run in parallel, that is, the N edge computing centers can process their respective meter distribution node data in parallel, greatly improving the overall operation efficiency. For example, when predicting the power consumption of multiple nodes at the same time, the calculation can be completed quickly and the results can be generated, so as to make an accurate prediction of the power demand of the entire substation in a short time, improving the computing efficiency and overall processing speed.

[0024] Step 5: According to the first electric meter's preset time zone predicted power consumption until the Nth electric meter's preset time zone predicted power consumption, perform a station area power consumption forecast based on the user distribution topology information to obtain the preset time zone station area power consumption.

[0025] Specifically, the predicted power consumption in a preset time zone refers to the prediction of the expected power consumption of each meter distribution node in a specific time zone in the future by analyzing historical data and user behavior patterns. Next, the predicted data of these nodes are combined with the user distribution topology information of the entire substation area, taking into account the characteristics of power transmission in the substation area. For example, some meter distribution nodes may be located at the edge of the substation area, and certain losses will be generated during power transmission, while nodes located in the center of the substation area may be supplied by multiple power lines, and the transmission loss is relatively small. Taking these factors into consideration, the predicted power consumption of each node is adjusted, so as to calculate a more accurate power consumption of the substation area in the preset time zone.

[0026] For example, if the total power consumption of 10 meter distribution nodes in a substation is predicted to be 1,000 kWh in a specific time zone, then after a comprehensive analysis of the topology information and taking into account the transmission loss between the nodes, the final power consumption forecast result of the substation may be slightly higher, such as 1,050 kWh, to ensure that the power supply can cover the actual needs of the entire substation, thereby optimizing distribution management and ensuring the rational allocation and efficient use of power resources.

[0027] Step 6: Perform power distribution management of the area according to the power consumption of the area in the preset time zone.

[0028] Specifically, after calculating the power consumption of the substation in the preset time zone, the power distribution management of the substation is carried out according to the power consumption of the substation in the preset time zone, that is, appropriate power is allocated to each meter distribution node based on the power consumption of the substation in the preset time zone, so as to ensure that all users in the preset time zone can obtain a stable power supply, reasonably and efficiently allocate power resources, and ensure that all users in the substation can obtain a stable power supply in different time zones, thereby realizing the optimized operation of the power system.

[0029] Further, step three of this application includes: According to the first meter user type and the first meter user scale, a meter big data analysis request is constructed and sent to the first edge computing center to perform the following steps: according to the first meter user type, weekly electricity consumption data of users of the same type are collected; the weekly electricity consumption data of users of the same type are divided into time zones to obtain the peak time zone of the first meter user, the average time zone of the first meter user and the valley time zone of the first meter user; according to the peak time zone of the first meter user, the average time zone of the first meter user and the valley time zone of the first meter user, the preset time zone to be predicted is divided to obtain the peak time zone to be predicted of the first meter user, the average time zone to be predicted of the first meter user and the valley time zone to be predicted of the first meter user; according to the first meter user type and the first meter user scale, meter big data analysis is performed on the peak time zone to be predicted of the first meter user, the average time zone to be predicted of the first meter user and the valley time zone to be predicted of the first meter user to obtain the predicted running power in the preset time zone of the first meter.

[0030] Specifically, in the process of performing an electric meter big data analysis on the first electric meter user type and the first electric meter user scale to obtain the predicted power consumption of the first electric meter in the preset time zone, it is first necessary to construct an electric meter big data analysis request according to the first electric meter user type and the first electric meter user scale, and send it to the first edge computing center. Then, the power consumption prediction is performed through the first edge computing center, and the specific steps are described in detail as follows.

[0031] According to the user type of the first meter, the historical weekly electricity consumption data of other users of the same type are collected as the weekly electricity consumption data of the same type of users. The same type of users refers to the user group with similar electricity consumption patterns as the first meter user. For example, if the first meter user type belongs to a residential user, then the electricity consumption data of other similar residential users in a week are collected. These data can reflect the electricity consumption patterns of residential users in different time periods in a week.

[0032] Next, the weekly electricity consumption data of the same type of users is segmented into time zones. Time zone segmentation refers to dividing the data within a week into different time zones according to time periods, usually including peak time zones, average time zones and valley time zones, so as to obtain the peak time zone of the first meter user, the average time zone of the first meter user and the valley time zone of the first meter user. The peak time zone usually represents the time period with the highest electricity consumption, the average time zone is the time period with relatively average electricity consumption, and the valley time zone is the time period with the lowest electricity consumption. For example, the electricity consumption of residential users is the highest between 7 pm and 10 pm, and this period is defined as the peak time zone. Further, according to the time intervals corresponding to the peak time zone of the first meter user, the average time zone of the first meter user and the valley time zone of the first meter user, the preset time zone to be predicted is segmented to generate the peak time zone to be predicted for the first meter user, the average time zone to be predicted for the first meter user and the valley time zone to be predicted for the first meter user. The preset time zone to be predicted is the preset time zone in the future. For example, the peak time zone of the first electricity meter user is from 18:00 to 22:00 on Saturday night, and the time zone from 18:00 to 22:00 on Saturday night in the preset time zone to be predicted is separated out as the peak time zone to be predicted for the first electricity meter user.

[0033] Finally, based on the type of the first meter user and the scale of the first meter user, meter big data analysis is performed on the peak time zone to be predicted for the first meter user, the average time zone to be predicted for the first meter user, and the valley time zone to be predicted for the first meter user, and finally the predicted power consumption of the first meter preset time zone is obtained. The predicted power consumption of the first meter preset time zone represents the predicted value of the power consumption of the first meter user of the first meter distribution node in the preset time zone to be predicted, and can provide accurate power consumption prediction for each meter distribution node, thereby providing data support for subsequent power distribution and optimization management.

[0034] Furthermore, the present application also includes the following steps: Obtain a computing power demand threshold, wherein the computing power demand threshold is the minimum computing power preset by the user for performing electric meter big data analysis; obtain a list of edge computing centers, a list of edge computing center distribution distances, and a list of edge computing center service logs that meet the computing power demand threshold; sort the list of edge computing centers according to the list of edge computing center distribution distances and the list of edge computing center service logs to obtain the first edge computing center.

[0035] Specifically, according to the first meter user type and the first meter user scale, the specific steps of constructing a meter big data analysis request and sending it to the first edge computing center include: first, it is necessary to determine the computing power demand threshold. The computing power demand threshold refers to the minimum computing power required to perform meter big data analysis, such as one billion floating-point operations per second. This threshold is defined according to the user's preset analysis requirements to ensure that the edge computing center can complete the computing task within the specified time. In other words, the minimum computing power required for running power prediction through the edge computing center can be determined by professional and technical personnel in this field through actual testing. Then, the edge computing centers that can meet the computing power demand threshold are selected from all available edge computing centers to form a list of edge computing centers. At the same time, the corresponding edge computing center distribution distance list and edge computing center service log list are generated.

[0036] The edge computing center distribution distance list records the physical distance between each computing center in the edge computing center list and the first meter distribution node; the edge computing center service log list contains the historical service records of each computing center in the edge computing center list, such as the performance and accuracy when performing similar tasks. Next, the edge computing center list is sorted according to the edge computing center distribution distance list and the edge computing center service log list. Specifically, the computing center that is closer to the first meter distribution node will be prioritized according to the edge computing center distribution distance list. Computing centers with closer distances can usually provide faster response times and reduce data transmission delays. Subsequently, the computing centers with closer distances are further screened according to the edge computing center service log list, focusing on historical computing accuracy and service stability. For example, if a computing center has performed well in past tasks, has high computing accuracy and stable services, it will take priority in the screening.

[0037] Through these two steps of screening, an optimal edge computing center, namely the first edge computing center, will be finally determined. This computing center not only meets the user's preset minimum computing power requirements, but also has advantages in distance and service quality, thus providing efficient and reliable computing support for meter big data analysis.

[0038] Furthermore, the present application also includes the following steps: According to the edge computing center service log list, extract the historical computing accuracy record value list of the edge computing center in the specified historical time zone; traverse the historical computing accuracy record value list of the edge computing center to perform mean statistics, and obtain the computing accuracy characteristic value list of the edge computing center; according to the computing accuracy characteristic value list of the edge computing center, sort the first-level sorting results of the edge computing center that meet the computing accuracy characteristic threshold from the edge computing center list; according to the edge computing center distribution distance list, perform the nearest sorting on the first-level sorting results of the edge computing center to obtain the first edge computing center.

[0039] Specifically, the edge computing center list is sorted according to the edge computing center distribution distance list and the edge computing center service log list, that is, the edge computing center list is sorted by reviewing the performance of the edge computing center when performing similar tasks in the past, especially the calculation accuracy in a specific time zone. First, the historical calculation accuracy record value of each computing center in the specified historical time zone is extracted from the edge computing center service log list, and a list of historical calculation accuracy record values ​​of the edge computing center is established. The historical calculation accuracy record value reflects the calculation accuracy of the edge computing center when processing meter data in the past. For example, if the calculation accuracy of a computing center is 98% in a certain period of time, this indicates that the error between its calculation results and the actual results in this period is very small, and it has high reliability.

[0040] Next, traverse the historical computing accuracy record value list of the edge computing center and perform mean statistics on it, that is, calculate the average value of the computing accuracy of each computing center in multiple time periods to generate a list of computing accuracy characteristic values ​​of the edge computing center. According to the computing accuracy characteristic value list of the edge computing center, preliminary screening is performed to extract the edge computing centers that meet the computing accuracy characteristic threshold to form the first-level sorting results of the edge computing center. The computing accuracy characteristic threshold is a standard preset by professional and technical personnel in this field based on actual needs. For example, if the set threshold is 95%, then only edge computing centers with computing accuracy characteristic values ​​greater than 95% will be retained in the first-level sorting results of the edge computing center.

[0041] Furthermore, the first-level sorting results are further screened according to the edge computing center distribution distance list, and the nearest computing center is selected as the first edge computing center to reduce the delay in data transmission, thereby improving computing efficiency and response speed.

[0042] Furthermore, the present application also includes the following steps: The weekly electricity consumption data of the same type of users includes the weekly electricity consumption data of the first same type of users to the weekly electricity consumption data of the Mth same type of users; the weekly electricity consumption data of the first same type of users is divided into time zones to obtain the peak time zone of the first same type of users, the average time zone of the first same type of users and the valley time zone of the first same type of users; until the weekly electricity consumption data of the Mth same type of users is divided into time zones to obtain the peak time zone of the Mth same type of users, the average time zone of the Mth same type of users and the valley time zone of the Mth same type of users; the first same type of user peak time zone, the first same type of user average time zone and the first same type of user valley time zone are subjected to frequency fusion analysis until the Mth same type user peak time zone, the Mth same type user average time zone and the Mth same type user valley time zone are subjected to frequency fusion analysis to obtain the first meter user peak time zone, the first meter user average time zone and the first meter user valley time zone.

[0043] Specifically, the specific steps of performing time zone segmentation on the weekly electricity consumption data of the same type of users, and obtaining the peak time zone of the first meter user, the average time zone of the first meter user, and the valley time zone of the first meter user include: in the aforementioned steps, the weekly electricity consumption data of the same type of users are collected, and the weekly electricity consumption data of multiple users of the same type are collected, that is, the weekly electricity consumption data of the first user of the same type until the weekly electricity consumption data of the Mth user of the same type, where M is the total number of users of the same type. Next, the weekly electricity consumption data of each user of the same type is segmented by time zone, that is, the electricity consumption data within a week is divided according to time periods, usually into peak time zones, average time zones, and valley time zones. These time zones reflect the electricity consumption characteristics of users in different time periods within a week. This process will be carried out for all users of the same type until the time zone segmentation of the weekly electricity consumption data of the Mth user of the same type is completed. At this point, the respective peak time zones, average time zones, and valley time zones have been obtained for all users of the same type.

[0044] After completing the time zone segmentation, frequency fusion analysis is performed on the first same-type user peak time zone, the first same-type user average time zone, and the first same-type user valley time zone, until the Mth same-type user peak time zone, the Mth same-type user average time zone, and the Mth same-type user valley time zone. Frequency fusion analysis refers to comparing the peak, average, and valley time zones of all the same-type users to find the time period with the highest frequency in these time zones. For example, the peak time zones of most same-type users are concentrated between 9 and 11 a.m., so this time period is used as the peak time zone. The same analysis also applies to average time zones and valley time zones.

[0045] Through this series of steps, the peak time zone of the first meter user, the average time zone of the first meter user and the valley time zone of the first meter user can be finally obtained, so as to more accurately predict the electricity demand of the target user in different time periods, better meet the actual needs of the user during electricity distribution, and provide strong data support for subsequent electricity optimization and management, thereby improving the refinement of power distribution management.

[0046] Furthermore, the present application also includes the following steps: Based on the preset time unit, the weekly electricity consumption data of the first same type of user is segmented to obtain the weekly electricity consumption time series information of the first same type of user; the central trend mean analysis is performed on the weekly electricity consumption time series information of the first same type of user to obtain the preset unit time characteristic electricity consumption; based on the preset electricity consumption deviation, the bandwidth of the preset unit time characteristic electricity consumption is configured to obtain the average electricity consumption characteristic interval; according to the average electricity consumption characteristic interval, the weekly electricity consumption data of the first same type of user is segmented by time zone to obtain the peak time zone of the first same type of user, the average time zone of the first same type of user and the valley time zone of the first same type of user.

[0047] Specifically, the specific steps of dividing the weekly electricity consumption data of the first same type of user into time zones and obtaining the peak time zone of the first same type of user, the average time zone of the first same type of user and the valley time zone of the first same type of user include: first, it is necessary to determine the preset time unit, which refers to the time interval for segmenting the user's electricity consumption data (for example, every hour or every 15 minutes), which is set by professional and technical personnel in this field, so as to segment the weekly electricity consumption data of the first same type of user and generate the weekly electricity consumption time series information of the first same type of user. The weekly electricity consumption time series information of the first same type of user reflects the electricity consumption of the first same type of user in each time unit throughout the week, forming a series of ordered electricity consumption data points.

[0048] Next, a centralized tendency mean analysis is performed on the weekly electricity consumption time series information of the first user of the same type. The centralized tendency mean analysis refers to the calculation of the mean value after counting the electricity consumption within each preset time unit, and the result is used as the preset unit time characteristic electricity consumption. Further, the bandwidth is configured based on the preset electricity consumption deviation and the preset unit time characteristic electricity consumption, wherein the preset electricity consumption deviation refers to the allowable range of electricity consumption fluctuation. According to the preset electricity consumption deviation, the bandwidth is configured for the preset unit time characteristic electricity consumption, and the average electricity consumption characteristic interval can be defined. For example, if the preset electricity consumption deviation is ±10%, the preset unit time characteristic electricity consumption is 200 kilowatts, and the average electricity consumption characteristic interval is 180 kWh to 220 kWh.

[0049] The weekly electricity consumption data of the first same-type user is divided into time zones according to the average power consumption characteristic interval, that is, the time period when the power consumption is higher than the upper limit of the average value characteristic interval is marked as the peak time zone, the time period when the power consumption is lower than the lower limit of the average value characteristic interval is marked as the valley time zone, and the time period falling within the average value characteristic interval is marked as the average time zone, thereby generating the first same-type user peak time zone, the first same-type user average time zone and the first same-type user valley time zone. Through accurate time zone division, resources can be better allocated, thereby improving the refinement of power distribution management.

[0050] Furthermore, the present application also includes the following steps: According to the first same-type user peak time zone, the first same-type user average time zone and the first same-type user valley time zone, until the Mth same-type user peak time zone, the Mth same-type user average time zone and the Mth same-type user valley time zone, obtain the first moment peak frequency, the first moment valley frequency and the first moment average frequency; when the first moment peak frequency is the largest, add the first moment to the first meter user peak time zone; when the first moment valley frequency is the largest, add the first moment to the first meter user valley time zone; when the first moment average frequency is the largest, add the first moment to the first meter user average time zone.

[0051] Specifically, frequency fusion analysis is performed on the peak time zone of the first same-type user, the average time zone of the first same-type user and the valley time zone of the first same-type user, until the peak time zone of the Mth same-type user, the average time zone of the Mth same-type user and the valley time zone of the Mth same-type user, and the specific steps of obtaining the peak time zone of the first meter user, the average time zone of the first meter user and the valley time zone of the first meter user include: first, according to the peak time zone, average time zone and valley time zone of each same-type user, the peak frequency, valley frequency and average frequency at each moment are calculated. Simply put, these frequency values ​​correspond to the number of times the same moment is marked as a peak, average or valley time zone by the same-type user.

[0052] For example, the peak time zones, average time zones, and valley time zones of 100 similar users are counted, and a time is randomly selected, such as 10:00 to 11:00 a.m. on Monday. If this time period is the peak time zone of 90 of the users, the peak frequency corresponding to this time is 90. Similarly, according to the first peak time zone of the same user, the first average time zone of the same user, and the first valley time zone of the same user, until the Mth peak time zone of the same user, the Mth average time zone of the same user, and the Mth valley time zone of the same user, the peak frequency, valley frequency, and average frequency at each time are calculated to generate the first time peak frequency, the first time valley frequency, and the first time average frequency, where the first time refers to any time, that is, each time is traversed.

[0053] Furthermore, the peak frequency, valley frequency and average frequency at the same moment are compared. If the peak frequency at a certain moment is the largest, it means that the power consumption of most users of the same type has reached a peak at this moment. At this time, this moment is classified as a peak time zone. If the valley frequency at a certain moment is the largest, it indicates that the power consumption of most users of the same type during this period is low, and this moment is classified as a valley time zone. Similarly, if the average frequency at a certain moment is the largest, this moment is divided into an average time zone. Based on this, by comparing the peak frequency at the first moment, the valley frequency at the first moment and the average frequency at the first moment, the classification time zone corresponding to the first moment is determined. After traversing and analyzing each moment, the peak time zone of the first meter user, the average time zone of the first meter user and the valley time zone of the first meter user can be obtained. Through the time zone division based on frequency analysis, the power consumption characteristic time zone of the target meter user can be predicted more accurately, which improves the reliability and accuracy of the prediction and provides important reference data for subsequent power distribution optimization.

[0054] Furthermore, the present application also includes the following steps: According to the type of the first meter user and the scale of the first meter user, obtain a data set of electricity consumption records per unit time in the peak time zone, a data set of electricity consumption records per unit time in the average time zone, and a data set of electricity consumption records per unit time in the valley time zone of the same type and mode users; traverse the data set of electricity consumption records per unit time in the peak time zone, the data set of electricity consumption records per unit time in the average time zone, and the data set of electricity consumption records per unit time in the valley time zone of the same type and mode users to perform de-outlier mean evaluation to obtain characteristic electricity consumption per unit time in the peak time zone, characteristic electricity consumption per unit time in the average time zone, and characteristic electricity consumption per unit time in the valley time zone; according to the characteristic electricity consumption per unit time in the peak time zone, the characteristic electricity consumption per unit time in the average time zone, and the characteristic electricity consumption per unit time in the valley time zone, fit the peak time zone to be predicted by the first meter user, the average time zone to be predicted by the first meter user, and the valley time zone to be predicted by the first meter user to obtain the predicted running electricity in the preset time zone of the first meter.

[0055] Specifically, according to the first meter user type and the first meter user scale, the meter big data analysis is performed on the first meter user's predicted peak time zone, the first meter user's predicted average time zone and the first meter user's predicted valley time zone, and the step of obtaining the predicted running power of the first meter preset time zone includes: first, according to the first meter user type and the first meter user scale, the same type and model users are determined, and the same type and model users refer to users with the same type and scale as the first meter user. Collect the unit time power consumption data of the same type and model users in the peak time zone, average time zone and valley time zone, and establish a data set of unit time power consumption records of the same type and model users in the peak time zone, a data set of unit time power consumption records of the same type and model users in the average time zone and a data set of unit time power consumption records of the same type and model users in the valley time zone. Traverse these data sets and perform de-outlier mean evaluation. De-outlierization refers to removing abnormal data points that obviously deviate from the normal range from the data set to avoid these outliers from having too much impact on the calculation of the average value. Then, the mean is calculated for the cleaned data set to obtain the characteristic electricity consumption per unit time in the peak time zone, the characteristic electricity consumption per unit time in the flat time zone, and the characteristic electricity consumption per unit time in the valley time zone.

[0056] De-outlierization is an existing data processing method. Common methods include the interquartile range method, standard score, etc. Taking the interquartile range method as an example, the peak time zone unit duration electricity consumption record data set of the same type and mode user, the average time zone unit duration electricity consumption record data set of the same type and mode user, and the valley time zone unit duration electricity consumption record data set of the same type and mode user are traversed, and each data set is sorted respectively, and then the first quartile and the third quartile are calculated. The first quartile is the value at the 25% position in the sorted data, and the third quartile is the value at the 75% position. Calculate the interquartile range, that is, the difference between the third quartile and the first quartile, and then determine the range of outliers, with the first quartile minus 1.5 times the interquartile range as the lower bound, and the third quartile plus 1.5 times the interquartile range as the upper bound. If the data point is lower than the lower bound or higher than the upper bound, it is regarded as an outlier. Data points that exceed the above upper and lower bounds are marked as outliers and removed, and only normal data within the range is retained.

[0057] Finally, the characteristic electricity consumption per unit time in the peak time zone, the characteristic electricity consumption per unit time in the flat time zone and the characteristic electricity consumption per unit time in the valley time zone of the same type and mode users are applied to the peak time zone to be predicted by the first meter user, the flat time zone to be predicted by the first meter user and the valley time zone to be predicted by the first meter user, and the electricity consumption corresponding to the peak time zone to be predicted by the first meter user, the flat time zone to be predicted by the first meter user and the valley time zone to be predicted by the first meter user are obtained, and then added up and calculated, and the result obtained is the predicted electricity consumption in the preset time zone of the first meter, thereby generating accurate electricity consumption forecast data for each meter user, thereby optimizing power dispatching and management, and thus improving the refinement of power distribution management.

[0058] Further, as attached Figure 2 As shown, step five of this application includes: According to the historical distribution data of the first electricity meter, the distribution loss mode value statistics of the first electricity meter distribution node are performed to obtain a first distribution loss ratio; until the historical distribution data of the Nth electricity meter are used to perform distribution loss mode value statistics on the Nth electricity meter distribution node to obtain an Nth distribution loss ratio; according to the first distribution loss ratio up to the Nth distribution loss ratio, the predicted power consumption of the first electricity meter in the preset time zone up to the predicted power consumption of the Nth electricity meter in the preset time zone are reverse calibrated and added to obtain the power consumption of the station in the preset time zone.

[0059] Specifically, the historical power distribution data of the first meter refers to the historical data related to power distribution recorded for the first meter distribution node in the past period of time, which usually includes the power consumption recorded by the meter, the power supply from the substation or power supply source to the meter distribution node, and the energy loss that may occur during the power distribution process. These historical data are used to analyze the typical power loss of the meter distribution node during the power distribution process and calculate the distribution loss ratio. Distribution loss refers to the energy lost during the transmission and distribution of electricity due to line resistance, transformer loss and other power equipment. Distribution loss is usually manifested as the difference between the power supply and the power consumption. The distribution loss ratio is usually expressed as a percentage, which is the proportion of distribution loss to power supply, reflecting the degree of power loss in the entire distribution process. The distribution loss ratio is equal to the ratio of distribution loss to power supply.

[0060] For example, the first meter distribution node recorded a power supply of 1000 kWh in historical time, while the actual power consumption recorded by the meter was 950 kWh. At this time, the power distribution loss is 50 kWh, and the power distribution loss ratio is 5%, which means that the meter distribution node lost 5% of the power during the power distribution process. Specifically, the calculation of the historical power distribution data of the first meter includes multiple groups of data, and the power distribution loss ratio can be calculated multiple times and then the mode value is counted to obtain the first power distribution loss ratio. Similarly, according to the historical power distribution data of the Nth meter, the power distribution loss mode value of the Nth meter distribution node is counted to obtain the Nth power distribution loss ratio.

[0061] Next, the predicted power consumption in the preset time zone of each meter distribution node is reversely calibrated according to the first distribution loss ratio to the Nth distribution loss ratio. Reverse calibration refers to reversely adjusting the predicted power consumption through the distribution loss ratio to better reflect the actual power consumption of the meter distribution node. For example, if the predicted power consumption of a meter distribution node is 100 kWh, but its distribution loss ratio is 5%, then its actual power consumption is adjusted to 105 kWh to compensate for this part of the distribution loss. Finally, the reverse-calibrated power consumption of all meter distribution nodes is added up to obtain the preset time zone power consumption of the entire substation, so as to accurately predict the total power demand of the substation, while taking into account the distribution loss of each node, providing key data support for power dispatch and optimization management, and improving the refinement of power distribution.

[0062] In summary, the multi-dimensional power distribution optimization method based on electric meter big data and edge computing provided by this application has the following technical effects: Receive user distribution topology information of a substation, wherein the user distribution topology information includes a first meter distribution node up to an Nth meter distribution node; traverse the first meter distribution node up to the Nth meter distribution node, extract a first meter user type and a first meter user scale, up to an Nth meter user type and an Nth meter user scale; perform meter big data analysis on the first meter user type and the first meter user scale through a first edge computing center, and obtain a predicted power outage amount of the first meter in a preset time zone; perform meter big data analysis on the Nth meter user type and the Nth meter user scale through an Nth edge computing center, and obtain a predicted power outage amount of the Nth meter in a preset time zone, wherein the first edge computing center and the Nth edge computing center operate simultaneously; perform substation power consumption prediction based on the user distribution topology information according to the predicted power outage amount of the first meter in a preset time zone up to the predicted power outage amount of the Nth meter in a preset time zone, and obtain the substation power consumption in a preset time zone; perform power distribution management in the substation according to the substation power consumption in the preset time zone. By receiving the topological information of user distribution in the substation, conducting big data analysis based on user type and scale, and combining the edge computing center to parallelly process the power consumption forecast of different meter distribution nodes, while taking distribution losses into consideration, the forecast results are corrected through reverse calibration, thereby achieving the technical effect of improving the power consumption forecast and distribution refinement of the substation, thereby improving the efficiency and accuracy of distribution management in the entire substation.

[0063] 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.

[0064] 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. A multi-dimensional power distribution optimization method based on electric meter big data and edge computing, characterized in that: include: Receiving user distribution topology information of the substation, wherein the user distribution topology information includes a first electricity meter distribution node to an Nth electricity meter distribution node; Traversing the first electricity meter distribution node until the Nth electricity meter distribution node, extracting the first electricity meter user type and the first electricity meter user scale until the Nth electricity meter user type and the Nth electricity meter user scale; Through the first edge computing center, an electric meter big data analysis is performed on the user type and user scale of the first electric meter to obtain a predicted running power of the first electric meter in a preset time zone; By means of an Nth edge computing center, an electric meter big data analysis is performed on the user type and the user scale of the Nth electric meter to obtain a predicted running power of the Nth electric meter in a preset time zone, wherein the first edge computing center and the Nth edge computing center are operated simultaneously; According to the first electric meter's preset time zone predicted power consumption until the Nth electric meter's preset time zone predicted power consumption, based on the user distribution topology information, perform a station area power consumption prediction to obtain a preset time zone station area power consumption; The power distribution management of the substation is performed according to the power consumption of the substation in the preset time zone.

2. The multi-dimensional power distribution optimization method based on electric meter big data and edge computing according to claim 1 is characterized in that: The first edge computing center performs an electric meter big data analysis on the user type and user scale of the first electric meter to obtain a predicted running power of the first electric meter in a preset time zone, including: According to the first electricity meter user type and the first electricity meter user scale, an electricity meter big data analysis request is constructed and sent to the first edge computing center to perform the following steps: According to the user type of the first electric meter, weekly electricity consumption data of users of the same type are collected; Performing time zone segmentation on the weekly electricity consumption data of the same type of users to obtain a peak time zone of a first electricity meter user, a flat time zone of a first electricity meter user, and a valley time zone of a first electricity meter user; According to the peak time zone of the first meter user, the average time zone of the first meter user and the valley time zone of the first meter user, the preset time zone to be predicted is divided to obtain the peak time zone to be predicted for the first meter user, the average time zone to be predicted for the first meter user and the valley time zone to be predicted for the first meter user; According to the type of the first meter user and the scale of the first meter user, meter big data analysis is performed on the first meter user's predicted peak time zone, the first meter user's predicted average time zone and the first meter user's predicted valley time zone to obtain the predicted running power of the first meter in the preset time zone.

3. The multi-dimensional power distribution optimization method based on electric meter big data and edge computing as claimed in claim 2, characterized in that: According to the first electricity meter user type and the first electricity meter user scale, construct an electricity meter big data analysis request and send it to the first edge computing center, including: Obtaining a computing power requirement threshold, wherein the computing power requirement threshold is a minimum computing power preset by a user for performing electric meter big data analysis; Obtain a list of edge computing centers that meet the computing power requirement threshold, a list of edge computing center distribution distances, and a list of edge computing center service logs; According to the edge computing center distribution distance list and the edge computing center service log list, the edge computing center list is sorted to obtain the first edge computing center.

4. The multi-dimensional power distribution optimization method based on electric meter big data and edge computing as claimed in claim 3 is characterized in that: According to the edge computing center distribution distance list and the edge computing center service log list, the edge computing center list is sorted to obtain the first edge computing center, including: According to the edge computing center service log list, extract the edge computing center historical calculation accuracy record value list in the specified historical time zone; Traversing the historical computing accuracy record value list of the edge computing center to perform average statistics, and obtaining a computing accuracy characteristic value list of the edge computing center; According to the edge computing power center computing precision characteristic value list, first-level sorting results of edge computing power centers that meet the computing precision characteristic threshold are sorted from the edge computing power center list; The first-level sorting results of the edge computing power centers are sorted closest according to the edge computing power center distribution distance list to obtain the first edge computing power center.

5. The multi-dimensional power distribution optimization method based on electric meter big data and edge computing as claimed in claim 2, characterized in that: The weekly electricity consumption data of the same type of users is divided into time zones to obtain a peak time zone of a first meter user, a flat time zone of a first meter user, and a valley time zone of a first meter user, including: The weekly electricity consumption data of the same type of users includes the weekly electricity consumption data of the first same type of users to the weekly electricity consumption data of the Mth same type of users; Performing time zone segmentation on the weekly electricity consumption data of the first same-type users to obtain a peak time zone of the first same-type users, an average time zone of the first same-type users, and a valley time zone of the first same-type users; Until the weekly electricity consumption data of the Mth user of the same type is divided into time zones to obtain the peak time zone of the Mth user of the same type, the average time zone of the Mth user of the same type, and the valley time zone of the Mth user of the same type; Frequency fusion analysis is performed on the first same-type user peak time zone, the first same-type user average time zone and the first same-type user valley time zone, until the Mth same-type user peak time zone, the Mth same-type user average time zone and the Mth same-type user valley time zone, to obtain the first meter user peak time zone, the first meter user average time zone and the first meter user valley time zone.

6. The multi-dimensional power distribution optimization method based on electric meter big data and edge computing according to claim 5 is characterized in that: The weekly electricity consumption data of the first same-type user is divided into time zones to obtain the first same-type user peak time zone, the first same-type user average time zone and the first same-type user valley time zone, including: Based on a preset time unit, the weekly electricity consumption data of the first same-type user is segmented to obtain the weekly electricity consumption time series information of the first same-type user; Performing a central tendency mean analysis on the weekly power consumption time series information of the first same type of user to obtain a preset characteristic power consumption per unit time; Based on the preset power consumption deviation, bandwidth configuration is performed on the preset characteristic power consumption per unit time to obtain a mean power consumption characteristic interval; The weekly electricity consumption data of the first same-type user is divided into time zones according to the average electricity consumption characteristic interval to obtain the first same-type user peak time zone, the first same-type user average time zone and the first same-type user valley time zone.

7. The multi-dimensional power distribution optimization method based on electric meter big data and edge computing according to claim 5, characterized in that: Performing frequency fusion analysis on the first same-type user peak time zone, the first same-type user average time zone, and the first same-type user valley time zone, until the Mth same-type user peak time zone, the Mth same-type user average time zone, and the Mth same-type user valley time zone, to obtain the first meter user peak time zone, the first meter user average time zone, and the first meter user valley time zone, including: According to the first same-type user peak time zone, the first same-type user average time zone and the first same-type user valley time zone, until the Mth same-type user peak time zone, the Mth same-type user average time zone and the Mth same-type user valley time zone, obtain a first moment peak frequency, a first moment valley frequency and a first moment average frequency; When the peak frequency at the first moment is the largest, adding the first moment into the user peak time zone of the first electricity meter; When the valley frequency of the first moment is the largest, adding the first moment into the valley time zone of the user of the first electric meter; When the average value frequency of the first moment is the largest, the first moment is added into the average value time zone of the first electricity meter user.

8. The multi-dimensional power distribution optimization method based on electric meter big data and edge computing according to claim 2, characterized in that: According to the type of the first meter user and the scale of the first meter user, an electric meter big data analysis is performed on the peak time zone to be predicted by the first meter user, the average time zone to be predicted by the first meter user, and the valley time zone to be predicted by the first meter user to obtain the predicted running power of the first meter in the preset time zone, including: According to the type of the first electricity meter user and the scale of the first electricity meter user, obtain a data set of electricity consumption records per unit time in a peak time zone for users of the same type and mode, a data set of electricity consumption records per unit time in a flat time zone for users of the same type and mode, and a data set of electricity consumption records per unit time in a valley time zone for users of the same type and mode; Traverse the peak time zone unit duration electricity consumption record data set, the average time zone unit duration electricity consumption record data set, and the valley time zone unit duration electricity consumption record data set for the same type and mode users to perform de-outlierized mean evaluation, and obtain the peak time zone unit duration characteristic electricity consumption, the average time zone unit duration characteristic electricity consumption, and the valley time zone unit duration characteristic electricity consumption; According to the characteristic electricity consumption per unit time in the peak time zone, the characteristic electricity consumption per unit time in the average time zone and the characteristic electricity consumption per unit time in the valley time zone, the peak time zone to be predicted by the user of the first meter, the average time zone to be predicted by the user of the first meter and the valley time zone to be predicted by the user of the first meter are fitted to obtain the predicted running electricity in the preset time zone of the first meter.

9. The multi-dimensional power distribution optimization method based on electric meter big data and edge computing according to claim 1, characterized in that: According to the first electric meter's preset time zone predicted power consumption until the Nth electric meter's preset time zone predicted power consumption, based on the user distribution topology information, performing a station area power consumption prediction to obtain the preset time zone station area power consumption, including: Performing power distribution loss mode statistics on the first power meter distribution node according to the historical power distribution data of the first power meter to obtain a first power distribution loss ratio; Until the distribution loss mode value statistics of the Nth electricity meter distribution node are performed according to the historical distribution data of the Nth electricity meter to obtain the Nth distribution loss ratio; According to the first power distribution loss ratio to the Nth power distribution loss ratio, the predicted power consumption of the first meter in the preset time zone to the predicted power consumption of the Nth meter in the preset time zone are reversely calibrated and summed to obtain the power consumption of the preset time zone.

Images